Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / llvm / 2c651fe6f445724627dcc48064797dca2aa4aedc / . / docs / LangRef.html
blob: 3a474a554d4c5de0151f7aeec055c976a44d81ec [file] [log] [blame]
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3<html>
4<head>
5 <title>LLVM Assembly Language Reference Manual</title>
Reid Spencer3921c742004-08-26 20:44:00 +0000[diff] [blame]6 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
7 <meta name="author" content="Chris Lattner">
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]8 <meta name="description"
Reid Spencer3921c742004-08-26 20:44:00 +0000[diff] [blame]9 content="LLVM Assembly Language Reference Manual.">
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]10 <link rel="stylesheet" href="llvm.css" type="text/css">
11</head>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]12
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]13<body>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]14
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]15<h1>LLVM Language Reference Manual</h1>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]16<ol>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]17 <li><a href="#abstract">Abstract</a></li>
18 <li><a href="#introduction">Introduction</a></li>
19 <li><a href="#identifiers">Identifiers</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]20 <li><a href="#highlevel">High Level Structure</a>
21 <ol>
22 <li><a href="#modulestructure">Module Structure</a></li>
Bill Wendling3d10a5a2009-07-20 01:03:30 +0000[diff] [blame]23 <li><a href="#linkage">Linkage Types</a>
24 <ol>
Bill Wendling987e7eb2009-07-20 02:41:50 +0000[diff] [blame]25 <li><a href="#linkage_private">'<tt>private</tt>' Linkage</a></li>
26 <li><a href="#linkage_linker_private">'<tt>linker_private</tt>' Linkage</a></li>
Bill Wendling5e721d72010-07-01 21:55:59 +0000[diff] [blame]27 <li><a href="#linkage_linker_private_weak">'<tt>linker_private_weak</tt>' Linkage</a></li>
Bill Wendling55ae5152010-08-20 22:05:50 +0000[diff] [blame]28 <li><a href="#linkage_linker_private_weak_def_auto">'<tt>linker_private_weak_def_auto</tt>' Linkage</a></li>
Bill Wendling987e7eb2009-07-20 02:41:50 +0000[diff] [blame]29 <li><a href="#linkage_internal">'<tt>internal</tt>' Linkage</a></li>
30 <li><a href="#linkage_available_externally">'<tt>available_externally</tt>' Linkage</a></li>
31 <li><a href="#linkage_linkonce">'<tt>linkonce</tt>' Linkage</a></li>
32 <li><a href="#linkage_common">'<tt>common</tt>' Linkage</a></li>
33 <li><a href="#linkage_weak">'<tt>weak</tt>' Linkage</a></li>
34 <li><a href="#linkage_appending">'<tt>appending</tt>' Linkage</a></li>
35 <li><a href="#linkage_externweak">'<tt>extern_weak</tt>' Linkage</a></li>
Chris Lattner5a2d8752009-10-10 18:26:06 +0000[diff] [blame]36 <li><a href="#linkage_linkonce_odr">'<tt>linkonce_odr</tt>' Linkage</a></li>
Bill Wendling987e7eb2009-07-20 02:41:50 +0000[diff] [blame]37 <li><a href="#linkage_weak">'<tt>weak_odr</tt>' Linkage</a></li>
Bill Wendlingf7f06102011-10-11 06:41:28 +0000[diff] [blame]38 <li><a href="#linkage_external">'<tt>external</tt>' Linkage</a></li>
Bill Wendling987e7eb2009-07-20 02:41:50 +0000[diff] [blame]39 <li><a href="#linkage_dllimport">'<tt>dllimport</tt>' Linkage</a></li>
40 <li><a href="#linkage_dllexport">'<tt>dllexport</tt>' Linkage</a></li>
Bill Wendling3d10a5a2009-07-20 01:03:30 +0000[diff] [blame]41 </ol>
42 </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]43 <li><a href="#callingconv">Calling Conventions</a></li>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]44 <li><a href="#namedtypes">Named Types</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]45 <li><a href="#globalvars">Global Variables</a></li>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]46 <li><a href="#functionstructure">Functions</a></li>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]47 <li><a href="#aliasstructure">Aliases</a></li>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]48 <li><a href="#namedmetadatastructure">Named Metadata</a></li>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]49 <li><a href="#paramattrs">Parameter Attributes</a></li>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]50 <li><a href="#fnattrs">Function Attributes</a></li>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]51 <li><a href="#gc">Garbage Collector Names</a></li>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]52 <li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]53 <li><a href="#datalayout">Data Layout</a></li>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]54 <li><a href="#pointeraliasing">Pointer Aliasing Rules</a></li>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]55 <li><a href="#volatile">Volatile Memory Accesses</a></li>
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]56 <li><a href="#memmodel">Memory Model for Concurrent Operations</a></li>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]57 <li><a href="#ordering">Atomic Memory Ordering Constraints</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]58 </ol>
59 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]60 <li><a href="#typesystem">Type System</a>
61 <ol>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]62 <li><a href="#t_classifications">Type Classifications</a></li>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]63 <li><a href="#t_primitive">Primitive Types</a>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]64 <ol>
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]65 <li><a href="#t_integer">Integer Type</a></li>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]66 <li><a href="#t_floating">Floating Point Types</a></li>
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]67 <li><a href="#t_x86mmx">X86mmx Type</a></li>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]68 <li><a href="#t_void">Void Type</a></li>
69 <li><a href="#t_label">Label Type</a></li>
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]70 <li><a href="#t_metadata">Metadata Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]71 </ol>
72 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]73 <li><a href="#t_derived">Derived Types</a>
74 <ol>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]75 <li><a href="#t_aggregate">Aggregate Types</a>
76 <ol>
77 <li><a href="#t_array">Array Type</a></li>
78 <li><a href="#t_struct">Structure Type</a></li>
Chris Lattner628ed392011-07-23 19:59:08 +0000[diff] [blame]79 <li><a href="#t_opaque">Opaque Structure Types</a></li>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]80 <li><a href="#t_vector">Vector Type</a></li>
81 </ol>
82 </li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]83 <li><a href="#t_function">Function Type</a></li>
84 <li><a href="#t_pointer">Pointer Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]85 </ol>
86 </li>
87 </ol>
88 </li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]89 <li><a href="#constants">Constants</a>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]90 <ol>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]91 <li><a href="#simpleconstants">Simple Constants</a></li>
Chris Lattner70882792009-02-28 18:32:25 +0000[diff] [blame]92 <li><a href="#complexconstants">Complex Constants</a></li>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]93 <li><a href="#globalconstants">Global Variable and Function Addresses</a></li>
94 <li><a href="#undefvalues">Undefined Values</a></li>
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]95 <li><a href="#poisonvalues">Poison Values</a></li>
Chris Lattnerf9d078e2009-10-27 21:19:13 +0000[diff] [blame]96 <li><a href="#blockaddress">Addresses of Basic Blocks</a></li>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]97 <li><a href="#constantexprs">Constant Expressions</a></li>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]98 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]99 </li>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]100 <li><a href="#othervalues">Other Values</a>
101 <ol>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]102 <li><a href="#inlineasm">Inline Assembler Expressions</a></li>
Peter Collingbourne249d9532011-10-27 19:19:07 +0000[diff] [blame]103 <li><a href="#metadata">Metadata Nodes and Metadata Strings</a>
104 <ol>
105 <li><a href="#tbaa">'<tt>tbaa</tt>' Metadata</a></li>
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>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]260 </ol>
261 </li>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]262 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]263 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]264 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]265 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
266 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
267 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]268 </ol>
269 </li>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]270 <li><a href="#int_overflow">Arithmetic with Overflow Intrinsics</a>
271 <ol>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]272 <li><a href="#int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt> Intrinsics</a></li>
273 <li><a href="#int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt> Intrinsics</a></li>
274 <li><a href="#int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt> Intrinsics</a></li>
275 <li><a href="#int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt> Intrinsics</a></li>
276 <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]277 <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]278 </ol>
279 </li>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]280 <li><a href="#int_fp16">Half Precision Floating Point Intrinsics</a>
281 <ol>
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]282 <li><a href="#int_convert_to_fp16">'<tt>llvm.convert.to.fp16</tt>' Intrinsic</a></li>
283 <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]284 </ol>
285 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]286 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]287 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]288 <li><a href="#int_trampoline">Trampoline Intrinsics</a>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]289 <ol>
290 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]291 <li><a href="#int_at">'<tt>llvm.adjust.trampoline</tt>' Intrinsic</a></li>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]292 </ol>
293 </li>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]294 <li><a href="#int_memorymarkers">Memory Use Markers</a>
295 <ol>
Jakub Staszak8e1b12a2011-12-04 20:44:25 +0000[diff] [blame]296 <li><a href="#int_lifetime_start">'<tt>llvm.lifetime.start</tt>' Intrinsic</a></li>
297 <li><a href="#int_lifetime_end">'<tt>llvm.lifetime.end</tt>' Intrinsic</a></li>
298 <li><a href="#int_invariant_start">'<tt>llvm.invariant.start</tt>' Intrinsic</a></li>
299 <li><a href="#int_invariant_end">'<tt>llvm.invariant.end</tt>' Intrinsic</a></li>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]300 </ol>
301 </li>
Reid Spencer20677642007-07-20 19:59:11 +0000[diff] [blame]302 <li><a href="#int_general">General intrinsics</a>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]303 <ol>
Reid Spencer20677642007-07-20 19:59:11 +0000[diff] [blame]304 <li><a href="#int_var_annotation">
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]305 '<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]306 <li><a href="#int_annotation">
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]307 '<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]308 <li><a href="#int_trap">
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]309 '<tt>llvm.trap</tt>' Intrinsic</a></li>
310 <li><a href="#int_stackprotector">
311 '<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]312 <li><a href="#int_objectsize">
313 '<tt>llvm.objectsize</tt>' Intrinsic</a></li>
Jakub Staszakb170e2d2011-12-04 18:29:26 +0000[diff] [blame]314 <li><a href="#int_expect">
315 '<tt>llvm.expect</tt>' Intrinsic</a></li>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]316 </ol>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]317 </li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]318 </ol>
319 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]320</ol>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]321
322<div class="doc_author">
323 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
324 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]325</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]326
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]327<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]328<h2><a name="abstract">Abstract</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]329<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]330
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]331<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]332
333<p>This document is a reference manual for the LLVM assembly language. LLVM is
334 a Static Single Assignment (SSA) based representation that provides type
335 safety, low-level operations, flexibility, and the capability of representing
336 'all' high-level languages cleanly. It is the common code representation
337 used throughout all phases of the LLVM compilation strategy.</p>
338
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]339</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]340
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]341<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]342<h2><a name="introduction">Introduction</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]343<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]344
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]345<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]346
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]347<p>The LLVM code representation is designed to be used in three different forms:
348 as an in-memory compiler IR, as an on-disk bitcode representation (suitable
349 for fast loading by a Just-In-Time compiler), and as a human readable
350 assembly language representation. This allows LLVM to provide a powerful
351 intermediate representation for efficient compiler transformations and
352 analysis, while providing a natural means to debug and visualize the
353 transformations. The three different forms of LLVM are all equivalent. This
354 document describes the human readable representation and notation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]355
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]356<p>The LLVM representation aims to be light-weight and low-level while being
357 expressive, typed, and extensible at the same time. It aims to be a
358 "universal IR" of sorts, by being at a low enough level that high-level ideas
359 may be cleanly mapped to it (similar to how microprocessors are "universal
360 IR's", allowing many source languages to be mapped to them). By providing
361 type information, LLVM can be used as the target of optimizations: for
362 example, through pointer analysis, it can be proven that a C automatic
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]363 variable is never accessed outside of the current function, allowing it to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]364 be promoted to a simple SSA value instead of a memory location.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]365
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]366<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]367<h4>
368 <a name="wellformed">Well-Formedness</a>
369</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]370
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]371<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]372
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]373<p>It is important to note that this document describes 'well formed' LLVM
374 assembly language. There is a difference between what the parser accepts and
375 what is considered 'well formed'. For example, the following instruction is
376 syntactically okay, but not well formed:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]377
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]378<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]379%x = <a href="#i_add">add</a> i32 1, %x
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]380</pre>
381
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]382<p>because the definition of <tt>%x</tt> does not dominate all of its uses. The
383 LLVM infrastructure provides a verification pass that may be used to verify
384 that an LLVM module is well formed. This pass is automatically run by the
385 parser after parsing input assembly and by the optimizer before it outputs
386 bitcode. The violations pointed out by the verifier pass indicate bugs in
387 transformation passes or input to the parser.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]388
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]389</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]390
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]391</div>
392
Chris Lattnercc689392007-10-03 17:34:29 +0000[diff] [blame]393<!-- Describe the typesetting conventions here. -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]394
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]395<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]396<h2><a name="identifiers">Identifiers</a></h2>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]397<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]398
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]399<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]400
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]401<p>LLVM identifiers come in two basic types: global and local. Global
402 identifiers (functions, global variables) begin with the <tt>'@'</tt>
403 character. Local identifiers (register names, types) begin with
404 the <tt>'%'</tt> character. Additionally, there are three different formats
405 for identifiers, for different purposes:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]406
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]407<ol>
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]408 <li>Named values are represented as a string of characters with their prefix.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]409 For example, <tt>%foo</tt>, <tt>@DivisionByZero</tt>,
410 <tt>%a.really.long.identifier</tt>. The actual regular expression used is
411 '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'. Identifiers which require
412 other characters in their names can be surrounded with quotes. Special
413 characters may be escaped using <tt>"\xx"</tt> where <tt>xx</tt> is the
414 ASCII code for the character in hexadecimal. In this way, any character
415 can be used in a name value, even quotes themselves.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]416
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]417 <li>Unnamed values are represented as an unsigned numeric value with their
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]418 prefix. For example, <tt>%12</tt>, <tt>@2</tt>, <tt>%44</tt>.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]419
Reid Spencercc16dc32004-12-09 18:02:53 +0000[diff] [blame]420 <li>Constants, which are described in a <a href="#constants">section about
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]421 constants</a>, below.</li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]422</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]423
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]424<p>LLVM requires that values start with a prefix for two reasons: Compilers
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]425 don't need to worry about name clashes with reserved words, and the set of
426 reserved words may be expanded in the future without penalty. Additionally,
427 unnamed identifiers allow a compiler to quickly come up with a temporary
428 variable without having to avoid symbol table conflicts.</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]429
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]430<p>Reserved words in LLVM are very similar to reserved words in other
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]431 languages. There are keywords for different opcodes
432 ('<tt><a href="#i_add">add</a></tt>',
433 '<tt><a href="#i_bitcast">bitcast</a></tt>',
434 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names
435 ('<tt><a href="#t_void">void</a></tt>',
436 '<tt><a href="#t_primitive">i32</a></tt>', etc...), and others. These
437 reserved words cannot conflict with variable names, because none of them
438 start with a prefix character (<tt>'%'</tt> or <tt>'@'</tt>).</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]439
440<p>Here is an example of LLVM code to multiply the integer variable
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]441 '<tt>%X</tt>' by 8:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]442
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]443<p>The easy way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]444
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]445<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]446%result = <a href="#i_mul">mul</a> i32 %X, 8
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]447</pre>
448
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]449<p>After strength reduction:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]450
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]451<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]452%result = <a href="#i_shl">shl</a> i32 %X, i8 3
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]453</pre>
454
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]455<p>And the hard way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]456
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]457<pre class="doc_code">
Gabor Greifec58f752009-10-28 13:05:07 +0000[diff] [blame]458%0 = <a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
459%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]460%result = <a href="#i_add">add</a> i32 %1, %1
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]461</pre>
462
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]463<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several important
464 lexical features of LLVM:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]465
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]466<ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]467 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]468 line.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]469
470 <li>Unnamed temporaries are created when the result of a computation is not
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]471 assigned to a named value.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]472
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]473 <li>Unnamed temporaries are numbered sequentially</li>
474</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]475
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]476<p>It also shows a convention that we follow in this document. When
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]477 demonstrating instructions, we will follow an instruction with a comment that
478 defines the type and name of value produced. Comments are shown in italic
479 text.</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]480
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]481</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]482
483<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]484<h2><a name="highlevel">High Level Structure</a></h2>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]485<!-- *********************************************************************** -->
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]486<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]487<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]488<h3>
489 <a name="modulestructure">Module Structure</a>
490</h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]491
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]492<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]493
Bill Wendling4cc2be62012-03-14 08:07:43 +0000[diff] [blame]494<p>LLVM programs are composed of <tt>Module</tt>s, each of which is a
495 translation unit of the input programs. Each module consists of functions,
496 global variables, and symbol table entries. Modules may be combined together
497 with the LLVM linker, which merges function (and global variable)
498 definitions, resolves forward declarations, and merges symbol table
499 entries. Here is an example of the "hello world" module:</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]500
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]501<pre class="doc_code">
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]502<i>; Declare the string constant as a global constant.</i>&nbsp;
Bill Wendling4cc2be62012-03-14 08:07:43 +0000[diff] [blame]503<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]504
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]505<i>; External declaration of the puts function</i>&nbsp;
Bill Wendling4cc2be62012-03-14 08:07:43 +0000[diff] [blame]506<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]507
508<i>; Definition of main function</i>
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]509define i32 @main() { <i>; i32()* </i>&nbsp;
510 <i>; Convert [13 x i8]* to i8 *...</i>&nbsp;
Bill Wendling4cc2be62012-03-14 08:07:43 +0000[diff] [blame]511 %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]512
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]513 <i>; Call puts function to write out the string to stdout.</i>&nbsp;
Bill Wendling4cc2be62012-03-14 08:07:43 +0000[diff] [blame]514 <a href="#i_call">call</a> i32 @puts(i8* %cast210)
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]515 <a href="#i_ret">ret</a> i32 0&nbsp;
516}
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]517
518<i>; Named metadata</i>
Bill Wendling4cc2be62012-03-14 08:07:43 +0000[diff] [blame]519!1 = metadata !{i32 42}
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]520!foo = !{!1, null}
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]521</pre>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]522
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]523<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]524 "<tt>.str</tt>", an external declaration of the "<tt>puts</tt>" function,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]525 a <a href="#functionstructure">function definition</a> for
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]526 "<tt>main</tt>" and <a href="#namedmetadatastructure">named metadata</a>
Bill Wendling4cc2be62012-03-14 08:07:43 +0000[diff] [blame]527 "<tt>foo</tt>".</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]528
Bill Wendling4cc2be62012-03-14 08:07:43 +0000[diff] [blame]529<p>In general, a module is made up of a list of global values (where both
530 functions and global variables are global values). Global values are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]531 represented by a pointer to a memory location (in this case, a pointer to an
532 array of char, and a pointer to a function), and have one of the
533 following <a href="#linkage">linkage types</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]534
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]535</div>
536
537<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]538<h3>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]539 <a name="linkage">Linkage Types</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]540</h3>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]541
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]542<div>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]543
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]544<p>All Global Variables and Functions have one of the following types of
545 linkage:</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]546
547<dl>
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]548 <dt><tt><b><a name="linkage_private">private</a></b></tt></dt>
Bill Wendling5e721d72010-07-01 21:55:59 +0000[diff] [blame]549 <dd>Global values with "<tt>private</tt>" linkage are only directly accessible
550 by objects in the current module. In particular, linking code into a
551 module with an private global value may cause the private to be renamed as
552 necessary to avoid collisions. Because the symbol is private to the
553 module, all references can be updated. This doesn't show up in any symbol
554 table in the object file.</dd>
Rafael Espindolabb46f522009-01-15 20:18:42 +0000[diff] [blame]555
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]556 <dt><tt><b><a name="linkage_linker_private">linker_private</a></b></tt></dt>
Bill Wendling5e721d72010-07-01 21:55:59 +0000[diff] [blame]557 <dd>Similar to <tt>private</tt>, but the symbol is passed through the
558 assembler and evaluated by the linker. Unlike normal strong symbols, they
559 are removed by the linker from the final linked image (executable or
560 dynamic library).</dd>
561
562 <dt><tt><b><a name="linkage_linker_private_weak">linker_private_weak</a></b></tt></dt>
563 <dd>Similar to "<tt>linker_private</tt>", but the symbol is weak. Note that
564 <tt>linker_private_weak</tt> symbols are subject to coalescing by the
565 linker. The symbols are removed by the linker from the final linked image
566 (executable or dynamic library).</dd>
Bill Wendling3d10a5a2009-07-20 01:03:30 +0000[diff] [blame]567
Bill Wendling55ae5152010-08-20 22:05:50 +0000[diff] [blame]568 <dt><tt><b><a name="linkage_linker_private_weak_def_auto">linker_private_weak_def_auto</a></b></tt></dt>
569 <dd>Similar to "<tt>linker_private_weak</tt>", but it's known that the address
570 of the object is not taken. For instance, functions that had an inline
571 definition, but the compiler decided not to inline it. Note,
572 unlike <tt>linker_private</tt> and <tt>linker_private_weak</tt>,
573 <tt>linker_private_weak_def_auto</tt> may have only <tt>default</tt>
574 visibility. The symbols are removed by the linker from the final linked
575 image (executable or dynamic library).</dd>
576
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]577 <dt><tt><b><a name="linkage_internal">internal</a></b></tt></dt>
Bill Wendling07d31772010-06-29 22:34:52 +0000[diff] [blame]578 <dd>Similar to private, but the value shows as a local symbol
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]579 (<tt>STB_LOCAL</tt> in the case of ELF) in the object file. This
580 corresponds to the notion of the '<tt>static</tt>' keyword in C.</dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]581
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]582 <dt><tt><b><a name="linkage_available_externally">available_externally</a></b></tt></dt>
Chris Lattner266c7bb2009-04-13 05:44:34 +0000[diff] [blame]583 <dd>Globals with "<tt>available_externally</tt>" linkage are never emitted
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]584 into the object file corresponding to the LLVM module. They exist to
585 allow inlining and other optimizations to take place given knowledge of
586 the definition of the global, which is known to be somewhere outside the
587 module. Globals with <tt>available_externally</tt> linkage are allowed to
588 be discarded at will, and are otherwise the same as <tt>linkonce_odr</tt>.
589 This linkage type is only allowed on definitions, not declarations.</dd>
Chris Lattner266c7bb2009-04-13 05:44:34 +0000[diff] [blame]590
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]591 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt></dt>
Chris Lattner4887bd82007-01-14 06:51:48 +0000[diff] [blame]592 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
Chris Lattner873187c2010-01-09 19:15:14 +0000[diff] [blame]593 the same name when linkage occurs. This can be used to implement
594 some forms of inline functions, templates, or other code which must be
595 generated in each translation unit that uses it, but where the body may
596 be overridden with a more definitive definition later. Unreferenced
597 <tt>linkonce</tt> globals are allowed to be discarded. Note that
598 <tt>linkonce</tt> linkage does not actually allow the optimizer to
599 inline the body of this function into callers because it doesn't know if
600 this definition of the function is the definitive definition within the
601 program or whether it will be overridden by a stronger definition.
602 To enable inlining and other optimizations, use "<tt>linkonce_odr</tt>"
603 linkage.</dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]604
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]605 <dt><tt><b><a name="linkage_weak">weak</a></b></tt></dt>
Chris Lattner26d054d2009-08-05 05:21:07 +0000[diff] [blame]606 <dd>"<tt>weak</tt>" linkage has the same merging semantics as
607 <tt>linkonce</tt> linkage, except that unreferenced globals with
608 <tt>weak</tt> linkage may not be discarded. This is used for globals that
609 are declared "weak" in C source code.</dd>
610
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]611 <dt><tt><b><a name="linkage_common">common</a></b></tt></dt>
Chris Lattner26d054d2009-08-05 05:21:07 +0000[diff] [blame]612 <dd>"<tt>common</tt>" linkage is most similar to "<tt>weak</tt>" linkage, but
613 they are used for tentative definitions in C, such as "<tt>int X;</tt>" at
614 global scope.
615 Symbols with "<tt>common</tt>" linkage are merged in the same way as
616 <tt>weak symbols</tt>, and they may not be deleted if unreferenced.
Chris Lattnercd81f5d2009-08-05 05:41:44 +0000[diff] [blame]617 <tt>common</tt> symbols may not have an explicit section,
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]618 must have a zero initializer, and may not be marked '<a
Chris Lattnercd81f5d2009-08-05 05:41:44 +0000[diff] [blame]619 href="#globalvars"><tt>constant</tt></a>'. Functions and aliases may not
620 have common linkage.</dd>
Chris Lattner26d054d2009-08-05 05:21:07 +0000[diff] [blame]621
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]622
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]623 <dt><tt><b><a name="linkage_appending">appending</a></b></tt></dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]624 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]625 pointer to array type. When two global variables with appending linkage
626 are linked together, the two global arrays are appended together. This is
627 the LLVM, typesafe, equivalent of having the system linker append together
628 "sections" with identical names when .o files are linked.</dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]629
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]630 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]631 <dd>The semantics of this linkage follow the ELF object file model: the symbol
632 is weak until linked, if not linked, the symbol becomes null instead of
633 being an undefined reference.</dd>
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000[diff] [blame]634
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]635 <dt><tt><b><a name="linkage_linkonce_odr">linkonce_odr</a></b></tt></dt>
636 <dt><tt><b><a name="linkage_weak_odr">weak_odr</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]637 <dd>Some languages allow differing globals to be merged, such as two functions
638 with different semantics. Other languages, such as <tt>C++</tt>, ensure
Bill Wendling5e721d72010-07-01 21:55:59 +0000[diff] [blame]639 that only equivalent globals are ever merged (the "one definition rule"
640 &mdash; "ODR"). Such languages can use the <tt>linkonce_odr</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]641 and <tt>weak_odr</tt> linkage types to indicate that the global will only
642 be merged with equivalent globals. These linkage types are otherwise the
643 same as their non-<tt>odr</tt> versions.</dd>
Duncan Sands667d4b82009-03-07 15:45:40 +0000[diff] [blame]644
Bill Wendling5c3a9f72011-11-04 20:40:41 +0000[diff] [blame]645 <dt><tt><b><a name="linkage_external">external</a></b></tt></dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]646 <dd>If none of the above identifiers are used, the global is externally
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]647 visible, meaning that it participates in linkage and can be used to
648 resolve external symbol references.</dd>
Reid Spencerc8910842007-04-11 23:49:50 +0000[diff] [blame]649</dl>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]650
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]651<p>The next two types of linkage are targeted for Microsoft Windows platform
652 only. They are designed to support importing (exporting) symbols from (to)
653 DLLs (Dynamic Link Libraries).</p>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]654
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]655<dl>
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]656 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt></dt>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]657 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]658 or variable via a global pointer to a pointer that is set up by the DLL
659 exporting the symbol. On Microsoft Windows targets, the pointer name is
660 formed by combining <code>__imp_</code> and the function or variable
661 name.</dd>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]662
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]663 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt></dt>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]664 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]665 pointer to a pointer in a DLL, so that it can be referenced with the
666 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
667 name is formed by combining <code>__imp_</code> and the function or
668 variable name.</dd>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]669</dl>
670
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]671<p>For example, since the "<tt>.LC0</tt>" variable is defined to be internal, if
672 another module defined a "<tt>.LC0</tt>" variable and was linked with this
673 one, one of the two would be renamed, preventing a collision. Since
674 "<tt>main</tt>" and "<tt>puts</tt>" are external (i.e., lacking any linkage
675 declarations), they are accessible outside of the current module.</p>
676
677<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]678 other than <tt>external</tt>, <tt>dllimport</tt>
679 or <tt>extern_weak</tt>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]680
Duncan Sands667d4b82009-03-07 15:45:40 +0000[diff] [blame]681<p>Aliases can have only <tt>external</tt>, <tt>internal</tt>, <tt>weak</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]682 or <tt>weak_odr</tt> linkages.</p>
683
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]684</div>
685
686<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]687<h3>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]688 <a name="callingconv">Calling Conventions</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]689</h3>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]690
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]691<div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]692
693<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]694 and <a href="#i_invoke">invokes</a> can all have an optional calling
695 convention specified for the call. The calling convention of any pair of
696 dynamic caller/callee must match, or the behavior of the program is
697 undefined. The following calling conventions are supported by LLVM, and more
698 may be added in the future:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]699
700<dl>
701 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]702 <dd>This calling convention (the default if no other calling convention is
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]703 specified) matches the target C calling conventions. This calling
704 convention supports varargs function calls and tolerates some mismatch in
705 the declared prototype and implemented declaration of the function (as
706 does normal C).</dd>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]707
708 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]709 <dd>This calling convention attempts to make calls as fast as possible
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]710 (e.g. by passing things in registers). This calling convention allows the
711 target to use whatever tricks it wants to produce fast code for the
712 target, without having to conform to an externally specified ABI
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]713 (Application Binary Interface).
714 <a href="CodeGenerator.html#tailcallopt">Tail calls can only be optimized
Chris Lattner29689432010-03-11 00:22:57 +0000[diff] [blame]715 when this or the GHC convention is used.</a> This calling convention
716 does not support varargs and requires the prototype of all callees to
717 exactly match the prototype of the function definition.</dd>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]718
719 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]720 <dd>This calling convention attempts to make code in the caller as efficient
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]721 as possible under the assumption that the call is not commonly executed.
722 As such, these calls often preserve all registers so that the call does
723 not break any live ranges in the caller side. This calling convention
724 does not support varargs and requires the prototype of all callees to
725 exactly match the prototype of the function definition.</dd>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]726
Chris Lattner29689432010-03-11 00:22:57 +0000[diff] [blame]727 <dt><b>"<tt>cc <em>10</em></tt>" - GHC convention</b>:</dt>
728 <dd>This calling convention has been implemented specifically for use by the
729 <a href="http://www.haskell.org/ghc">Glasgow Haskell Compiler (GHC)</a>.
730 It passes everything in registers, going to extremes to achieve this by
731 disabling callee save registers. This calling convention should not be
732 used lightly but only for specific situations such as an alternative to
733 the <em>register pinning</em> performance technique often used when
734 implementing functional programming languages.At the moment only X86
735 supports this convention and it has the following limitations:
736 <ul>
737 <li>On <em>X86-32</em> only supports up to 4 bit type parameters. No
738 floating point types are supported.</li>
739 <li>On <em>X86-64</em> only supports up to 10 bit type parameters and
740 6 floating point parameters.</li>
741 </ul>
742 This calling convention supports
743 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> but
744 requires both the caller and callee are using it.
745 </dd>
746
Chris Lattnercfe6b372005-05-07 01:46:40 +0000[diff] [blame]747 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]748 <dd>Any calling convention may be specified by number, allowing
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]749 target-specific calling conventions to be used. Target specific calling
750 conventions start at 64.</dd>
Chris Lattnercfe6b372005-05-07 01:46:40 +0000[diff] [blame]751</dl>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]752
753<p>More calling conventions can be added/defined on an as-needed basis, to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]754 support Pascal conventions or any other well-known target-independent
755 convention.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]756
757</div>
758
759<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]760<h3>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]761 <a name="visibility">Visibility Styles</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]762</h3>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]763
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]764<div>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]765
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]766<p>All Global Variables and Functions have one of the following visibility
767 styles:</p>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]768
769<dl>
770 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]771 <dd>On targets that use the ELF object file format, default visibility means
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]772 that the declaration is visible to other modules and, in shared libraries,
773 means that the declared entity may be overridden. On Darwin, default
774 visibility means that the declaration is visible to other modules. Default
775 visibility corresponds to "external linkage" in the language.</dd>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]776
777 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]778 <dd>Two declarations of an object with hidden visibility refer to the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]779 object if they are in the same shared object. Usually, hidden visibility
780 indicates that the symbol will not be placed into the dynamic symbol
781 table, so no other module (executable or shared library) can reference it
782 directly.</dd>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]783
Anton Korobeynikov6f9896f2007-04-29 18:35:00 +0000[diff] [blame]784 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
Anton Korobeynikov6f9896f2007-04-29 18:35:00 +0000[diff] [blame]785 <dd>On ELF, protected visibility indicates that the symbol will be placed in
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]786 the dynamic symbol table, but that references within the defining module
787 will bind to the local symbol. That is, the symbol cannot be overridden by
788 another module.</dd>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]789</dl>
790
791</div>
792
793<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]794<h3>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]795 <a name="namedtypes">Named Types</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]796</h3>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]797
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]798<div>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]799
800<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]801 it easier to read the IR and make the IR more condensed (particularly when
802 recursive types are involved). An example of a name specification is:</p>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]803
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]804<pre class="doc_code">
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]805%mytype = type { %mytype*, i32 }
806</pre>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]807
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]808<p>You may give a name to any <a href="#typesystem">type</a> except
Chris Lattnerdc65f222010-08-17 23:26:04 +0000[diff] [blame]809 "<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]810 is expected with the syntax "%mytype".</p>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]811
812<p>Note that type names are aliases for the structural type that they indicate,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]813 and that you can therefore specify multiple names for the same type. This
814 often leads to confusing behavior when dumping out a .ll file. Since LLVM IR
815 uses structural typing, the name is not part of the type. When printing out
816 LLVM IR, the printer will pick <em>one name</em> to render all types of a
817 particular shape. This means that if you have code where two different
818 source types end up having the same LLVM type, that the dumper will sometimes
819 print the "wrong" or unexpected type. This is an important design point and
820 isn't going to change.</p>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]821
822</div>
823
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]824<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]825<h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]826 <a name="globalvars">Global Variables</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]827</h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]828
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]829<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]830
Chris Lattner3689a342005-02-12 19:30:21 +0000[diff] [blame]831<p>Global variables define regions of memory allocated at compilation time
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]832 instead of run-time. Global variables may optionally be initialized, may
833 have an explicit section to be placed in, and may have an optional explicit
834 alignment specified. A variable may be defined as "thread_local", which
835 means that it will not be shared by threads (each thread will have a
836 separated copy of the variable). A variable may be defined as a global
837 "constant," which indicates that the contents of the variable
838 will <b>never</b> be modified (enabling better optimization, allowing the
839 global data to be placed in the read-only section of an executable, etc).
840 Note that variables that need runtime initialization cannot be marked
841 "constant" as there is a store to the variable.</p>
Chris Lattner3689a342005-02-12 19:30:21 +0000[diff] [blame]842
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]843<p>LLVM explicitly allows <em>declarations</em> of global variables to be marked
844 constant, even if the final definition of the global is not. This capability
845 can be used to enable slightly better optimization of the program, but
846 requires the language definition to guarantee that optimizations based on the
847 'constantness' are valid for the translation units that do not include the
848 definition.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]849
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]850<p>As SSA values, global variables define pointer values that are in scope
851 (i.e. they dominate) all basic blocks in the program. Global variables
852 always define a pointer to their "content" type because they describe a
853 region of memory, and all memory objects in LLVM are accessed through
854 pointers.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]855
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]856<p>Global variables can be marked with <tt>unnamed_addr</tt> which indicates
857 that the address is not significant, only the content. Constants marked
Rafael Espindolaa5eaa862011-01-15 08:20:57 +0000[diff] [blame]858 like this can be merged with other constants if they have the same
859 initializer. Note that a constant with significant address <em>can</em>
860 be merged with a <tt>unnamed_addr</tt> constant, the result being a
861 constant whose address is significant.</p>
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]862
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]863<p>A global variable may be declared to reside in a target-specific numbered
864 address space. For targets that support them, address spaces may affect how
865 optimizations are performed and/or what target instructions are used to
866 access the variable. The default address space is zero. The address space
867 qualifier must precede any other attributes.</p>
Christopher Lamb284d9922007-12-11 09:31:00 +0000[diff] [blame]868
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]869<p>LLVM allows an explicit section to be specified for globals. If the target
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]870 supports it, it will emit globals to the section specified.</p>
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]871
Chris Lattnerce99fa92010-04-28 00:13:42 +0000[diff] [blame]872<p>An explicit alignment may be specified for a global, which must be a power
873 of 2. If not present, or if the alignment is set to zero, the alignment of
874 the global is set by the target to whatever it feels convenient. If an
875 explicit alignment is specified, the global is forced to have exactly that
Chris Lattner2d4b8ee2010-04-28 00:31:12 +0000[diff] [blame]876 alignment. Targets and optimizers are not allowed to over-align the global
877 if the global has an assigned section. In this case, the extra alignment
878 could be observable: for example, code could assume that the globals are
879 densely packed in their section and try to iterate over them as an array,
880 alignment padding would break this iteration.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]881
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]882<p>For example, the following defines a global in a numbered address space with
883 an initializer, section, and alignment:</p>
Chris Lattner68027ea2007-01-14 00:27:09 +0000[diff] [blame]884
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]885<pre class="doc_code">
Dan Gohman398873c2009-01-11 00:40:00 +0000[diff] [blame]886@G = addrspace(5) constant float 1.0, section "foo", align 4
Chris Lattner68027ea2007-01-14 00:27:09 +0000[diff] [blame]887</pre>
888
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]889</div>
890
891
892<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]893<h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]894 <a name="functionstructure">Functions</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]895</h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]896
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]897<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]898
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]899<p>LLVM function definitions consist of the "<tt>define</tt>" keyword, an
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]900 optional <a href="#linkage">linkage type</a>, an optional
901 <a href="#visibility">visibility style</a>, an optional
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]902 <a href="#callingconv">calling convention</a>,
903 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]904 <a href="#paramattrs">parameter attribute</a> for the return type, a function
905 name, a (possibly empty) argument list (each with optional
906 <a href="#paramattrs">parameter attributes</a>), optional
907 <a href="#fnattrs">function attributes</a>, an optional section, an optional
908 alignment, an optional <a href="#gc">garbage collector name</a>, an opening
909 curly brace, a list of basic blocks, and a closing curly brace.</p>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]910
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]911<p>LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
912 optional <a href="#linkage">linkage type</a>, an optional
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]913 <a href="#visibility">visibility style</a>, an optional
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]914 <a href="#callingconv">calling convention</a>,
915 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]916 <a href="#paramattrs">parameter attribute</a> for the return type, a function
917 name, a possibly empty list of arguments, an optional alignment, and an
918 optional <a href="#gc">garbage collector name</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]919
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]920<p>A function definition contains a list of basic blocks, forming the CFG
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]921 (Control Flow Graph) for the function. Each basic block may optionally start
922 with a label (giving the basic block a symbol table entry), contains a list
923 of instructions, and ends with a <a href="#terminators">terminator</a>
924 instruction (such as a branch or function return).</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]925
Chris Lattner4a3c9012007-06-08 16:52:14 +0000[diff] [blame]926<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]927 executed on entrance to the function, and it is not allowed to have
928 predecessor basic blocks (i.e. there can not be any branches to the entry
929 block of a function). Because the block can have no predecessors, it also
930 cannot have any <a href="#i_phi">PHI nodes</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]931
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]932<p>LLVM allows an explicit section to be specified for functions. If the target
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]933 supports it, it will emit functions to the section specified.</p>
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]934
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]935<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]936 the alignment is set to zero, the alignment of the function is set by the
937 target to whatever it feels convenient. If an explicit alignment is
938 specified, the function is forced to have at least that much alignment. All
939 alignments must be a power of 2.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]940
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]941<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]942 be significant and two identical functions can be merged.</p>
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]943
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]944<h5>Syntax:</h5>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]945<pre class="doc_code">
Chris Lattner50ad45c2008-10-13 16:55:18 +0000[diff] [blame]946define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]947 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
948 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
949 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
950 [<a href="#gc">gc</a>] { ... }
951</pre>
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]952
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]953</div>
954
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]955<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]956<h3>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]957 <a name="aliasstructure">Aliases</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]958</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]959
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]960<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]961
962<p>Aliases act as "second name" for the aliasee value (which can be either
963 function, global variable, another alias or bitcast of global value). Aliases
964 may have an optional <a href="#linkage">linkage type</a>, and an
965 optional <a href="#visibility">visibility style</a>.</p>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]966
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]967<h5>Syntax:</h5>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]968<pre class="doc_code">
Duncan Sands0b23ac12008-09-12 20:48:21 +0000[diff] [blame]969@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Bill Wendlingaac388b2007-05-29 09:42:13 +0000[diff] [blame]970</pre>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]971
972</div>
973
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]974<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]975<h3>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]976 <a name="namedmetadatastructure">Named Metadata</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]977</h3>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]978
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]979<div>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]980
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]981<p>Named metadata is a collection of metadata. <a href="#metadata">Metadata
Dan Gohman872814a2010-07-21 18:54:18 +0000[diff] [blame]982 nodes</a> (but not metadata strings) are the only valid operands for
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]983 a named metadata.</p>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]984
985<h5>Syntax:</h5>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]986<pre class="doc_code">
Dan Gohman872814a2010-07-21 18:54:18 +0000[diff] [blame]987; Some unnamed metadata nodes, which are referenced by the named metadata.
988!0 = metadata !{metadata !"zero"}
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]989!1 = metadata !{metadata !"one"}
Dan Gohman872814a2010-07-21 18:54:18 +0000[diff] [blame]990!2 = metadata !{metadata !"two"}
Dan Gohman1005bc52010-07-13 19:48:13 +0000[diff] [blame]991; A named metadata.
Dan Gohman872814a2010-07-21 18:54:18 +0000[diff] [blame]992!name = !{!0, !1, !2}
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]993</pre>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]994
995</div>
996
997<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]998<h3>
999 <a name="paramattrs">Parameter Attributes</a>
1000</h3>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1001
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1002<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1003
1004<p>The return type and each parameter of a function type may have a set of
1005 <i>parameter attributes</i> associated with them. Parameter attributes are
1006 used to communicate additional information about the result or parameters of
1007 a function. Parameter attributes are considered to be part of the function,
1008 not of the function type, so functions with different parameter attributes
1009 can have the same function type.</p>
1010
1011<p>Parameter attributes are simple keywords that follow the type specified. If
1012 multiple parameter attributes are needed, they are space separated. For
1013 example:</p>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1014
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1015<pre class="doc_code">
Nick Lewyckyb6a7d252009-02-15 23:06:14 +0000[diff] [blame]1016declare i32 @printf(i8* noalias nocapture, ...)
Chris Lattner66d922c2008-10-04 18:33:34 +0000[diff] [blame]1017declare i32 @atoi(i8 zeroext)
1018declare signext i8 @returns_signed_char()
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1019</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1020
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1021<p>Note that any attributes for the function result (<tt>nounwind</tt>,
1022 <tt>readonly</tt>) come immediately after the argument list.</p>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1023
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1024<p>Currently, only the following parameter attributes are defined:</p>
Chris Lattner47507de2008-01-11 06:20:47 +0000[diff] [blame]1025
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1026<dl>
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1027 <dt><tt><b>zeroext</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1028 <dd>This indicates to the code generator that the parameter or return value
Cameron Zwarichebe81732011-03-16 22:20:18 +0000[diff] [blame]1029 should be zero-extended to the extent required by the target's ABI (which
1030 is usually 32-bits, but is 8-bits for a i1 on x86-64) by the caller (for a
1031 parameter) or the callee (for a return value).</dd>
Chris Lattner47507de2008-01-11 06:20:47 +0000[diff] [blame]1032
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1033 <dt><tt><b>signext</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1034 <dd>This indicates to the code generator that the parameter or return value
Cameron Zwarich9e69ff92011-03-17 14:21:58 +0000[diff] [blame]1035 should be sign-extended to the extent required by the target's ABI (which
1036 is usually 32-bits) by the caller (for a parameter) or the callee (for a
1037 return value).</dd>
Chris Lattner47507de2008-01-11 06:20:47 +0000[diff] [blame]1038
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1039 <dt><tt><b>inreg</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1040 <dd>This indicates that this parameter or return value should be treated in a
1041 special target-dependent fashion during while emitting code for a function
1042 call or return (usually, by putting it in a register as opposed to memory,
1043 though some targets use it to distinguish between two different kinds of
1044 registers). Use of this attribute is target-specific.</dd>
Chris Lattner47507de2008-01-11 06:20:47 +0000[diff] [blame]1045
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1046 <dt><tt><b><a name="byval">byval</a></b></tt></dt>
Chris Lattnera6fd81d2010-11-20 23:49:06 +0000[diff] [blame]1047 <dd><p>This indicates that the pointer parameter should really be passed by
1048 value to the function. The attribute implies that a hidden copy of the
1049 pointee
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1050 is made between the caller and the callee, so the callee is unable to
1051 modify the value in the callee. This attribute is only valid on LLVM
1052 pointer arguments. It is generally used to pass structs and arrays by
1053 value, but is also valid on pointers to scalars. The copy is considered
1054 to belong to the caller not the callee (for example,
1055 <tt><a href="#readonly">readonly</a></tt> functions should not write to
1056 <tt>byval</tt> parameters). This is not a valid attribute for return
Chris Lattnera6fd81d2010-11-20 23:49:06 +0000[diff] [blame]1057 values.</p>
1058
1059 <p>The byval attribute also supports specifying an alignment with
1060 the align attribute. It indicates the alignment of the stack slot to
1061 form and the known alignment of the pointer specified to the call site. If
1062 the alignment is not specified, then the code generator makes a
1063 target-specific assumption.</p></dd>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1064
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]1065 <dt><tt><b><a name="sret">sret</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1066 <dd>This indicates that the pointer parameter specifies the address of a
1067 structure that is the return value of the function in the source program.
1068 This pointer must be guaranteed by the caller to be valid: loads and
1069 stores to the structure may be assumed by the callee to not to trap. This
1070 may only be applied to the first parameter. This is not a valid attribute
1071 for return values. </dd>
1072
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]1073 <dt><tt><b><a name="noalias">noalias</a></b></tt></dt>
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1074 <dd>This indicates that pointer values
1075 <a href="#pointeraliasing"><i>based</i></a> on the argument or return
Dan Gohmanefca7f92010-07-02 23:46:54 +0000[diff] [blame]1076 value do not alias pointer values which are not <i>based</i> on it,
1077 ignoring certain "irrelevant" dependencies.
1078 For a call to the parent function, dependencies between memory
1079 references from before or after the call and from those during the call
1080 are "irrelevant" to the <tt>noalias</tt> keyword for the arguments and
1081 return value used in that call.
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1082 The caller shares the responsibility with the callee for ensuring that
1083 these requirements are met.
1084 For further details, please see the discussion of the NoAlias response in
Dan Gohmanff70fe42010-07-06 15:26:33 +0000[diff] [blame]1085 <a href="AliasAnalysis.html#MustMayNo">alias analysis</a>.<br>
1086<br>
John McCall191d4ee2010-07-06 21:07:14 +0000[diff] [blame]1087 Note that this definition of <tt>noalias</tt> is intentionally
1088 similar to the definition of <tt>restrict</tt> in C99 for function
Chris Lattner211244a2010-07-06 20:51:35 +0000[diff] [blame]1089 arguments, though it is slightly weaker.
Dan Gohmanff70fe42010-07-06 15:26:33 +0000[diff] [blame]1090<br>
1091 For function return values, C99's <tt>restrict</tt> is not meaningful,
1092 while LLVM's <tt>noalias</tt> is.
1093 </dd>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1094
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]1095 <dt><tt><b><a name="nocapture">nocapture</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1096 <dd>This indicates that the callee does not make any copies of the pointer
1097 that outlive the callee itself. This is not a valid attribute for return
1098 values.</dd>
1099
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]1100 <dt><tt><b><a name="nest">nest</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1101 <dd>This indicates that the pointer parameter can be excised using the
1102 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
1103 attribute for return values.</dd>
1104</dl>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1105
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1106</div>
1107
1108<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1109<h3>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1110 <a name="gc">Garbage Collector Names</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1111</h3>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1112
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1113<div>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1114
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1115<p>Each function may specify a garbage collector name, which is simply a
1116 string:</p>
1117
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1118<pre class="doc_code">
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1119define void @f() gc "name" { ... }
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1120</pre>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1121
1122<p>The compiler declares the supported values of <i>name</i>. Specifying a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1123 collector which will cause the compiler to alter its output in order to
1124 support the named garbage collection algorithm.</p>
1125
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1126</div>
1127
1128<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1129<h3>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1130 <a name="fnattrs">Function Attributes</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1131</h3>
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1132
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1133<div>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1134
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1135<p>Function attributes are set to communicate additional information about a
1136 function. Function attributes are considered to be part of the function, not
1137 of the function type, so functions with different parameter attributes can
1138 have the same function type.</p>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1139
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1140<p>Function attributes are simple keywords that follow the type specified. If
1141 multiple attributes are needed, they are space separated. For example:</p>
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1142
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1143<pre class="doc_code">
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1144define void @f() noinline { ... }
1145define void @f() alwaysinline { ... }
1146define void @f() alwaysinline optsize { ... }
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1147define void @f() optsize { ... }
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1148</pre>
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1149
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1150<dl>
Kostya Serebryany164b86b2012-01-20 17:56:17 +0000[diff] [blame]1151 <dt><tt><b>address_safety</b></tt></dt>
1152 <dd>This attribute indicates that the address safety analysis
1153 is enabled for this function. </dd>
1154
Charles Davis1e063d12010-02-12 00:31:15 +0000[diff] [blame]1155 <dt><tt><b>alignstack(&lt;<em>n</em>&gt;)</b></tt></dt>
1156 <dd>This attribute indicates that, when emitting the prologue and epilogue,
1157 the backend should forcibly align the stack pointer. Specify the
1158 desired alignment, which must be a power of two, in parentheses.
1159
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1160 <dt><tt><b>alwaysinline</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1161 <dd>This attribute indicates that the inliner should attempt to inline this
1162 function into callers whenever possible, ignoring any active inlining size
1163 threshold for this caller.</dd>
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1164
Dan Gohman129bd562011-06-16 16:03:13 +0000[diff] [blame]1165 <dt><tt><b>nonlazybind</b></tt></dt>
1166 <dd>This attribute suppresses lazy symbol binding for the function. This
1167 may make calls to the function faster, at the cost of extra program
1168 startup time if the function is not called during program startup.</dd>
1169
Jakob Stoklund Olesen570a4a52010-02-06 01:16:28 +0000[diff] [blame]1170 <dt><tt><b>inlinehint</b></tt></dt>
1171 <dd>This attribute indicates that the source code contained a hint that inlining
1172 this function is desirable (such as the "inline" keyword in C/C++). It
1173 is just a hint; it imposes no requirements on the inliner.</dd>
1174
Nick Lewycky76ec37a2010-07-06 18:24:09 +0000[diff] [blame]1175 <dt><tt><b>naked</b></tt></dt>
1176 <dd>This attribute disables prologue / epilogue emission for the function.
1177 This can have very system-specific consequences.</dd>
1178
1179 <dt><tt><b>noimplicitfloat</b></tt></dt>
1180 <dd>This attributes disables implicit floating point instructions.</dd>
1181
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1182 <dt><tt><b>noinline</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1183 <dd>This attribute indicates that the inliner should never inline this
1184 function in any situation. This attribute may not be used together with
1185 the <tt>alwaysinline</tt> attribute.</dd>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1186
Nick Lewycky76ec37a2010-07-06 18:24:09 +0000[diff] [blame]1187 <dt><tt><b>noredzone</b></tt></dt>
1188 <dd>This attribute indicates that the code generator should not use a red
1189 zone, even if the target-specific ABI normally permits it.</dd>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1190
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1191 <dt><tt><b>noreturn</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1192 <dd>This function attribute indicates that the function never returns
1193 normally. This produces undefined behavior at runtime if the function
1194 ever does dynamically return.</dd>
Bill Wendling31359ba2008-11-13 01:02:51 +0000[diff] [blame]1195
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1196 <dt><tt><b>nounwind</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1197 <dd>This function attribute indicates that the function never returns with an
1198 unwind or exceptional control flow. If the function does unwind, its
1199 runtime behavior is undefined.</dd>
Bill Wendlingfbaa7ed2008-11-26 19:07:40 +0000[diff] [blame]1200
Nick Lewycky76ec37a2010-07-06 18:24:09 +0000[diff] [blame]1201 <dt><tt><b>optsize</b></tt></dt>
1202 <dd>This attribute suggests that optimization passes and code generator passes
1203 make choices that keep the code size of this function low, and otherwise
1204 do optimizations specifically to reduce code size.</dd>
1205
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1206 <dt><tt><b>readnone</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1207 <dd>This attribute indicates that the function computes its result (or decides
1208 to unwind an exception) based strictly on its arguments, without
1209 dereferencing any pointer arguments or otherwise accessing any mutable
1210 state (e.g. memory, control registers, etc) visible to caller functions.
1211 It does not write through any pointer arguments
1212 (including <tt><a href="#byval">byval</a></tt> arguments) and never
1213 changes any state visible to callers. This means that it cannot unwind
Bill Wendling7b9e5392012-02-06 21:57:33 +0000[diff] [blame]1214 exceptions by calling the <tt>C++</tt> exception throwing methods.</dd>
Devang Patel5d96fda2009-06-12 19:45:19 +0000[diff] [blame]1215
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1216 <dt><tt><b><a name="readonly">readonly</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1217 <dd>This attribute indicates that the function does not write through any
1218 pointer arguments (including <tt><a href="#byval">byval</a></tt>
1219 arguments) or otherwise modify any state (e.g. memory, control registers,
1220 etc) visible to caller functions. It may dereference pointer arguments
1221 and read state that may be set in the caller. A readonly function always
1222 returns the same value (or unwinds an exception identically) when called
1223 with the same set of arguments and global state. It cannot unwind an
Bill Wendling7b9e5392012-02-06 21:57:33 +0000[diff] [blame]1224 exception by calling the <tt>C++</tt> exception throwing methods.</dd>
Anton Korobeynikovc5ec8a72009-07-17 18:07:26 +0000[diff] [blame]1225
Bill Wendling9bd5d042011-12-05 21:27:54 +0000[diff] [blame]1226 <dt><tt><b><a name="returns_twice">returns_twice</a></b></tt></dt>
1227 <dd>This attribute indicates that this function can return twice. The
1228 C <code>setjmp</code> is an example of such a function. The compiler
1229 disables some optimizations (like tail calls) in the caller of these
1230 functions.</dd>
1231
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1232 <dt><tt><b><a name="ssp">ssp</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1233 <dd>This attribute indicates that the function should emit a stack smashing
1234 protector. It is in the form of a "canary"&mdash;a random value placed on
1235 the stack before the local variables that's checked upon return from the
1236 function to see if it has been overwritten. A heuristic is used to
1237 determine if a function needs stack protectors or not.<br>
1238<br>
1239 If a function that has an <tt>ssp</tt> attribute is inlined into a
1240 function that doesn't have an <tt>ssp</tt> attribute, then the resulting
1241 function will have an <tt>ssp</tt> attribute.</dd>
1242
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1243 <dt><tt><b>sspreq</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1244 <dd>This attribute indicates that the function should <em>always</em> emit a
1245 stack smashing protector. This overrides
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1246 the <tt><a href="#ssp">ssp</a></tt> function attribute.<br>
1247<br>
1248 If a function that has an <tt>sspreq</tt> attribute is inlined into a
1249 function that doesn't have an <tt>sspreq</tt> attribute or which has
1250 an <tt>ssp</tt> attribute, then the resulting function will have
1251 an <tt>sspreq</tt> attribute.</dd>
Rafael Espindolafbff0ec2011-07-25 15:27:59 +0000[diff] [blame]1252
1253 <dt><tt><b><a name="uwtable">uwtable</a></b></tt></dt>
1254 <dd>This attribute indicates that the ABI being targeted requires that
1255 an unwind table entry be produce for this function even if we can
1256 show that no exceptions passes by it. This is normally the case for
1257 the ELF x86-64 abi, but it can be disabled for some compilation
1258 units.</dd>
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1259</dl>
1260
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1261</div>
1262
1263<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1264<h3>
Chris Lattner1eeeb0c2006-04-08 04:40:53 +0000[diff] [blame]1265 <a name="moduleasm">Module-Level Inline Assembly</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1266</h3>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1267
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1268<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1269
1270<p>Modules may contain "module-level inline asm" blocks, which corresponds to
1271 the GCC "file scope inline asm" blocks. These blocks are internally
1272 concatenated by LLVM and treated as a single unit, but may be separated in
1273 the <tt>.ll</tt> file if desired. The syntax is very simple:</p>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1274
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1275<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1276module asm "inline asm code goes here"
1277module asm "more can go here"
1278</pre>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1279
1280<p>The strings can contain any character by escaping non-printable characters.
1281 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1282 for the number.</p>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1283
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1284<p>The inline asm code is simply printed to the machine code .s file when
1285 assembly code is generated.</p>
1286
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1287</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1288
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1289<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1290<h3>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1291 <a name="datalayout">Data Layout</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1292</h3>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1293
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1294<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1295
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1296<p>A module may specify a target specific data layout string that specifies how
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1297 data is to be laid out in memory. The syntax for the data layout is
1298 simply:</p>
1299
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1300<pre class="doc_code">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1301target datalayout = "<i>layout specification</i>"
1302</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1303
1304<p>The <i>layout specification</i> consists of a list of specifications
1305 separated by the minus sign character ('-'). Each specification starts with
1306 a letter and may include other information after the letter to define some
1307 aspect of the data layout. The specifications accepted are as follows:</p>
1308
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1309<dl>
1310 <dt><tt>E</tt></dt>
1311 <dd>Specifies that the target lays out data in big-endian form. That is, the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1312 bits with the most significance have the lowest address location.</dd>
1313
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1314 <dt><tt>e</tt></dt>
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]1315 <dd>Specifies that the target lays out data in little-endian form. That is,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1316 the bits with the least significance have the lowest address
1317 location.</dd>
1318
Lang Hamesbb5b3f32011-10-10 23:42:08 +0000[diff] [blame]1319 <dt><tt>S<i>size</i></tt></dt>
1320 <dd>Specifies the natural alignment of the stack in bits. Alignment promotion
1321 of stack variables is limited to the natural stack alignment to avoid
1322 dynamic stack realignment. The stack alignment must be a multiple of
Lang Hames5f119a62011-10-11 17:50:14 +0000[diff] [blame]1323 8-bits. If omitted, the natural stack alignment defaults to "unspecified",
1324 which does not prevent any alignment promotions.</dd>
Lang Hamesbb5b3f32011-10-10 23:42:08 +0000[diff] [blame]1325
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1326 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1327 <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]1328 <i>preferred</i> alignments. All sizes are in bits. Specifying
1329 the <i>pref</i> alignment is optional. If omitted, the
1330 preceding <tt>:</tt> should be omitted too.</dd>
1331
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1332 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1333 <dd>This specifies the alignment for an integer type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1334 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1335
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1336 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1337 <dd>This specifies the alignment for a vector type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1338 <i>size</i>.</dd>
1339
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1340 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1341 <dd>This specifies the alignment for a floating point type of a given bit
Dale Johannesen9d8d2212010-05-28 18:54:47 +0000[diff] [blame]1342 <i>size</i>. Only values of <i>size</i> that are supported by the target
1343 will work. 32 (float) and 64 (double) are supported on all targets;
1344 80 or 128 (different flavors of long double) are also supported on some
1345 targets.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1346
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1347 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1348 <dd>This specifies the alignment for an aggregate type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1349 <i>size</i>.</dd>
1350
Daniel Dunbar87bde0b2009-06-08 22:17:53 +0000[diff] [blame]1351 <dt><tt>s<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1352 <dd>This specifies the alignment for a stack object of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1353 <i>size</i>.</dd>
Chris Lattnere82bdc42009-11-07 09:35:34 +0000[diff] [blame]1354
1355 <dt><tt>n<i>size1</i>:<i>size2</i>:<i>size3</i>...</tt></dt>
1356 <dd>This specifies a set of native integer widths for the target CPU
1357 in bits. For example, it might contain "n32" for 32-bit PowerPC,
1358 "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]1359 this set are considered to support most general arithmetic
Chris Lattnere82bdc42009-11-07 09:35:34 +0000[diff] [blame]1360 operations efficiently.</dd>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1361</dl>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1362
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1363<p>When constructing the data layout for a given target, LLVM starts with a
Dan Gohman1c70c002010-04-28 00:36:01 +0000[diff] [blame]1364 default set of specifications which are then (possibly) overridden by the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1365 specifications in the <tt>datalayout</tt> keyword. The default specifications
1366 are given in this list:</p>
1367
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1368<ul>
1369 <li><tt>E</tt> - big endian</li>
Dan Gohmanfdf2e8c2010-02-23 02:44:03 +0000[diff] [blame]1370 <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]1371 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1372 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1373 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1374 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]1375 <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]1376 alignment of 64-bits</li>
1377 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1378 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1379 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1380 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1381 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
Daniel Dunbar87bde0b2009-06-08 22:17:53 +0000[diff] [blame]1382 <li><tt>s0:64:64</tt> - stack objects are 64-bit aligned</li>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1383</ul>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1384
1385<p>When LLVM is determining the alignment for a given type, it uses the
1386 following rules:</p>
1387
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1388<ol>
1389 <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]1390 specification is used.</li>
1391
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1392 <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]1393 smallest integer type that is larger than the bitwidth of the sought type
1394 is used. If none of the specifications are larger than the bitwidth then
1395 the the largest integer type is used. For example, given the default
1396 specifications above, the i7 type will use the alignment of i8 (next
1397 largest) while both i65 and i256 will use the alignment of i64 (largest
1398 specified).</li>
1399
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1400 <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]1401 largest vector type that is smaller than the sought vector type will be
1402 used as a fall back. This happens because &lt;128 x double&gt; can be
1403 implemented in terms of 64 &lt;2 x double&gt;, for example.</li>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1404</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1405
Chris Lattner6509f502011-10-11 23:01:39 +0000[diff] [blame]1406<p>The function of the data layout string may not be what you expect. Notably,
1407 this is not a specification from the frontend of what alignment the code
1408 generator should use.</p>
1409
1410<p>Instead, if specified, the target data layout is required to match what the
1411 ultimate <em>code generator</em> expects. This string is used by the
1412 mid-level optimizers to
1413 improve code, and this only works if it matches what the ultimate code
1414 generator uses. If you would like to generate IR that does not embed this
1415 target-specific detail into the IR, then you don't have to specify the
1416 string. This will disable some optimizations that require precise layout
1417 information, but this also prevents those optimizations from introducing
1418 target specificity into the IR.</p>
1419
1420
1421
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1422</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1423
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1424<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1425<h3>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1426 <a name="pointeraliasing">Pointer Aliasing Rules</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1427</h3>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1428
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1429<div>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1430
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1431<p>Any memory access must be done through a pointer value associated
Andreas Bolka99a82052009-07-27 20:37:10 +0000[diff] [blame]1432with an address range of the memory access, otherwise the behavior
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1433is undefined. Pointer values are associated with address ranges
1434according to the following rules:</p>
1435
1436<ul>
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1437 <li>A pointer value is associated with the addresses associated with
1438 any value it is <i>based</i> on.
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1439 <li>An address of a global variable is associated with the address
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1440 range of the variable's storage.</li>
1441 <li>The result value of an allocation instruction is associated with
1442 the address range of the allocated storage.</li>
1443 <li>A null pointer in the default address-space is associated with
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1444 no address.</li>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1445 <li>An integer constant other than zero or a pointer value returned
1446 from a function not defined within LLVM may be associated with address
1447 ranges allocated through mechanisms other than those provided by
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1448 LLVM. Such ranges shall not overlap with any ranges of addresses
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1449 allocated by mechanisms provided by LLVM.</li>
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1450</ul>
1451
1452<p>A pointer value is <i>based</i> on another pointer value according
1453 to the following rules:</p>
1454
1455<ul>
1456 <li>A pointer value formed from a
1457 <tt><a href="#i_getelementptr">getelementptr</a></tt> operation
1458 is <i>based</i> on the first operand of the <tt>getelementptr</tt>.</li>
1459 <li>The result value of a
1460 <tt><a href="#i_bitcast">bitcast</a></tt> is <i>based</i> on the operand
1461 of the <tt>bitcast</tt>.</li>
1462 <li>A pointer value formed by an
1463 <tt><a href="#i_inttoptr">inttoptr</a></tt> is <i>based</i> on all
1464 pointer values that contribute (directly or indirectly) to the
1465 computation of the pointer's value.</li>
1466 <li>The "<i>based</i> on" relationship is transitive.</li>
1467</ul>
1468
1469<p>Note that this definition of <i>"based"</i> is intentionally
1470 similar to the definition of <i>"based"</i> in C99, though it is
1471 slightly weaker.</p>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1472
1473<p>LLVM IR does not associate types with memory. The result type of a
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1474<tt><a href="#i_load">load</a></tt> merely indicates the size and
1475alignment of the memory from which to load, as well as the
Dan Gohmanc22c0f32010-06-17 19:23:50 +0000[diff] [blame]1476interpretation of the value. The first operand type of a
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1477<tt><a href="#i_store">store</a></tt> similarly only indicates the size
1478and alignment of the store.</p>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1479
1480<p>Consequently, type-based alias analysis, aka TBAA, aka
1481<tt>-fstrict-aliasing</tt>, is not applicable to general unadorned
1482LLVM IR. <a href="#metadata">Metadata</a> may be used to encode
1483additional information which specialized optimization passes may use
1484to implement type-based alias analysis.</p>
1485
1486</div>
1487
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1488<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1489<h3>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1490 <a name="volatile">Volatile Memory Accesses</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1491</h3>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1492
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1493<div>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1494
1495<p>Certain memory accesses, such as <a href="#i_load"><tt>load</tt></a>s, <a
1496href="#i_store"><tt>store</tt></a>s, and <a
1497href="#int_memcpy"><tt>llvm.memcpy</tt></a>s may be marked <tt>volatile</tt>.
1498The optimizers must not change the number of volatile operations or change their
1499order of execution relative to other volatile operations. The optimizers
1500<i>may</i> change the order of volatile operations relative to non-volatile
1501operations. This is not Java's "volatile" and has no cross-thread
1502synchronization behavior.</p>
1503
1504</div>
1505
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1506<!-- ======================================================================= -->
1507<h3>
1508 <a name="memmodel">Memory Model for Concurrent Operations</a>
1509</h3>
1510
1511<div>
1512
1513<p>The LLVM IR does not define any way to start parallel threads of execution
1514or to register signal handlers. Nonetheless, there are platform-specific
1515ways to create them, and we define LLVM IR's behavior in their presence. This
1516model is inspired by the C++0x memory model.</p>
1517
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1518<p>For a more informal introduction to this model, see the
1519<a href="Atomics.html">LLVM Atomic Instructions and Concurrency Guide</a>.
1520
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1521<p>We define a <i>happens-before</i> partial order as the least partial order
1522that</p>
1523<ul>
1524 <li>Is a superset of single-thread program order, and</li>
1525 <li>When a <i>synchronizes-with</i> <tt>b</tt>, includes an edge from
1526 <tt>a</tt> to <tt>b</tt>. <i>Synchronizes-with</i> pairs are introduced
1527 by platform-specific techniques, like pthread locks, thread
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1528 creation, thread joining, etc., and by atomic instructions.
1529 (See also <a href="#ordering">Atomic Memory Ordering Constraints</a>).
1530 </li>
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1531</ul>
1532
1533<p>Note that program order does not introduce <i>happens-before</i> edges
1534between a thread and signals executing inside that thread.</p>
1535
1536<p>Every (defined) read operation (load instructions, memcpy, atomic
1537loads/read-modify-writes, etc.) <var>R</var> reads a series of bytes written by
1538(defined) write operations (store instructions, atomic
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1539stores/read-modify-writes, memcpy, etc.). For the purposes of this section,
1540initialized globals are considered to have a write of the initializer which is
1541atomic and happens before any other read or write of the memory in question.
1542For each byte of a read <var>R</var>, <var>R<sub>byte</sub></var> may see
1543any write to the same byte, except:</p>
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1544
1545<ul>
1546 <li>If <var>write<sub>1</sub></var> happens before
1547 <var>write<sub>2</sub></var>, and <var>write<sub>2</sub></var> happens
1548 before <var>R<sub>byte</sub></var>, then <var>R<sub>byte</sub></var>
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1549 does not see <var>write<sub>1</sub></var>.
Bill Wendling0246bb72011-07-31 06:45:03 +0000[diff] [blame]1550 <li>If <var>R<sub>byte</sub></var> happens before
1551 <var>write<sub>3</sub></var>, then <var>R<sub>byte</sub></var> does not
1552 see <var>write<sub>3</sub></var>.
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1553</ul>
1554
1555<p>Given that definition, <var>R<sub>byte</sub></var> is defined as follows:
1556<ul>
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1557 <li>If <var>R</var> is volatile, the result is target-dependent. (Volatile
1558 is supposed to give guarantees which can support
1559 <code>sig_atomic_t</code> in C/C++, and may be used for accesses to
1560 addresses which do not behave like normal memory. It does not generally
1561 provide cross-thread synchronization.)
1562 <li>Otherwise, if there is no write to the same byte that happens before
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1563 <var>R<sub>byte</sub></var>, <var>R<sub>byte</sub></var> returns
1564 <tt>undef</tt> for that byte.
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1565 <li>Otherwise, if <var>R<sub>byte</sub></var> may see exactly one write,
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1566 <var>R<sub>byte</sub></var> returns the value written by that
1567 write.</li>
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1568 <li>Otherwise, if <var>R</var> is atomic, and all the writes
1569 <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]1570 values written. See the <a href="#ordering">Atomic Memory Ordering
1571 Constraints</a> section for additional constraints on how the choice
1572 is made.
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1573 <li>Otherwise <var>R<sub>byte</sub></var> returns <tt>undef</tt>.</li>
1574</ul>
1575
1576<p><var>R</var> returns the value composed of the series of bytes it read.
1577This implies that some bytes within the value may be <tt>undef</tt>
1578<b>without</b> the entire value being <tt>undef</tt>. Note that this only
1579defines the semantics of the operation; it doesn't mean that targets will
1580emit more than one instruction to read the series of bytes.</p>
1581
1582<p>Note that in cases where none of the atomic intrinsics are used, this model
1583places only one restriction on IR transformations on top of what is required
1584for single-threaded execution: introducing a store to a byte which might not
Eli Friedman101c81d2011-08-02 01:15:34 +0000[diff] [blame]1585otherwise be stored is not allowed in general. (Specifically, in the case
1586where another thread might write to and read from an address, introducing a
1587store can change a load that may see exactly one write into a load that may
1588see multiple writes.)</p>
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1589
1590<!-- FIXME: This model assumes all targets where concurrency is relevant have
1591a byte-size store which doesn't affect adjacent bytes. As far as I can tell,
1592none of the backends currently in the tree fall into this category; however,
1593there might be targets which care. If there are, we want a paragraph
1594like the following:
1595
1596Targets may specify that stores narrower than a certain width are not
1597available; on such a target, for the purposes of this model, treat any
1598non-atomic write with an alignment or width less than the minimum width
1599as if it writes to the relevant surrounding bytes.
1600-->
1601
1602</div>
1603
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1604<!-- ======================================================================= -->
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]1605<h3>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1606 <a name="ordering">Atomic Memory Ordering Constraints</a>
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]1607</h3>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1608
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]1609<div>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1610
1611<p>Atomic instructions (<a href="#i_cmpxchg"><code>cmpxchg</code></a>,
Eli Friedman21006d42011-08-09 23:02:53 +0000[diff] [blame]1612<a href="#i_atomicrmw"><code>atomicrmw</code></a>,
1613<a href="#i_fence"><code>fence</code></a>,
1614<a href="#i_load"><code>atomic load</code></a>, and
Eli Friedman8fa281a2011-08-09 23:26:12 +0000[diff] [blame]1615<a href="#i_store"><code>atomic store</code></a>) take an ordering parameter
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1616that determines which other atomic instructions on the same address they
1617<i>synchronize with</i>. These semantics are borrowed from Java and C++0x,
1618but are somewhat more colloquial. If these descriptions aren't precise enough,
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1619check those specs (see spec references in the
Nick Lewycky300a2632012-01-23 08:47:21 +0000[diff] [blame]1620<a href="Atomics.html#introduction">atomics guide</a>).
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1621<a href="#i_fence"><code>fence</code></a> instructions
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1622treat these orderings somewhat differently since they don't take an address.
1623See that instruction's documentation for details.</p>
1624
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1625<p>For a simpler introduction to the ordering constraints, see the
1626<a href="Atomics.html">LLVM Atomic Instructions and Concurrency Guide</a>.</p>
1627
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1628<dl>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1629<dt><code>unordered</code></dt>
1630<dd>The set of values that can be read is governed by the happens-before
1631partial order. A value cannot be read unless some operation wrote it.
1632This is intended to provide a guarantee strong enough to model Java's
1633non-volatile shared variables. This ordering cannot be specified for
1634read-modify-write operations; it is not strong enough to make them atomic
1635in any interesting way.</dd>
1636<dt><code>monotonic</code></dt>
1637<dd>In addition to the guarantees of <code>unordered</code>, there is a single
1638total order for modifications by <code>monotonic</code> operations on each
1639address. All modification orders must be compatible with the happens-before
1640order. There is no guarantee that the modification orders can be combined to
1641a global total order for the whole program (and this often will not be
1642possible). The read in an atomic read-modify-write operation
1643(<a href="#i_cmpxchg"><code>cmpxchg</code></a> and
1644<a href="#i_atomicrmw"><code>atomicrmw</code></a>)
1645reads the value in the modification order immediately before the value it
1646writes. If one atomic read happens before another atomic read of the same
1647address, the later read must see the same value or a later value in the
1648address's modification order. This disallows reordering of
1649<code>monotonic</code> (or stronger) operations on the same address. If an
1650address is written <code>monotonic</code>ally by one thread, and other threads
1651<code>monotonic</code>ally read that address repeatedly, the other threads must
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1652eventually see the write. This corresponds to the C++0x/C1x
1653<code>memory_order_relaxed</code>.</dd>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1654<dt><code>acquire</code></dt>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1655<dd>In addition to the guarantees of <code>monotonic</code>,
Eli Friedmanc264b2f2011-08-24 20:28:39 +0000[diff] [blame]1656a <i>synchronizes-with</i> edge may be formed with a <code>release</code>
1657operation. This is intended to model C++'s <code>memory_order_acquire</code>.</dd>
1658<dt><code>release</code></dt>
1659<dd>In addition to the guarantees of <code>monotonic</code>, if this operation
1660writes a value which is subsequently read by an <code>acquire</code> operation,
1661it <i>synchronizes-with</i> that operation. (This isn't a complete
1662description; see the C++0x definition of a release sequence.) This corresponds
1663to the C++0x/C1x <code>memory_order_release</code>.</dd>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1664<dt><code>acq_rel</code> (acquire+release)</dt><dd>Acts as both an
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1665<code>acquire</code> and <code>release</code> operation on its address.
1666This corresponds to the C++0x/C1x <code>memory_order_acq_rel</code>.</dd>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1667<dt><code>seq_cst</code> (sequentially consistent)</dt><dd>
1668<dd>In addition to the guarantees of <code>acq_rel</code>
1669(<code>acquire</code> for an operation which only reads, <code>release</code>
1670for an operation which only writes), there is a global total order on all
1671sequentially-consistent operations on all addresses, which is consistent with
1672the <i>happens-before</i> partial order and with the modification orders of
1673all the affected addresses. Each sequentially-consistent read sees the last
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1674preceding write to the same address in this global order. This corresponds
1675to the C++0x/C1x <code>memory_order_seq_cst</code> and Java volatile.</dd>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1676</dl>
1677
1678<p id="singlethread">If an atomic operation is marked <code>singlethread</code>,
1679it only <i>synchronizes with</i> or participates in modification and seq_cst
1680total orderings with other operations running in the same thread (for example,
1681in signal handlers).</p>
1682
1683</div>
1684
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1685</div>
1686
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1687<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1688<h2><a name="typesystem">Type System</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1689<!-- *********************************************************************** -->
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1690
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1691<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1692
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1693<p>The LLVM type system is one of the most important features of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1694 intermediate representation. Being typed enables a number of optimizations
1695 to be performed on the intermediate representation directly, without having
1696 to do extra analyses on the side before the transformation. A strong type
1697 system makes it easier to read the generated code and enables novel analyses
1698 and transformations that are not feasible to perform on normal three address
1699 code representations.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1700
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1701<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1702<h3>
1703 <a name="t_classifications">Type Classifications</a>
1704</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1705
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1706<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1707
1708<p>The types fall into a few useful classifications:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1709
1710<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1711 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1712 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1713 <tr>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1714 <td><a href="#t_integer">integer</a></td>
Reid Spencer2b916312007-05-16 18:44:01 +0000[diff] [blame]1715 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1716 </tr>
1717 <tr>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1718 <td><a href="#t_floating">floating point</a></td>
Dan Gohmance163392011-12-17 00:04:22 +0000[diff] [blame]1719 <td><tt>half, float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1720 </tr>
1721 <tr>
1722 <td><a name="t_firstclass">first class</a></td>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1723 <td><a href="#t_integer">integer</a>,
1724 <a href="#t_floating">floating point</a>,
1725 <a href="#t_pointer">pointer</a>,
Dan Gohman0066db62008-06-18 18:42:13 +0000[diff] [blame]1726 <a href="#t_vector">vector</a>,
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]1727 <a href="#t_struct">structure</a>,
1728 <a href="#t_array">array</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1729 <a href="#t_label">label</a>,
1730 <a href="#t_metadata">metadata</a>.
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1731 </td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1732 </tr>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1733 <tr>
1734 <td><a href="#t_primitive">primitive</a></td>
1735 <td><a href="#t_label">label</a>,
1736 <a href="#t_void">void</a>,
Tobias Grosser05387292010-12-28 20:29:31 +0000[diff] [blame]1737 <a href="#t_integer">integer</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1738 <a href="#t_floating">floating point</a>,
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1739 <a href="#t_x86mmx">x86mmx</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1740 <a href="#t_metadata">metadata</a>.</td>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1741 </tr>
1742 <tr>
1743 <td><a href="#t_derived">derived</a></td>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1744 <td><a href="#t_array">array</a>,
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1745 <a href="#t_function">function</a>,
1746 <a href="#t_pointer">pointer</a>,
1747 <a href="#t_struct">structure</a>,
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1748 <a href="#t_vector">vector</a>,
1749 <a href="#t_opaque">opaque</a>.
Dan Gohman01ac1012008-10-14 16:32:04 +0000[diff] [blame]1750 </td>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1751 </tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1752 </tbody>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1753</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1754
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1755<p>The <a href="#t_firstclass">first class</a> types are perhaps the most
1756 important. Values of these types are the only ones which can be produced by
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1757 instructions.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1758
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1759</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1760
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1761<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1762<h3>
1763 <a name="t_primitive">Primitive Types</a>
1764</h3>
Chris Lattner8f8c7b72008-01-04 04:34:14 +0000[diff] [blame]1765
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1766<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1767
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1768<p>The primitive types are the fundamental building blocks of the LLVM
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1769 system.</p>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1770
1771<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1772<h4>
1773 <a name="t_integer">Integer Type</a>
1774</h4>
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1775
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1776<div>
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1777
1778<h5>Overview:</h5>
1779<p>The integer type is a very simple type that simply specifies an arbitrary
1780 bit width for the integer type desired. Any bit width from 1 bit to
1781 2<sup>23</sup>-1 (about 8 million) can be specified.</p>
1782
1783<h5>Syntax:</h5>
1784<pre>
1785 iN
1786</pre>
1787
1788<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1789 value.</p>
1790
1791<h5>Examples:</h5>
1792<table class="layout">
1793 <tr class="layout">
1794 <td class="left"><tt>i1</tt></td>
1795 <td class="left">a single-bit integer.</td>
1796 </tr>
1797 <tr class="layout">
1798 <td class="left"><tt>i32</tt></td>
1799 <td class="left">a 32-bit integer.</td>
1800 </tr>
1801 <tr class="layout">
1802 <td class="left"><tt>i1942652</tt></td>
1803 <td class="left">a really big integer of over 1 million bits.</td>
1804 </tr>
1805</table>
1806
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1807</div>
1808
1809<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1810<h4>
1811 <a name="t_floating">Floating Point Types</a>
1812</h4>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1813
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1814<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1815
1816<table>
1817 <tbody>
1818 <tr><th>Type</th><th>Description</th></tr>
Dan Gohmance163392011-12-17 00:04:22 +0000[diff] [blame]1819 <tr><td><tt>half</tt></td><td>16-bit floating point value</td></tr>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1820 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1821 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1822 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1823 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1824 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1825 </tbody>
1826</table>
1827
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1828</div>
1829
1830<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1831<h4>
1832 <a name="t_x86mmx">X86mmx Type</a>
1833</h4>
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1834
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1835<div>
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1836
1837<h5>Overview:</h5>
1838<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>
1839
1840<h5>Syntax:</h5>
1841<pre>
Dale Johannesen473a8c82010-10-01 01:07:02 +0000[diff] [blame]1842 x86mmx
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1843</pre>
1844
1845</div>
1846
1847<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1848<h4>
1849 <a name="t_void">Void Type</a>
1850</h4>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1851
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1852<div>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1853
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1854<h5>Overview:</h5>
1855<p>The void type does not represent any value and has no size.</p>
1856
1857<h5>Syntax:</h5>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1858<pre>
1859 void
1860</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1861
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1862</div>
1863
1864<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1865<h4>
1866 <a name="t_label">Label Type</a>
1867</h4>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1868
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1869<div>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1870
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1871<h5>Overview:</h5>
1872<p>The label type represents code labels.</p>
1873
1874<h5>Syntax:</h5>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1875<pre>
1876 label
1877</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1878
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1879</div>
1880
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1881<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1882<h4>
1883 <a name="t_metadata">Metadata Type</a>
1884</h4>
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1885
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1886<div>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1887
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1888<h5>Overview:</h5>
Nick Lewyckyc261df92009-09-27 23:27:42 +0000[diff] [blame]1889<p>The metadata type represents embedded metadata. No derived types may be
1890 created from metadata except for <a href="#t_function">function</a>
1891 arguments.
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1892
1893<h5>Syntax:</h5>
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1894<pre>
1895 metadata
1896</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1897
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1898</div>
1899
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1900</div>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1901
1902<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1903<h3>
1904 <a name="t_derived">Derived Types</a>
1905</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1906
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1907<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1908
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1909<p>The real power in LLVM comes from the derived types in the system. This is
1910 what allows a programmer to represent arrays, functions, pointers, and other
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1911 useful types. Each of these types contain one or more element types which
1912 may be a primitive type, or another derived type. For example, it is
1913 possible to have a two dimensional array, using an array as the element type
1914 of another array.</p>
Dan Gohmand8791e52009-01-24 15:58:40 +0000[diff] [blame]1915
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1916<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1917<h4>
1918 <a name="t_aggregate">Aggregate Types</a>
1919</h4>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1920
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1921<div>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1922
1923<p>Aggregate Types are a subset of derived types that can contain multiple
Duncan Sands20536b52011-12-14 15:44:20 +0000[diff] [blame]1924 member types. <a href="#t_array">Arrays</a> and
1925 <a href="#t_struct">structs</a> are aggregate types.
1926 <a href="#t_vector">Vectors</a> are not considered to be aggregate types.</p>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1927
1928</div>
1929
Reid Spencer2b916312007-05-16 18:44:01 +0000[diff] [blame]1930<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1931<h4>
1932 <a name="t_array">Array Type</a>
1933</h4>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1934
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1935<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1936
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1937<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1938<p>The array type is a very simple derived type that arranges elements
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1939 sequentially in memory. The array type requires a size (number of elements)
1940 and an underlying data type.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1941
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1942<h5>Syntax:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1943<pre>
1944 [&lt;# elements&gt; x &lt;elementtype&gt;]
1945</pre>
1946
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1947<p>The number of elements is a constant integer value; <tt>elementtype</tt> may
1948 be any type with a size.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1949
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1950<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1951<table class="layout">
1952 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]1953 <td class="left"><tt>[40 x i32]</tt></td>
1954 <td class="left">Array of 40 32-bit integer values.</td>
1955 </tr>
1956 <tr class="layout">
1957 <td class="left"><tt>[41 x i32]</tt></td>
1958 <td class="left">Array of 41 32-bit integer values.</td>
1959 </tr>
1960 <tr class="layout">
1961 <td class="left"><tt>[4 x i8]</tt></td>
1962 <td class="left">Array of 4 8-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1963 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1964</table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1965<p>Here are some examples of multidimensional arrays:</p>
1966<table class="layout">
1967 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]1968 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1969 <td class="left">3x4 array of 32-bit integer values.</td>
1970 </tr>
1971 <tr class="layout">
1972 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1973 <td class="left">12x10 array of single precision floating point values.</td>
1974 </tr>
1975 <tr class="layout">
1976 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1977 <td class="left">2x3x4 array of 16-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1978 </tr>
1979</table>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]1980
Dan Gohman7657f6b2009-11-09 19:01:53 +0000[diff] [blame]1981<p>There is no restriction on indexing beyond the end of the array implied by
1982 a static type (though there are restrictions on indexing beyond the bounds
1983 of an allocated object in some cases). This means that single-dimension
1984 'variable sized array' addressing can be implemented in LLVM with a zero
1985 length array type. An implementation of 'pascal style arrays' in LLVM could
1986 use the type "<tt>{ i32, [0 x float]}</tt>", for example.</p>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]1987
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1988</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1989
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1990<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1991<h4>
1992 <a name="t_function">Function Type</a>
1993</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1994
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1995<div>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]1996
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1997<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1998<p>The function type can be thought of as a function signature. It consists of
1999 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]2000 function type is a first class type or a void type.</p>
Devang Patelc3fc6df2008-03-10 20:49:15 +0000[diff] [blame]2001
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2002<h5>Syntax:</h5>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]2003<pre>
Nick Lewycky51386942009-09-27 07:55:32 +0000[diff] [blame]2004 &lt;returntype&gt; (&lt;parameter list&gt;)
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]2005</pre>
2006
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]2007<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]2008 specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
2009 which indicates that the function takes a variable number of arguments.
2010 Variable argument functions can access their arguments with
2011 the <a href="#int_varargs">variable argument handling intrinsic</a>
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]2012 functions. '<tt>&lt;returntype&gt;</tt>' is any type except
Nick Lewyckyc261df92009-09-27 23:27:42 +0000[diff] [blame]2013 <a href="#t_label">label</a>.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]2014
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2015<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2016<table class="layout">
2017 <tr class="layout">
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]2018 <td class="left"><tt>i32 (i32)</tt></td>
2019 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2020 </td>
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]2021 </tr><tr class="layout">
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]2022 <td class="left"><tt>float&nbsp;(i16,&nbsp;i32&nbsp;*)&nbsp;*
Reid Spencerf17a0b72006-12-31 07:20:23 +0000[diff] [blame]2023 </tt></td>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2024 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]2025 an <tt>i16</tt> and a <a href="#t_pointer">pointer</a> to <tt>i32</tt>,
2026 returning <tt>float</tt>.
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]2027 </td>
2028 </tr><tr class="layout">
2029 <td class="left"><tt>i32 (i8*, ...)</tt></td>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2030 <td class="left">A vararg function that takes at least one
2031 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
2032 which returns an integer. This is the signature for <tt>printf</tt> in
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]2033 LLVM.
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2034 </td>
Devang Patela582f402008-03-24 05:35:41 +0000[diff] [blame]2035 </tr><tr class="layout">
2036 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Nick Lewycky51386942009-09-27 07:55:32 +0000[diff] [blame]2037 <td class="left">A function taking an <tt>i32</tt>, returning a
2038 <a href="#t_struct">structure</a> containing two <tt>i32</tt> values
Devang Patela582f402008-03-24 05:35:41 +0000[diff] [blame]2039 </td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2040 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2041</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]2042
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2043</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2044
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2045<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2046<h4>
2047 <a name="t_struct">Structure Type</a>
2048</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2049
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2050<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2051
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2052<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2053<p>The structure type is used to represent a collection of data members together
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2054 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]2055
Jeffrey Yasskin7a088cf2010-01-11 19:19:26 +0000[diff] [blame]2056<p>Structures in memory are accessed using '<tt><a href="#i_load">load</a></tt>'
2057 and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field
2058 with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
2059 Structures in registers are accessed using the
2060 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' and
2061 '<tt><a href="#i_insertvalue">insertvalue</a></tt>' instructions.</p>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2062
2063<p>Structures may optionally be "packed" structures, which indicate that the
2064 alignment of the struct is one byte, and that there is no padding between
Chris Lattner2c38d652011-08-12 17:31:02 +0000[diff] [blame]2065 the elements. In non-packed structs, padding between field types is inserted
2066 as defined by the TargetData string in the module, which is required to match
Chris Lattnere4617b02011-10-11 23:02:17 +0000[diff] [blame]2067 what the underlying code generator expects.</p>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2068
Chris Lattner2c38d652011-08-12 17:31:02 +0000[diff] [blame]2069<p>Structures can either be "literal" or "identified". A literal structure is
2070 defined inline with other types (e.g. <tt>{i32, i32}*</tt>) whereas identified
2071 types are always defined at the top level with a name. Literal types are
2072 uniqued by their contents and can never be recursive or opaque since there is
Chris Lattneraa175c32011-08-12 18:12:40 +0000[diff] [blame]2073 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]2074 never uniqued.
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2075</p>
2076
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2077<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2078<pre>
Chris Lattner2c38d652011-08-12 17:31:02 +0000[diff] [blame]2079 %T1 = type { &lt;type list&gt; } <i>; Identified normal struct type</i>
2080 %T2 = type &lt;{ &lt;type list&gt; }&gt; <i>; Identified packed struct type</i>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2081</pre>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2082
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2083<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2084<table class="layout">
2085 <tr class="layout">
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2086 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
2087 <td class="left">A triple of three <tt>i32</tt> values</td>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2088 </tr>
2089 <tr class="layout">
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2090 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
2091 <td class="left">A pair, where the first element is a <tt>float</tt> and the
2092 second element is a <a href="#t_pointer">pointer</a> to a
2093 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
2094 an <tt>i32</tt>.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2095 </tr>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2096 <tr class="layout">
2097 <td class="left"><tt>&lt;{ i8, i32 }&gt;</tt></td>
2098 <td class="left">A packed struct known to be 5 bytes in size.</td>
2099 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2100</table>
Dan Gohmand8791e52009-01-24 15:58:40 +0000[diff] [blame]2101
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2102</div>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2103
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2104<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2105<h4>
Chris Lattner628ed392011-07-23 19:59:08 +0000[diff] [blame]2106 <a name="t_opaque">Opaque Structure Types</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2107</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2108
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2109<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2110
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2111<h5>Overview:</h5>
Chris Lattner628ed392011-07-23 19:59:08 +0000[diff] [blame]2112<p>Opaque structure types are used to represent named structure types that do
2113 not have a body specified. This corresponds (for example) to the C notion of
2114 a forward declared structure.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2115
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2116<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2117<pre>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2118 %X = type opaque
2119 %52 = type opaque
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2120</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2121
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2122<h5>Examples:</h5>
2123<table class="layout">
2124 <tr class="layout">
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2125 <td class="left"><tt>opaque</tt></td>
2126 <td class="left">An opaque type.</td>
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2127 </tr>
2128</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2129
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2130</div>
2131
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2132
2133
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2134<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2135<h4>
2136 <a name="t_pointer">Pointer Type</a>
2137</h4>
Chris Lattner0fd4a272009-02-08 19:53:29 +0000[diff] [blame]2138
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2139<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2140
2141<h5>Overview:</h5>
Dan Gohmanff3ef322010-02-25 16:50:07 +0000[diff] [blame]2142<p>The pointer type is used to specify memory locations.
2143 Pointers are commonly used to reference objects in memory.</p>
2144
2145<p>Pointer types may have an optional address space attribute defining the
2146 numbered address space where the pointed-to object resides. The default
2147 address space is number zero. The semantics of non-zero address
2148 spaces are target-specific.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2149
2150<p>Note that LLVM does not permit pointers to void (<tt>void*</tt>) nor does it
2151 permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</tt> instead.</p>
Chris Lattner0fd4a272009-02-08 19:53:29 +0000[diff] [blame]2152
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2153<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2154<pre>
2155 &lt;type&gt; *
2156</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2157
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2158<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2159<table class="layout">
2160 <tr class="layout">
Dan Gohman2a08c532009-01-04 23:44:43 +0000[diff] [blame]2161 <td class="left"><tt>[4 x i32]*</tt></td>
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2162 <td class="left">A <a href="#t_pointer">pointer</a> to <a
2163 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
2164 </tr>
2165 <tr class="layout">
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]2166 <td class="left"><tt>i32 (i32*) *</tt></td>
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2167 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2168 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2169 <tt>i32</tt>.</td>
2170 </tr>
2171 <tr class="layout">
2172 <td class="left"><tt>i32 addrspace(5)*</tt></td>
2173 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
2174 that resides in address space #5.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2175 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2176</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2177
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2178</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2179
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2180<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2181<h4>
2182 <a name="t_vector">Vector Type</a>
2183</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2184
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2185<div>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2186
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2187<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2188<p>A vector type is a simple derived type that represents a vector of elements.
2189 Vector types are used when multiple primitive data are operated in parallel
2190 using a single instruction (SIMD). A vector type requires a size (number of
Duncan Sandsd40d14e2009-11-27 13:38:03 +0000[diff] [blame]2191 elements) and an underlying primitive data type. Vector types are considered
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2192 <a href="#t_firstclass">first class</a>.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2193
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2194<h5>Syntax:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2195<pre>
2196 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
2197</pre>
2198
Chris Lattner7d2e7be2010-10-10 18:20:35 +0000[diff] [blame]2199<p>The number of elements is a constant integer value larger than 0; elementtype
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]2200 may be any integer or floating point type, or a pointer to these types.
2201 Vectors of size zero are not allowed. </p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2202
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2203<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2204<table class="layout">
2205 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2206 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
2207 <td class="left">Vector of 4 32-bit integer values.</td>
2208 </tr>
2209 <tr class="layout">
2210 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
2211 <td class="left">Vector of 8 32-bit floating-point values.</td>
2212 </tr>
2213 <tr class="layout">
2214 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
2215 <td class="left">Vector of 2 64-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2216 </tr>
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]2217 <tr class="layout">
2218 <td class="left"><tt>&lt;4 x i64*&gt;</tt></td>
2219 <td class="left">Vector of 4 pointers to 64-bit integer values.</td>
2220 </tr>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2221</table>
Dan Gohmand8791e52009-01-24 15:58:40 +0000[diff] [blame]2222
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2223</div>
2224
Bill Wendlingaf75f0c2011-07-31 06:47:33 +0000[diff] [blame]2225</div>
2226
NAKAMURA Takumi4b2e07a2011-10-31 13:04:26 +0000[diff] [blame]2227</div>
2228
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2229<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2230<h2><a name="constants">Constants</a></h2>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2231<!-- *********************************************************************** -->
2232
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2233<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2234
2235<p>LLVM has several different basic types of constants. This section describes
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2236 them all and their syntax.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2237
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2238<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2239<h3>
2240 <a name="simpleconstants">Simple Constants</a>
2241</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2242
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2243<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2244
2245<dl>
2246 <dt><b>Boolean constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2247 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]2248 constants of the <tt><a href="#t_integer">i1</a></tt> type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2249
2250 <dt><b>Integer constants</b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2251 <dd>Standard integers (such as '4') are constants of
2252 the <a href="#t_integer">integer</a> type. Negative numbers may be used
2253 with integer types.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2254
2255 <dt><b>Floating point constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2256 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2257 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
2258 notation (see below). The assembler requires the exact decimal value of a
2259 floating-point constant. For example, the assembler accepts 1.25 but
2260 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
2261 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2262
2263 <dt><b>Null pointer constants</b></dt>
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]2264 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2265 and must be of <a href="#t_pointer">pointer type</a>.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2266</dl>
2267
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2268<p>The one non-intuitive notation for constants is the hexadecimal form of
2269 floating point constants. For example, the form '<tt>double
2270 0x432ff973cafa8000</tt>' is equivalent to (but harder to read than)
2271 '<tt>double 4.5e+15</tt>'. The only time hexadecimal floating point
2272 constants are required (and the only time that they are generated by the
2273 disassembler) is when a floating point constant must be emitted but it cannot
2274 be represented as a decimal floating point number in a reasonable number of
2275 digits. For example, NaN's, infinities, and other special values are
2276 represented in their IEEE hexadecimal format so that assembly and disassembly
2277 do not cause any bits to change in the constants.</p>
2278
Dan Gohmance163392011-12-17 00:04:22 +0000[diff] [blame]2279<p>When using the hexadecimal form, constants of types half, float, and double are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2280 represented using the 16-digit form shown above (which matches the IEEE754
Dan Gohmance163392011-12-17 00:04:22 +0000[diff] [blame]2281 representation for double); half and float values must, however, be exactly
2282 representable as IEE754 half and single precision, respectively.
2283 Hexadecimal format is always used
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2284 for long double, and there are three forms of long double. The 80-bit format
2285 used by x86 is represented as <tt>0xK</tt> followed by 20 hexadecimal digits.
2286 The 128-bit format used by PowerPC (two adjacent doubles) is represented
2287 by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit format
2288 is represented by <tt>0xL</tt> followed by 32 hexadecimal digits; no
2289 currently supported target uses this format. Long doubles will only work if
2290 they match the long double format on your target. All hexadecimal formats
2291 are big-endian (sign bit at the left).</p>
2292
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]2293<p>There are no constants of type x86mmx.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2294</div>
2295
2296<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2297<h3>
Bill Wendlingd9fe2982009-07-20 02:32:41 +0000[diff] [blame]2298<a name="aggregateconstants"></a> <!-- old anchor -->
2299<a name="complexconstants">Complex Constants</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2300</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2301
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2302<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2303
Chris Lattner70882792009-02-28 18:32:25 +0000[diff] [blame]2304<p>Complex constants are a (potentially recursive) combination of simple
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2305 constants and smaller complex constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2306
2307<dl>
2308 <dt><b>Structure constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2309 <dd>Structure constants are represented with notation similar to structure
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2310 type definitions (a comma separated list of elements, surrounded by braces
2311 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
2312 where "<tt>@G</tt>" is declared as "<tt>@G = external global i32</tt>".
2313 Structure constants must have <a href="#t_struct">structure type</a>, and
2314 the number and types of elements must match those specified by the
2315 type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2316
2317 <dt><b>Array constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2318 <dd>Array constants are represented with notation similar to array type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2319 definitions (a comma separated list of elements, surrounded by square
2320 brackets (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74
2321 ]</tt>". Array constants must have <a href="#t_array">array type</a>, and
2322 the number and types of elements must match those specified by the
2323 type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2324
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2325 <dt><b>Vector constants</b></dt>
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2326 <dd>Vector constants are represented with notation similar to vector type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2327 definitions (a comma separated list of elements, surrounded by
2328 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32
2329 42, i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must
2330 have <a href="#t_vector">vector type</a>, and the number and types of
2331 elements must match those specified by the type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2332
2333 <dt><b>Zero initialization</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2334 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]2335 value to zero of <em>any</em> type, including scalar and
2336 <a href="#t_aggregate">aggregate</a> types.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2337 This is often used to avoid having to print large zero initializers
2338 (e.g. for large arrays) and is always exactly equivalent to using explicit
2339 zero initializers.</dd>
Nick Lewycky21cc4462009-04-04 07:22:01 +0000[diff] [blame]2340
2341 <dt><b>Metadata node</b></dt>
Nick Lewycky1e8c7a62009-05-30 16:08:30 +0000[diff] [blame]2342 <dd>A metadata node is a structure-like constant with
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2343 <a href="#t_metadata">metadata type</a>. For example: "<tt>metadata !{
2344 i32 0, metadata !"test" }</tt>". Unlike other constants that are meant to
2345 be interpreted as part of the instruction stream, metadata is a place to
2346 attach additional information such as debug info.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2347</dl>
2348
2349</div>
2350
2351<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2352<h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2353 <a name="globalconstants">Global Variable and Function Addresses</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2354</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2355
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2356<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2357
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2358<p>The addresses of <a href="#globalvars">global variables</a>
2359 and <a href="#functionstructure">functions</a> are always implicitly valid
2360 (link-time) constants. These constants are explicitly referenced when
2361 the <a href="#identifiers">identifier for the global</a> is used and always
2362 have <a href="#t_pointer">pointer</a> type. For example, the following is a
2363 legal LLVM file:</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2364
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2365<pre class="doc_code">
Chris Lattnera18a4242007-06-06 18:28:13 +0000[diff] [blame]2366@X = global i32 17
2367@Y = global i32 42
2368@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2369</pre>
2370
2371</div>
2372
2373<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2374<h3>
2375 <a name="undefvalues">Undefined Values</a>
2376</h3>
2377
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2378<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2379
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2380<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]2381 indicates that the user of the value may receive an unspecified bit-pattern.
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2382 Undefined values may be of any type (other than '<tt>label</tt>'
2383 or '<tt>void</tt>') and be used anywhere a constant is permitted.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2384
Chris Lattnerc608cb12009-09-11 01:49:31 +0000[diff] [blame]2385<p>Undefined values are useful because they indicate to the compiler that the
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2386 program is well defined no matter what value is used. This gives the
2387 compiler more freedom to optimize. Here are some examples of (potentially
2388 surprising) transformations that are valid (in pseudo IR):</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2389
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2390
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2391<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2392 %A = add %X, undef
2393 %B = sub %X, undef
2394 %C = xor %X, undef
2395Safe:
2396 %A = undef
2397 %B = undef
2398 %C = undef
2399</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2400
2401<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]2402 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]2403
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2404<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2405 %A = or %X, undef
2406 %B = and %X, undef
2407Safe:
2408 %A = -1
2409 %B = 0
2410Unsafe:
2411 %A = undef
2412 %B = undef
2413</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2414
2415<p>These logical operations have bits that are not always affected by the input.
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2416 For example, if <tt>%X</tt> has a zero bit, then the output of the
2417 '<tt>and</tt>' operation will always be a zero for that bit, no matter what
2418 the corresponding bit from the '<tt>undef</tt>' is. As such, it is unsafe to
2419 optimize or assume that the result of the '<tt>and</tt>' is '<tt>undef</tt>'.
2420 However, it is safe to assume that all bits of the '<tt>undef</tt>' could be
2421 0, and optimize the '<tt>and</tt>' to 0. Likewise, it is safe to assume that
2422 all the bits of the '<tt>undef</tt>' operand to the '<tt>or</tt>' could be
2423 set, allowing the '<tt>or</tt>' to be folded to -1.</p>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2424
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2425<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2426 %A = select undef, %X, %Y
2427 %B = select undef, 42, %Y
2428 %C = select %X, %Y, undef
2429Safe:
2430 %A = %X (or %Y)
2431 %B = 42 (or %Y)
2432 %C = %Y
2433Unsafe:
2434 %A = undef
2435 %B = undef
2436 %C = undef
2437</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2438
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2439<p>This set of examples shows that undefined '<tt>select</tt>' (and conditional
2440 branch) conditions can go <em>either way</em>, but they have to come from one
2441 of the two operands. In the <tt>%A</tt> example, if <tt>%X</tt> and
2442 <tt>%Y</tt> were both known to have a clear low bit, then <tt>%A</tt> would
2443 have to have a cleared low bit. However, in the <tt>%C</tt> example, the
2444 optimizer is allowed to assume that the '<tt>undef</tt>' operand could be the
2445 same as <tt>%Y</tt>, allowing the whole '<tt>select</tt>' to be
2446 eliminated.</p>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2447
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2448<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2449 %A = xor undef, undef
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2450
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2451 %B = undef
2452 %C = xor %B, %B
2453
2454 %D = undef
2455 %E = icmp lt %D, 4
2456 %F = icmp gte %D, 4
2457
2458Safe:
2459 %A = undef
2460 %B = undef
2461 %C = undef
2462 %D = undef
2463 %E = undef
2464 %F = undef
2465</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2466
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2467<p>This example points out that two '<tt>undef</tt>' operands are not
2468 necessarily the same. This can be surprising to people (and also matches C
2469 semantics) where they assume that "<tt>X^X</tt>" is always zero, even
2470 if <tt>X</tt> is undefined. This isn't true for a number of reasons, but the
2471 short answer is that an '<tt>undef</tt>' "variable" can arbitrarily change
2472 its value over its "live range". This is true because the variable doesn't
2473 actually <em>have a live range</em>. Instead, the value is logically read
2474 from arbitrary registers that happen to be around when needed, so the value
2475 is not necessarily consistent over time. In fact, <tt>%A</tt> and <tt>%C</tt>
2476 need to have the same semantics or the core LLVM "replace all uses with"
2477 concept would not hold.</p>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2478
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2479<pre class="doc_code">
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2480 %A = fdiv undef, %X
2481 %B = fdiv %X, undef
2482Safe:
2483 %A = undef
2484b: unreachable
2485</pre>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2486
2487<p>These examples show the crucial difference between an <em>undefined
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2488 value</em> and <em>undefined behavior</em>. An undefined value (like
2489 '<tt>undef</tt>') is allowed to have an arbitrary bit-pattern. This means that
2490 the <tt>%A</tt> operation can be constant folded to '<tt>undef</tt>', because
2491 the '<tt>undef</tt>' could be an SNaN, and <tt>fdiv</tt> is not (currently)
2492 defined on SNaN's. However, in the second example, we can make a more
2493 aggressive assumption: because the <tt>undef</tt> is allowed to be an
2494 arbitrary value, we are allowed to assume that it could be zero. Since a
2495 divide by zero has <em>undefined behavior</em>, we are allowed to assume that
2496 the operation does not execute at all. This allows us to delete the divide and
2497 all code after it. Because the undefined operation "can't happen", the
2498 optimizer can assume that it occurs in dead code.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2499
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2500<pre class="doc_code">
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2501a: store undef -> %X
2502b: store %X -> undef
2503Safe:
2504a: &lt;deleted&gt;
2505b: unreachable
2506</pre>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2507
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2508<p>These examples reiterate the <tt>fdiv</tt> example: a store <em>of</em> an
2509 undefined value can be assumed to not have any effect; we can assume that the
2510 value is overwritten with bits that happen to match what was already there.
2511 However, a store <em>to</em> an undefined location could clobber arbitrary
2512 memory, therefore, it has undefined behavior.</p>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2513
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2514</div>
2515
2516<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2517<h3>
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2518 <a name="poisonvalues">Poison Values</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2519</h3>
2520
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2521<div>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2522
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2523<p>Poison values are similar to <a href="#undefvalues">undef values</a>, however
Dan Gohmane1a29842011-12-06 03:35:58 +0000[diff] [blame]2524 they also represent the fact that an instruction or constant expression which
2525 cannot evoke side effects has nevertheless detected a condition which results
2526 in undefined behavior.</p>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2527
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2528<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]2529 only exist when produced by operations such as
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2530 <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]2531
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2532<p>Poison value behavior is defined in terms of value <i>dependence</i>:</p>
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2533
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2534<ul>
2535<li>Values other than <a href="#i_phi"><tt>phi</tt></a> nodes depend on
2536 their operands.</li>
2537
2538<li><a href="#i_phi"><tt>Phi</tt></a> nodes depend on the operand corresponding
2539 to their dynamic predecessor basic block.</li>
2540
2541<li>Function arguments depend on the corresponding actual argument values in
2542 the dynamic callers of their functions.</li>
2543
2544<li><a href="#i_call"><tt>Call</tt></a> instructions depend on the
2545 <a href="#i_ret"><tt>ret</tt></a> instructions that dynamically transfer
2546 control back to them.</li>
2547
Dan Gohmanb5328162010-05-03 14:55:22 +0000[diff] [blame]2548<li><a href="#i_invoke"><tt>Invoke</tt></a> instructions depend on the
Bill Wendling7b9e5392012-02-06 21:57:33 +0000[diff] [blame]2549 <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]2550 or exception-throwing call instructions that dynamically transfer control
2551 back to them.</li>
2552
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2553<li>Non-volatile loads and stores depend on the most recent stores to all of the
2554 referenced memory addresses, following the order in the IR
2555 (including loads and stores implied by intrinsics such as
2556 <a href="#int_memcpy"><tt>@llvm.memcpy</tt></a>.)</li>
2557
Dan Gohman7c24ff12010-05-03 14:59:34 +0000[diff] [blame]2558<!-- TODO: In the case of multiple threads, this only applies if the store
2559 "happens-before" the load or store. -->
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2560
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2561<!-- TODO: floating-point exception state -->
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2562
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2563<li>An instruction with externally visible side effects depends on the most
2564 recent preceding instruction with externally visible side effects, following
Dan Gohmanff70fe42010-07-06 15:26:33 +0000[diff] [blame]2565 the order in the IR. (This includes
2566 <a href="#volatile">volatile operations</a>.)</li>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2567
Dan Gohmanb5328162010-05-03 14:55:22 +0000[diff] [blame]2568<li>An instruction <i>control-depends</i> on a
2569 <a href="#terminators">terminator instruction</a>
2570 if the terminator instruction has multiple successors and the instruction
2571 is always executed when control transfers to one of the successors, and
Chris Lattner7a2bdde2011-04-15 05:18:47 +0000[diff] [blame]2572 may not be executed when control is transferred to another.</li>
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2573
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2574<li>Additionally, an instruction also <i>control-depends</i> on a terminator
2575 instruction if the set of instructions it otherwise depends on would be
Chris Lattner7a2bdde2011-04-15 05:18:47 +0000[diff] [blame]2576 different if the terminator had transferred control to a different
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2577 successor.</li>
2578
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2579<li>Dependence is transitive.</li>
2580
2581</ul>
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2582
Dan Gohmane1a29842011-12-06 03:35:58 +0000[diff] [blame]2583<p>Poison Values have the same behavior as <a href="#undefvalues">undef values</a>,
2584 with the additional affect that any instruction which has a <i>dependence</i>
2585 on a poison value has undefined behavior.</p>
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2586
2587<p>Here are some examples:</p>
Dan Gohmanc30f6e12010-04-26 20:54:53 +0000[diff] [blame]2588
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2589<pre class="doc_code">
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2590entry:
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2591 %poison = sub nuw i32 0, 1 ; Results in a poison value.
Dan Gohmane1a29842011-12-06 03:35:58 +0000[diff] [blame]2592 %still_poison = and i32 %poison, 0 ; 0, but also poison.
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2593 %poison_yet_again = getelementptr i32* @h, i32 %still_poison
Dan Gohmane1a29842011-12-06 03:35:58 +0000[diff] [blame]2594 store i32 0, i32* %poison_yet_again ; memory at @h[0] is poisoned
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2595
Dan Gohmane1a29842011-12-06 03:35:58 +0000[diff] [blame]2596 store i32 %poison, i32* @g ; Poison value stored to memory.
2597 %poison2 = load i32* @g ; Poison value loaded back from memory.
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2598
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2599 store volatile i32 %poison, i32* @g ; External observation; undefined behavior.
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2600
2601 %narrowaddr = bitcast i32* @g to i16*
2602 %wideaddr = bitcast i32* @g to i64*
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2603 %poison3 = load i16* %narrowaddr ; Returns a poison value.
2604 %poison4 = load i64* %wideaddr ; Returns a poison value.
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2605
Dan Gohman5cdc51e2011-12-06 03:31:14 +0000[diff] [blame]2606 %cmp = icmp slt i32 %poison, 0 ; Returns a poison value.
2607 br i1 %cmp, label %true, label %end ; Branch to either destination.
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2608
2609true:
Dan Gohman5cdc51e2011-12-06 03:31:14 +0000[diff] [blame]2610 store volatile i32 0, i32* @g ; This is control-dependent on %cmp, so
2611 ; it has undefined behavior.
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2612 br label %end
2613
2614end:
2615 %p = phi i32 [ 0, %entry ], [ 1, %true ]
Dan Gohman5cdc51e2011-12-06 03:31:14 +0000[diff] [blame]2616 ; Both edges into this PHI are
2617 ; control-dependent on %cmp, so this
2618 ; always results in a poison value.
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2619
Dan Gohman5cdc51e2011-12-06 03:31:14 +0000[diff] [blame]2620 store volatile i32 0, i32* @g ; This would depend on the store in %true
2621 ; if %cmp is true, or the store in %entry
2622 ; otherwise, so this is undefined behavior.
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2623
Nick Lewycky64f9fb12011-05-16 19:29:30 +0000[diff] [blame]2624 br i1 %cmp, label %second_true, label %second_end
Dan Gohman5cdc51e2011-12-06 03:31:14 +0000[diff] [blame]2625 ; The same branch again, but this time the
2626 ; true block doesn't have side effects.
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2627
2628second_true:
2629 ; No side effects!
Nick Lewycky64f9fb12011-05-16 19:29:30 +0000[diff] [blame]2630 ret void
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2631
2632second_end:
Dan Gohman5cdc51e2011-12-06 03:31:14 +0000[diff] [blame]2633 store volatile i32 0, i32* @g ; This time, the instruction always depends
2634 ; on the store in %end. Also, it is
2635 ; control-equivalent to %end, so this is
Dan Gohmane1a29842011-12-06 03:35:58 +0000[diff] [blame]2636 ; well-defined (ignoring earlier undefined
2637 ; behavior in this example).
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2638</pre>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2639
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2640</div>
2641
2642<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2643<h3>
2644 <a name="blockaddress">Addresses of Basic Blocks</a>
2645</h3>
2646
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2647<div>
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2648
Chris Lattnercdfc9402009-11-01 01:27:45 +0000[diff] [blame]2649<p><b><tt>blockaddress(@function, %block)</tt></b></p>
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2650
2651<p>The '<tt>blockaddress</tt>' constant computes the address of the specified
Chris Lattner2dfdf2a2009-10-27 21:49:40 +0000[diff] [blame]2652 basic block in the specified function, and always has an i8* type. Taking
Chris Lattnercdfc9402009-11-01 01:27:45 +0000[diff] [blame]2653 the address of the entry block is illegal.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2654
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2655<p>This value only has defined behavior when used as an operand to the
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2656 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>' instruction, or for
2657 comparisons against null. Pointer equality tests between labels addresses
2658 results in undefined behavior &mdash; though, again, comparison against null
2659 is ok, and no label is equal to the null pointer. This may be passed around
2660 as an opaque pointer sized value as long as the bits are not inspected. This
2661 allows <tt>ptrtoint</tt> and arithmetic to be performed on these values so
2662 long as the original value is reconstituted before the <tt>indirectbr</tt>
2663 instruction.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2664
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2665<p>Finally, some targets may provide defined semantics when using the value as
2666 the operand to an inline assembly, but that is target specific.</p>
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2667
2668</div>
2669
2670
2671<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2672<h3>
2673 <a name="constantexprs">Constant Expressions</a>
2674</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2675
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2676<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2677
2678<p>Constant expressions are used to allow expressions involving other constants
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2679 to be used as constants. Constant expressions may be of
2680 any <a href="#t_firstclass">first class</a> type and may involve any LLVM
2681 operation that does not have side effects (e.g. load and call are not
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2682 supported). The following is the syntax for constant expressions:</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2683
2684<dl>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2685 <dt><b><tt>trunc (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2686 <dd>Truncate a constant to another type. The bit size of CST must be larger
2687 than the bit size of TYPE. Both types must be integers.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2688
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2689 <dt><b><tt>zext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2690 <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]2691 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2692
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2693 <dt><b><tt>sext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2694 <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]2695 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2696
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2697 <dt><b><tt>fptrunc (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2698 <dd>Truncate a floating point constant to another floating point type. The
2699 size of CST must be larger than the size of TYPE. Both types must be
2700 floating point.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2701
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2702 <dt><b><tt>fpext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2703 <dd>Floating point extend a constant to another type. The size of CST must be
2704 smaller or equal to the size of TYPE. Both types must be floating
2705 point.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2706
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2707 <dt><b><tt>fptoui (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2708 <dd>Convert a floating point constant to the corresponding unsigned integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2709 constant. TYPE must be a scalar or vector integer type. CST must be of
2710 scalar or vector floating point type. Both CST and TYPE must be scalars,
2711 or vectors of the same number of elements. If the value won't fit in the
2712 integer type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2713
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2714 <dt><b><tt>fptosi (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2715 <dd>Convert a floating point constant to the corresponding signed integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2716 constant. TYPE must be a scalar or vector integer type. CST must be of
2717 scalar or vector floating point type. Both CST and TYPE must be scalars,
2718 or vectors of the same number of elements. If the value won't fit in the
2719 integer type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2720
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2721 <dt><b><tt>uitofp (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2722 <dd>Convert an unsigned integer constant to the corresponding floating point
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2723 constant. TYPE must be a scalar or vector floating point type. CST must be
2724 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2725 vectors of the same number of elements. If the value won't fit in the
2726 floating point type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2727
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2728 <dt><b><tt>sitofp (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2729 <dd>Convert a signed integer constant to the corresponding floating point
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2730 constant. TYPE must be a scalar or vector floating point type. CST must be
2731 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2732 vectors of the same number of elements. If the value won't fit in the
2733 floating point type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2734
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2735 <dt><b><tt>ptrtoint (CST to TYPE)</tt></b></dt>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]2736 <dd>Convert a pointer typed constant to the corresponding integer constant
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2737 <tt>TYPE</tt> must be an integer type. <tt>CST</tt> must be of pointer
2738 type. The <tt>CST</tt> value is zero extended, truncated, or unchanged to
2739 make it fit in <tt>TYPE</tt>.</dd>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]2740
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2741 <dt><b><tt>inttoptr (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2742 <dd>Convert a integer constant to a pointer constant. TYPE must be a pointer
2743 type. CST must be of integer type. The CST value is zero extended,
2744 truncated, or unchanged to make it fit in a pointer size. This one is
2745 <i>really</i> dangerous!</dd>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]2746
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2747 <dt><b><tt>bitcast (CST to TYPE)</tt></b></dt>
Chris Lattner03bbad62009-02-28 18:27:03 +0000[diff] [blame]2748 <dd>Convert a constant, CST, to another TYPE. The constraints of the operands
2749 are the same as those for the <a href="#i_bitcast">bitcast
2750 instruction</a>.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2751
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2752 <dt><b><tt>getelementptr (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
2753 <dt><b><tt>getelementptr inbounds (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2754 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2755 constants. As with the <a href="#i_getelementptr">getelementptr</a>
2756 instruction, the index list may have zero or more indexes, which are
2757 required to make sense for the type of "CSTPTR".</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2758
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2759 <dt><b><tt>select (COND, VAL1, VAL2)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2760 <dd>Perform the <a href="#i_select">select operation</a> on constants.</dd>
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]2761
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2762 <dt><b><tt>icmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]2763 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
2764
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2765 <dt><b><tt>fcmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]2766 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
Robert Bocchino9fbe1452006-01-10 19:31:34 +0000[diff] [blame]2767
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2768 <dt><b><tt>extractelement (VAL, IDX)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2769 <dd>Perform the <a href="#i_extractelement">extractelement operation</a> on
2770 constants.</dd>
Robert Bocchino9fbe1452006-01-10 19:31:34 +0000[diff] [blame]2771
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2772 <dt><b><tt>insertelement (VAL, ELT, IDX)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2773 <dd>Perform the <a href="#i_insertelement">insertelement operation</a> on
2774 constants.</dd>
Chris Lattnerc1989542006-04-08 00:13:41 +0000[diff] [blame]2775
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2776 <dt><b><tt>shufflevector (VEC1, VEC2, IDXMASK)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2777 <dd>Perform the <a href="#i_shufflevector">shufflevector operation</a> on
2778 constants.</dd>
Chris Lattnerc1989542006-04-08 00:13:41 +0000[diff] [blame]2779
Nick Lewycky9e130ce2010-05-29 06:44:15 +0000[diff] [blame]2780 <dt><b><tt>extractvalue (VAL, IDX0, IDX1, ...)</tt></b></dt>
2781 <dd>Perform the <a href="#i_extractvalue">extractvalue operation</a> on
2782 constants. The index list is interpreted in a similar manner as indices in
2783 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2784 index value must be specified.</dd>
2785
2786 <dt><b><tt>insertvalue (VAL, ELT, IDX0, IDX1, ...)</tt></b></dt>
2787 <dd>Perform the <a href="#i_insertvalue">insertvalue operation</a> on
2788 constants. The index list is interpreted in a similar manner as indices in
2789 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2790 index value must be specified.</dd>
2791
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2792 <dt><b><tt>OPCODE (LHS, RHS)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2793 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
2794 be any of the <a href="#binaryops">binary</a>
2795 or <a href="#bitwiseops">bitwise binary</a> operations. The constraints
2796 on operands are the same as those for the corresponding instruction
2797 (e.g. no bitwise operations on floating point values are allowed).</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2798</dl>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2799
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2800</div>
Chris Lattner9ee5d222004-03-08 16:49:10 +0000[diff] [blame]2801
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2802</div>
2803
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2804<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2805<h2><a name="othervalues">Other Values</a></h2>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2806<!-- *********************************************************************** -->
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2807<div>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2808<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2809<h3>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2810<a name="inlineasm">Inline Assembler Expressions</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2811</h3>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2812
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2813<div>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2814
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2815<p>LLVM supports inline assembler expressions (as opposed
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2816 to <a href="#moduleasm">Module-Level Inline Assembly</a>) through the use of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2817 a special value. This value represents the inline assembler as a string
2818 (containing the instructions to emit), a list of operand constraints (stored
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2819 as a string), a flag that indicates whether or not the inline asm
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +0000[diff] [blame]2820 expression has side effects, and a flag indicating whether the function
2821 containing the asm needs to align its stack conservatively. An example
2822 inline assembler expression is:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2823
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2824<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2825i32 (i32) asm "bswap $0", "=r,r"
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2826</pre>
2827
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2828<p>Inline assembler expressions may <b>only</b> be used as the callee operand of
2829 a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we
2830 have:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2831
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2832<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2833%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]2834</pre>
2835
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2836<p>Inline asms with side effects not visible in the constraint list must be
2837 marked as having side effects. This is done through the use of the
2838 '<tt>sideeffect</tt>' keyword, like so:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2839
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2840<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2841call void asm sideeffect "eieio", ""()
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2842</pre>
2843
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +0000[diff] [blame]2844<p>In some cases inline asms will contain code that will not work unless the
2845 stack is aligned in some way, such as calls or SSE instructions on x86,
2846 yet will not contain code that does that alignment within the asm.
2847 The compiler should make conservative assumptions about what the asm might
2848 contain and should generate its usual stack alignment code in the prologue
2849 if the '<tt>alignstack</tt>' keyword is present:</p>
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2850
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2851<pre class="doc_code">
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +0000[diff] [blame]2852call void asm alignstack "eieio", ""()
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2853</pre>
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2854
2855<p>If both keywords appear the '<tt>sideeffect</tt>' keyword must come
2856 first.</p>
2857
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2858<!--
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2859<p>TODO: The format of the asm and constraints string still need to be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2860 documented here. Constraints on what can be done (e.g. duplication, moving,
2861 etc need to be documented). This is probably best done by reference to
2862 another document that covers inline asm from a holistic perspective.</p>
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2863 -->
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2864
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2865<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2866<h4>
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2867 <a name="inlineasm_md">Inline Asm Metadata</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2868</h4>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2869
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2870<div>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2871
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2872<p>The call instructions that wrap inline asm nodes may have a
2873 "<tt>!srcloc</tt>" MDNode attached to it that contains a list of constant
2874 integers. If present, the code generator will use the integer as the
2875 location cookie value when report errors through the <tt>LLVMContext</tt>
2876 error reporting mechanisms. This allows a front-end to correlate backend
2877 errors that occur with inline asm back to the source code that produced it.
2878 For example:</p>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2879
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2880<pre class="doc_code">
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2881call void asm sideeffect "something bad", ""()<b>, !srcloc !42</b>
2882...
2883!42 = !{ i32 1234567 }
2884</pre>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2885
2886<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]2887 IR. If the MDNode contains multiple constants, the code generator will use
Chris Lattnerce1b9ad2010-11-17 08:20:42 +0000[diff] [blame]2888 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]2889
2890</div>
2891
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2892</div>
2893
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2894<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2895<h3>
2896 <a name="metadata">Metadata Nodes and Metadata Strings</a>
2897</h3>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2898
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2899<div>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2900
2901<p>LLVM IR allows metadata to be attached to instructions in the program that
2902 can convey extra information about the code to the optimizers and code
2903 generator. One example application of metadata is source-level debug
2904 information. There are two metadata primitives: strings and nodes. All
2905 metadata has the <tt>metadata</tt> type and is identified in syntax by a
2906 preceding exclamation point ('<tt>!</tt>').</p>
2907
2908<p>A metadata string is a string surrounded by double quotes. It can contain
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]2909 any character by escaping non-printable characters with "<tt>\xx</tt>" where
2910 "<tt>xx</tt>" is the two digit hex code. For example:
2911 "<tt>!"test\00"</tt>".</p>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2912
2913<p>Metadata nodes are represented with notation similar to structure constants
2914 (a comma separated list of elements, surrounded by braces and preceded by an
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]2915 exclamation point). Metadata nodes can have any values as their operand. For
2916 example:</p>
2917
2918<div class="doc_code">
2919<pre>
2920!{ metadata !"test\00", i32 10}
2921</pre>
2922</div>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2923
2924<p>A <a href="#namedmetadatastructure">named metadata</a> is a collection of
2925 metadata nodes, which can be looked up in the module symbol table. For
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]2926 example:</p>
2927
2928<div class="doc_code">
2929<pre>
2930!foo = metadata !{!4, !3}
2931</pre>
2932</div>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2933
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2934<p>Metadata can be used as function arguments. Here <tt>llvm.dbg.value</tt>
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]2935 function is using two metadata arguments:</p>
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2936
Bill Wendling9ff5de92011-03-02 02:17:11 +0000[diff] [blame]2937<div class="doc_code">
2938<pre>
2939call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2940</pre>
2941</div>
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2942
2943<p>Metadata can be attached with an instruction. Here metadata <tt>!21</tt> is
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]2944 attached to the <tt>add</tt> instruction using the <tt>!dbg</tt>
2945 identifier:</p>
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2946
Bill Wendling9ff5de92011-03-02 02:17:11 +0000[diff] [blame]2947<div class="doc_code">
2948<pre>
2949%indvar.next = add i64 %indvar, 1, !dbg !21
2950</pre>
2951</div>
2952
Peter Collingbourne249d9532011-10-27 19:19:07 +0000[diff] [blame]2953<p>More information about specific metadata nodes recognized by the optimizers
2954 and code generator is found below.</p>
2955
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]2956<!-- _______________________________________________________________________ -->
Peter Collingbourne249d9532011-10-27 19:19:07 +0000[diff] [blame]2957<h4>
2958 <a name="tbaa">'<tt>tbaa</tt>' Metadata</a>
2959</h4>
2960
2961<div>
2962
2963<p>In LLVM IR, memory does not have types, so LLVM's own type system is not
2964 suitable for doing TBAA. Instead, metadata is added to the IR to describe
2965 a type system of a higher level language. This can be used to implement
2966 typical C/C++ TBAA, but it can also be used to implement custom alias
2967 analysis behavior for other languages.</p>
2968
2969<p>The current metadata format is very simple. TBAA metadata nodes have up to
2970 three fields, e.g.:</p>
2971
2972<div class="doc_code">
2973<pre>
2974!0 = metadata !{ metadata !"an example type tree" }
2975!1 = metadata !{ metadata !"int", metadata !0 }
2976!2 = metadata !{ metadata !"float", metadata !0 }
2977!3 = metadata !{ metadata !"const float", metadata !2, i64 1 }
2978</pre>
2979</div>
2980
2981<p>The first field is an identity field. It can be any value, usually
2982 a metadata string, which uniquely identifies the type. The most important
2983 name in the tree is the name of the root node. Two trees with
2984 different root node names are entirely disjoint, even if they
2985 have leaves with common names.</p>
2986
2987<p>The second field identifies the type's parent node in the tree, or
2988 is null or omitted for a root node. A type is considered to alias
2989 all of its descendants and all of its ancestors in the tree. Also,
2990 a type is considered to alias all types in other trees, so that
2991 bitcode produced from multiple front-ends is handled conservatively.</p>
2992
2993<p>If the third field is present, it's an integer which if equal to 1
2994 indicates that the type is "constant" (meaning
2995 <tt>pointsToConstantMemory</tt> should return true; see
2996 <a href="AliasAnalysis.html#OtherItfs">other useful
2997 <tt>AliasAnalysis</tt> methods</a>).</p>
2998
2999</div>
3000
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]3001<!-- _______________________________________________________________________ -->
Peter Collingbourne999f90b2011-10-27 19:19:14 +0000[diff] [blame]3002<h4>
Duncan Sands5e5c5f82012-04-14 12:36:06 +0000[diff] [blame]3003 <a name="fpmath">'<tt>fpmath</tt>' Metadata</a>
Peter Collingbourne999f90b2011-10-27 19:19:14 +0000[diff] [blame]3004</h4>
3005
3006<div>
3007
Duncan Sands5e5c5f82012-04-14 12:36:06 +0000[diff] [blame]3008<p><tt>fpmath</tt> metadata may be attached to any instruction of floating point
Duncan Sands8883c432012-04-16 16:28:59 +0000[diff] [blame]3009 type. It can be used to express the maximum acceptable error in the result of
3010 that instruction, in ULPs, thus potentially allowing the compiler to use a
3011 more efficient but less accurate method of computing it. The number of ULPs
3012 may also be the string <tt>"fast"</tt>, which tells the compiler that speed
3013 matters more than accuracy, so any fairly accurate method of computation is
3014 fine as long as it is quick. ULP is defined as follows:</p>
Peter Collingbourne999f90b2011-10-27 19:19:14 +0000[diff] [blame]3015
Bill Wendling0656e252011-11-09 19:33:56 +0000[diff] [blame]3016<blockquote>
3017
3018<p>If <tt>x</tt> is a real number that lies between two finite consecutive
3019 floating-point numbers <tt>a</tt> and <tt>b</tt>, without being equal to one
3020 of them, then <tt>ulp(x) = |b - a|</tt>, otherwise <tt>ulp(x)</tt> is the
3021 distance between the two non-equal finite floating-point numbers nearest
3022 <tt>x</tt>. Moreover, <tt>ulp(NaN)</tt> is <tt>NaN</tt>.</p>
3023
3024</blockquote>
Peter Collingbourne999f90b2011-10-27 19:19:14 +0000[diff] [blame]3025
Duncan Sands8883c432012-04-16 16:28:59 +0000[diff] [blame]3026<p>The metadata node shall consist of a single positive floating point number
3027 representing the maximum relative error, or the string <tt>"fast"</tt>.
3028 For example:</p>
Peter Collingbourne999f90b2011-10-27 19:19:14 +0000[diff] [blame]3029
3030<div class="doc_code">
3031<pre>
Duncan Sands8883c432012-04-16 16:28:59 +0000[diff] [blame]3032!0 = metadata !{ float 2.5 } ; maximum acceptable inaccuracy is 2.5 ULPs
3033!1 = metadata !{ !metadata !"fast" } ; potentially unbounded inaccuracy
Peter Collingbourne999f90b2011-10-27 19:19:14 +0000[diff] [blame]3034</pre>
3035</div>
3036
NAKAMURA Takumi9c55f592012-03-27 11:25:16 +0000[diff] [blame]3037</div>
3038
Rafael Espindola39dd3282012-03-24 00:14:51 +0000[diff] [blame]3039<!-- _______________________________________________________________________ -->
3040<h4>
3041 <a name="range">'<tt>range</tt>' Metadata</a>
3042</h4>
3043
3044<div>
3045<p><tt>range</tt> metadata may be attached only to loads of integer types. It
3046 expresses the possible ranges the loaded value is in. The ranges are
3047 represented with a flattened list of integers. The loaded value is known to
3048 be in the union of the ranges defined by each consecutive pair. Each pair
3049 has the following properties:</p>
3050<ul>
3051 <li>The type must match the type loaded by the instruction.</li>
3052 <li>The pair <tt>a,b</tt> represents the range <tt>[a,b)</tt>.</li>
3053 <li>Both <tt>a</tt> and <tt>b</tt> are constants.</li>
3054 <li>The range is allowed to wrap.</li>
3055 <li>The range should not represent the full or empty set. That is,
3056 <tt>a!=b</tt>. </li>
3057</ul>
3058
3059<p>Examples:</p>
3060<div class="doc_code">
3061<pre>
3062 %a = load i8* %x, align 1, !range !0 ; Can only be 0 or 1
3063 %b = load i8* %y, align 1, !range !1 ; Can only be 255 (-1), 0 or 1
3064 %c = load i8* %z, align 1, !range !2 ; Can only be 0, 1, 3, 4 or 5
3065...
3066!0 = metadata !{ i8 0, i8 2 }
3067!1 = metadata !{ i8 255, i8 2 }
3068!2 = metadata !{ i8 0, i8 2, i8 3, i8 6 }
3069</pre>
3070</div>
3071</div>
Peter Collingbourne999f90b2011-10-27 19:19:14 +0000[diff] [blame]3072</div>
3073
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]3074</div>
3075
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3076<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3077<h2>
Bill Wendlingb9d75a92012-02-11 11:59:36 +0000[diff] [blame]3078 <a name="module_flags">Module Flags Metadata</a>
3079</h2>
3080<!-- *********************************************************************** -->
3081
3082<div>
3083
3084<p>Information about the module as a whole is difficult to convey to LLVM's
3085 subsystems. The LLVM IR isn't sufficient to transmit this
3086 information. The <tt>llvm.module.flags</tt> named metadata exists in order to
3087 facilitate this. These flags are in the form of key / value pairs &mdash;
3088 much like a dictionary &mdash; making it easy for any subsystem who cares
3089 about a flag to look it up.</p>
3090
3091<p>The <tt>llvm.module.flags</tt> metadata contains a list of metadata
3092 triplets. Each triplet has the following form:</p>
3093
3094<ul>
3095 <li>The first element is a <i>behavior</i> flag, which specifies the behavior
3096 when two (or more) modules are merged together, and it encounters two (or
3097 more) metadata with the same ID. The supported behaviors are described
3098 below.</li>
3099
3100 <li>The second element is a metadata string that is a unique ID for the
3101 metadata. How each ID is interpreted is documented below.</li>
3102
3103 <li>The third element is the value of the flag.</li>
3104</ul>
3105
3106<p>When two (or more) modules are merged together, the resulting
3107 <tt>llvm.module.flags</tt> metadata is the union of the
3108 modules' <tt>llvm.module.flags</tt> metadata. The only exception being a flag
3109 with the <i>Override</i> behavior, which may override another flag's value
3110 (see below).</p>
3111
3112<p>The following behaviors are supported:</p>
3113
3114<table border="1" cellspacing="0" cellpadding="4">
3115 <tbody>
3116 <tr>
3117 <th>Value</th>
3118 <th>Behavior</th>
3119 </tr>
3120 <tr>
3121 <td>1</td>
3122 <td align="left">
Bill Wendlinga0edecf2012-03-06 09:17:04 +0000[diff] [blame]3123 <dl>
3124 <dt><b>Error</b></dt>
3125 <dd>Emits an error if two values disagree. It is an error to have an ID
3126 with both an Error and a Warning behavior.</dd>
3127 </dl>
Bill Wendlingb9d75a92012-02-11 11:59:36 +0000[diff] [blame]3128 </td>
3129 </tr>
3130 <tr>
3131 <td>2</td>
3132 <td align="left">
Bill Wendlinga0edecf2012-03-06 09:17:04 +0000[diff] [blame]3133 <dl>
3134 <dt><b>Warning</b></dt>
3135 <dd>Emits a warning if two values disagree.</dd>
3136 </dl>
Bill Wendlingb9d75a92012-02-11 11:59:36 +0000[diff] [blame]3137 </td>
3138 </tr>
3139 <tr>
3140 <td>3</td>
3141 <td align="left">
Bill Wendlinga0edecf2012-03-06 09:17:04 +0000[diff] [blame]3142 <dl>
3143 <dt><b>Require</b></dt>
3144 <dd>Emits an error when the specified value is not present or doesn't
3145 have the specified value. It is an error for two (or more)
3146 <tt>llvm.module.flags</tt> with the same ID to have the Require
3147 behavior but different values. There may be multiple Require flags
3148 per ID.</dd>
3149 </dl>
Bill Wendlingb9d75a92012-02-11 11:59:36 +0000[diff] [blame]3150 </td>
3151 </tr>
3152 <tr>
3153 <td>4</td>
3154 <td align="left">
Bill Wendlinga0edecf2012-03-06 09:17:04 +0000[diff] [blame]3155 <dl>
3156 <dt><b>Override</b></dt>
3157 <dd>Uses the specified value if the two values disagree. It is an
3158 error for two (or more) <tt>llvm.module.flags</tt> with the same
3159 ID to have the Override behavior but different values.</dd>
3160 </dl>
Bill Wendlingb9d75a92012-02-11 11:59:36 +0000[diff] [blame]3161 </td>
3162 </tr>
3163 </tbody>
3164</table>
3165
3166<p>An example of module flags:</p>
3167
3168<pre class="doc_code">
3169!0 = metadata !{ i32 1, metadata !"foo", i32 1 }
3170!1 = metadata !{ i32 4, metadata !"bar", i32 37 }
3171!2 = metadata !{ i32 2, metadata !"qux", i32 42 }
3172!3 = metadata !{ i32 3, metadata !"qux",
3173 metadata !{
3174 metadata !"foo", i32 1
3175 }
3176}
3177!llvm.module.flags = !{ !0, !1, !2, !3 }
3178</pre>
3179
3180<ul>
3181 <li><p>Metadata <tt>!0</tt> has the ID <tt>!"foo"</tt> and the value '1'. The
3182 behavior if two or more <tt>!"foo"</tt> flags are seen is to emit an
3183 error if their values are not equal.</p></li>
3184
3185 <li><p>Metadata <tt>!1</tt> has the ID <tt>!"bar"</tt> and the value '37'. The
3186 behavior if two or more <tt>!"bar"</tt> flags are seen is to use the
3187 value '37' if their values are not equal.</p></li>
3188
3189 <li><p>Metadata <tt>!2</tt> has the ID <tt>!"qux"</tt> and the value '42'. The
3190 behavior if two or more <tt>!"qux"</tt> flags are seen is to emit a
3191 warning if their values are not equal.</p></li>
3192
3193 <li><p>Metadata <tt>!3</tt> has the ID <tt>!"qux"</tt> and the value:</p>
3194
3195<pre class="doc_code">
3196metadata !{ metadata !"foo", i32 1 }
3197</pre>
Bill Wendlingf7b367c2012-02-16 01:10:50 +0000[diff] [blame]3198
Bill Wendlingb9d75a92012-02-11 11:59:36 +0000[diff] [blame]3199 <p>The behavior is to emit an error if the <tt>llvm.module.flags</tt> does
3200 not contain a flag with the ID <tt>!"foo"</tt> that has the value
3201 '1'. If two or more <tt>!"qux"</tt> flags exist, then they must have
3202 the same value or an error will be issued.</p></li>
3203</ul>
3204
Bill Wendlingf7b367c2012-02-16 01:10:50 +0000[diff] [blame]3205
3206<!-- ======================================================================= -->
3207<h3>
3208<a name="objc_gc_flags">Objective-C Garbage Collection Module Flags Metadata</a>
3209</h3>
3210
3211<div>
3212
3213<p>On the Mach-O platform, Objective-C stores metadata about garbage collection
3214 in a special section called "image info". The metadata consists of a version
3215 number and a bitmask specifying what types of garbage collection are
3216 supported (if any) by the file. If two or more modules are linked together
3217 their garbage collection metadata needs to be merged rather than appended
3218 together.</p>
3219
3220<p>The Objective-C garbage collection module flags metadata consists of the
3221 following key-value pairs:</p>
3222
3223<table border="1" cellspacing="0" cellpadding="4">
Bill Wendlingb3ef2232012-03-06 09:23:25 +0000[diff] [blame]3224 <col width="30%">
Bill Wendlingf7b367c2012-02-16 01:10:50 +0000[diff] [blame]3225 <tbody>
3226 <tr>
Bill Wendlinga0edecf2012-03-06 09:17:04 +0000[diff] [blame]3227 <th>Key</th>
Bill Wendlingf7b367c2012-02-16 01:10:50 +0000[diff] [blame]3228 <th>Value</th>
3229 </tr>
3230 <tr>
3231 <td><tt>Objective-C&nbsp;Version</tt></td>
3232 <td align="left"><b>[Required]</b> &mdash; The Objective-C ABI
3233 version. Valid values are 1 and 2.</td>
3234 </tr>
3235 <tr>
3236 <td><tt>Objective-C&nbsp;Image&nbsp;Info&nbsp;Version</tt></td>
3237 <td align="left"><b>[Required]</b> &mdash; The version of the image info
3238 section. Currently always 0.</td>
3239 </tr>
3240 <tr>
3241 <td><tt>Objective-C&nbsp;Image&nbsp;Info&nbsp;Section</tt></td>
3242 <td align="left"><b>[Required]</b> &mdash; The section to place the
3243 metadata. Valid values are <tt>"__OBJC, __image_info, regular"</tt> for
3244 Objective-C ABI version 1, and <tt>"__DATA,__objc_imageinfo, regular,
3245 no_dead_strip"</tt> for Objective-C ABI version 2.</td>
3246 </tr>
3247 <tr>
3248 <td><tt>Objective-C&nbsp;Garbage&nbsp;Collection</tt></td>
3249 <td align="left"><b>[Required]</b> &mdash; Specifies whether garbage
3250 collection is supported or not. Valid values are 0, for no garbage
3251 collection, and 2, for garbage collection supported.</td>
3252 </tr>
3253 <tr>
3254 <td><tt>Objective-C&nbsp;GC&nbsp;Only</tt></td>
3255 <td align="left"><b>[Optional]</b> &mdash; Specifies that only garbage
3256 collection is supported. If present, its value must be 6. This flag
3257 requires that the <tt>Objective-C Garbage Collection</tt> flag have the
3258 value 2.</td>
3259 </tr>
3260 </tbody>
3261</table>
3262
3263<p>Some important flag interactions:</p>
3264
3265<ul>
3266 <li>If a module with <tt>Objective-C Garbage Collection</tt> set to 0 is
3267 merged with a module with <tt>Objective-C Garbage Collection</tt> set to
3268 2, then the resulting module has the <tt>Objective-C Garbage
3269 Collection</tt> flag set to 0.</li>
3270
3271 <li>A module with <tt>Objective-C Garbage Collection</tt> set to 0 cannot be
3272 merged with a module with <tt>Objective-C GC Only</tt> set to 6.</li>
3273</ul>
3274
3275</div>
3276
Bill Wendlingb9d75a92012-02-11 11:59:36 +0000[diff] [blame]3277</div>
3278
3279<!-- *********************************************************************** -->
3280<h2>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3281 <a name="intrinsic_globals">Intrinsic Global Variables</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3282</h2>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3283<!-- *********************************************************************** -->
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3284<div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3285<p>LLVM has a number of "magic" global variables that contain data that affect
3286code generation or other IR semantics. These are documented here. All globals
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]3287of this sort should have a section specified as "<tt>llvm.metadata</tt>". This
3288section and all globals that start with "<tt>llvm.</tt>" are reserved for use
3289by LLVM.</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3290
3291<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3292<h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3293<a name="intg_used">The '<tt>llvm.used</tt>' Global Variable</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3294</h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3295
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3296<div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3297
3298<p>The <tt>@llvm.used</tt> global is an array with i8* element type which has <a
3299href="#linkage_appending">appending linkage</a>. This array contains a list of
3300pointers to global variables and functions which may optionally have a pointer
3301cast formed of bitcast or getelementptr. For example, a legal use of it is:</p>
3302
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3303<div class="doc_code">
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3304<pre>
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3305@X = global i8 4
3306@Y = global i32 123
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3307
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3308@llvm.used = appending global [2 x i8*] [
3309 i8* @X,
3310 i8* bitcast (i32* @Y to i8*)
3311], section "llvm.metadata"
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3312</pre>
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3313</div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3314
3315<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]3316 compiler, assembler, and linker are required to treat the symbol as if there
3317 is a reference to the global that it cannot see. For example, if a variable
3318 has internal linkage and no references other than that from
3319 the <tt>@llvm.used</tt> list, it cannot be deleted. This is commonly used to
3320 represent references from inline asms and other things the compiler cannot
3321 "see", and corresponds to "<tt>attribute((used))</tt>" in GNU C.</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3322
3323<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]3324 object file to prevent the assembler and linker from molesting the
3325 symbol.</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3326
3327</div>
3328
3329<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3330<h3>
3331 <a name="intg_compiler_used">
3332 The '<tt>llvm.compiler.used</tt>' Global Variable
3333 </a>
3334</h3>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]3335
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3336<div>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]3337
3338<p>The <tt>@llvm.compiler.used</tt> directive is the same as the
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3339 <tt>@llvm.used</tt> directive, except that it only prevents the compiler from
3340 touching the symbol. On targets that support it, this allows an intelligent
3341 linker to optimize references to the symbol without being impeded as it would
3342 be by <tt>@llvm.used</tt>.</p>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]3343
3344<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]3345 should not be exposed to source languages.</p>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]3346
3347</div>
3348
3349<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3350<h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3351<a name="intg_global_ctors">The '<tt>llvm.global_ctors</tt>' Global Variable</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3352</h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3353
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3354<div>
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3355
3356<div class="doc_code">
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]3357<pre>
3358%0 = type { i32, void ()* }
David Chisnall27195a52010-04-30 19:27:35 +0000[diff] [blame]3359@llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor }]
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]3360</pre>
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3361</div>
3362
3363<p>The <tt>@llvm.global_ctors</tt> array contains a list of constructor
3364 functions and associated priorities. The functions referenced by this array
3365 will be called in ascending order of priority (i.e. lowest first) when the
3366 module is loaded. The order of functions with the same priority is not
3367 defined.</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3368
3369</div>
3370
3371<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3372<h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3373<a name="intg_global_dtors">The '<tt>llvm.global_dtors</tt>' Global Variable</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3374</h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3375
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3376<div>
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3377
3378<div class="doc_code">
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]3379<pre>
3380%0 = type { i32, void ()* }
David Chisnall27195a52010-04-30 19:27:35 +0000[diff] [blame]3381@llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor }]
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]3382</pre>
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3383</div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3384
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3385<p>The <tt>@llvm.global_dtors</tt> array contains a list of destructor functions
3386 and associated priorities. The functions referenced by this array will be
3387 called in descending order of priority (i.e. highest first) when the module
3388 is loaded. The order of functions with the same priority is not defined.</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3389
3390</div>
3391
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3392</div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3393
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]3394<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3395<h2><a name="instref">Instruction Reference</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3396<!-- *********************************************************************** -->
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3397
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3398<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3399
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3400<p>The LLVM instruction set consists of several different classifications of
3401 instructions: <a href="#terminators">terminator
3402 instructions</a>, <a href="#binaryops">binary instructions</a>,
3403 <a href="#bitwiseops">bitwise binary instructions</a>,
3404 <a href="#memoryops">memory instructions</a>, and
3405 <a href="#otherops">other instructions</a>.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3406
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3407<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3408<h3>
3409 <a name="terminators">Terminator Instructions</a>
3410</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3411
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3412<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3413
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3414<p>As mentioned <a href="#functionstructure">previously</a>, every basic block
3415 in a program ends with a "Terminator" instruction, which indicates which
3416 block should be executed after the current block is finished. These
3417 terminator instructions typically yield a '<tt>void</tt>' value: they produce
3418 control flow, not values (the one exception being the
3419 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
3420
Chris Lattner6445ecb2011-08-02 20:29:13 +0000[diff] [blame]3421<p>The terminator instructions are:
3422 '<a href="#i_ret"><tt>ret</tt></a>',
3423 '<a href="#i_br"><tt>br</tt></a>',
3424 '<a href="#i_switch"><tt>switch</tt></a>',
3425 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>',
3426 '<a href="#i_invoke"><tt>invoke</tt></a>',
Chris Lattner6445ecb2011-08-02 20:29:13 +0000[diff] [blame]3427 '<a href="#i_resume"><tt>resume</tt></a>', and
3428 '<a href="#i_unreachable"><tt>unreachable</tt></a>'.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3429
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3430<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3431<h4>
3432 <a name="i_ret">'<tt>ret</tt>' Instruction</a>
3433</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3434
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3435<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3436
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3437<h5>Syntax:</h5>
Dan Gohmanb1e6b962008-10-04 19:00:07 +0000[diff] [blame]3438<pre>
3439 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]3440 ret void <i>; Return from void function</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3441</pre>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3442
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3443<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3444<p>The '<tt>ret</tt>' instruction is used to return control flow (and optionally
3445 a value) from a function back to the caller.</p>
3446
3447<p>There are two forms of the '<tt>ret</tt>' instruction: one that returns a
3448 value and then causes control flow, and one that just causes control flow to
3449 occur.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3450
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3451<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3452<p>The '<tt>ret</tt>' instruction optionally accepts a single argument, the
3453 return value. The type of the return value must be a
3454 '<a href="#t_firstclass">first class</a>' type.</p>
Dan Gohmanb1e6b962008-10-04 19:00:07 +0000[diff] [blame]3455
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3456<p>A function is not <a href="#wellformed">well formed</a> if it it has a
3457 non-void return type and contains a '<tt>ret</tt>' instruction with no return
3458 value or a return value with a type that does not match its type, or if it
3459 has a void return type and contains a '<tt>ret</tt>' instruction with a
3460 return value.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3461
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3462<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3463<p>When the '<tt>ret</tt>' instruction is executed, control flow returns back to
3464 the calling function's context. If the caller is a
3465 "<a href="#i_call"><tt>call</tt></a>" instruction, execution continues at the
3466 instruction after the call. If the caller was an
3467 "<a href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues at
3468 the beginning of the "normal" destination block. If the instruction returns
3469 a value, that value shall set the call or invoke instruction's return
3470 value.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3471
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3472<h5>Example:</h5>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3473<pre>
3474 ret i32 5 <i>; Return an integer value of 5</i>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]3475 ret void <i>; Return from a void function</i>
Bill Wendling0a4bbbf2009-02-28 22:12:54 +0000[diff] [blame]3476 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]3477</pre>
Dan Gohmanf3e60bd2009-01-12 23:12:39 +0000[diff] [blame]3478
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3479</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3480<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3481<h4>
3482 <a name="i_br">'<tt>br</tt>' Instruction</a>
3483</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3484
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3485<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3486
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3487<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3488<pre>
Bill Wendlingb3aa4712011-07-26 10:41:15 +0000[diff] [blame]3489 br i1 &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;
3490 br label &lt;dest&gt; <i>; Unconditional branch</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3491</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3492
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3493<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3494<p>The '<tt>br</tt>' instruction is used to cause control flow to transfer to a
3495 different basic block in the current function. There are two forms of this
3496 instruction, corresponding to a conditional branch and an unconditional
3497 branch.</p>
3498
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3499<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3500<p>The conditional branch form of the '<tt>br</tt>' instruction takes a single
3501 '<tt>i1</tt>' value and two '<tt>label</tt>' values. The unconditional form
3502 of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>' value as a
3503 target.</p>
3504
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3505<h5>Semantics:</h5>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3506<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3507 argument is evaluated. If the value is <tt>true</tt>, control flows to the
3508 '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
3509 control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
3510
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3511<h5>Example:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]3512<pre>
3513Test:
3514 %cond = <a href="#i_icmp">icmp</a> eq i32 %a, %b
3515 br i1 %cond, label %IfEqual, label %IfUnequal
3516IfEqual:
3517 <a href="#i_ret">ret</a> i32 1
3518IfUnequal:
3519 <a href="#i_ret">ret</a> i32 0
3520</pre>
3521
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3522</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3523
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3524<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3525<h4>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3526 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3527</h4>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3528
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3529<div>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3530
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3531<h5>Syntax:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3532<pre>
3533 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
3534</pre>
3535
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3536<h5>Overview:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3537<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]3538 several different places. It is a generalization of the '<tt>br</tt>'
3539 instruction, allowing a branch to occur to one of many possible
3540 destinations.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3541
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3542<h5>Arguments:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3543<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3544 comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination,
3545 and an array of pairs of comparison value constants and '<tt>label</tt>'s.
3546 The table is not allowed to contain duplicate constant entries.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3547
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3548<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3549<p>The <tt>switch</tt> instruction specifies a table of values and
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3550 destinations. When the '<tt>switch</tt>' instruction is executed, this table
3551 is searched for the given value. If the value is found, control flow is
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]3552 transferred to the corresponding destination; otherwise, control flow is
3553 transferred to the default destination.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3554
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3555<h5>Implementation:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3556<p>Depending on properties of the target machine and the particular
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3557 <tt>switch</tt> instruction, this instruction may be code generated in
3558 different ways. For example, it could be generated as a series of chained
3559 conditional branches or with a lookup table.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3560
3561<h5>Example:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3562<pre>
3563 <i>; Emulate a conditional br instruction</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3564 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Dan Gohman2a08c532009-01-04 23:44:43 +0000[diff] [blame]3565 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3566
3567 <i>; Emulate an unconditional br instruction</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3568 switch i32 0, label %dest [ ]
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3569
3570 <i>; Implement a jump table:</i>
Dan Gohman2a08c532009-01-04 23:44:43 +0000[diff] [blame]3571 switch i32 %val, label %otherwise [ i32 0, label %onzero
3572 i32 1, label %onone
3573 i32 2, label %ontwo ]
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3574</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3575
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3576</div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3577
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3578
3579<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3580<h4>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3581 <a name="i_indirectbr">'<tt>indirectbr</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3582</h4>
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3583
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3584<div>
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3585
3586<h5>Syntax:</h5>
3587<pre>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3588 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]3589</pre>
3590
3591<h5>Overview:</h5>
3592
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3593<p>The '<tt>indirectbr</tt>' instruction implements an indirect branch to a label
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3594 within the current function, whose address is specified by
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]3595 "<tt>address</tt>". Address must be derived from a <a
3596 href="#blockaddress">blockaddress</a> constant.</p>
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3597
3598<h5>Arguments:</h5>
3599
3600<p>The '<tt>address</tt>' argument is the address of the label to jump to. The
3601 rest of the arguments indicate the full set of possible destinations that the
3602 address may point to. Blocks are allowed to occur multiple times in the
3603 destination list, though this isn't particularly useful.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3604
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3605<p>This destination list is required so that dataflow analysis has an accurate
3606 understanding of the CFG.</p>
3607
3608<h5>Semantics:</h5>
3609
3610<p>Control transfers to the block specified in the address argument. All
3611 possible destination blocks must be listed in the label list, otherwise this
3612 instruction has undefined behavior. This implies that jumps to labels
3613 defined in other functions have undefined behavior as well.</p>
3614
3615<h5>Implementation:</h5>
3616
3617<p>This is typically implemented with a jump through a register.</p>
3618
3619<h5>Example:</h5>
3620<pre>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3621 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3622</pre>
3623
3624</div>
3625
3626
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3627<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3628<h4>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3629 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3630</h4>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3631
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3632<div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3633
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3634<h5>Syntax:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3635<pre>
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]3636 &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]3637 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3638</pre>
3639
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]3640<h5>Overview:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3641<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3642 function, with the possibility of control flow transfer to either the
3643 '<tt>normal</tt>' label or the '<tt>exception</tt>' label. If the callee
3644 function returns with the "<tt><a href="#i_ret">ret</a></tt>" instruction,
3645 control flow will return to the "normal" label. If the callee (or any
Bill Wendling7b9e5392012-02-06 21:57:33 +0000[diff] [blame]3646 indirect callees) returns via the "<a href="#i_resume"><tt>resume</tt></a>"
3647 instruction or other exception handling mechanism, control is interrupted and
3648 continued at the dynamically nearest "exception" label.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3649
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]3650<p>The '<tt>exception</tt>' label is a
3651 <i><a href="ExceptionHandling.html#overview">landing pad</a></i> for the
3652 exception. As such, '<tt>exception</tt>' label is required to have the
3653 "<a href="#i_landingpad"><tt>landingpad</tt></a>" instruction, which contains
Chad Rosier85f5a1a2011-12-09 02:00:44 +0000[diff] [blame]3654 the information about the behavior of the program after unwinding
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]3655 happens, as its first non-PHI instruction. The restrictions on the
3656 "<tt>landingpad</tt>" instruction's tightly couples it to the
3657 "<tt>invoke</tt>" instruction, so that the important information contained
3658 within the "<tt>landingpad</tt>" instruction can't be lost through normal
3659 code motion.</p>
3660
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3661<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3662<p>This instruction requires several arguments:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3663
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3664<ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3665 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
3666 convention</a> the call should use. If none is specified, the call
3667 defaults to using C calling conventions.</li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]3668
3669 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3670 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
3671 '<tt>inreg</tt>' attributes are valid here.</li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]3672
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3673 <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]3674 function value being invoked. In most cases, this is a direct function
3675 invocation, but indirect <tt>invoke</tt>s are just as possible, branching
3676 off an arbitrary pointer to function value.</li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3677
3678 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3679 function to be invoked. </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3680
3681 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]3682 signature argument types and parameter attributes. All arguments must be
3683 of <a href="#t_firstclass">first class</a> type. If the function
3684 signature indicates the function accepts a variable number of arguments,
3685 the extra arguments can be specified.</li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3686
3687 <li>'<tt>normal label</tt>': the label reached when the called function
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3688 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3689
Bill Wendling7b9e5392012-02-06 21:57:33 +0000[diff] [blame]3690 <li>'<tt>exception label</tt>': the label reached when a callee returns via
3691 the <a href="#i_resume"><tt>resume</tt></a> instruction or other exception
3692 handling mechanism.</li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3693
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]3694 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3695 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
3696 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3697</ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3698
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3699<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3700<p>This instruction is designed to operate as a standard
3701 '<tt><a href="#i_call">call</a></tt>' instruction in most regards. The
3702 primary difference is that it establishes an association with a label, which
3703 is used by the runtime library to unwind the stack.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3704
3705<p>This instruction is used in languages with destructors to ensure that proper
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3706 cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
3707 exception. Additionally, this is important for implementation of
3708 '<tt>catch</tt>' clauses in high-level languages that support them.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3709
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3710<p>For the purposes of the SSA form, the definition of the value returned by the
3711 '<tt>invoke</tt>' instruction is deemed to occur on the edge from the current
3712 block to the "normal" label. If the callee unwinds then no return value is
3713 available.</p>
Dan Gohmanf96a4992009-05-22 21:47:08 +0000[diff] [blame]3714
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3715<h5>Example:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3716<pre>
Nick Lewyckyd703f652008-03-16 07:18:12 +0000[diff] [blame]3717 %retval = invoke i32 @Test(i32 15) to label %Continue
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]3718 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewyckyd703f652008-03-16 07:18:12 +0000[diff] [blame]3719 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]3720 unwind label %TestCleanup <i>; {i32}:retval set</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3721</pre>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3722
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3723</div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3724
Bill Wendlingdccc03b2011-07-31 06:30:59 +0000[diff] [blame]3725 <!-- _______________________________________________________________________ -->
3726
3727<h4>
3728 <a name="i_resume">'<tt>resume</tt>' Instruction</a>
3729</h4>
3730
3731<div>
3732
3733<h5>Syntax:</h5>
3734<pre>
3735 resume &lt;type&gt; &lt;value&gt;
3736</pre>
3737
3738<h5>Overview:</h5>
3739<p>The '<tt>resume</tt>' instruction is a terminator instruction that has no
3740 successors.</p>
3741
3742<h5>Arguments:</h5>
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]3743<p>The '<tt>resume</tt>' instruction requires one argument, which must have the
Bill Wendlinge4ad50b2011-08-03 18:37:32 +0000[diff] [blame]3744 same type as the result of any '<tt>landingpad</tt>' instruction in the same
3745 function.</p>
Bill Wendlingdccc03b2011-07-31 06:30:59 +0000[diff] [blame]3746
3747<h5>Semantics:</h5>
3748<p>The '<tt>resume</tt>' instruction resumes propagation of an existing
3749 (in-flight) exception whose unwinding was interrupted with
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]3750 a <a href="#i_landingpad"><tt>landingpad</tt></a> instruction.</p>
Bill Wendlingdccc03b2011-07-31 06:30:59 +0000[diff] [blame]3751
3752<h5>Example:</h5>
3753<pre>
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]3754 resume { i8*, i32 } %exn
Bill Wendlingdccc03b2011-07-31 06:30:59 +0000[diff] [blame]3755</pre>
3756
3757</div>
3758
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3759<!-- _______________________________________________________________________ -->
3760
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3761<h4>
3762 <a name="i_unreachable">'<tt>unreachable</tt>' Instruction</a>
3763</h4>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3764
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3765<div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3766
3767<h5>Syntax:</h5>
3768<pre>
3769 unreachable
3770</pre>
3771
3772<h5>Overview:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3773<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3774 instruction is used to inform the optimizer that a particular portion of the
3775 code is not reachable. This can be used to indicate that the code after a
3776 no-return function cannot be reached, and other facts.</p>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3777
3778<h5>Semantics:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3779<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3780
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3781</div>
3782
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3783</div>
3784
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3785<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3786<h3>
3787 <a name="binaryops">Binary Operations</a>
3788</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3789
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3790<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3791
3792<p>Binary operators are used to do most of the computation in a program. They
3793 require two operands of the same type, execute an operation on them, and
3794 produce a single value. The operands might represent multiple data, as is
3795 the case with the <a href="#t_vector">vector</a> data type. The result value
3796 has the same type as its operands.</p>
3797
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3798<p>There are several different binary operators:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3799
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3800<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3801<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3802 <a name="i_add">'<tt>add</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3803</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3804
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3805<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3806
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3807<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3808<pre>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3809 &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]3810 &lt;result&gt; = add nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3811 &lt;result&gt; = add nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3812 &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]3813</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3814
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3815<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3816<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3817
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3818<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3819<p>The two arguments to the '<tt>add</tt>' instruction must
3820 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3821 integer values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3822
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3823<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3824<p>The value produced is the integer sum of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3825
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3826<p>If the sum has unsigned overflow, the result returned is the mathematical
3827 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]3828
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3829<p>Because LLVM integers use a two's complement representation, this instruction
3830 is appropriate for both signed and unsigned integers.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3831
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3832<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3833 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3834 <tt>nsw</tt> keywords are present, the result value of the <tt>add</tt>
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]3835 is a <a href="#poisonvalues">poison value</a> if unsigned and/or signed overflow,
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]3836 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3837
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3838<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3839<pre>
3840 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3841</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3842
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3843</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3844
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3845<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3846<h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3847 <a name="i_fadd">'<tt>fadd</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3848</h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3849
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3850<div>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3851
3852<h5>Syntax:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3853<pre>
3854 &lt;result&gt; = fadd &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3855</pre>
3856
3857<h5>Overview:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3858<p>The '<tt>fadd</tt>' instruction returns the sum of its two operands.</p>
3859
3860<h5>Arguments:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3861<p>The two arguments to the '<tt>fadd</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3862 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3863 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3864
3865<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3866<p>The value produced is the floating point sum of the two operands.</p>
3867
3868<h5>Example:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3869<pre>
3870 &lt;result&gt; = fadd float 4.0, %var <i>; yields {float}:result = 4.0 + %var</i>
3871</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3872
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3873</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3874
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3875<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3876<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3877 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3878</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3879
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3880<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3881
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3882<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3883<pre>
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3884 &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]3885 &lt;result&gt; = sub nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3886 &lt;result&gt; = sub nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3887 &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]3888</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3889
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3890<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3891<p>The '<tt>sub</tt>' instruction returns the difference of its two
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3892 operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3893
3894<p>Note that the '<tt>sub</tt>' instruction is used to represent the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3895 '<tt>neg</tt>' instruction present in most other intermediate
3896 representations.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3897
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3898<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3899<p>The two arguments to the '<tt>sub</tt>' instruction must
3900 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3901 integer values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3902
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3903<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3904<p>The value produced is the integer difference of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3905
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3906<p>If the difference has unsigned overflow, the result returned is the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3907 mathematical result modulo 2<sup>n</sup>, where n is the bit width of the
3908 result.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3909
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3910<p>Because LLVM integers use a two's complement representation, this instruction
3911 is appropriate for both signed and unsigned integers.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3912
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3913<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3914 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3915 <tt>nsw</tt> keywords are present, the result value of the <tt>sub</tt>
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]3916 is a <a href="#poisonvalues">poison value</a> if unsigned and/or signed overflow,
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]3917 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3918
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3919<h5>Example:</h5>
Bill Wendlingaac388b2007-05-29 09:42:13 +0000[diff] [blame]3920<pre>
3921 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3922 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3923</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3924
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3925</div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3926
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3927<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3928<h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3929 <a name="i_fsub">'<tt>fsub</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3930</h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3931
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3932<div>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3933
3934<h5>Syntax:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3935<pre>
3936 &lt;result&gt; = fsub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3937</pre>
3938
3939<h5>Overview:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3940<p>The '<tt>fsub</tt>' instruction returns the difference of its two
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3941 operands.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3942
3943<p>Note that the '<tt>fsub</tt>' instruction is used to represent the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3944 '<tt>fneg</tt>' instruction present in most other intermediate
3945 representations.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3946
3947<h5>Arguments:</h5>
Bill Wendlingd9fe2982009-07-20 02:32:41 +0000[diff] [blame]3948<p>The two arguments to the '<tt>fsub</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3949 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3950 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3951
3952<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3953<p>The value produced is the floating point difference of the two operands.</p>
3954
3955<h5>Example:</h5>
3956<pre>
3957 &lt;result&gt; = fsub float 4.0, %var <i>; yields {float}:result = 4.0 - %var</i>
3958 &lt;result&gt; = fsub float -0.0, %val <i>; yields {float}:result = -%var</i>
3959</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3960
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3961</div>
3962
3963<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3964<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3965 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3966</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3967
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3968<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3969
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3970<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3971<pre>
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3972 &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]3973 &lt;result&gt; = mul nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3974 &lt;result&gt; = mul nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3975 &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]3976</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3977
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3978<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3979<p>The '<tt>mul</tt>' instruction returns the product of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3980
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3981<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3982<p>The two arguments to the '<tt>mul</tt>' instruction must
3983 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3984 integer values. Both arguments must have identical types.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3985
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3986<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3987<p>The value produced is the integer product of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3988
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3989<p>If the result of the multiplication has unsigned overflow, the result
3990 returned is the mathematical result modulo 2<sup>n</sup>, where n is the bit
3991 width of the result.</p>
3992
3993<p>Because LLVM integers use a two's complement representation, and the result
3994 is the same width as the operands, this instruction returns the correct
3995 result for both signed and unsigned integers. If a full product
3996 (e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands should
3997 be sign-extended or zero-extended as appropriate to the width of the full
3998 product.</p>
3999
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]4000<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
4001 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
4002 <tt>nsw</tt> keywords are present, the result value of the <tt>mul</tt>
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]4003 is a <a href="#poisonvalues">poison value</a> if unsigned and/or signed overflow,
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]4004 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]4005
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4006<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4007<pre>
4008 &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4009</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4010
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4011</div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4012
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4013<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4014<h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]4015 <a name="i_fmul">'<tt>fmul</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4016</h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]4017
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4018<div>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]4019
4020<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4021<pre>
4022 &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]4023</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4024
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]4025<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4026<p>The '<tt>fmul</tt>' instruction returns the product of its two operands.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]4027
4028<h5>Arguments:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]4029<p>The two arguments to the '<tt>fmul</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4030 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
4031 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]4032
4033<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]4034<p>The value produced is the floating point product of the two operands.</p>
4035
4036<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4037<pre>
4038 &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]4039</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4040
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]4041</div>
4042
4043<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4044<h4>
4045 <a name="i_udiv">'<tt>udiv</tt>' Instruction</a>
4046</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4047
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4048<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4049
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4050<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4051<pre>
Chris Lattner35bda892011-02-06 21:44:57 +0000[diff] [blame]4052 &lt;result&gt; = udiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4053 &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]4054</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4055
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4056<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4057<p>The '<tt>udiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4058
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4059<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4060<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4061 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4062 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4063
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4064<h5>Semantics:</h5>
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4065<p>The value produced is the unsigned integer quotient of the two operands.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4066
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4067<p>Note that unsigned integer division and signed integer division are distinct
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4068 operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
4069
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4070<p>Division by zero leads to undefined behavior.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4071
Chris Lattner35bda892011-02-06 21:44:57 +0000[diff] [blame]4072<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]4073 <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]4074 multiple of %op2 (as such, "((a udiv exact b) mul b) == a").</p>
4075
4076
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4077<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4078<pre>
4079 &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4080</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4081
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4082</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4083
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4084<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4085<h4>
4086 <a name="i_sdiv">'<tt>sdiv</tt>' Instruction</a>
4087</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4088
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4089<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4090
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4091<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4092<pre>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]4093 &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]4094 &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]4095</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4096
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4097<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4098<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4099
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4100<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4101<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4102 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4103 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4104
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4105<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4106<p>The value produced is the signed integer quotient of the two operands rounded
4107 towards zero.</p>
4108
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4109<p>Note that signed integer division and unsigned integer division are distinct
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4110 operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
4111
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4112<p>Division by zero leads to undefined behavior. Overflow also leads to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4113 undefined behavior; this is a rare case, but can occur, for example, by doing
4114 a 32-bit division of -2147483648 by -1.</p>
4115
Dan Gohman9c5beed2009-07-22 00:04:19 +0000[diff] [blame]4116<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]4117 <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]4118 be rounded.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]4119
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4120<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4121<pre>
4122 &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4123</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4124
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4125</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4126
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4127<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4128<h4>
4129 <a name="i_fdiv">'<tt>fdiv</tt>' Instruction</a>
4130</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4131
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4132<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4133
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4134<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4135<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]4136 &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]4137</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4138
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4139<h5>Overview:</h5>
4140<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4141
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4142<h5>Arguments:</h5>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]4143<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4144 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
4145 floating point values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4146
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4147<h5>Semantics:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4148<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4149
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4150<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4151<pre>
4152 &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]4153</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4154
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4155</div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4156
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4157<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4158<h4>
4159 <a name="i_urem">'<tt>urem</tt>' Instruction</a>
4160</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4161
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4162<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4163
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4164<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4165<pre>
4166 &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]4167</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4168
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4169<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4170<p>The '<tt>urem</tt>' instruction returns the remainder from the unsigned
4171 division of its two arguments.</p>
4172
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4173<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4174<p>The two arguments to the '<tt>urem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4175 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4176 values. Both arguments must have identical types.</p>
4177
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4178<h5>Semantics:</h5>
4179<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4180 This instruction always performs an unsigned division to get the
4181 remainder.</p>
4182
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4183<p>Note that unsigned integer remainder and signed integer remainder are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4184 distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
4185
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4186<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4187
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4188<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4189<pre>
4190 &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4191</pre>
4192
4193</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4194
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4195<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4196<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4197 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4198</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4199
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4200<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4201
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4202<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4203<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]4204 &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]4205</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4206
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4207<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4208<p>The '<tt>srem</tt>' instruction returns the remainder from the signed
4209 division of its two operands. This instruction can also take
4210 <a href="#t_vector">vector</a> versions of the values in which case the
4211 elements must be integers.</p>
Chris Lattnerc7d3ab32008-01-04 04:33:49 +0000[diff] [blame]4212
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4213<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4214<p>The two arguments to the '<tt>srem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4215 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4216 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4217
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4218<h5>Semantics:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4219<p>This instruction returns the <i>remainder</i> of a division (where the result
Duncan Sandsdea3a5e2011-03-07 09:12:24 +0000[diff] [blame]4220 is either zero or has the same sign as the dividend, <tt>op1</tt>), not the
4221 <i>modulo</i> operator (where the result is either zero or has the same sign
4222 as the divisor, <tt>op2</tt>) of a value.
4223 For more information about the difference,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4224 see <a href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
4225 Math Forum</a>. For a table of how this is implemented in various languages,
4226 please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
4227 Wikipedia: modulo operation</a>.</p>
4228
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4229<p>Note that signed integer remainder and unsigned integer remainder are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4230 distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
4231
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4232<p>Taking the remainder of a division by zero leads to undefined behavior.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4233 Overflow also leads to undefined behavior; this is a rare case, but can
4234 occur, for example, by taking the remainder of a 32-bit division of
4235 -2147483648 by -1. (The remainder doesn't actually overflow, but this rule
4236 lets srem be implemented using instructions that return both the result of
4237 the division and the remainder.)</p>
4238
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4239<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4240<pre>
4241 &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4242</pre>
4243
4244</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4245
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4246<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4247<h4>
4248 <a name="i_frem">'<tt>frem</tt>' Instruction</a>
4249</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4250
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4251<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4252
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4253<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4254<pre>
4255 &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]4256</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4257
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4258<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4259<p>The '<tt>frem</tt>' instruction returns the remainder from the division of
4260 its two operands.</p>
4261
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4262<h5>Arguments:</h5>
4263<p>The two arguments to the '<tt>frem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4264 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
4265 floating point values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4266
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4267<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4268<p>This instruction returns the <i>remainder</i> of a division. The remainder
4269 has the same sign as the dividend.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4270
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4271<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4272<pre>
4273 &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]4274</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4275
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4276</div>
Robert Bocchino7b81c752006-02-17 21:18:08 +0000[diff] [blame]4277
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4278</div>
4279
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]4280<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4281<h3>
4282 <a name="bitwiseops">Bitwise Binary Operations</a>
4283</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4284
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4285<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4286
4287<p>Bitwise binary operators are used to do various forms of bit-twiddling in a
4288 program. They are generally very efficient instructions and can commonly be
4289 strength reduced from other instructions. They require two operands of the
4290 same type, execute an operation on them, and produce a single value. The
4291 resulting value is the same type as its operands.</p>
4292
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4293<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4294<h4>
4295 <a name="i_shl">'<tt>shl</tt>' Instruction</a>
4296</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4297
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4298<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4299
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4300<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4301<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4302 &lt;result&gt; = shl &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4303 &lt;result&gt; = shl nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4304 &lt;result&gt; = shl nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4305 &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]4306</pre>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4307
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4308<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4309<p>The '<tt>shl</tt>' instruction returns the first operand shifted to the left
4310 a specified number of bits.</p>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4311
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4312<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4313<p>Both arguments to the '<tt>shl</tt>' instruction must be the
4314 same <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
4315 integer type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4316
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4317<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4318<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod
4319 2<sup>n</sup>, where <tt>n</tt> is the width of the result. If <tt>op2</tt>
4320 is (statically or dynamically) negative or equal to or larger than the number
4321 of bits in <tt>op1</tt>, the result is undefined. If the arguments are
4322 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
4323 shift amount in <tt>op2</tt>.</p>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4324
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4325<p>If the <tt>nuw</tt> keyword is present, then the shift produces a
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]4326 <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]4327 the <tt>nsw</tt> keyword is present, then the shift produces a
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]4328 <a href="#poisonvalues">poison value</a> if it shifts out any bits that disagree
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4329 with the resultant sign bit. As such, NUW/NSW have the same semantics as
4330 they would if the shift were expressed as a mul instruction with the same
4331 nsw/nuw bits in (mul %op1, (shl 1, %op2)).</p>
4332
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4333<h5>Example:</h5>
4334<pre>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4335 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
4336 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
4337 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4338 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +0000[diff] [blame]4339 &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]4340</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4341
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4342</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4343
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4344<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4345<h4>
4346 <a name="i_lshr">'<tt>lshr</tt>' Instruction</a>
4347</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4348
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4349<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4350
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4351<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4352<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4353 &lt;result&gt; = lshr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4354 &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]4355</pre>
4356
4357<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4358<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
4359 operand shifted to the right a specified number of bits with zero fill.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4360
4361<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4362<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4363 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4364 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4365
4366<h5>Semantics:</h5>
4367<p>This instruction always performs a logical shift right operation. The most
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4368 significant bits of the result will be filled with zero bits after the shift.
4369 If <tt>op2</tt> is (statically or dynamically) equal to or larger than the
4370 number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
4371 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
4372 shift amount in <tt>op2</tt>.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4373
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4374<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]4375 <tt>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]4376 shifted out are non-zero.</p>
4377
4378
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4379<h5>Example:</h5>
4380<pre>
4381 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
4382 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
4383 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
4384 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4385 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +0000[diff] [blame]4386 &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]4387</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4388
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4389</div>
4390
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]4391<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4392<h4>
4393 <a name="i_ashr">'<tt>ashr</tt>' Instruction</a>
4394</h4>
4395
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4396<div>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4397
4398<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4399<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4400 &lt;result&gt; = ashr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4401 &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]4402</pre>
4403
4404<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4405<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
4406 operand shifted to the right a specified number of bits with sign
4407 extension.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4408
4409<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4410<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4411 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4412 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4413
4414<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4415<p>This instruction always performs an arithmetic shift right operation, The
4416 most significant bits of the result will be filled with the sign bit
4417 of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
4418 larger than the number of bits in <tt>op1</tt>, the result is undefined. If
4419 the arguments are vectors, each vector element of <tt>op1</tt> is shifted by
4420 the corresponding shift amount in <tt>op2</tt>.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4421
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4422<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]4423 <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]4424 shifted out are non-zero.</p>
4425
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4426<h5>Example:</h5>
4427<pre>
4428 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
4429 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
4430 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
4431 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4432 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +0000[diff] [blame]4433 &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]4434</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4435
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4436</div>
4437
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4438<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4439<h4>
4440 <a name="i_and">'<tt>and</tt>' Instruction</a>
4441</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4442
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4443<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4444
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4445<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4446<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]4447 &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]4448</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4449
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4450<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4451<p>The '<tt>and</tt>' instruction returns the bitwise logical and of its two
4452 operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4453
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4454<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4455<p>The two arguments to the '<tt>and</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4456 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4457 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4458
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4459<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4460<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4461
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4462<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4463 <tbody>
4464 <tr>
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]4465 <th>In0</th>
4466 <th>In1</th>
4467 <th>Out</th>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4468 </tr>
4469 <tr>
4470 <td>0</td>
4471 <td>0</td>
4472 <td>0</td>
4473 </tr>
4474 <tr>
4475 <td>0</td>
4476 <td>1</td>
4477 <td>0</td>
4478 </tr>
4479 <tr>
4480 <td>1</td>
4481 <td>0</td>
4482 <td>0</td>
4483 </tr>
4484 <tr>
4485 <td>1</td>
4486 <td>1</td>
4487 <td>1</td>
4488 </tr>
4489 </tbody>
4490</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4491
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4492<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4493<pre>
4494 &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]4495 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
4496 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4497</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4498</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4499<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4500<h4>
4501 <a name="i_or">'<tt>or</tt>' Instruction</a>
4502</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4503
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4504<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4505
4506<h5>Syntax:</h5>
4507<pre>
4508 &lt;result&gt; = or &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4509</pre>
4510
4511<h5>Overview:</h5>
4512<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive or of its
4513 two operands.</p>
4514
4515<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4516<p>The two arguments to the '<tt>or</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4517 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4518 values. Both arguments must have identical types.</p>
4519
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4520<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4521<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4522
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4523<table border="1" cellspacing="0" cellpadding="4">
4524 <tbody>
4525 <tr>
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]4526 <th>In0</th>
4527 <th>In1</th>
4528 <th>Out</th>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4529 </tr>
4530 <tr>
4531 <td>0</td>
4532 <td>0</td>
4533 <td>0</td>
4534 </tr>
4535 <tr>
4536 <td>0</td>
4537 <td>1</td>
4538 <td>1</td>
4539 </tr>
4540 <tr>
4541 <td>1</td>
4542 <td>0</td>
4543 <td>1</td>
4544 </tr>
4545 <tr>
4546 <td>1</td>
4547 <td>1</td>
4548 <td>1</td>
4549 </tr>
4550 </tbody>
4551</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4552
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4553<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4554<pre>
4555 &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4556 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
4557 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4558</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4559
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4560</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4561
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4562<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4563<h4>
4564 <a name="i_xor">'<tt>xor</tt>' Instruction</a>
4565</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4566
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4567<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4568
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4569<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4570<pre>
4571 &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]4572</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4573
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4574<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4575<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive or of
4576 its two operands. The <tt>xor</tt> is used to implement the "one's
4577 complement" operation, which is the "~" operator in C.</p>
4578
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4579<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4580<p>The two arguments to the '<tt>xor</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4581 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4582 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4583
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4584<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4585<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4586
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4587<table border="1" cellspacing="0" cellpadding="4">
4588 <tbody>
4589 <tr>
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]4590 <th>In0</th>
4591 <th>In1</th>
4592 <th>Out</th>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4593 </tr>
4594 <tr>
4595 <td>0</td>
4596 <td>0</td>
4597 <td>0</td>
4598 </tr>
4599 <tr>
4600 <td>0</td>
4601 <td>1</td>
4602 <td>1</td>
4603 </tr>
4604 <tr>
4605 <td>1</td>
4606 <td>0</td>
4607 <td>1</td>
4608 </tr>
4609 <tr>
4610 <td>1</td>
4611 <td>1</td>
4612 <td>0</td>
4613 </tr>
4614 </tbody>
4615</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4616
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4617<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4618<pre>
4619 &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4620 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
4621 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
4622 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4623</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4624
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4625</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4626
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4627</div>
4628
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4629<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4630<h3>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4631 <a name="vectorops">Vector Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4632</h3>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4633
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4634<div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4635
4636<p>LLVM supports several instructions to represent vector operations in a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4637 target-independent manner. These instructions cover the element-access and
4638 vector-specific operations needed to process vectors effectively. While LLVM
4639 does directly support these vector operations, many sophisticated algorithms
4640 will want to use target-specific intrinsics to take full advantage of a
4641 specific target.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4642
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4643<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4644<h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4645 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4646</h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4647
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4648<div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4649
4650<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4651<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4652 &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]4653</pre>
4654
4655<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4656<p>The '<tt>extractelement</tt>' instruction extracts a single scalar element
4657 from a vector at a specified index.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4658
4659
4660<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4661<p>The first operand of an '<tt>extractelement</tt>' instruction is a value
4662 of <a href="#t_vector">vector</a> type. The second operand is an index
4663 indicating the position from which to extract the element. The index may be
4664 a variable.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4665
4666<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4667<p>The result is a scalar of the same type as the element type of
4668 <tt>val</tt>. Its value is the value at position <tt>idx</tt> of
4669 <tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4670 results are undefined.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4671
4672<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4673<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +0000[diff] [blame]4674 &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]4675</pre>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4676
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4677</div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4678
4679<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4680<h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4681 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4682</h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4683
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4684<div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4685
4686<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4687<pre>
Dan Gohmanf3480b92008-05-12 23:38:42 +0000[diff] [blame]4688 &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]4689</pre>
4690
4691<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4692<p>The '<tt>insertelement</tt>' instruction inserts a scalar element into a
4693 vector at a specified index.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4694
4695<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4696<p>The first operand of an '<tt>insertelement</tt>' instruction is a value
4697 of <a href="#t_vector">vector</a> type. The second operand is a scalar value
4698 whose type must equal the element type of the first operand. The third
4699 operand is an index indicating the position at which to insert the value.
4700 The index may be a variable.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4701
4702<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4703<p>The result is a vector of the same type as <tt>val</tt>. Its element values
4704 are those of <tt>val</tt> except at position <tt>idx</tt>, where it gets the
4705 value <tt>elt</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4706 results are undefined.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4707
4708<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4709<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +0000[diff] [blame]4710 &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]4711</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4712
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4713</div>
4714
4715<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4716<h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4717 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4718</h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4719
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4720<div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4721
4722<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4723<pre>
Mon P Wangaeb06d22008-11-10 04:46:22 +0000[diff] [blame]4724 &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]4725</pre>
4726
4727<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4728<p>The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
4729 from two input vectors, returning a vector with the same element type as the
4730 input and length that is the same as the shuffle mask.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4731
4732<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4733<p>The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
4734 with types that match each other. The third argument is a shuffle mask whose
4735 element type is always 'i32'. The result of the instruction is a vector
4736 whose length is the same as the shuffle mask and whose element type is the
4737 same as the element type of the first two operands.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4738
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4739<p>The shuffle mask operand is required to be a constant vector with either
4740 constant integer or undef values.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4741
4742<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4743<p>The elements of the two input vectors are numbered from left to right across
4744 both of the vectors. The shuffle mask operand specifies, for each element of
4745 the result vector, which element of the two input vectors the result element
4746 gets. The element selector may be undef (meaning "don't care") and the
4747 second operand may be undef if performing a shuffle from only one vector.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4748
4749<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4750<pre>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4751 &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]4752 &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]4753 &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]4754 &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]4755 &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]4756 &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]4757 &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]4758 &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]4759</pre>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4760
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4761</div>
Tanya Lattner09474292006-04-14 19:24:33 +0000[diff] [blame]4762
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4763</div>
4764
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4765<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4766<h3>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4767 <a name="aggregateops">Aggregate Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4768</h3>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4769
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4770<div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4771
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4772<p>LLVM supports several instructions for working with
4773 <a href="#t_aggregate">aggregate</a> values.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4774
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4775<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4776<h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4777 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4778</h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4779
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4780<div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4781
4782<h5>Syntax:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4783<pre>
4784 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
4785</pre>
4786
4787<h5>Overview:</h5>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4788<p>The '<tt>extractvalue</tt>' instruction extracts the value of a member field
4789 from an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4790
4791<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4792<p>The first operand of an '<tt>extractvalue</tt>' instruction is a value
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]4793 of <a href="#t_struct">struct</a> or
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4794 <a href="#t_array">array</a> type. The operands are constant indices to
4795 specify which value to extract in a similar manner as indices in a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4796 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
Frits van Bommel13242892010-12-05 20:54:38 +0000[diff] [blame]4797 <p>The major differences to <tt>getelementptr</tt> indexing are:</p>
4798 <ul>
4799 <li>Since the value being indexed is not a pointer, the first index is
4800 omitted and assumed to be zero.</li>
4801 <li>At least one index must be specified.</li>
4802 <li>Not only struct indices but also array indices must be in
4803 bounds.</li>
4804 </ul>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4805
4806<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4807<p>The result is the value at the position in the aggregate specified by the
4808 index operands.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4809
4810<h5>Example:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4811<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +0000[diff] [blame]4812 &lt;result&gt; = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4813</pre>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4814
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4815</div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4816
4817<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4818<h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4819 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4820</h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4821
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4822<div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4823
4824<h5>Syntax:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4825<pre>
Bill Wendling194229e2011-07-26 20:42:28 +0000[diff] [blame]4826 &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]4827</pre>
4828
4829<h5>Overview:</h5>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4830<p>The '<tt>insertvalue</tt>' instruction inserts a value into a member field
4831 in an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4832
4833<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4834<p>The first operand of an '<tt>insertvalue</tt>' instruction is a value
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]4835 of <a href="#t_struct">struct</a> or
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4836 <a href="#t_array">array</a> type. The second operand is a first-class
4837 value to insert. The following operands are constant indices indicating
4838 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]4839 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' instruction. The
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4840 value to insert must have the same type as the value identified by the
4841 indices.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4842
4843<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4844<p>The result is an aggregate of the same type as <tt>val</tt>. Its value is
4845 that of <tt>val</tt> except that the value at the position specified by the
4846 indices is that of <tt>elt</tt>.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4847
4848<h5>Example:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4849<pre>
Chris Lattner8645d1a2011-05-22 07:18:08 +0000[diff] [blame]4850 %agg1 = insertvalue {i32, float} undef, i32 1, 0 <i>; yields {i32 1, float undef}</i>
4851 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 <i>; yields {i32 1, float %val}</i>
4852 %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]4853</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4854
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4855</div>
4856
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4857</div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4858
4859<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4860<h3>
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]4861 <a name="memoryops">Memory Access and Addressing Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4862</h3>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4863
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4864<div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4865
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4866<p>A key design point of an SSA-based representation is how it represents
4867 memory. In LLVM, no memory locations are in SSA form, which makes things
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]4868 very simple. This section describes how to read, write, and allocate
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4869 memory in LLVM.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4870
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4871<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4872<h4>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4873 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4874</h4>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4875
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4876<div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4877
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4878<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4879<pre>
Dan Gohmanf75a7d32010-05-28 01:14:11 +0000[diff] [blame]4880 &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]4881</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4882
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4883<h5>Overview:</h5>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]4884<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4885 currently executing function, to be automatically released when this function
4886 returns to its caller. The object is always allocated in the generic address
4887 space (address space zero).</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4888
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4889<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4890<p>The '<tt>alloca</tt>' instruction
4891 allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt> bytes of memory on the
4892 runtime stack, returning a pointer of the appropriate type to the program.
4893 If "NumElements" is specified, it is the number of elements allocated,
4894 otherwise "NumElements" is defaulted to be one. If a constant alignment is
4895 specified, the value result of the allocation is guaranteed to be aligned to
4896 at least that boundary. If not specified, or if zero, the target can choose
4897 to align the allocation on any convenient boundary compatible with the
4898 type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4899
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4900<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4901
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4902<h5>Semantics:</h5>
Bill Wendling871eb0a2009-05-08 20:49:29 +0000[diff] [blame]4903<p>Memory is allocated; a pointer is returned. The operation is undefined if
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4904 there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
4905 memory is automatically released when the function returns. The
4906 '<tt>alloca</tt>' instruction is commonly used to represent automatic
4907 variables that must have an address available. When the function returns
4908 (either with the <tt><a href="#i_ret">ret</a></tt>
Bill Wendling7b9e5392012-02-06 21:57:33 +0000[diff] [blame]4909 or <tt><a href="#i_resume">resume</a></tt> instructions), the memory is
Nick Lewycky84a1d232012-02-29 08:26:44 +0000[diff] [blame]4910 reclaimed. Allocating zero bytes is legal, but the result is undefined.
4911 The order in which memory is allocated (ie., which way the stack grows) is
Nick Lewycky75d05e62012-03-18 09:35:50 +0000[diff] [blame]4912 not specified.</p>
Nick Lewycky84a1d232012-02-29 08:26:44 +0000[diff] [blame]4913
4914<p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4915
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4916<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4917<pre>
Dan Gohman81e21672009-01-04 23:49:44 +0000[diff] [blame]4918 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
4919 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
4920 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
4921 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4922</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4923
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4924</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4925
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4926<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4927<h4>
4928 <a name="i_load">'<tt>load</tt>' Instruction</a>
4929</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4930
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4931<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4932
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4933<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4934<pre>
Pete Cooperf95acc62012-02-10 18:13:54 +0000[diff] [blame]4935 &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]4936 &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]4937 !&lt;index&gt; = !{ i32 1 }
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4938</pre>
4939
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4940<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4941<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4942
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4943<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4944<p>The argument to the '<tt>load</tt>' instruction specifies the memory address
4945 from which to load. The pointer must point to
4946 a <a href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
4947 marked as <tt>volatile</tt>, then the optimizer is not allowed to modify the
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]4948 number or order of execution of this <tt>load</tt> with other <a
4949 href="#volatile">volatile operations</a>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4950
Eli Friedman21006d42011-08-09 23:02:53 +0000[diff] [blame]4951<p>If the <code>load</code> is marked as <code>atomic</code>, it takes an extra
4952 <a href="#ordering">ordering</a> and optional <code>singlethread</code>
4953 argument. The <code>release</code> and <code>acq_rel</code> orderings are
4954 not valid on <code>load</code> instructions. Atomic loads produce <a
4955 href="#memorymodel">defined</a> results when they may see multiple atomic
4956 stores. The type of the pointee must be an integer type whose bit width
4957 is a power of two greater than or equal to eight and less than or equal
4958 to a target-specific size limit. <code>align</code> must be explicitly
4959 specified on atomic loads, and the load has undefined behavior if the
4960 alignment is not set to a value which is at least the size in bytes of
4961 the pointee. <code>!nontemporal</code> does not have any defined semantics
4962 for atomic loads.</p>
4963
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4964<p>The optional constant <tt>align</tt> argument specifies the alignment of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4965 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]4966 omitted <tt>align</tt> argument means that the operation has the preferential
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4967 alignment for the target. It is the responsibility of the code emitter to
4968 ensure that the alignment information is correct. Overestimating the
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4969 alignment results in undefined behavior. Underestimating the alignment may
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4970 produce less efficient code. An alignment of 1 is always safe.</p>
4971
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4972<p>The optional <tt>!nontemporal</tt> metadata must reference a single
4973 metatadata name &lt;index&gt; corresponding to a metadata node with
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]4974 one <tt>i32</tt> entry of value 1. The existence of
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4975 the <tt>!nontemporal</tt> metatadata on the instruction tells the optimizer
4976 and code generator that this load is not expected to be reused in the cache.
4977 The code generator may select special instructions to save cache bandwidth,
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]4978 such as the <tt>MOVNT</tt> instruction on x86.</p>
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]4979
Pete Cooperf95acc62012-02-10 18:13:54 +0000[diff] [blame]4980<p>The optional <tt>!invariant.load</tt> metadata must reference a single
4981 metatadata name &lt;index&gt; corresponding to a metadata node with no
4982 entries. The existence of the <tt>!invariant.load</tt> metatadata on the
4983 instruction tells the optimizer and code generator that this load address
4984 points to memory which does not change value during program execution.
4985 The optimizer may then move this load around, for example, by hoisting it
4986 out of loops using loop invariant code motion.</p>
4987
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4988<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4989<p>The location of memory pointed to is loaded. If the value being loaded is of
4990 scalar type then the number of bytes read does not exceed the minimum number
4991 of bytes needed to hold all bits of the type. For example, loading an
4992 <tt>i24</tt> reads at most three bytes. When loading a value of a type like
4993 <tt>i20</tt> with a size that is not an integral number of bytes, the result
4994 is undefined if the value was not originally written using a store of the
4995 same type.</p>
4996
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4997<h5>Examples:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4998<pre>
4999 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
5000 <a href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5001 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]5002</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5003
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5004</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5005
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]5006<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5007<h4>
5008 <a name="i_store">'<tt>store</tt>' Instruction</a>
5009</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5010
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5011<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5012
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]5013<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5014<pre>
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]5015 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>
5016 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]5017</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5018
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]5019<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5020<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5021
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]5022<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5023<p>There are two arguments to the '<tt>store</tt>' instruction: a value to store
5024 and an address at which to store it. The type of the
5025 '<tt>&lt;pointer&gt;</tt>' operand must be a pointer to
5026 the <a href="#t_firstclass">first class</a> type of the
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]5027 '<tt>&lt;value&gt;</tt>' operand. If the <tt>store</tt> is marked as
5028 <tt>volatile</tt>, then the optimizer is not allowed to modify the number or
5029 order of execution of this <tt>store</tt> with other <a
5030 href="#volatile">volatile operations</a>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5031
Eli Friedman21006d42011-08-09 23:02:53 +0000[diff] [blame]5032<p>If the <code>store</code> is marked as <code>atomic</code>, it takes an extra
5033 <a href="#ordering">ordering</a> and optional <code>singlethread</code>
5034 argument. The <code>acquire</code> and <code>acq_rel</code> orderings aren't
5035 valid on <code>store</code> instructions. Atomic loads produce <a
5036 href="#memorymodel">defined</a> results when they may see multiple atomic
5037 stores. The type of the pointee must be an integer type whose bit width
5038 is a power of two greater than or equal to eight and less than or equal
5039 to a target-specific size limit. <code>align</code> must be explicitly
5040 specified on atomic stores, and the store has undefined behavior if the
5041 alignment is not set to a value which is at least the size in bytes of
5042 the pointee. <code>!nontemporal</code> does not have any defined semantics
5043 for atomic stores.</p>
5044
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5045<p>The optional constant "align" argument specifies the alignment of the
5046 operation (that is, the alignment of the memory address). A value of 0 or an
5047 omitted "align" argument means that the operation has the preferential
5048 alignment for the target. It is the responsibility of the code emitter to
5049 ensure that the alignment information is correct. Overestimating the
5050 alignment results in an undefined behavior. Underestimating the alignment may
5051 produce less efficient code. An alignment of 1 is always safe.</p>
5052
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]5053<p>The optional !nontemporal metadata must reference a single metatadata
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]5054 name &lt;index&gt; corresponding to a metadata node with one i32 entry of
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]5055 value 1. The existence of the !nontemporal metatadata on the
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]5056 instruction tells the optimizer and code generator that this load is
5057 not expected to be reused in the cache. The code generator may
5058 select special instructions to save cache bandwidth, such as the
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]5059 MOVNT instruction on x86.</p>
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]5060
5061
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]5062<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5063<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>' at the
5064 location specified by the '<tt>&lt;pointer&gt;</tt>' operand. If
5065 '<tt>&lt;value&gt;</tt>' is of scalar type then the number of bytes written
5066 does not exceed the minimum number of bytes needed to hold all bits of the
5067 type. For example, storing an <tt>i24</tt> writes at most three bytes. When
5068 writing a value of a type like <tt>i20</tt> with a size that is not an
5069 integral number of bytes, it is unspecified what happens to the extra bits
5070 that do not belong to the type, but they will typically be overwritten.</p>
5071
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]5072<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5073<pre>
5074 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling8c6c72d2007-10-22 05:10:05 +0000[diff] [blame]5075 store i32 3, i32* %ptr <i>; yields {void}</i>
5076 %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]5077</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5078
Reid Spencer47ce1792006-11-09 21:15:49 +0000[diff] [blame]5079</div>
5080
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]5081<!-- _______________________________________________________________________ -->
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]5082<h4>
5083<a name="i_fence">'<tt>fence</tt>' Instruction</a>
5084</h4>
Eli Friedman47f35132011-07-25 23:16:38 +0000[diff] [blame]5085
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]5086<div>
Eli Friedman47f35132011-07-25 23:16:38 +0000[diff] [blame]5087
5088<h5>Syntax:</h5>
5089<pre>
5090 fence [singlethread] &lt;ordering&gt; <i>; yields {void}</i>
5091</pre>
5092
5093<h5>Overview:</h5>
5094<p>The '<tt>fence</tt>' instruction is used to introduce happens-before edges
5095between operations.</p>
5096
5097<h5>Arguments:</h5> <p>'<code>fence</code>' instructions take an <a
5098href="#ordering">ordering</a> argument which defines what
5099<i>synchronizes-with</i> edges they add. They can only be given
5100<code>acquire</code>, <code>release</code>, <code>acq_rel</code>, and
5101<code>seq_cst</code> orderings.</p>
5102
5103<h5>Semantics:</h5>
5104<p>A fence <var>A</var> which has (at least) <code>release</code> ordering
5105semantics <i>synchronizes with</i> a fence <var>B</var> with (at least)
5106<code>acquire</code> ordering semantics if and only if there exist atomic
5107operations <var>X</var> and <var>Y</var>, both operating on some atomic object
5108<var>M</var>, such that <var>A</var> is sequenced before <var>X</var>,
5109<var>X</var> modifies <var>M</var> (either directly or through some side effect
5110of a sequence headed by <var>X</var>), <var>Y</var> is sequenced before
5111<var>B</var>, and <var>Y</var> observes <var>M</var>. This provides a
5112<i>happens-before</i> dependency between <var>A</var> and <var>B</var>. Rather
5113than an explicit <code>fence</code>, one (but not both) of the atomic operations
5114<var>X</var> or <var>Y</var> might provide a <code>release</code> or
5115<code>acquire</code> (resp.) ordering constraint and still
5116<i>synchronize-with</i> the explicit <code>fence</code> and establish the
5117<i>happens-before</i> edge.</p>
5118
5119<p>A <code>fence</code> which has <code>seq_cst</code> ordering, in addition to
5120having both <code>acquire</code> and <code>release</code> semantics specified
5121above, participates in the global program order of other <code>seq_cst</code>
5122operations and/or fences.</p>
5123
5124<p>The optional "<a href="#singlethread"><code>singlethread</code></a>" argument
5125specifies that the fence only synchronizes with other fences in the same
5126thread. (This is useful for interacting with signal handlers.)</p>
5127
Eli Friedman47f35132011-07-25 23:16:38 +0000[diff] [blame]5128<h5>Example:</h5>
5129<pre>
5130 fence acquire <i>; yields {void}</i>
5131 fence singlethread seq_cst <i>; yields {void}</i>
5132</pre>
5133
5134</div>
5135
5136<!-- _______________________________________________________________________ -->
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]5137<h4>
5138<a name="i_cmpxchg">'<tt>cmpxchg</tt>' Instruction</a>
5139</h4>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]5140
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]5141<div>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]5142
5143<h5>Syntax:</h5>
5144<pre>
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]5145 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]5146</pre>
5147
5148<h5>Overview:</h5>
5149<p>The '<tt>cmpxchg</tt>' instruction is used to atomically modify memory.
5150It loads a value in memory and compares it to a given value. If they are
5151equal, it stores a new value into the memory.</p>
5152
5153<h5>Arguments:</h5>
5154<p>There are three arguments to the '<code>cmpxchg</code>' instruction: an
5155address to operate on, a value to compare to the value currently be at that
5156address, and a new value to place at that address if the compared values are
5157equal. The type of '<var>&lt;cmp&gt;</var>' must be an integer type whose
5158bit width is a power of two greater than or equal to eight and less than
5159or equal to a target-specific size limit. '<var>&lt;cmp&gt;</var>' and
5160'<var>&lt;new&gt;</var>' must have the same type, and the type of
5161'<var>&lt;pointer&gt;</var>' must be a pointer to that type. If the
5162<code>cmpxchg</code> is marked as <code>volatile</code>, then the
5163optimizer is not allowed to modify the number or order of execution
5164of this <code>cmpxchg</code> with other <a href="#volatile">volatile
5165operations</a>.</p>
5166
5167<!-- FIXME: Extend allowed types. -->
5168
5169<p>The <a href="#ordering"><var>ordering</var></a> argument specifies how this
5170<code>cmpxchg</code> synchronizes with other atomic operations.</p>
5171
5172<p>The optional "<code>singlethread</code>" argument declares that the
5173<code>cmpxchg</code> is only atomic with respect to code (usually signal
5174handlers) running in the same thread as the <code>cmpxchg</code>. Otherwise the
5175cmpxchg is atomic with respect to all other code in the system.</p>
5176
5177<p>The pointer passed into cmpxchg must have alignment greater than or equal to
5178the size in memory of the operand.
5179
5180<h5>Semantics:</h5>
5181<p>The contents of memory at the location specified by the
5182'<tt>&lt;pointer&gt;</tt>' operand is read and compared to
5183'<tt>&lt;cmp&gt;</tt>'; if the read value is the equal,
5184'<tt>&lt;new&gt;</tt>' is written. The original value at the location
5185is returned.
5186
5187<p>A successful <code>cmpxchg</code> is a read-modify-write instruction for the
5188purpose of identifying <a href="#release_sequence">release sequences</a>. A
5189failed <code>cmpxchg</code> is equivalent to an atomic load with an ordering
5190parameter determined by dropping any <code>release</code> part of the
5191<code>cmpxchg</code>'s ordering.</p>
5192
5193<!--
5194FIXME: Is compare_exchange_weak() necessary? (Consider after we've done
5195optimization work on ARM.)
5196
5197FIXME: Is a weaker ordering constraint on failure helpful in practice?
5198-->
5199
5200<h5>Example:</h5>
5201<pre>
5202entry:
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]5203 %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]5204 <a href="#i_br">br</a> label %loop
5205
5206loop:
5207 %cmp = <a href="#i_phi">phi</a> i32 [ %orig, %entry ], [%old, %loop]
5208 %squared = <a href="#i_mul">mul</a> i32 %cmp, %cmp
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]5209 %old = cmpxchg i32* %ptr, i32 %cmp, i32 %squared <i>; yields {i32}</i>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]5210 %success = <a href="#i_icmp">icmp</a> eq i32 %cmp, %old
5211 <a href="#i_br">br</a> i1 %success, label %done, label %loop
5212
5213done:
5214 ...
5215</pre>
5216
5217</div>
5218
5219<!-- _______________________________________________________________________ -->
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]5220<h4>
5221<a name="i_atomicrmw">'<tt>atomicrmw</tt>' Instruction</a>
5222</h4>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]5223
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]5224<div>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]5225
5226<h5>Syntax:</h5>
5227<pre>
Eli Friedmanf03bb262011-08-12 22:50:01 +0000[diff] [blame]5228 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]5229</pre>
5230
5231<h5>Overview:</h5>
5232<p>The '<tt>atomicrmw</tt>' instruction is used to atomically modify memory.</p>
5233
5234<h5>Arguments:</h5>
5235<p>There are three arguments to the '<code>atomicrmw</code>' instruction: an
5236operation to apply, an address whose value to modify, an argument to the
5237operation. The operation must be one of the following keywords:</p>
5238<ul>
5239 <li>xchg</li>
5240 <li>add</li>
5241 <li>sub</li>
5242 <li>and</li>
5243 <li>nand</li>
5244 <li>or</li>
5245 <li>xor</li>
5246 <li>max</li>
5247 <li>min</li>
5248 <li>umax</li>
5249 <li>umin</li>
5250</ul>
5251
5252<p>The type of '<var>&lt;value&gt;</var>' must be an integer type whose
5253bit width is a power of two greater than or equal to eight and less than
5254or equal to a target-specific size limit. The type of the
5255'<code>&lt;pointer&gt;</code>' operand must be a pointer to that type.
5256If the <code>atomicrmw</code> is marked as <code>volatile</code>, then the
5257optimizer is not allowed to modify the number or order of execution of this
5258<code>atomicrmw</code> with other <a href="#volatile">volatile
5259 operations</a>.</p>
5260
5261<!-- FIXME: Extend allowed types. -->
5262
5263<h5>Semantics:</h5>
5264<p>The contents of memory at the location specified by the
5265'<tt>&lt;pointer&gt;</tt>' operand are atomically read, modified, and written
5266back. The original value at the location is returned. The modification is
5267specified by the <var>operation</var> argument:</p>
5268
5269<ul>
5270 <li>xchg: <code>*ptr = val</code></li>
5271 <li>add: <code>*ptr = *ptr + val</code></li>
5272 <li>sub: <code>*ptr = *ptr - val</code></li>
5273 <li>and: <code>*ptr = *ptr &amp; val</code></li>
5274 <li>nand: <code>*ptr = ~(*ptr &amp; val)</code></li>
5275 <li>or: <code>*ptr = *ptr | val</code></li>
5276 <li>xor: <code>*ptr = *ptr ^ val</code></li>
5277 <li>max: <code>*ptr = *ptr &gt; val ? *ptr : val</code> (using a signed comparison)</li>
5278 <li>min: <code>*ptr = *ptr &lt; val ? *ptr : val</code> (using a signed comparison)</li>
5279 <li>umax: <code>*ptr = *ptr &gt; val ? *ptr : val</code> (using an unsigned comparison)</li>
5280 <li>umin: <code>*ptr = *ptr &lt; val ? *ptr : val</code> (using an unsigned comparison)</li>
5281</ul>
5282
5283<h5>Example:</h5>
5284<pre>
5285 %old = atomicrmw add i32* %ptr, i32 1 acquire <i>; yields {i32}</i>
5286</pre>
5287
5288</div>
5289
5290<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5291<h4>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5292 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5293</h4>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5294
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5295<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5296
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]5297<h5>Syntax:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5298<pre>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]5299 &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]5300 &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]5301 &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]5302</pre>
5303
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]5304<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5305<p>The '<tt>getelementptr</tt>' instruction is used to get the address of a
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]5306 subelement of an <a href="#t_aggregate">aggregate</a> data structure.
5307 It performs address calculation only and does not access memory.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5308
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]5309<h5>Arguments:</h5>
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5310<p>The first argument is always a pointer or a vector of pointers,
5311 and forms the basis of the
Chris Lattnerc8eef442009-07-29 06:44:13 +0000[diff] [blame]5312 calculation. The remaining arguments are indices that indicate which of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5313 elements of the aggregate object are indexed. The interpretation of each
5314 index is dependent on the type being indexed into. The first index always
5315 indexes the pointer value given as the first argument, the second index
5316 indexes a value of the type pointed to (not necessarily the value directly
5317 pointed to, since the first index can be non-zero), etc. The first type
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]5318 indexed into must be a pointer value, subsequent types can be arrays,
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]5319 vectors, and structs. Note that subsequent types being indexed into
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]5320 can never be pointers, since that would require loading the pointer before
5321 continuing calculation.</p>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]5322
5323<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]5324 When indexing into a (optionally packed) structure, only <tt>i32</tt>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]5325 integer <b>constants</b> are allowed. When indexing into an array, pointer
5326 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]5327 constant. These integers are treated as signed values where relevant.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5328
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5329<p>For example, let's consider a C code fragment and how it gets compiled to
5330 LLVM:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5331
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]5332<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5333struct RT {
5334 char A;
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]5335 int B[10][20];
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5336 char C;
5337};
5338struct ST {
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]5339 int X;
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5340 double Y;
5341 struct RT Z;
5342};
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5343
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]5344int *foo(struct ST *s) {
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5345 return &amp;s[1].Z.B[5][13];
5346}
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5347</pre>
5348
Bill Wendlinga3495392011-12-13 01:07:07 +0000[diff] [blame]5349<p>The LLVM code generated by Clang is:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5350
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]5351<pre class="doc_code">
Bill Wendlinga3495392011-12-13 01:07:07 +0000[diff] [blame]5352%struct.RT = <a href="#namedtypes">type</a> { i8, [10 x [20 x i32]], i8 }
5353%struct.ST = <a href="#namedtypes">type</a> { i32, double, %struct.RT }
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5354
Bill Wendlinga3495392011-12-13 01:07:07 +0000[diff] [blame]5355define i32* @foo(%struct.ST* %s) nounwind uwtable readnone optsize ssp {
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5356entry:
Bill Wendlinga3495392011-12-13 01:07:07 +0000[diff] [blame]5357 %arrayidx = getelementptr inbounds %struct.ST* %s, i64 1, i32 2, i32 1, i64 5, i64 13
5358 ret i32* %arrayidx
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5359}
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5360</pre>
5361
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]5362<h5>Semantics:</h5>
Bill Wendlinga3495392011-12-13 01:07:07 +0000[diff] [blame]5363<p>In the example above, the first index is indexing into the
5364 '<tt>%struct.ST*</tt>' type, which is a pointer, yielding a
5365 '<tt>%struct.ST</tt>' = '<tt>{ i32, double, %struct.RT }</tt>' type, a
5366 structure. The second index indexes into the third element of the structure,
5367 yielding a '<tt>%struct.RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]], i8 }</tt>'
5368 type, another structure. The third index indexes into the second element of
5369 the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an array. The
5370 two dimensions of the array are subscripted into, yielding an '<tt>i32</tt>'
5371 type. The '<tt>getelementptr</tt>' instruction returns a pointer to this
5372 element, thus computing a value of '<tt>i32*</tt>' type.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5373
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5374<p>Note that it is perfectly legal to index partially through a structure,
5375 returning a pointer to an inner element. Because of this, the LLVM code for
5376 the given testcase is equivalent to:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5377
Bill Wendlinga3495392011-12-13 01:07:07 +0000[diff] [blame]5378<pre class="doc_code">
5379define i32* @foo(%struct.ST* %s) {
5380 %t1 = getelementptr %struct.ST* %s, i32 1 <i>; yields %struct.ST*:%t1</i>
5381 %t2 = getelementptr %struct.ST* %t1, i32 0, i32 2 <i>; yields %struct.RT*:%t2</i>
5382 %t3 = getelementptr %struct.RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
5383 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
5384 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
5385 ret i32* %t5
5386}
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5387</pre>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]5388
Dan Gohmandd8004d2009-07-27 21:53:46 +0000[diff] [blame]5389<p>If the <tt>inbounds</tt> keyword is present, the result value of the
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]5390 <tt>getelementptr</tt> is a <a href="#poisonvalues">poison value</a> if the
Dan Gohman27ef9972010-04-23 15:23:32 +0000[diff] [blame]5391 base pointer is not an <i>in bounds</i> address of an allocated object,
5392 or if any of the addresses that would be formed by successive addition of
5393 the offsets implied by the indices to the base address with infinitely
Eli Friedman266246c2011-08-12 23:37:55 +0000[diff] [blame]5394 precise signed arithmetic are not an <i>in bounds</i> address of that
5395 allocated object. The <i>in bounds</i> addresses for an allocated object
5396 are all the addresses that point into the object, plus the address one
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5397 byte past the end.
5398 In cases where the base is a vector of pointers the <tt>inbounds</tt> keyword
5399 applies to each of the computations element-wise. </p>
Dan Gohmandd8004d2009-07-27 21:53:46 +0000[diff] [blame]5400
5401<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]5402 the base address with silently-wrapping two's complement arithmetic. If the
5403 offsets have a different width from the pointer, they are sign-extended or
5404 truncated to the width of the pointer. The result value of the
5405 <tt>getelementptr</tt> may be outside the object pointed to by the base
5406 pointer. The result value may not necessarily be used to access memory
5407 though, even if it happens to point into allocated storage. See the
5408 <a href="#pointeraliasing">Pointer Aliasing Rules</a> section for more
5409 information.</p>
Dan Gohmandd8004d2009-07-27 21:53:46 +0000[diff] [blame]5410
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5411<p>The getelementptr instruction is often confusing. For some more insight into
5412 how it works, see <a href="GetElementPtr.html">the getelementptr FAQ</a>.</p>
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]5413
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]5414<h5>Example:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5415<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5416 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]5417 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
5418 <i>; yields i8*:vptr</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5419 %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]5420 <i>; yields i8*:eptr</i>
5421 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Sanjiv Gupta9f805c22009-04-25 07:27:44 +0000[diff] [blame]5422 <i>; yields i32*:iptr</i>
Sanjiv Gupta16ffa802009-04-24 16:38:13 +0000[diff] [blame]5423 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5424</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5425
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5426<p>In cases where the pointer argument is a vector of pointers, only a
5427 single index may be used, and the number of vector elements has to be
5428 the same. For example: </p>
5429<pre class="doc_code">
5430 %A = getelementptr <4 x i8*> %ptrs, <4 x i64> %offsets,
5431</pre>
5432
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5433</div>
Reid Spencer47ce1792006-11-09 21:15:49 +0000[diff] [blame]5434
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5435</div>
5436
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]5437<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5438<h3>
5439 <a name="convertops">Conversion Operations</a>
5440</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5441
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5442<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5443
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5444<p>The instructions in this category are the conversion instructions (casting)
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5445 which all take a single operand and a type. They perform various bit
5446 conversions on the operand.</p>
5447
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5448<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5449<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5450 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5451</h4>
5452
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5453<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5454
5455<h5>Syntax:</h5>
5456<pre>
5457 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5458</pre>
5459
5460<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5461<p>The '<tt>trunc</tt>' instruction truncates its operand to the
5462 type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5463
5464<h5>Arguments:</h5>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]5465<p>The '<tt>trunc</tt>' instruction takes a value to trunc, and a type to trunc it to.
5466 Both types must be of <a href="#t_integer">integer</a> types, or vectors
5467 of the same number of integers.
5468 The bit size of the <tt>value</tt> must be larger than
5469 the bit size of the destination type, <tt>ty2</tt>.
5470 Equal sized types are not allowed.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5471
5472<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5473<p>The '<tt>trunc</tt>' instruction truncates the high order bits
5474 in <tt>value</tt> and converts the remaining bits to <tt>ty2</tt>. Since the
5475 source size must be larger than the destination size, <tt>trunc</tt> cannot
5476 be a <i>no-op cast</i>. It will always truncate bits.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5477
5478<h5>Example:</h5>
5479<pre>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]5480 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
5481 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
5482 %Z = trunc i32 122 to i1 <i>; yields i1:false</i>
5483 %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]5484</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5485
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5486</div>
5487
5488<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5489<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5490 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5491</h4>
5492
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5493<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5494
5495<h5>Syntax:</h5>
5496<pre>
5497 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5498</pre>
5499
5500<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5501<p>The '<tt>zext</tt>' instruction zero extends its operand to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5502 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5503
5504
5505<h5>Arguments:</h5>
Nadav Rotemed9b9342011-02-20 12:37:50 +0000[diff] [blame]5506<p>The '<tt>zext</tt>' instruction takes a value to cast, and a type to cast it to.
5507 Both types must be of <a href="#t_integer">integer</a> types, or vectors
5508 of the same number of integers.
5509 The bit size of the <tt>value</tt> must be smaller than
5510 the bit size of the destination type,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5511 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5512
5513<h5>Semantics:</h5>
5514<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]5515 bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5516
Reid Spencerb5929522007-01-12 15:46:11 +0000[diff] [blame]5517<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]5518
5519<h5>Example:</h5>
5520<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5521 %X = zext i32 257 to i64 <i>; yields i64:257</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]5522 %Y = zext i1 true to i32 <i>; yields i32:1</i>
Nadav Rotemed9b9342011-02-20 12:37:50 +0000[diff] [blame]5523 %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]5524</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5525
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5526</div>
5527
5528<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5529<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5530 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5531</h4>
5532
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5533<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5534
5535<h5>Syntax:</h5>
5536<pre>
5537 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5538</pre>
5539
5540<h5>Overview:</h5>
5541<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
5542
5543<h5>Arguments:</h5>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]5544<p>The '<tt>sext</tt>' instruction takes a value to cast, and a type to cast it to.
5545 Both types must be of <a href="#t_integer">integer</a> types, or vectors
5546 of the same number of integers.
5547 The bit size of the <tt>value</tt> must be smaller than
5548 the bit size of the destination type,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5549 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5550
5551<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5552<p>The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
5553 bit (highest order bit) of the <tt>value</tt> until it reaches the bit size
5554 of the type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5555
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]5556<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]5557
5558<h5>Example:</h5>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5559<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5560 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]5561 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]5562 %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]5563</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5564
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5565</div>
5566
5567<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5568<h4>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5569 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5570</h4>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5571
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5572<div>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5573
5574<h5>Syntax:</h5>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5575<pre>
5576 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5577</pre>
5578
5579<h5>Overview:</h5>
5580<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5581 <tt>ty2</tt>.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5582
5583<h5>Arguments:</h5>
5584<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5585 point</a> value to cast and a <a href="#t_floating">floating point</a> type
5586 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]5587 <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]5588 <i>no-op cast</i>.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5589
5590<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5591<p>The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5592 <a href="#t_floating">floating point</a> type to a smaller
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5593 <a href="#t_floating">floating point</a> type. If the value cannot fit
5594 within the destination type, <tt>ty2</tt>, then the results are
5595 undefined.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5596
5597<h5>Example:</h5>
5598<pre>
5599 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
5600 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
5601</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5602
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5603</div>
5604
5605<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5606<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5607 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5608</h4>
5609
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5610<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5611
5612<h5>Syntax:</h5>
5613<pre>
5614 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5615</pre>
5616
5617<h5>Overview:</h5>
5618<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]5619 floating point value.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5620
5621<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5622<p>The '<tt>fpext</tt>' instruction takes a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5623 <a href="#t_floating">floating point</a> <tt>value</tt> to cast, and
5624 a <a href="#t_floating">floating point</a> type to cast it to. The source
5625 type must be smaller than the destination type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5626
5627<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5628<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5629 <a href="#t_floating">floating point</a> type to a larger
5630 <a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
5631 used to make a <i>no-op cast</i> because it always changes bits. Use
5632 <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]5633
5634<h5>Example:</h5>
5635<pre>
Nick Lewycky5bb3ece2011-03-31 18:20:19 +0000[diff] [blame]5636 %X = fpext float 3.125 to double <i>; yields double:3.125000e+00</i>
5637 %Y = fpext double %X to fp128 <i>; yields fp128:0xL00000000000000004000900000000000</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5638</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5639
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5640</div>
5641
5642<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5643<h4>
Reid Spencer24d6da52007-01-21 00:29:26 +0000[diff] [blame]5644 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5645</h4>
5646
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5647<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5648
5649<h5>Syntax:</h5>
5650<pre>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]5651 &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]5652</pre>
5653
5654<h5>Overview:</h5>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]5655<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5656 unsigned integer equivalent of type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5657
5658<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5659<p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
5660 scalar or vector <a href="#t_floating">floating point</a> value, and a type
5661 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
5662 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
5663 vector integer type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5664
5665<h5>Semantics:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5666<p>The '<tt>fptoui</tt>' instruction converts its
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5667 <a href="#t_floating">floating point</a> operand into the nearest (rounding
5668 towards zero) unsigned integer value. If the value cannot fit
5669 in <tt>ty2</tt>, the results are undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5670
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5671<h5>Example:</h5>
5672<pre>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]5673 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner88519042007-09-22 03:17:52 +0000[diff] [blame]5674 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Gabor Greif6a292012009-10-28 09:21:30 +0000[diff] [blame]5675 %Z = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5676</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5677
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5678</div>
5679
5680<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5681<h4>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5682 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5683</h4>
5684
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5685<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5686
5687<h5>Syntax:</h5>
5688<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5689 &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]5690</pre>
5691
5692<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5693<p>The '<tt>fptosi</tt>' instruction converts
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5694 <a href="#t_floating">floating point</a> <tt>value</tt> to
5695 type <tt>ty2</tt>.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5696
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5697<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5698<p>The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
5699 scalar or vector <a href="#t_floating">floating point</a> value, and a type
5700 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
5701 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
5702 vector integer type with the same number of elements as <tt>ty</tt></p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5703
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5704<h5>Semantics:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5705<p>The '<tt>fptosi</tt>' instruction converts its
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5706 <a href="#t_floating">floating point</a> operand into the nearest (rounding
5707 towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
5708 the results are undefined.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5709
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]5710<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5711<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]5712 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner88519042007-09-22 03:17:52 +0000[diff] [blame]5713 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Gabor Greif6a292012009-10-28 09:21:30 +0000[diff] [blame]5714 %Z = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5715</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5716
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5717</div>
5718
5719<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5720<h4>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5721 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5722</h4>
5723
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5724<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5725
5726<h5>Syntax:</h5>
5727<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5728 &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]5729</pre>
5730
5731<h5>Overview:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5732<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5733 integer and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5734
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5735<h5>Arguments:</h5>
Nate Begemanb348d182007-11-17 03:58:34 +0000[diff] [blame]5736<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]5737 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
5738 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
5739 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
5740 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5741
5742<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5743<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5744 integer quantity and converts it to the corresponding floating point
5745 value. If the value cannot fit in the floating point value, the results are
5746 undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5747
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5748<h5>Example:</h5>
5749<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5750 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5751 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5752</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5753
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5754</div>
5755
5756<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5757<h4>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5758 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5759</h4>
5760
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5761<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5762
5763<h5>Syntax:</h5>
5764<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5765 &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]5766</pre>
5767
5768<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5769<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed integer
5770 and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5771
5772<h5>Arguments:</h5>
Nate Begemanb348d182007-11-17 03:58:34 +0000[diff] [blame]5773<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]5774 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
5775 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
5776 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
5777 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5778
5779<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5780<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed integer
5781 quantity and converts it to the corresponding floating point value. If the
5782 value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5783
5784<h5>Example:</h5>
5785<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5786 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5787 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5788</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5789
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5790</div>
5791
5792<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5793<h4>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5794 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5795</h4>
5796
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5797<div>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5798
5799<h5>Syntax:</h5>
5800<pre>
5801 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5802</pre>
5803
5804<h5>Overview:</h5>
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5805<p>The '<tt>ptrtoint</tt>' instruction converts the pointer or a vector of
5806 pointers <tt>value</tt> to
5807 the integer (or vector of integers) type <tt>ty2</tt>.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5808
5809<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5810<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5811 must be a a value of type <a href="#t_pointer">pointer</a> or a vector of
5812 pointers, and a type to cast it to
5813 <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> or a vector
5814 of integers type.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5815
5816<h5>Semantics:</h5>
5817<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5818 <tt>ty2</tt> by interpreting the pointer value as an integer and either
5819 truncating or zero extending that value to the size of the integer type. If
5820 <tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
5821 <tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
5822 are the same size, then nothing is done (<i>no-op cast</i>) other than a type
5823 change.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5824
5825<h5>Example:</h5>
5826<pre>
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5827 %X = ptrtoint i32* %P to i8 <i>; yields truncation on 32-bit architecture</i>
5828 %Y = ptrtoint i32* %P to i64 <i>; yields zero extension on 32-bit architecture</i>
5829 %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]5830</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5831
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5832</div>
5833
5834<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5835<h4>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5836 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5837</h4>
5838
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5839<div>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5840
5841<h5>Syntax:</h5>
5842<pre>
5843 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5844</pre>
5845
5846<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5847<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to a
5848 pointer type, <tt>ty2</tt>.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5849
5850<h5>Arguments:</h5>
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]5851<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5852 value to cast, and a type to cast it to, which must be a
5853 <a href="#t_pointer">pointer</a> type.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5854
5855<h5>Semantics:</h5>
5856<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5857 <tt>ty2</tt> by applying either a zero extension or a truncation depending on
5858 the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
5859 size of a pointer then a truncation is done. If <tt>value</tt> is smaller
5860 than the size of a pointer then a zero extension is done. If they are the
5861 same size, nothing is done (<i>no-op cast</i>).</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5862
5863<h5>Example:</h5>
5864<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5865 %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]5866 %Y = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
5867 %Z = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5868 %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]5869</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5870
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5871</div>
5872
5873<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5874<h4>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5875 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5876</h4>
5877
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5878<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5879
5880<h5>Syntax:</h5>
5881<pre>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5882 &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]5883</pre>
5884
5885<h5>Overview:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5886<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5887 <tt>ty2</tt> without changing any bits.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5888
5889<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5890<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be a
5891 non-aggregate first class value, and a type to cast it to, which must also be
5892 a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes
5893 of <tt>value</tt> and the destination type, <tt>ty2</tt>, must be
5894 identical. If the source type is a pointer, the destination type must also be
5895 a pointer. This instruction supports bitwise conversion of vectors to
5896 integers and to vectors of other types (as long as they have the same
5897 size).</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5898
5899<h5>Semantics:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5900<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5901 <tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
5902 this conversion. The conversion is done as if the <tt>value</tt> had been
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5903 stored to memory and read back as type <tt>ty2</tt>.
5904 Pointer (or vector of pointers) types may only be converted to other pointer
5905 (or vector of pointers) types with this instruction. To convert
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5906 pointers to other types, use the <a href="#i_inttoptr">inttoptr</a> or
5907 <a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5908
5909<h5>Example:</h5>
5910<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5911 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5912 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5913 %Z = bitcast &lt;2 x int&gt; %V to i64; <i>; yields i64: %V</i>
5914 %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]5915</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5916
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5917</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5918
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5919</div>
5920
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5921<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5922<h3>
5923 <a name="otherops">Other Operations</a>
5924</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5925
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5926<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5927
5928<p>The instructions in this category are the "miscellaneous" instructions, which
5929 defy better classification.</p>
5930
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5931<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5932<h4>
5933 <a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
5934</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5935
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5936<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5937
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5938<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5939<pre>
5940 &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]5941</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5942
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5943<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5944<p>The '<tt>icmp</tt>' instruction returns a boolean value or a vector of
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5945 boolean values based on comparison of its two integer, integer vector,
5946 pointer, or pointer vector operands.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5947
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5948<h5>Arguments:</h5>
5949<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5950 the condition code indicating the kind of comparison to perform. It is not a
5951 value, just a keyword. The possible condition code are:</p>
5952
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5953<ol>
5954 <li><tt>eq</tt>: equal</li>
5955 <li><tt>ne</tt>: not equal </li>
5956 <li><tt>ugt</tt>: unsigned greater than</li>
5957 <li><tt>uge</tt>: unsigned greater or equal</li>
5958 <li><tt>ult</tt>: unsigned less than</li>
5959 <li><tt>ule</tt>: unsigned less or equal</li>
5960 <li><tt>sgt</tt>: signed greater than</li>
5961 <li><tt>sge</tt>: signed greater or equal</li>
5962 <li><tt>slt</tt>: signed less than</li>
5963 <li><tt>sle</tt>: signed less or equal</li>
5964</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5965
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]5966<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5967 <a href="#t_pointer">pointer</a> or integer <a href="#t_vector">vector</a>
5968 typed. They must also be identical types.</p>
5969
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5970<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5971<p>The '<tt>icmp</tt>' compares <tt>op1</tt> and <tt>op2</tt> according to the
5972 condition code given as <tt>cond</tt>. The comparison performed always yields
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]5973 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]5974 result, as follows:</p>
5975
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5976<ol>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5977 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5978 <tt>false</tt> otherwise. No sign interpretation is necessary or
5979 performed.</li>
5980
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5981 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5982 <tt>false</tt> otherwise. No sign interpretation is necessary or
5983 performed.</li>
5984
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5985 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5986 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5987
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5988 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5989 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5990 to <tt>op2</tt>.</li>
5991
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5992 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5993 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5994
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5995 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5996 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5997
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5998 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5999 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
6000
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6001 <li><tt>sge</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6002 <tt>true</tt> if <tt>op1</tt> is greater than or equal
6003 to <tt>op2</tt>.</li>
6004
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6005 <li><tt>slt</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6006 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
6007
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6008 <li><tt>sle</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6009 <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]6010</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6011
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6012<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6013 values are compared as if they were integers.</p>
6014
6015<p>If the operands are integer vectors, then they are compared element by
6016 element. The result is an <tt>i1</tt> vector with the same number of elements
6017 as the values being compared. Otherwise, the result is an <tt>i1</tt>.</p>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6018
6019<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6020<pre>
6021 &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]6022 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
6023 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
6024 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
6025 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
6026 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6027</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +0000[diff] [blame]6028
6029<p>Note that the code generator does not yet support vector types with
6030 the <tt>icmp</tt> instruction.</p>
6031
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6032</div>
6033
6034<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6035<h4>
6036 <a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
6037</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6038
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6039<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6040
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6041<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6042<pre>
6043 &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]6044</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6045
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6046<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6047<p>The '<tt>fcmp</tt>' instruction returns a boolean value or vector of boolean
6048 values based on comparison of its operands.</p>
6049
6050<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]6051(<a href="#t_integer"><tt>i1</tt></a>).</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6052
6053<p>If the operands are floating point vectors, then the result type is a vector
6054 of boolean with the same number of elements as the operands being
6055 compared.</p>
6056
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6057<h5>Arguments:</h5>
6058<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6059 the condition code indicating the kind of comparison to perform. It is not a
6060 value, just a keyword. The possible condition code are:</p>
6061
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6062<ol>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]6063 <li><tt>false</tt>: no comparison, always returns false</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6064 <li><tt>oeq</tt>: ordered and equal</li>
6065 <li><tt>ogt</tt>: ordered and greater than </li>
6066 <li><tt>oge</tt>: ordered and greater than or equal</li>
6067 <li><tt>olt</tt>: ordered and less than </li>
6068 <li><tt>ole</tt>: ordered and less than or equal</li>
6069 <li><tt>one</tt>: ordered and not equal</li>
6070 <li><tt>ord</tt>: ordered (no nans)</li>
6071 <li><tt>ueq</tt>: unordered or equal</li>
6072 <li><tt>ugt</tt>: unordered or greater than </li>
6073 <li><tt>uge</tt>: unordered or greater than or equal</li>
6074 <li><tt>ult</tt>: unordered or less than </li>
6075 <li><tt>ule</tt>: unordered or less than or equal</li>
6076 <li><tt>une</tt>: unordered or not equal</li>
6077 <li><tt>uno</tt>: unordered (either nans)</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]6078 <li><tt>true</tt>: no comparison, always returns true</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6079</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6080
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6081<p><i>Ordered</i> means that neither operand is a QNAN while
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6082 <i>unordered</i> means that either operand may be a QNAN.</p>
6083
6084<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be either
6085 a <a href="#t_floating">floating point</a> type or
6086 a <a href="#t_vector">vector</a> of floating point type. They must have
6087 identical types.</p>
6088
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6089<h5>Semantics:</h5>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]6090<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6091 according to the condition code given as <tt>cond</tt>. If the operands are
6092 vectors, then the vectors are compared element by element. Each comparison
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]6093 performed always yields an <a href="#t_integer">i1</a> result, as
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6094 follows:</p>
6095
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6096<ol>
6097 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6098
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6099 <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]6100 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
6101
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]6102 <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]6103 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6104
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6105 <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]6106 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
6107
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6108 <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]6109 <tt>op1</tt> is less than <tt>op2</tt>.</li>
6110
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6111 <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]6112 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
6113
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6114 <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]6115 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
6116
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]6117 <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]6118
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6119 <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]6120 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
6121
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6122 <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]6123 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
6124
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6125 <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]6126 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
6127
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6128 <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]6129 <tt>op1</tt> is less than <tt>op2</tt>.</li>
6130
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6131 <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]6132 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
6133
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6134 <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]6135 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
6136
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]6137 <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]6138
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6139 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
6140</ol>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6141
6142<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6143<pre>
6144 &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]6145 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
6146 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
6147 &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]6148</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +0000[diff] [blame]6149
6150<p>Note that the code generator does not yet support vector types with
6151 the <tt>fcmp</tt> instruction.</p>
6152
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6153</div>
6154
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]6155<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6156<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]6157 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6158</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]6159
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6160<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]6161
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]6162<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6163<pre>
6164 &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...
6165</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]6166
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]6167<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6168<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in the
6169 SSA graph representing the function.</p>
6170
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]6171<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6172<p>The type of the incoming values is specified with the first type field. After
6173 this, the '<tt>phi</tt>' instruction takes a list of pairs as arguments, with
6174 one pair for each predecessor basic block of the current block. Only values
6175 of <a href="#t_firstclass">first class</a> type may be used as the value
6176 arguments to the PHI node. Only labels may be used as the label
6177 arguments.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]6178
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6179<p>There must be no non-phi instructions between the start of a basic block and
6180 the PHI instructions: i.e. PHI instructions must be first in a basic
6181 block.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]6182
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6183<p>For the purposes of the SSA form, the use of each incoming value is deemed to
6184 occur on the edge from the corresponding predecessor block to the current
6185 block (but after any definition of an '<tt>invoke</tt>' instruction's return
6186 value on the same edge).</p>
Jay Foadd2449092009-06-03 10:20:10 +0000[diff] [blame]6187
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]6188<h5>Semantics:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6189<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6190 specified by the pair corresponding to the predecessor basic block that
6191 executed just prior to the current block.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]6192
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]6193<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]6194<pre>
6195Loop: ; Infinite loop that counts from 0 on up...
6196 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
6197 %nextindvar = add i32 %indvar, 1
6198 br label %Loop
6199</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6200
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]6201</div>
6202
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6203<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6204<h4>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6205 <a name="i_select">'<tt>select</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6206</h4>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6207
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6208<div>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6209
6210<h5>Syntax:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6211<pre>
Dan Gohmanf72fb672008-09-09 01:02:47 +0000[diff] [blame]6212 &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>
6213
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]6214 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6215</pre>
6216
6217<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6218<p>The '<tt>select</tt>' instruction is used to choose one value based on a
6219 condition, without branching.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6220
6221
6222<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6223<p>The '<tt>select</tt>' instruction requires an 'i1' value or a vector of 'i1'
6224 values indicating the condition, and two values of the
6225 same <a href="#t_firstclass">first class</a> type. If the val1/val2 are
6226 vectors and the condition is a scalar, then entire vectors are selected, not
6227 individual elements.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6228
6229<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6230<p>If the condition is an i1 and it evaluates to 1, the instruction returns the
6231 first value argument; otherwise, it returns the second value argument.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6232
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6233<p>If the condition is a vector of i1, then the value arguments must be vectors
6234 of the same size, and the selection is done element by element.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6235
6236<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6237<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]6238 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6239</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +0000[diff] [blame]6240
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6241</div>
6242
Robert Bocchino05ccd702006-01-15 20:48:27 +0000[diff] [blame]6243<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6244<h4>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6245 <a name="i_call">'<tt>call</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6246</h4>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6247
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6248<div>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6249
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]6250<h5>Syntax:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6251<pre>
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]6252 &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]6253</pre>
6254
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]6255<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6256<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6257
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]6258<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6259<p>This instruction requires several arguments:</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6260
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]6261<ol>
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]6262 <li>The optional "tail" marker indicates that the callee function does not
6263 access any allocas or varargs in the caller. Note that calls may be
6264 marked "tail" even if they do not occur before
6265 a <a href="#i_ret"><tt>ret</tt></a> instruction. If the "tail" marker is
6266 present, the function call is eligible for tail call optimization,
6267 but <a href="CodeGenerator.html#tailcallopt">might not in fact be
Evan Chengdc444e92010-03-08 21:05:02 +0000[diff] [blame]6268 optimized into a jump</a>. The code generator may optimize calls marked
6269 "tail" with either 1) automatic <a href="CodeGenerator.html#sibcallopt">
6270 sibling call optimization</a> when the caller and callee have
6271 matching signatures, or 2) forced tail call optimization when the
6272 following extra requirements are met:
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]6273 <ul>
6274 <li>Caller and callee both have the calling
6275 convention <tt>fastcc</tt>.</li>
6276 <li>The call is in tail position (ret immediately follows call and ret
6277 uses value of call or is void).</li>
6278 <li>Option <tt>-tailcallopt</tt> is enabled,
Dan Gohmanfbbee8d2010-03-02 01:08:11 +0000[diff] [blame]6279 or <code>llvm::GuaranteedTailCallOpt</code> is <code>true</code>.</li>
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]6280 <li><a href="CodeGenerator.html#tailcallopt">Platform specific
6281 constraints are met.</a></li>
6282 </ul>
6283 </li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]6284
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6285 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
6286 convention</a> the call should use. If none is specified, the call
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]6287 defaults to using C calling conventions. The calling convention of the
6288 call must match the calling convention of the target function, or else the
6289 behavior is undefined.</li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]6290
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6291 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
6292 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
6293 '<tt>inreg</tt>' attributes are valid here.</li>
6294
6295 <li>'<tt>ty</tt>': the type of the call instruction itself which is also the
6296 type of the return value. Functions that return no value are marked
6297 <tt><a href="#t_void">void</a></tt>.</li>
6298
6299 <li>'<tt>fnty</tt>': shall be the signature of the pointer to function value
6300 being invoked. The argument types must match the types implied by this
6301 signature. This type can be omitted if the function is not varargs and if
6302 the function type does not return a pointer to a function.</li>
6303
6304 <li>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
6305 be invoked. In most cases, this is a direct function invocation, but
6306 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
6307 to function value.</li>
6308
6309 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]6310 signature argument types and parameter attributes. All arguments must be
6311 of <a href="#t_firstclass">first class</a> type. If the function
6312 signature indicates the function accepts a variable number of arguments,
6313 the extra arguments can be specified.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6314
6315 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
6316 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
6317 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]6318</ol>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6319
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]6320<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6321<p>The '<tt>call</tt>' instruction is used to cause control flow to transfer to
6322 a specified function, with its incoming arguments bound to the specified
6323 values. Upon a '<tt><a href="#i_ret">ret</a></tt>' instruction in the called
6324 function, control flow continues with the instruction after the function
6325 call, and the return value of the function is bound to the result
6326 argument.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6327
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]6328<h5>Example:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6329<pre>
Nick Lewyckydb7e3c92007-09-08 13:57:50 +0000[diff] [blame]6330 %retval = call i32 @test(i32 %argc)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6331 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) <i>; yields i32</i>
Chris Lattner772fccf2008-03-21 17:24:17 +0000[diff] [blame]6332 %X = tail call i32 @foo() <i>; yields i32</i>
6333 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
6334 call void %foo(i8 97 signext)
Devang Patelc3fc6df2008-03-10 20:49:15 +0000[diff] [blame]6335
6336 %struct.A = type { i32, i8 }
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]6337 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohmanb1e6b962008-10-04 19:00:07 +0000[diff] [blame]6338 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
6339 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattner85a350f2008-10-08 06:26:11 +0000[diff] [blame]6340 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijmancb73d192008-10-07 10:03:45 +0000[diff] [blame]6341 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6342</pre>
6343
Dale Johannesen07de8d12009-09-24 18:38:21 +0000[diff] [blame]6344<p>llvm treats calls to some functions with names and arguments that match the
Dale Johannesen9f8380b2009-09-25 17:04:42 +0000[diff] [blame]6345standard C99 library as being the C99 library functions, and may perform
6346optimizations or generate code for them under that assumption. This is
6347something we'd like to change in the future to provide better support for
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]6348freestanding environments and non-C-based languages.</p>
Dale Johannesen07de8d12009-09-24 18:38:21 +0000[diff] [blame]6349
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6350</div>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6351
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6352<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6353<h4>
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000[diff] [blame]6354 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6355</h4>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6356
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6357<div>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6358
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]6359<h5>Syntax:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6360<pre>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]6361 &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]6362</pre>
6363
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]6364<h5>Overview:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]6365<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6366 the "variable argument" area of a function call. It is used to implement the
6367 <tt>va_arg</tt> macro in C.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6368
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]6369<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6370<p>This instruction takes a <tt>va_list*</tt> value and the type of the
6371 argument. It returns a value of the specified argument type and increments
6372 the <tt>va_list</tt> to point to the next argument. The actual type
6373 of <tt>va_list</tt> is target specific.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6374
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]6375<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6376<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified type
6377 from the specified <tt>va_list</tt> and causes the <tt>va_list</tt> to point
6378 to the next argument. For more information, see the variable argument
6379 handling <a href="#int_varargs">Intrinsic Functions</a>.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6380
6381<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]6382 take a variable number of arguments, for example, the <tt>vfprintf</tt>
6383 function.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6384
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6385<p><tt>va_arg</tt> is an LLVM instruction instead of
6386 an <a href="#intrinsics">intrinsic function</a> because it takes a type as an
6387 argument.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6388
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]6389<h5>Example:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6390<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
6391
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6392<p>Note that the code generator does not yet fully support va_arg on many
6393 targets. Also, it does not currently support va_arg with aggregate types on
6394 any target.</p>
Dan Gohmanf3e60bd2009-01-12 23:12:39 +0000[diff] [blame]6395
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6396</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6397
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6398<!-- _______________________________________________________________________ -->
6399<h4>
6400 <a name="i_landingpad">'<tt>landingpad</tt>' Instruction</a>
6401</h4>
6402
6403<div>
6404
6405<h5>Syntax:</h5>
6406<pre>
Duncan Sands8d6796b2012-01-13 19:59:16 +0000[diff] [blame]6407 &lt;resultval&gt; = landingpad &lt;resultty&gt; personality &lt;type&gt; &lt;pers_fn&gt; &lt;clause&gt;+
6408 &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]6409
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6410 &lt;clause&gt; := catch &lt;type&gt; &lt;value&gt;
Bill Wendlinge6e88262011-08-12 20:24:12 +0000[diff] [blame]6411 &lt;clause&gt; := filter &lt;array constant type&gt; &lt;array constant&gt;
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6412</pre>
6413
6414<h5>Overview:</h5>
6415<p>The '<tt>landingpad</tt>' instruction is used by
6416 <a href="ExceptionHandling.html#overview">LLVM's exception handling
6417 system</a> to specify that a basic block is a landing pad &mdash; one where
6418 the exception lands, and corresponds to the code found in the
6419 <i><tt>catch</tt></i> portion of a <i><tt>try/catch</tt></i> sequence. It
6420 defines values supplied by the personality function (<tt>pers_fn</tt>) upon
6421 re-entry to the function. The <tt>resultval</tt> has the
Duncan Sands8d6796b2012-01-13 19:59:16 +0000[diff] [blame]6422 type <tt>resultty</tt>.</p>
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6423
6424<h5>Arguments:</h5>
6425<p>This instruction takes a <tt>pers_fn</tt> value. This is the personality
6426 function associated with the unwinding mechanism. The optional
6427 <tt>cleanup</tt> flag indicates that the landing pad block is a cleanup.</p>
6428
6429<p>A <tt>clause</tt> begins with the clause type &mdash; <tt>catch</tt>
Bill Wendlinge6e88262011-08-12 20:24:12 +0000[diff] [blame]6430 or <tt>filter</tt> &mdash; and contains the global variable representing the
6431 "type" that may be caught or filtered respectively. Unlike the
6432 <tt>catch</tt> clause, the <tt>filter</tt> clause takes an array constant as
6433 its argument. Use "<tt>[0 x i8**] undef</tt>" for a filter which cannot
6434 throw. The '<tt>landingpad</tt>' instruction must contain <em>at least</em>
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6435 one <tt>clause</tt> or the <tt>cleanup</tt> flag.</p>
6436
6437<h5>Semantics:</h5>
6438<p>The '<tt>landingpad</tt>' instruction defines the values which are set by the
6439 personality function (<tt>pers_fn</tt>) upon re-entry to the function, and
6440 therefore the "result type" of the <tt>landingpad</tt> instruction. As with
6441 calling conventions, how the personality function results are represented in
6442 LLVM IR is target specific.</p>
6443
Bill Wendlingb7a01352011-08-03 17:17:06 +0000[diff] [blame]6444<p>The clauses are applied in order from top to bottom. If two
6445 <tt>landingpad</tt> instructions are merged together through inlining, the
Duncan Sands8d6796b2012-01-13 19:59:16 +0000[diff] [blame]6446 clauses from the calling function are appended to the list of clauses.
6447 When the call stack is being unwound due to an exception being thrown, the
6448 exception is compared against each <tt>clause</tt> in turn. If it doesn't
6449 match any of the clauses, and the <tt>cleanup</tt> flag is not set, then
6450 unwinding continues further up the call stack.</p>
Bill Wendlingb7a01352011-08-03 17:17:06 +0000[diff] [blame]6451
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6452<p>The <tt>landingpad</tt> instruction has several restrictions:</p>
6453
6454<ul>
6455 <li>A landing pad block is a basic block which is the unwind destination of an
6456 '<tt>invoke</tt>' instruction.</li>
6457 <li>A landing pad block must have a '<tt>landingpad</tt>' instruction as its
6458 first non-PHI instruction.</li>
6459 <li>There can be only one '<tt>landingpad</tt>' instruction within the landing
6460 pad block.</li>
6461 <li>A basic block that is not a landing pad block may not include a
6462 '<tt>landingpad</tt>' instruction.</li>
6463 <li>All '<tt>landingpad</tt>' instructions in a function must have the same
6464 personality function.</li>
6465</ul>
6466
6467<h5>Example:</h5>
6468<pre>
6469 ;; A landing pad which can catch an integer.
6470 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6471 catch i8** @_ZTIi
6472 ;; A landing pad that is a cleanup.
6473 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
Bill Wendlinge6e88262011-08-12 20:24:12 +0000[diff] [blame]6474 cleanup
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6475 ;; A landing pad which can catch an integer and can only throw a double.
6476 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6477 catch i8** @_ZTIi
Bill Wendlinge6e88262011-08-12 20:24:12 +0000[diff] [blame]6478 filter [1 x i8**] [@_ZTId]
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6479</pre>
6480
6481</div>
6482
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6483</div>
6484
6485</div>
6486
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6487<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6488<h2><a name="intrinsics">Intrinsic Functions</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]6489<!-- *********************************************************************** -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6490
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6491<div>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6492
6493<p>LLVM supports the notion of an "intrinsic function". These functions have
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6494 well known names and semantics and are required to follow certain
6495 restrictions. Overall, these intrinsics represent an extension mechanism for
6496 the LLVM language that does not require changing all of the transformations
6497 in LLVM when adding to the language (or the bitcode reader/writer, the
6498 parser, etc...).</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6499
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]6500<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6501 prefix is reserved in LLVM for intrinsic names; thus, function names may not
6502 begin with this prefix. Intrinsic functions must always be external
6503 functions: you cannot define the body of intrinsic functions. Intrinsic
6504 functions may only be used in call or invoke instructions: it is illegal to
6505 take the address of an intrinsic function. Additionally, because intrinsic
6506 functions are part of the LLVM language, it is required if any are added that
6507 they be documented here.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6508
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6509<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents a
6510 family of functions that perform the same operation but on different data
6511 types. Because LLVM can represent over 8 million different integer types,
6512 overloading is used commonly to allow an intrinsic function to operate on any
6513 integer type. One or more of the argument types or the result type can be
6514 overloaded to accept any integer type. Argument types may also be defined as
6515 exactly matching a previous argument's type or the result type. This allows
6516 an intrinsic function which accepts multiple arguments, but needs all of them
6517 to be of the same type, to only be overloaded with respect to a single
6518 argument or the result.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6519
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6520<p>Overloaded intrinsics will have the names of its overloaded argument types
6521 encoded into its function name, each preceded by a period. Only those types
6522 which are overloaded result in a name suffix. Arguments whose type is matched
6523 against another type do not. For example, the <tt>llvm.ctpop</tt> function
6524 can take an integer of any width and returns an integer of exactly the same
6525 integer width. This leads to a family of functions such as
6526 <tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29
6527 %val)</tt>. Only one type, the return type, is overloaded, and only one type
6528 suffix is required. Because the argument's type is matched against the return
6529 type, it does not require its own name suffix.</p>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]6530
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6531<p>To learn how to add an intrinsic function, please see the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6532 <a href="ExtendingLLVM.html">Extending LLVM Guide</a>.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6533
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6534<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6535<h3>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6536 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6537</h3>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6538
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6539<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6540
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6541<p>Variable argument support is defined in LLVM with
6542 the <a href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
6543 intrinsic functions. These functions are related to the similarly named
6544 macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6545
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6546<p>All of these functions operate on arguments that use a target-specific value
6547 type "<tt>va_list</tt>". The LLVM assembly language reference manual does
6548 not define what this type is, so all transformations should be prepared to
6549 handle these functions regardless of the type used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6550
Chris Lattner374ab302006-05-15 17:26:46 +0000[diff] [blame]6551<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]6552 instruction and the variable argument handling intrinsic functions are
6553 used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6554
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]6555<pre class="doc_code">
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6556define i32 @test(i32 %X, ...) {
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6557 ; Initialize variable argument processing
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6558 %ap = alloca i8*
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6559 %ap2 = bitcast i8** %ap to i8*
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6560 call void @llvm.va_start(i8* %ap2)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6561
6562 ; Read a single integer argument
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6563 %tmp = va_arg i8** %ap, i32
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6564
6565 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6566 %aq = alloca i8*
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6567 %aq2 = bitcast i8** %aq to i8*
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6568 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6569 call void @llvm.va_end(i8* %aq2)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6570
6571 ; Stop processing of arguments.
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6572 call void @llvm.va_end(i8* %ap2)
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]6573 ret i32 %tmp
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6574}
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6575
6576declare void @llvm.va_start(i8*)
6577declare void @llvm.va_copy(i8*, i8*)
6578declare void @llvm.va_end(i8*)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6579</pre>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6580
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6581<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6582<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6583 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6584</h4>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6585
6586
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6587<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6588
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6589<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6590<pre>
6591 declare void %llvm.va_start(i8* &lt;arglist&gt;)
6592</pre>
6593
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6594<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6595<p>The '<tt>llvm.va_start</tt>' intrinsic initializes <tt>*&lt;arglist&gt;</tt>
6596 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]6597
6598<h5>Arguments:</h5>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]6599<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]6600
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6601<h5>Semantics:</h5>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]6602<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]6603 macro available in C. In a target-dependent way, it initializes
6604 the <tt>va_list</tt> element to which the argument points, so that the next
6605 call to <tt>va_arg</tt> will produce the first variable argument passed to
6606 the function. Unlike the C <tt>va_start</tt> macro, this intrinsic does not
6607 need to know the last argument of the function as the compiler can figure
6608 that out.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]6609
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6610</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6611
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6612<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6613<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6614 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6615</h4>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6616
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6617<div>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6618
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6619<h5>Syntax:</h5>
6620<pre>
6621 declare void @llvm.va_end(i8* &lt;arglist&gt;)
6622</pre>
6623
6624<h5>Overview:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6625<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6626 which has been initialized previously
6627 with <tt><a href="#int_va_start">llvm.va_start</a></tt>
6628 or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6629
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6630<h5>Arguments:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6631<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]6632
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6633<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6634<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]6635 macro available in C. In a target-dependent way, it destroys
6636 the <tt>va_list</tt> element to which the argument points. Calls
6637 to <a href="#int_va_start"><tt>llvm.va_start</tt></a>
6638 and <a href="#int_va_copy"> <tt>llvm.va_copy</tt></a> must be matched exactly
6639 with calls to <tt>llvm.va_end</tt>.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6640
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6641</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6642
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6643<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6644<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6645 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6646</h4>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6647
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6648<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6649
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6650<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6651<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]6652 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6653</pre>
6654
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6655<h5>Overview:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6656<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6657 from the source argument list to the destination argument list.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6658
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6659<h5>Arguments:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]6660<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]6661 The second argument is a pointer to a <tt>va_list</tt> element to copy
6662 from.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6663
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6664<h5>Semantics:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6665<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]6666 macro available in C. In a target-dependent way, it copies the
6667 source <tt>va_list</tt> element into the destination <tt>va_list</tt>
6668 element. This intrinsic is necessary because
6669 the <tt><a href="#int_va_start"> llvm.va_start</a></tt> intrinsic may be
6670 arbitrarily complex and require, for example, memory allocation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6671
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6672</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6673
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6674</div>
6675
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6676<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6677<h3>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6678 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6679</h3>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6680
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6681<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6682
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6683<p>LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]6684Collection</a> (GC) requires the implementation and generation of these
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6685intrinsics. These intrinsics allow identification of <a href="#int_gcroot">GC
6686roots on the stack</a>, as well as garbage collector implementations that
6687require <a href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a>
6688barriers. Front-ends for type-safe garbage collected languages should generate
6689these intrinsics to make use of the LLVM garbage collectors. For more details,
6690see <a href="GarbageCollection.html">Accurate Garbage Collection with
6691LLVM</a>.</p>
Christopher Lamb303dae92007-12-17 01:00:21 +0000[diff] [blame]6692
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6693<p>The garbage collection intrinsics only operate on objects in the generic
6694 address space (address space zero).</p>
Christopher Lamb303dae92007-12-17 01:00:21 +0000[diff] [blame]6695
Chris Lattnerd7923912004-05-23 21:06:01 +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_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6699</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +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
6703<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6704<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]6705 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6706</pre>
6707
6708<h5>Overview:</h5>
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]6709<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]6710 the code generator, and allows some metadata to be associated with it.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6711
6712<h5>Arguments:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6713<p>The first argument specifies the address of a stack object that contains the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6714 root pointer. The second pointer (which must be either a constant or a
6715 global value address) contains the meta-data to be associated with the
6716 root.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6717
6718<h5>Semantics:</h5>
Chris Lattner05d67092008-04-24 05:59:56 +0000[diff] [blame]6719<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]6720 location. At compile-time, the code generator generates information to allow
6721 the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
6722 intrinsic may only be used in a function which <a href="#gc">specifies a GC
6723 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6724
6725</div>
6726
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6727<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6728<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6729 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6730</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6731
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6732<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6733
6734<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6735<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]6736 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6737</pre>
6738
6739<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6740<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6741 locations, allowing garbage collector implementations that require read
6742 barriers.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6743
6744<h5>Arguments:</h5>
Chris Lattner80626e92006-03-14 20:02:51 +0000[diff] [blame]6745<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]6746 allocated from the garbage collector. The first object is a pointer to the
6747 start of the referenced object, if needed by the language runtime (otherwise
6748 null).</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6749
6750<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6751<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6752 instruction, but may be replaced with substantially more complex code by the
6753 garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
6754 may only be used in a function which <a href="#gc">specifies a GC
6755 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6756
6757</div>
6758
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6759<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6760<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6761 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6762</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6763
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6764<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6765
6766<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6767<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]6768 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6769</pre>
6770
6771<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6772<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6773 locations, allowing garbage collector implementations that require write
6774 barriers (such as generational or reference counting collectors).</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6775
6776<h5>Arguments:</h5>
Chris Lattner80626e92006-03-14 20:02:51 +0000[diff] [blame]6777<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]6778 object to store it to, and the third is the address of the field of Obj to
6779 store to. If the runtime does not require a pointer to the object, Obj may
6780 be null.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6781
6782<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6783<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6784 instruction, but may be replaced with substantially more complex code by the
6785 garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
6786 may only be used in a function which <a href="#gc">specifies a GC
6787 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6788
6789</div>
6790
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6791</div>
6792
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6793<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6794<h3>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6795 <a name="int_codegen">Code Generator Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6796</h3>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6797
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6798<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6799
6800<p>These intrinsics are provided by LLVM to expose special features that may
6801 only be implemented with code generator support.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6802
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6803<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6804<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6805 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6806</h4>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6807
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6808<div>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6809
6810<h5>Syntax:</h5>
6811<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]6812 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6813</pre>
6814
6815<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6816<p>The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
6817 target-specific value indicating the return address of the current function
6818 or one of its callers.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6819
6820<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6821<p>The argument to this intrinsic indicates which function to return the address
6822 for. Zero indicates the calling function, one indicates its caller, etc.
6823 The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6824
6825<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6826<p>The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer
6827 indicating the return address of the specified call frame, or zero if it
6828 cannot be identified. The value returned by this intrinsic is likely to be
6829 incorrect or 0 for arguments other than zero, so it should only be used for
6830 debugging purposes.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6831
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6832<p>Note that calling this intrinsic does not prevent function inlining or other
6833 aggressive transformations, so the value returned may not be that of the
6834 obvious source-language caller.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6835
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6836</div>
6837
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6838<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6839<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6840 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6841</h4>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6842
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6843<div>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6844
6845<h5>Syntax:</h5>
6846<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6847 declare i8* @llvm.frameaddress(i32 &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6848</pre>
6849
6850<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6851<p>The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
6852 target-specific frame pointer value for the specified stack frame.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6853
6854<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6855<p>The argument to this intrinsic indicates which function to return the frame
6856 pointer for. Zero indicates the calling function, one indicates its caller,
6857 etc. The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6858
6859<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6860<p>The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer
6861 indicating the frame address of the specified call frame, or zero if it
6862 cannot be identified. The value returned by this intrinsic is likely to be
6863 incorrect or 0 for arguments other than zero, so it should only be used for
6864 debugging purposes.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6865
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6866<p>Note that calling this intrinsic does not prevent function inlining or other
6867 aggressive transformations, so the value returned may not be that of the
6868 obvious source-language caller.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6869
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6870</div>
6871
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6872<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6873<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6874 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6875</h4>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6876
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6877<div>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6878
6879<h5>Syntax:</h5>
6880<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6881 declare i8* @llvm.stacksave()
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6882</pre>
6883
6884<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6885<p>The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state
6886 of the function stack, for use
6887 with <a href="#int_stackrestore"> <tt>llvm.stackrestore</tt></a>. This is
6888 useful for implementing language features like scoped automatic variable
6889 sized arrays in C99.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6890
6891<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6892<p>This intrinsic returns a opaque pointer value that can be passed
6893 to <a href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When
6894 an <tt>llvm.stackrestore</tt> intrinsic is executed with a value saved
6895 from <tt>llvm.stacksave</tt>, it effectively restores the state of the stack
6896 to the state it was in when the <tt>llvm.stacksave</tt> intrinsic executed.
6897 In practice, this pops any <a href="#i_alloca">alloca</a> blocks from the
6898 stack that were allocated after the <tt>llvm.stacksave</tt> was executed.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6899
6900</div>
6901
6902<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6903<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6904 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6905</h4>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6906
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6907<div>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6908
6909<h5>Syntax:</h5>
6910<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6911 declare void @llvm.stackrestore(i8* %ptr)
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6912</pre>
6913
6914<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6915<p>The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
6916 the function stack to the state it was in when the
6917 corresponding <a href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic
6918 executed. This is useful for implementing language features like scoped
6919 automatic variable sized arrays in C99.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6920
6921<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6922<p>See the description
6923 for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6924
6925</div>
6926
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6927<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6928<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6929 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6930</h4>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6931
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6932<div>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6933
6934<h5>Syntax:</h5>
6935<pre>
Bruno Cardoso Lopes9a767332011-06-14 04:58:37 +0000[diff] [blame]6936 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]6937</pre>
6938
6939<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6940<p>The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to
6941 insert a prefetch instruction if supported; otherwise, it is a noop.
6942 Prefetches have no effect on the behavior of the program but can change its
6943 performance characteristics.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6944
6945<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6946<p><tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the
6947 specifier determining if the fetch should be for a read (0) or write (1),
6948 and <tt>locality</tt> is a temporal locality specifier ranging from (0) - no
Bruno Cardoso Lopes9a767332011-06-14 04:58:37 +0000[diff] [blame]6949 locality, to (3) - extremely local keep in cache. The <tt>cache type</tt>
6950 specifies whether the prefetch is performed on the data (1) or instruction (0)
6951 cache. The <tt>rw</tt>, <tt>locality</tt> and <tt>cache type</tt> arguments
6952 must be constant integers.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6953
6954<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6955<p>This intrinsic does not modify the behavior of the program. In particular,
6956 prefetches cannot trap and do not produce a value. On targets that support
6957 this intrinsic, the prefetch can provide hints to the processor cache for
6958 better performance.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6959
6960</div>
6961
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6962<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6963<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6964 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6965</h4>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6966
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6967<div>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6968
6969<h5>Syntax:</h5>
6970<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]6971 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6972</pre>
6973
6974<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6975<p>The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program
6976 Counter (PC) in a region of code to simulators and other tools. The method
6977 is target specific, but it is expected that the marker will use exported
6978 symbols to transmit the PC of the marker. The marker makes no guarantees
6979 that it will remain with any specific instruction after optimizations. It is
6980 possible that the presence of a marker will inhibit optimizations. The
6981 intended use is to be inserted after optimizations to allow correlations of
6982 simulation runs.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6983
6984<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6985<p><tt>id</tt> is a numerical id identifying the marker.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6986
6987<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6988<p>This intrinsic does not modify the behavior of the program. Backends that do
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]6989 not support this intrinsic may ignore it.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6990
6991</div>
6992
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6993<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6994<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6995 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6996</h4>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6997
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6998<div>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6999
7000<h5>Syntax:</h5>
7001<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7002 declare i64 @llvm.readcyclecounter()
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]7003</pre>
7004
7005<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7006<p>The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
7007 counter register (or similar low latency, high accuracy clocks) on those
7008 targets that support it. On X86, it should map to RDTSC. On Alpha, it
7009 should map to RPCC. As the backing counters overflow quickly (on the order
7010 of 9 seconds on alpha), this should only be used for small timings.</p>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]7011
7012<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7013<p>When directly supported, reading the cycle counter should not modify any
7014 memory. Implementations are allowed to either return a application specific
7015 value or a system wide value. On backends without support, this is lowered
7016 to a constant 0.</p>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]7017
7018</div>
7019
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7020</div>
7021
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7022<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7023<h3>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7024 <a name="int_libc">Standard C Library Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7025</h3>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7026
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7027<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7028
7029<p>LLVM provides intrinsics for a few important standard C library functions.
7030 These intrinsics allow source-language front-ends to pass information about
7031 the alignment of the pointer arguments to the code generator, providing
7032 opportunity for more efficient code generation.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7033
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7034<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7035<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]7036 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7037</h4>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7038
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7039<div>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7040
7041<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7042<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]7043 integer bit width and for different address spaces. Not all targets support
7044 all bit widths however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7045
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7046<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]7047 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]7048 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]7049 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]7050 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7051</pre>
7052
7053<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7054<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
7055 source location to the destination location.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7056
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7057<p>Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7058 intrinsics do not return a value, takes extra alignment/isvolatile arguments
7059 and the pointers can be in specified address spaces.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7060
7061<h5>Arguments:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7062
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7063<p>The first argument is a pointer to the destination, the second is a pointer
7064 to the source. The third argument is an integer argument specifying the
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7065 number of bytes to copy, the fourth argument is the alignment of the
7066 source and destination locations, and the fifth is a boolean indicating a
7067 volatile access.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7068
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]7069<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]7070 then the caller guarantees that both the source and destination pointers are
7071 aligned to that boundary.</p>
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]7072
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]7073<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
7074 <tt>llvm.memcpy</tt> call is a <a href="#volatile">volatile operation</a>.
7075 The detailed access behavior is not very cleanly specified and it is unwise
7076 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7077
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7078<h5>Semantics:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7079
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7080<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
7081 source location to the destination location, which are not allowed to
7082 overlap. It copies "len" bytes of memory over. If the argument is known to
7083 be aligned to some boundary, this can be specified as the fourth argument,
7084 otherwise it should be set to 0 or 1.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7085
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7086</div>
7087
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7088<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7089<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]7090 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7091</h4>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7092
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7093<div>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7094
7095<h5>Syntax:</h5>
Chris Lattner824b9582008-11-21 16:42:48 +0000[diff] [blame]7096<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]7097 width and for different address space. Not all targets support all bit
7098 widths however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7099
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7100<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]7101 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]7102 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]7103 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]7104 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7105</pre>
7106
7107<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7108<p>The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the
7109 source location to the destination location. It is similar to the
7110 '<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to
7111 overlap.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7112
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7113<p>Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7114 intrinsics do not return a value, takes extra alignment/isvolatile arguments
7115 and the pointers can be in specified address spaces.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7116
7117<h5>Arguments:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7118
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7119<p>The first argument is a pointer to the destination, the second is a pointer
7120 to the source. The third argument is an integer argument specifying the
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7121 number of bytes to copy, the fourth argument is the alignment of the
7122 source and destination locations, and the fifth is a boolean indicating a
7123 volatile access.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7124
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]7125<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]7126 then the caller guarantees that the source and destination pointers are
7127 aligned to that boundary.</p>
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]7128
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]7129<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
7130 <tt>llvm.memmove</tt> call is a <a href="#volatile">volatile operation</a>.
7131 The detailed access behavior is not very cleanly specified and it is unwise
7132 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7133
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7134<h5>Semantics:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7135
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7136<p>The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the
7137 source location to the destination location, which may overlap. It copies
7138 "len" bytes of memory over. If the argument is known to be aligned to some
7139 boundary, this can be specified as the fourth argument, otherwise it should
7140 be set to 0 or 1.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7141
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7142</div>
7143
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7144<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7145<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]7146 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7147</h4>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7148
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7149<div>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7150
7151<h5>Syntax:</h5>
Chris Lattner824b9582008-11-21 16:42:48 +0000[diff] [blame]7152<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]7153 width and for different address spaces. However, not all targets support all
7154 bit widths.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7155
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7156<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]7157 declare void @llvm.memset.p0i8.i32(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattnerff35c3f2010-04-08 00:54:34 +0000[diff] [blame]7158 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]7159 declare void @llvm.memset.p0i8.i64(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattnerff35c3f2010-04-08 00:54:34 +0000[diff] [blame]7160 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7161</pre>
7162
7163<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7164<p>The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a
7165 particular byte value.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7166
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7167<p>Note that, unlike the standard libc function, the <tt>llvm.memset</tt>
John Criswellcdcbbfc2010-07-30 16:30:28 +0000[diff] [blame]7168 intrinsic does not return a value and takes extra alignment/volatile
7169 arguments. Also, the destination can be in an arbitrary address space.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7170
7171<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7172<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]7173 byte value with which to fill it, the third argument is an integer argument
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7174 specifying the number of bytes to fill, and the fourth argument is the known
John Criswellcdcbbfc2010-07-30 16:30:28 +0000[diff] [blame]7175 alignment of the destination location.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7176
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]7177<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]7178 then the caller guarantees that the destination pointer is aligned to that
7179 boundary.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7180
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]7181<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
7182 <tt>llvm.memset</tt> call is a <a href="#volatile">volatile operation</a>.
7183 The detailed access behavior is not very cleanly specified and it is unwise
7184 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7185
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7186<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7187<p>The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting
7188 at the destination location. If the argument is known to be aligned to some
7189 boundary, this can be specified as the fourth argument, otherwise it should
7190 be set to 0 or 1.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7191
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7192</div>
7193
Chris Lattner32006282004-06-11 02:28:03 +0000[diff] [blame]7194<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7195<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]7196 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7197</h4>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]7198
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7199<div>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]7200
7201<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7202<p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
7203 floating point or vector of floating point type. Not all targets support all
7204 types however.</p>
7205
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]7206<pre>
Dale Johannesen408f9c12007-10-02 17:47:38 +0000[diff] [blame]7207 declare float @llvm.sqrt.f32(float %Val)
7208 declare double @llvm.sqrt.f64(double %Val)
7209 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
7210 declare fp128 @llvm.sqrt.f128(fp128 %Val)
7211 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]7212</pre>
7213
7214<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7215<p>The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
7216 returning the same value as the libm '<tt>sqrt</tt>' functions would.
7217 Unlike <tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined
7218 behavior for negative numbers other than -0.0 (which allows for better
7219 optimization, because there is no need to worry about errno being
7220 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]7221
7222<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7223<p>The argument and return value are floating point numbers of the same
7224 type.</p>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]7225
7226<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7227<p>This function returns the sqrt of the specified operand if it is a
7228 nonnegative floating point number.</p>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]7229
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]7230</div>
7231
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7232<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7233<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]7234 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7235</h4>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7236
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7237<div>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7238
7239<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7240<p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
7241 floating point or vector of floating point type. Not all targets support all
7242 types however.</p>
7243
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7244<pre>
Dale Johannesen408f9c12007-10-02 17:47:38 +0000[diff] [blame]7245 declare float @llvm.powi.f32(float %Val, i32 %power)
7246 declare double @llvm.powi.f64(double %Val, i32 %power)
7247 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
7248 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
7249 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7250</pre>
7251
7252<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7253<p>The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
7254 specified (positive or negative) power. The order of evaluation of
7255 multiplications is not defined. When a vector of floating point type is
7256 used, the second argument remains a scalar integer value.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7257
7258<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7259<p>The second argument is an integer power, and the first is a value to raise to
7260 that power.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7261
7262<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7263<p>This function returns the first value raised to the second power with an
7264 unspecified sequence of rounding operations.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7265
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7266</div>
7267
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7268<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7269<h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7270 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7271</h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7272
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7273<div>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7274
7275<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7276<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
7277 floating point or vector of floating point type. Not all targets support all
7278 types however.</p>
7279
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7280<pre>
7281 declare float @llvm.sin.f32(float %Val)
7282 declare double @llvm.sin.f64(double %Val)
7283 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
7284 declare fp128 @llvm.sin.f128(fp128 %Val)
7285 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
7286</pre>
7287
7288<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7289<p>The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7290
7291<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7292<p>The argument and return value are floating point numbers of the same
7293 type.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7294
7295<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7296<p>This function returns the sine of the specified operand, returning the same
7297 values as the libm <tt>sin</tt> functions would, and handles error conditions
7298 in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7299
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7300</div>
7301
7302<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7303<h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7304 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7305</h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7306
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7307<div>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7308
7309<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7310<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
7311 floating point or vector of floating point type. Not all targets support all
7312 types however.</p>
7313
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7314<pre>
7315 declare float @llvm.cos.f32(float %Val)
7316 declare double @llvm.cos.f64(double %Val)
7317 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
7318 declare fp128 @llvm.cos.f128(fp128 %Val)
7319 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
7320</pre>
7321
7322<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7323<p>The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7324
7325<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7326<p>The argument and return value are floating point numbers of the same
7327 type.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7328
7329<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7330<p>This function returns the cosine of the specified operand, returning the same
7331 values as the libm <tt>cos</tt> functions would, and handles error conditions
7332 in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7333
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7334</div>
7335
7336<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7337<h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7338 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7339</h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7340
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7341<div>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7342
7343<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7344<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
7345 floating point or vector of floating point type. Not all targets support all
7346 types however.</p>
7347
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7348<pre>
7349 declare float @llvm.pow.f32(float %Val, float %Power)
7350 declare double @llvm.pow.f64(double %Val, double %Power)
7351 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
7352 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
7353 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
7354</pre>
7355
7356<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7357<p>The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
7358 specified (positive or negative) power.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7359
7360<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7361<p>The second argument is a floating point power, and the first is a value to
7362 raise to that power.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7363
7364<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7365<p>This function returns the first value raised to the second power, returning
7366 the same values as the libm <tt>pow</tt> functions would, and handles error
7367 conditions in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7368
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7369</div>
7370
Dan Gohman4e9011c2011-05-23 21:13:03 +0000[diff] [blame]7371<!-- _______________________________________________________________________ -->
7372<h4>
7373 <a name="int_exp">'<tt>llvm.exp.*</tt>' Intrinsic</a>
7374</h4>
7375
7376<div>
7377
7378<h5>Syntax:</h5>
7379<p>This is an overloaded intrinsic. You can use <tt>llvm.exp</tt> on any
7380 floating point or vector of floating point type. Not all targets support all
7381 types however.</p>
7382
7383<pre>
7384 declare float @llvm.exp.f32(float %Val)
7385 declare double @llvm.exp.f64(double %Val)
7386 declare x86_fp80 @llvm.exp.f80(x86_fp80 %Val)
7387 declare fp128 @llvm.exp.f128(fp128 %Val)
7388 declare ppc_fp128 @llvm.exp.ppcf128(ppc_fp128 %Val)
7389</pre>
7390
7391<h5>Overview:</h5>
7392<p>The '<tt>llvm.exp.*</tt>' intrinsics perform the exp function.</p>
7393
7394<h5>Arguments:</h5>
7395<p>The argument and return value are floating point numbers of the same
7396 type.</p>
7397
7398<h5>Semantics:</h5>
7399<p>This function returns the same values as the libm <tt>exp</tt> functions
7400 would, and handles error conditions in the same way.</p>
7401
7402</div>
7403
7404<!-- _______________________________________________________________________ -->
7405<h4>
7406 <a name="int_log">'<tt>llvm.log.*</tt>' Intrinsic</a>
7407</h4>
7408
7409<div>
7410
7411<h5>Syntax:</h5>
7412<p>This is an overloaded intrinsic. You can use <tt>llvm.log</tt> on any
7413 floating point or vector of floating point type. Not all targets support all
7414 types however.</p>
7415
7416<pre>
7417 declare float @llvm.log.f32(float %Val)
7418 declare double @llvm.log.f64(double %Val)
7419 declare x86_fp80 @llvm.log.f80(x86_fp80 %Val)
7420 declare fp128 @llvm.log.f128(fp128 %Val)
7421 declare ppc_fp128 @llvm.log.ppcf128(ppc_fp128 %Val)
7422</pre>
7423
7424<h5>Overview:</h5>
7425<p>The '<tt>llvm.log.*</tt>' intrinsics perform the log function.</p>
7426
7427<h5>Arguments:</h5>
7428<p>The argument and return value are floating point numbers of the same
7429 type.</p>
7430
7431<h5>Semantics:</h5>
7432<p>This function returns the same values as the libm <tt>log</tt> functions
7433 would, and handles error conditions in the same way.</p>
7434
Nick Lewycky1c929be2011-10-31 01:32:21 +0000[diff] [blame]7435</div>
7436
7437<!-- _______________________________________________________________________ -->
Cameron Zwarich33390842011-07-08 21:39:21 +0000[diff] [blame]7438<h4>
7439 <a name="int_fma">'<tt>llvm.fma.*</tt>' Intrinsic</a>
7440</h4>
7441
7442<div>
7443
7444<h5>Syntax:</h5>
7445<p>This is an overloaded intrinsic. You can use <tt>llvm.fma</tt> on any
7446 floating point or vector of floating point type. Not all targets support all
7447 types however.</p>
7448
7449<pre>
7450 declare float @llvm.fma.f32(float %a, float %b, float %c)
7451 declare double @llvm.fma.f64(double %a, double %b, double %c)
7452 declare x86_fp80 @llvm.fma.f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c)
7453 declare fp128 @llvm.fma.f128(fp128 %a, fp128 %b, fp128 %c)
7454 declare ppc_fp128 @llvm.fma.ppcf128(ppc_fp128 %a, ppc_fp128 %b, ppc_fp128 %c)
7455</pre>
7456
7457<h5>Overview:</h5>
Cameron Zwarichabc43e62011-07-08 22:13:55 +0000[diff] [blame]7458<p>The '<tt>llvm.fma.*</tt>' intrinsics perform the fused multiply-add
Cameron Zwarich33390842011-07-08 21:39:21 +0000[diff] [blame]7459 operation.</p>
7460
7461<h5>Arguments:</h5>
7462<p>The argument and return value are floating point numbers of the same
7463 type.</p>
7464
7465<h5>Semantics:</h5>
7466<p>This function returns the same values as the libm <tt>fma</tt> functions
7467 would.</p>
7468
Dan Gohman4e9011c2011-05-23 21:13:03 +0000[diff] [blame]7469</div>
7470
NAKAMURA Takumi4b2e07a2011-10-31 13:04:26 +0000[diff] [blame]7471</div>
7472
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7473<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7474<h3>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7475 <a name="int_manip">Bit Manipulation Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7476</h3>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7477
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7478<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7479
7480<p>LLVM provides intrinsics for a few important bit manipulation operations.
7481 These allow efficient code generation for some algorithms.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7482
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7483<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7484<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]7485 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7486</h4>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7487
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7488<div>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7489
7490<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]7491<p>This is an overloaded intrinsic function. You can use bswap on any integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7492 type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
7493
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7494<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]7495 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
7496 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
7497 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7498</pre>
7499
7500<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7501<p>The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
7502 values with an even number of bytes (positive multiple of 16 bits). These
7503 are useful for performing operations on data that is not in the target's
7504 native byte order.</p>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7505
7506<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7507<p>The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
7508 and low byte of the input i16 swapped. Similarly,
7509 the <tt>llvm.bswap.i32</tt> intrinsic returns an i32 value that has the four
7510 bytes of the input i32 swapped, so that if the input bytes are numbered 0, 1,
7511 2, 3 then the returned i32 will have its bytes in 3, 2, 1, 0 order.
7512 The <tt>llvm.bswap.i48</tt>, <tt>llvm.bswap.i64</tt> and other intrinsics
7513 extend this concept to additional even-byte lengths (6 bytes, 8 bytes and
7514 more, respectively).</p>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7515
7516</div>
7517
7518<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7519<h4>
Reid Spencer0b118202006-01-16 21:12:35 +0000[diff] [blame]7520 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7521</h4>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7522
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7523<div>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7524
7525<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]7526<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]7527 width, or on any vector with integer elements. Not all targets support all
7528 bit widths or vector types, however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7529
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7530<pre>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7531 declare i8 @llvm.ctpop.i8(i8 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]7532 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]7533 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]7534 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
7535 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]7536 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]7537</pre>
7538
7539<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7540<p>The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set
7541 in a value.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7542
7543<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7544<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]7545 integer type, or a vector with integer elements.
7546 The return type must match the argument type.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7547
7548<h5>Semantics:</h5>
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]7549<p>The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable, or within each
7550 element of a vector.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7551
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7552</div>
7553
7554<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7555<h4>
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]7556 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7557</h4>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7558
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7559<div>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7560
7561<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7562<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]7563 integer bit width, or any vector whose elements are integers. Not all
7564 targets support all bit widths or vector types, however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7565
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7566<pre>
Chandler Carruth48b0bbf2011-12-12 04:36:04 +0000[diff] [blame]7567 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7568 declare i16 @llvm.ctlz.i16 (i16 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7569 declare i32 @llvm.ctlz.i32 (i32 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7570 declare i64 @llvm.ctlz.i64 (i64 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7571 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7572 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]7573</pre>
7574
7575<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7576<p>The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
7577 leading zeros in a variable.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7578
7579<h5>Arguments:</h5>
Chandler Carruth48b0bbf2011-12-12 04:36:04 +0000[diff] [blame]7580<p>The first argument is the value to be counted. This argument may be of any
7581 integer type, or a vectory with integer element type. The return type
7582 must match the first argument type.</p>
7583
7584<p>The second argument must be a constant and is a flag to indicate whether the
7585 intrinsic should ensure that a zero as the first argument produces a defined
7586 result. Historically some architectures did not provide a defined result for
7587 zero values as efficiently, and many algorithms are now predicated on
7588 avoiding zero-value inputs.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7589
7590<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7591<p>The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant)
Chandler Carruth48b0bbf2011-12-12 04:36:04 +0000[diff] [blame]7592 zeros in a variable, or within each element of the vector.
7593 If <tt>src == 0</tt> then the result is the size in bits of the type of
7594 <tt>src</tt> if <tt>is_zero_undef == 0</tt> and <tt>undef</tt> otherwise.
7595 For example, <tt>llvm.ctlz(i32 2) = 30</tt>.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7596
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7597</div>
Chris Lattner32006282004-06-11 02:28:03 +0000[diff] [blame]7598
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7599<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7600<h4>
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]7601 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7602</h4>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7603
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7604<div>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7605
7606<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7607<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]7608 integer bit width, or any vector of integer elements. Not all targets
7609 support all bit widths or vector types, however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7610
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7611<pre>
Chandler Carruth48b0bbf2011-12-12 04:36:04 +0000[diff] [blame]7612 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7613 declare i16 @llvm.cttz.i16 (i16 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7614 declare i32 @llvm.cttz.i32 (i32 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7615 declare i64 @llvm.cttz.i64 (i64 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7616 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7617 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]7618</pre>
7619
7620<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7621<p>The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
7622 trailing zeros.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7623
7624<h5>Arguments:</h5>
Chandler Carruth48b0bbf2011-12-12 04:36:04 +0000[diff] [blame]7625<p>The first argument is the value to be counted. This argument may be of any
7626 integer type, or a vectory with integer element type. The return type
7627 must match the first argument type.</p>
7628
7629<p>The second argument must be a constant and is a flag to indicate whether the
7630 intrinsic should ensure that a zero as the first argument produces a defined
7631 result. Historically some architectures did not provide a defined result for
7632 zero values as efficiently, and many algorithms are now predicated on
7633 avoiding zero-value inputs.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7634
7635<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7636<p>The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant)
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]7637 zeros in a variable, or within each element of a vector.
Chandler Carruth48b0bbf2011-12-12 04:36:04 +0000[diff] [blame]7638 If <tt>src == 0</tt> then the result is the size in bits of the type of
7639 <tt>src</tt> if <tt>is_zero_undef == 0</tt> and <tt>undef</tt> otherwise.
7640 For example, <tt>llvm.cttz(2) = 1</tt>.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7641
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7642</div>
7643
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7644</div>
7645
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]7646<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7647<h3>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]7648 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7649</h3>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]7650
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7651<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7652
7653<p>LLVM provides intrinsics for some arithmetic with overflow operations.</p>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]7654
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7655<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7656<h4>
7657 <a name="int_sadd_overflow">
7658 '<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics
7659 </a>
7660</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7661
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7662<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7663
7664<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7665<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]7666 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7667
7668<pre>
7669 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
7670 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
7671 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
7672</pre>
7673
7674<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7675<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7676 a signed addition of the two arguments, and indicate whether an overflow
7677 occurred during the signed summation.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7678
7679<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7680<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]7681 be of integer types of any bit width, but they must have the same bit
7682 width. The second element of the result structure must be of
7683 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7684 undergo signed addition.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7685
7686<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7687<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7688 a signed addition of the two variables. They return a structure &mdash; the
7689 first element of which is the signed summation, and the second element of
7690 which is a bit specifying if the signed summation resulted in an
7691 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7692
7693<h5>Examples:</h5>
7694<pre>
7695 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
7696 %sum = extractvalue {i32, i1} %res, 0
7697 %obit = extractvalue {i32, i1} %res, 1
7698 br i1 %obit, label %overflow, label %normal
7699</pre>
7700
7701</div>
7702
7703<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7704<h4>
7705 <a name="int_uadd_overflow">
7706 '<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics
7707 </a>
7708</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7709
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7710<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7711
7712<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7713<p>This is an overloaded intrinsic. You can use <tt>llvm.uadd.with.overflow</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7714 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7715
7716<pre>
7717 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
7718 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
7719 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
7720</pre>
7721
7722<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7723<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7724 an unsigned addition of the two arguments, and indicate whether a carry
7725 occurred during the unsigned summation.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7726
7727<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7728<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7729 be of integer types of any bit width, but they must have the same bit
7730 width. The second element of the result structure must be of
7731 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7732 undergo unsigned addition.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7733
7734<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7735<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7736 an unsigned addition of the two arguments. They return a structure &mdash;
7737 the first element of which is the sum, and the second element of which is a
7738 bit specifying if the unsigned summation resulted in a carry.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7739
7740<h5>Examples:</h5>
7741<pre>
7742 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
7743 %sum = extractvalue {i32, i1} %res, 0
7744 %obit = extractvalue {i32, i1} %res, 1
7745 br i1 %obit, label %carry, label %normal
7746</pre>
7747
7748</div>
7749
7750<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7751<h4>
7752 <a name="int_ssub_overflow">
7753 '<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics
7754 </a>
7755</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7756
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7757<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7758
7759<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7760<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]7761 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7762
7763<pre>
7764 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
7765 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
7766 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
7767</pre>
7768
7769<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7770<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7771 a signed subtraction of the two arguments, and indicate whether an overflow
7772 occurred during the signed subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7773
7774<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7775<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]7776 be of integer types of any bit width, but they must have the same bit
7777 width. The second element of the result structure must be of
7778 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7779 undergo signed subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7780
7781<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7782<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7783 a signed subtraction of the two arguments. They return a structure &mdash;
7784 the first element of which is the subtraction, and the second element of
7785 which is a bit specifying if the signed subtraction resulted in an
7786 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7787
7788<h5>Examples:</h5>
7789<pre>
7790 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
7791 %sum = extractvalue {i32, i1} %res, 0
7792 %obit = extractvalue {i32, i1} %res, 1
7793 br i1 %obit, label %overflow, label %normal
7794</pre>
7795
7796</div>
7797
7798<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7799<h4>
7800 <a name="int_usub_overflow">
7801 '<tt>llvm.usub.with.overflow.*</tt>' Intrinsics
7802 </a>
7803</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7804
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7805<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7806
7807<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7808<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]7809 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7810
7811<pre>
7812 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
7813 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
7814 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
7815</pre>
7816
7817<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7818<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7819 an unsigned subtraction of the two arguments, and indicate whether an
7820 overflow occurred during the unsigned subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7821
7822<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7823<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]7824 be of integer types of any bit width, but they must have the same bit
7825 width. The second element of the result structure must be of
7826 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7827 undergo unsigned subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7828
7829<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7830<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7831 an unsigned subtraction of the two arguments. They return a structure &mdash;
7832 the first element of which is the subtraction, and the second element of
7833 which is a bit specifying if the unsigned subtraction resulted in an
7834 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7835
7836<h5>Examples:</h5>
7837<pre>
7838 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
7839 %sum = extractvalue {i32, i1} %res, 0
7840 %obit = extractvalue {i32, i1} %res, 1
7841 br i1 %obit, label %overflow, label %normal
7842</pre>
7843
7844</div>
7845
7846<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7847<h4>
7848 <a name="int_smul_overflow">
7849 '<tt>llvm.smul.with.overflow.*</tt>' Intrinsics
7850 </a>
7851</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7852
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7853<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7854
7855<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7856<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]7857 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7858
7859<pre>
7860 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
7861 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
7862 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
7863</pre>
7864
7865<h5>Overview:</h5>
7866
7867<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7868 a signed multiplication of the two arguments, and indicate whether an
7869 overflow occurred during the signed multiplication.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7870
7871<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7872<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]7873 be of integer types of any bit width, but they must have the same bit
7874 width. The second element of the result structure must be of
7875 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7876 undergo signed multiplication.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7877
7878<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7879<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7880 a signed multiplication of the two arguments. They return a structure &mdash;
7881 the first element of which is the multiplication, and the second element of
7882 which is a bit specifying if the signed multiplication resulted in an
7883 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7884
7885<h5>Examples:</h5>
7886<pre>
7887 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
7888 %sum = extractvalue {i32, i1} %res, 0
7889 %obit = extractvalue {i32, i1} %res, 1
7890 br i1 %obit, label %overflow, label %normal
7891</pre>
7892
Reid Spencerf86037f2007-04-11 23:23:49 +0000[diff] [blame]7893</div>
7894
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7895<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7896<h4>
7897 <a name="int_umul_overflow">
7898 '<tt>llvm.umul.with.overflow.*</tt>' Intrinsics
7899 </a>
7900</h4>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7901
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7902<div>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7903
7904<h5>Syntax:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7905<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]7906 on any integer bit width.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7907
7908<pre>
7909 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
7910 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7911 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
7912</pre>
7913
7914<h5>Overview:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7915<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7916 a unsigned multiplication of the two arguments, and indicate whether an
7917 overflow occurred during the unsigned multiplication.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7918
7919<h5>Arguments:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7920<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]7921 be of integer types of any bit width, but they must have the same bit
7922 width. The second element of the result structure must be of
7923 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7924 undergo unsigned multiplication.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7925
7926<h5>Semantics:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7927<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7928 an unsigned multiplication of the two arguments. They return a structure
7929 &mdash; the first element of which is the multiplication, and the second
7930 element of which is a bit specifying if the unsigned multiplication resulted
7931 in an overflow.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7932
7933<h5>Examples:</h5>
7934<pre>
7935 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7936 %sum = extractvalue {i32, i1} %res, 0
7937 %obit = extractvalue {i32, i1} %res, 1
7938 br i1 %obit, label %overflow, label %normal
7939</pre>
7940
7941</div>
7942
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7943</div>
7944
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]7945<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7946<h3>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7947 <a name="int_fp16">Half Precision Floating Point Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7948</h3>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7949
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7950<div>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7951
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]7952<p>Half precision floating point is a storage-only format. This means that it is
7953 a dense encoding (in memory) but does not support computation in the
7954 format.</p>
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]7955
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]7956<p>This means that code must first load the half-precision floating point
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]7957 value as an i16, then convert it to float with <a
7958 href="#int_convert_from_fp16"><tt>llvm.convert.from.fp16</tt></a>.
7959 Computation can then be performed on the float value (including extending to
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]7960 double etc). To store the value back to memory, it is first converted to
7961 float if needed, then converted to i16 with
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]7962 <a href="#int_convert_to_fp16"><tt>llvm.convert.to.fp16</tt></a>, then
7963 storing as an i16 value.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7964
7965<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7966<h4>
7967 <a name="int_convert_to_fp16">
7968 '<tt>llvm.convert.to.fp16</tt>' Intrinsic
7969 </a>
7970</h4>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7971
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7972<div>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7973
7974<h5>Syntax:</h5>
7975<pre>
7976 declare i16 @llvm.convert.to.fp16(f32 %a)
7977</pre>
7978
7979<h5>Overview:</h5>
7980<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
7981 a conversion from single precision floating point format to half precision
7982 floating point format.</p>
7983
7984<h5>Arguments:</h5>
7985<p>The intrinsic function contains single argument - the value to be
7986 converted.</p>
7987
7988<h5>Semantics:</h5>
7989<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
7990 a conversion from single precision floating point format to half precision
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]7991 floating point format. The return value is an <tt>i16</tt> which
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]7992 contains the converted number.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7993
7994<h5>Examples:</h5>
7995<pre>
7996 %res = call i16 @llvm.convert.to.fp16(f32 %a)
7997 store i16 %res, i16* @x, align 2
7998</pre>
7999
8000</div>
8001
8002<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8003<h4>
8004 <a name="int_convert_from_fp16">
8005 '<tt>llvm.convert.from.fp16</tt>' Intrinsic
8006 </a>
8007</h4>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]8008
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8009<div>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]8010
8011<h5>Syntax:</h5>
8012<pre>
8013 declare f32 @llvm.convert.from.fp16(i16 %a)
8014</pre>
8015
8016<h5>Overview:</h5>
8017<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs
8018 a conversion from half precision floating point format to single precision
8019 floating point format.</p>
8020
8021<h5>Arguments:</h5>
8022<p>The intrinsic function contains single argument - the value to be
8023 converted.</p>
8024
8025<h5>Semantics:</h5>
8026<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs a
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]8027 conversion from half single precision floating point format to single
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]8028 precision floating point format. The input half-float value is represented by
8029 an <tt>i16</tt> value.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]8030
8031<h5>Examples:</h5>
8032<pre>
8033 %a = load i16* @x, align 2
8034 %res = call f32 @llvm.convert.from.fp16(i16 %a)
8035</pre>
8036
8037</div>
8038
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8039</div>
8040
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]8041<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8042<h3>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]8043 <a name="int_debugger">Debugger Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8044</h3>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]8045
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8046<div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]8047
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8048<p>The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt>
8049 prefix), are described in
8050 the <a href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source
8051 Level Debugging</a> document.</p>
8052
8053</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]8054
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]8055<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8056<h3>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]8057 <a name="int_eh">Exception Handling Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8058</h3>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]8059
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8060<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8061
8062<p>The LLVM exception handling intrinsics (which all start with
8063 <tt>llvm.eh.</tt> prefix), are described in
8064 the <a href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
8065 Handling</a> document.</p>
8066
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]8067</div>
8068
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8069<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8070<h3>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]8071 <a name="int_trampoline">Trampoline Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8072</h3>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8073
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8074<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8075
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]8076<p>These intrinsics make it possible to excise one parameter, marked with
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]8077 the <a href="#nest"><tt>nest</tt></a> attribute, from a function.
8078 The result is a callable
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8079 function pointer lacking the nest parameter - the caller does not need to
8080 provide a value for it. Instead, the value to use is stored in advance in a
8081 "trampoline", a block of memory usually allocated on the stack, which also
8082 contains code to splice the nest value into the argument list. This is used
8083 to implement the GCC nested function address extension.</p>
8084
8085<p>For example, if the function is
8086 <tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
8087 pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as
8088 follows:</p>
8089
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]8090<pre class="doc_code">
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]8091 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
8092 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]8093 call i8* @llvm.init.trampoline(i8* %tramp1, i8* bitcast (i32 (i8*, i32, i32)* @f to i8*), i8* %nval)
8094 %p = call i8* @llvm.adjust.trampoline(i8* %tramp1)
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]8095 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8096</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8097
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8098<p>The call <tt>%val = call i32 %fp(i32 %x, i32 %y)</tt> is then equivalent
8099 to <tt>%val = call i32 %f(i8* %nval, i32 %x, i32 %y)</tt>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8100
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8101<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8102<h4>
8103 <a name="int_it">
8104 '<tt>llvm.init.trampoline</tt>' Intrinsic
8105 </a>
8106</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8107
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8108<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8109
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8110<h5>Syntax:</h5>
8111<pre>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]8112 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]8113</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8114
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8115<h5>Overview:</h5>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]8116<p>This fills the memory pointed to by <tt>tramp</tt> with executable code,
8117 turning it into a trampoline.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8118
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8119<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8120<p>The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
8121 pointers. The <tt>tramp</tt> argument must point to a sufficiently large and
8122 sufficiently aligned block of memory; this memory is written to by the
8123 intrinsic. Note that the size and the alignment are target-specific - LLVM
8124 currently provides no portable way of determining them, so a front-end that
8125 generates this intrinsic needs to have some target-specific knowledge.
8126 The <tt>func</tt> argument must hold a function bitcast to
8127 an <tt>i8*</tt>.</p>
8128
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8129<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8130<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]8131 dependent code, turning it into a function. Then <tt>tramp</tt> needs to be
8132 passed to <a href="#int_at">llvm.adjust.trampoline</a> to get a pointer
8133 which can be <a href="#int_trampoline">bitcast (to a new function) and
8134 called</a>. The new function's signature is the same as that of
8135 <tt>func</tt> with any arguments marked with the <tt>nest</tt> attribute
8136 removed. At most one such <tt>nest</tt> argument is allowed, and it must be of
8137 pointer type. Calling the new function is equivalent to calling <tt>func</tt>
8138 with the same argument list, but with <tt>nval</tt> used for the missing
8139 <tt>nest</tt> argument. If, after calling <tt>llvm.init.trampoline</tt>, the
8140 memory pointed to by <tt>tramp</tt> is modified, then the effect of any later call
8141 to the returned function pointer is undefined.</p>
8142</div>
8143
8144<!-- _______________________________________________________________________ -->
8145<h4>
8146 <a name="int_at">
8147 '<tt>llvm.adjust.trampoline</tt>' Intrinsic
8148 </a>
8149</h4>
8150
8151<div>
8152
8153<h5>Syntax:</h5>
8154<pre>
8155 declare i8* @llvm.adjust.trampoline(i8* &lt;tramp&gt;)
8156</pre>
8157
8158<h5>Overview:</h5>
8159<p>This performs any required machine-specific adjustment to the address of a
8160 trampoline (passed as <tt>tramp</tt>).</p>
8161
8162<h5>Arguments:</h5>
8163<p><tt>tramp</tt> must point to a block of memory which already has trampoline code
8164 filled in by a previous call to <a href="#int_it"><tt>llvm.init.trampoline</tt>
8165 </a>.</p>
8166
8167<h5>Semantics:</h5>
8168<p>On some architectures the address of the code to be executed needs to be
8169 different to the address where the trampoline is actually stored. This
8170 intrinsic returns the executable address corresponding to <tt>tramp</tt>
8171 after performing the required machine specific adjustments.
8172 The pointer returned can then be <a href="#int_trampoline"> bitcast and
8173 executed</a>.
8174</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8175
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8176</div>
8177
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8178</div>
8179
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8180<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8181<h3>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8182 <a name="int_memorymarkers">Memory Use Markers</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8183</h3>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8184
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8185<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8186
8187<p>This class of intrinsics exists to information about the lifetime of memory
8188 objects and ranges where variables are immutable.</p>
8189
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8190<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8191<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8192 <a name="int_lifetime_start">'<tt>llvm.lifetime.start</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8193</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8194
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8195<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8196
8197<h5>Syntax:</h5>
8198<pre>
8199 declare void @llvm.lifetime.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
8200</pre>
8201
8202<h5>Overview:</h5>
8203<p>The '<tt>llvm.lifetime.start</tt>' intrinsic specifies the start of a memory
8204 object's lifetime.</p>
8205
8206<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]8207<p>The first argument is a constant integer representing the size of the
8208 object, or -1 if it is variable sized. The second argument is a pointer to
8209 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8210
8211<h5>Semantics:</h5>
8212<p>This intrinsic indicates that before this point in the code, the value of the
8213 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]8214 never be used and has an undefined value. A load from the pointer that
8215 precedes this intrinsic can be replaced with
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8216 <tt>'<a href="#undefvalues">undef</a>'</tt>.</p>
8217
8218</div>
8219
8220<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8221<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8222 <a name="int_lifetime_end">'<tt>llvm.lifetime.end</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8223</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8224
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8225<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8226
8227<h5>Syntax:</h5>
8228<pre>
8229 declare void @llvm.lifetime.end(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
8230</pre>
8231
8232<h5>Overview:</h5>
8233<p>The '<tt>llvm.lifetime.end</tt>' intrinsic specifies the end of a memory
8234 object's lifetime.</p>
8235
8236<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]8237<p>The first argument is a constant integer representing the size of the
8238 object, or -1 if it is variable sized. The second argument is a pointer to
8239 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8240
8241<h5>Semantics:</h5>
8242<p>This intrinsic indicates that after this point in the code, the value of the
8243 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
8244 never be used and has an undefined value. Any stores into the memory object
8245 following this intrinsic may be removed as dead.
8246
8247</div>
8248
8249<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8250<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8251 <a name="int_invariant_start">'<tt>llvm.invariant.start</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8252</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8253
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8254<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8255
8256<h5>Syntax:</h5>
8257<pre>
Nick Lewycky29b6cb42010-11-30 04:13:41 +0000[diff] [blame]8258 declare {}* @llvm.invariant.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8259</pre>
8260
8261<h5>Overview:</h5>
8262<p>The '<tt>llvm.invariant.start</tt>' intrinsic specifies that the contents of
8263 a memory object will not change.</p>
8264
8265<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]8266<p>The first argument is a constant integer representing the size of the
8267 object, or -1 if it is variable sized. The second argument is a pointer to
8268 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8269
8270<h5>Semantics:</h5>
8271<p>This intrinsic indicates that until an <tt>llvm.invariant.end</tt> that uses
8272 the return value, the referenced memory location is constant and
8273 unchanging.</p>
8274
8275</div>
8276
8277<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8278<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8279 <a name="int_invariant_end">'<tt>llvm.invariant.end</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8280</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8281
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8282<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8283
8284<h5>Syntax:</h5>
8285<pre>
8286 declare void @llvm.invariant.end({}* &lt;start&gt;, i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
8287</pre>
8288
8289<h5>Overview:</h5>
8290<p>The '<tt>llvm.invariant.end</tt>' intrinsic specifies that the contents of
8291 a memory object are mutable.</p>
8292
8293<h5>Arguments:</h5>
8294<p>The first argument is the matching <tt>llvm.invariant.start</tt> intrinsic.
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]8295 The second argument is a constant integer representing the size of the
8296 object, or -1 if it is variable sized and the third argument is a pointer
8297 to the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8298
8299<h5>Semantics:</h5>
8300<p>This intrinsic indicates that the memory is mutable again.</p>
8301
8302</div>
8303
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8304</div>
8305
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]8306<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8307<h3>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8308 <a name="int_general">General Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8309</h3>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8310
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8311<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8312
8313<p>This class of intrinsics is designed to be generic and has no specific
8314 purpose.</p>
8315
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8316<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8317<h4>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8318 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8319</h4>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8320
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8321<div>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8322
8323<h5>Syntax:</h5>
8324<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8325 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]8326</pre>
8327
8328<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8329<p>The '<tt>llvm.var.annotation</tt>' intrinsic.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8330
8331<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8332<p>The first argument is a pointer to a value, the second is a pointer to a
8333 global string, the third is a pointer to a global string which is the source
8334 file name, and the last argument is the line number.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8335
8336<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8337<p>This intrinsic allows annotation of local variables with arbitrary strings.
8338 This can be useful for special purpose optimizations that want to look for
John Criswelle865c032011-08-19 16:57:55 +0000[diff] [blame]8339 these annotations. These have no other defined use; they are ignored by code
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8340 generation and optimization.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8341
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8342</div>
8343
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8344<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8345<h4>
Tanya Lattnere1a8da02007-09-21 23:57:59 +0000[diff] [blame]8346 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8347</h4>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8348
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8349<div>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8350
8351<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8352<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
8353 any integer bit width.</p>
8354
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8355<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8356 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8357 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8358 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8359 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8360 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]8361</pre>
8362
8363<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8364<p>The '<tt>llvm.annotation</tt>' intrinsic.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8365
8366<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8367<p>The first argument is an integer value (result of some expression), the
8368 second is a pointer to a global string, the third is a pointer to a global
8369 string which is the source file name, and the last argument is the line
8370 number. It returns the value of the first argument.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8371
8372<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8373<p>This intrinsic allows annotations to be put on arbitrary expressions with
8374 arbitrary strings. This can be useful for special purpose optimizations that
John Criswelle865c032011-08-19 16:57:55 +0000[diff] [blame]8375 want to look for these annotations. These have no other defined use; they
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8376 are ignored by code generation and optimization.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8377
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8378</div>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]8379
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8380<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8381<h4>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8382 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8383</h4>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8384
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8385<div>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8386
8387<h5>Syntax:</h5>
8388<pre>
8389 declare void @llvm.trap()
8390</pre>
8391
8392<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8393<p>The '<tt>llvm.trap</tt>' intrinsic.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8394
8395<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8396<p>None.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8397
8398<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8399<p>This intrinsics is lowered to the target dependent trap instruction. If the
8400 target does not have a trap instruction, this intrinsic will be lowered to
8401 the call of the <tt>abort()</tt> function.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8402
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8403</div>
8404
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8405<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8406<h4>
Misha Brukmandccb0252008-11-22 23:55:29 +0000[diff] [blame]8407 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8408</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8409
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8410<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8411
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8412<h5>Syntax:</h5>
8413<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8414 declare void @llvm.stackprotector(i8* &lt;guard&gt;, i8** &lt;slot&gt;)
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8415</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8416
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8417<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8418<p>The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and
8419 stores it onto the stack at <tt>slot</tt>. The stack slot is adjusted to
8420 ensure that it is placed on the stack before local variables.</p>
8421
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8422<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8423<p>The <tt>llvm.stackprotector</tt> intrinsic requires two pointer
8424 arguments. The first argument is the value loaded from the stack
8425 guard <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt>
8426 that has enough space to hold the value of the guard.</p>
8427
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8428<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8429<p>This intrinsic causes the prologue/epilogue inserter to force the position of
8430 the <tt>AllocaInst</tt> stack slot to be before local variables on the
8431 stack. This is to ensure that if a local variable on the stack is
8432 overwritten, it will destroy the value of the guard. When the function exits,
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8433 the guard on the stack is checked against the original guard. If they are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8434 different, then the program aborts by calling the <tt>__stack_chk_fail()</tt>
8435 function.</p>
8436
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8437</div>
8438
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8439<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8440<h4>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8441 <a name="int_objectsize">'<tt>llvm.objectsize</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8442</h4>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8443
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8444<div>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8445
8446<h5>Syntax:</h5>
8447<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8448 declare i32 @llvm.objectsize.i32(i8* &lt;object&gt;, i1 &lt;type&gt;)
8449 declare i64 @llvm.objectsize.i64(i8* &lt;object&gt;, i1 &lt;type&gt;)
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8450</pre>
8451
8452<h5>Overview:</h5>
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8453<p>The <tt>llvm.objectsize</tt> intrinsic is designed to provide information to
8454 the optimizers to determine at compile time whether a) an operation (like
8455 memcpy) will overflow a buffer that corresponds to an object, or b) that a
8456 runtime check for overflow isn't necessary. An object in this context means
8457 an allocation of a specific class, structure, array, or other object.</p>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8458
8459<h5>Arguments:</h5>
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8460<p>The <tt>llvm.objectsize</tt> intrinsic takes two arguments. The first
Eric Christopher8295a0a2009-12-23 00:29:49 +0000[diff] [blame]8461 argument is a pointer to or into the <tt>object</tt>. The second argument
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8462 is a boolean 0 or 1. This argument determines whether you want the
8463 maximum (0) or minimum (1) bytes remaining. This needs to be a literal 0 or
Eric Christopher8295a0a2009-12-23 00:29:49 +0000[diff] [blame]8464 1, variables are not allowed.</p>
8465
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8466<h5>Semantics:</h5>
8467<p>The <tt>llvm.objectsize</tt> intrinsic is lowered to either a constant
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8468 representing the size of the object concerned, or <tt>i32/i64 -1 or 0</tt>,
8469 depending on the <tt>type</tt> argument, if the size cannot be determined at
8470 compile time.</p>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8471
8472</div>
Jakub Staszakb170e2d2011-12-04 18:29:26 +0000[diff] [blame]8473<!-- _______________________________________________________________________ -->
8474<h4>
8475 <a name="int_expect">'<tt>llvm.expect</tt>' Intrinsic</a>
8476</h4>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8477
Jakub Staszakb170e2d2011-12-04 18:29:26 +0000[diff] [blame]8478<div>
8479
8480<h5>Syntax:</h5>
8481<pre>
8482 declare i32 @llvm.expect.i32(i32 &lt;val&gt;, i32 &lt;expected_val&gt;)
8483 declare i64 @llvm.expect.i64(i64 &lt;val&gt;, i64 &lt;expected_val&gt;)
8484</pre>
8485
8486<h5>Overview:</h5>
8487<p>The <tt>llvm.expect</tt> intrinsic provides information about expected (the
8488 most probable) value of <tt>val</tt>, which can be used by optimizers.</p>
8489
8490<h5>Arguments:</h5>
8491<p>The <tt>llvm.expect</tt> intrinsic takes two arguments. The first
8492 argument is a value. The second argument is an expected value, this needs to
8493 be a constant value, variables are not allowed.</p>
8494
8495<h5>Semantics:</h5>
8496<p>This intrinsic is lowered to the <tt>val</tt>.</p>
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8497</div>
8498
8499</div>
8500
Jakub Staszakb170e2d2011-12-04 18:29:26 +0000[diff] [blame]8501</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]8502<!-- *********************************************************************** -->
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]8503<hr>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]8504<address>
8505 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]8506 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]8507 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]8508 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]8509
8510 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
NAKAMURA Takumib9a33632011-04-09 02:13:37 +0000[diff] [blame]8511 <a href="http://llvm.org/">The LLVM Compiler Infrastructure</a><br>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]8512 Last modified: $Date$
8513</address>
Chris Lattnerc7d3ab32008-01-04 04:33:49 +0000[diff] [blame]8514
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]8515</body>
8516</html>