Gitiles
Code Review Sign In
gerrit-public.fairphone.software / toolchain / llvm-project / 3f6a3a277c148cd2c7eead3c12e2f6e656f4353c / . / llvm / docs / LangRef.html
blob: a71cbfc092715a4cabcd684dd04024ff935a83c7 [file] [log] [blame]
Misha Brukmanc501f552004-03-01 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 Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3<html>
4<head>
5 <title>LLVM Assembly Language Reference Manual</title>
Reid Spencercb84e432004-08-26 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 Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]8 <meta name="description"
Reid Spencercb84e432004-08-26 20:44:00 +0000[diff] [blame]9 content="LLVM Assembly Language Reference Manual.">
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]10 <link rel="stylesheet" href="llvm.css" type="text/css">
11</head>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]12
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]13<body>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]14
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]15<h1>LLVM Language Reference Manual</h1>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]16<ol>
Misha Brukman76307852003-11-08 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 Lattner6af02f32004-12-09 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 Wendlinga3c6f6b2009-07-20 01:03:30 +0000[diff] [blame]23 <li><a href="#linkage">Linkage Types</a>
24 <ol>
Bill Wendling8693ef82009-07-20 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 Wendling03bcd6e2010-07-01 21:55:59 +0000[diff] [blame]27 <li><a href="#linkage_linker_private_weak">'<tt>linker_private_weak</tt>' Linkage</a></li>
Bill Wendling578ee402010-08-20 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 Wendling8693ef82009-07-20 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 Lattner80d73c72009-10-10 18:26:06 +0000[diff] [blame]36 <li><a href="#linkage_linkonce_odr">'<tt>linkonce_odr</tt>' Linkage</a></li>
Bill Wendling8693ef82009-07-20 02:41:50 +0000[diff] [blame]37 <li><a href="#linkage_weak">'<tt>weak_odr</tt>' Linkage</a></li>
Bill Wendlingb4d076e2011-10-11 06:41:28 +0000[diff] [blame]38 <li><a href="#linkage_external">'<tt>external</tt>' Linkage</a></li>
Bill Wendling8693ef82009-07-20 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 Wendlinga3c6f6b2009-07-20 01:03:30 +0000[diff] [blame]41 </ol>
42 </li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]43 <li><a href="#callingconv">Calling Conventions</a></li>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]44 <li><a href="#namedtypes">Named Types</a></li>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]45 <li><a href="#globalvars">Global Variables</a></li>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]46 <li><a href="#functionstructure">Functions</a></li>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]47 <li><a href="#aliasstructure">Aliases</a></li>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]48 <li><a href="#namedmetadatastructure">Named Metadata</a></li>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]49 <li><a href="#paramattrs">Parameter Attributes</a></li>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]50 <li><a href="#fnattrs">Function Attributes</a></li>
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]51 <li><a href="#gc">Garbage Collector Names</a></li>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]52 <li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]53 <li><a href="#datalayout">Data Layout</a></li>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]54 <li><a href="#pointeraliasing">Pointer Aliasing Rules</a></li>
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]55 <li><a href="#volatile">Volatile Memory Accesses</a></li>
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]56 <li><a href="#memmodel">Memory Model for Concurrent Operations</a></li>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]57 <li><a href="#ordering">Atomic Memory Ordering Constraints</a></li>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]58 </ol>
59 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]60 <li><a href="#typesystem">Type System</a>
61 <ol>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]62 <li><a href="#t_classifications">Type Classifications</a></li>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]63 <li><a href="#t_primitive">Primitive Types</a>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]64 <ol>
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]65 <li><a href="#t_integer">Integer Type</a></li>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]66 <li><a href="#t_floating">Floating Point Types</a></li>
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]67 <li><a href="#t_x86mmx">X86mmx Type</a></li>
Chris Lattner7824d182008-01-04 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 Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]70 <li><a href="#t_metadata">Metadata Type</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]71 </ol>
72 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]73 <li><a href="#t_derived">Derived Types</a>
74 <ol>
Chris Lattner392be582010-02-12 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 Lattner2a843822011-07-23 19:59:08 +0000[diff] [blame]79 <li><a href="#t_opaque">Opaque Structure Types</a></li>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]80 <li><a href="#t_vector">Vector Type</a></li>
81 </ol>
82 </li>
Misha Brukman76307852003-11-08 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 Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]85 </ol>
86 </li>
87 </ol>
88 </li>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]89 <li><a href="#constants">Constants</a>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]90 <ol>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]91 <li><a href="#simpleconstants">Simple Constants</a></li>
Chris Lattner361bfcd2009-02-28 18:32:25 +0000[diff] [blame]92 <li><a href="#complexconstants">Complex Constants</a></li>
Dan Gohmanef9462f2008-10-14 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 Gohman9a2a0932011-12-06 03:18:47 +0000[diff] [blame]95 <li><a href="#poisonvalues">Poison Values</a></li>
Chris Lattner2bfd3202009-10-27 21:19:13 +0000[diff] [blame]96 <li><a href="#blockaddress">Addresses of Basic Blocks</a></li>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]97 <li><a href="#constantexprs">Constant Expressions</a></li>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]98 </ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]99 </li>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]100 <li><a href="#othervalues">Other Values</a>
101 <ol>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]102 <li><a href="#inlineasm">Inline Assembler Expressions</a></li>
Peter Collingbourneec9ff672011-10-27 19:19:07 +0000[diff] [blame]103 <li><a href="#metadata">Metadata Nodes and Metadata Strings</a>
104 <ol>
105 <li><a href="#tbaa">'<tt>tbaa</tt>' Metadata</a></li>
Peter Collingbournef7d1e7b2011-10-27 19:19:14 +0000[diff] [blame]106 <li><a href="#fpaccuracy">'<tt>fpaccuracy</tt>' Metadata</a></li>
Peter Collingbourneec9ff672011-10-27 19:19:07 +0000[diff] [blame]107 </ol>
108 </li>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]109 </ol>
110 </li>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]111 <li><a href="#intrinsic_globals">Intrinsic Global Variables</a>
112 <ol>
113 <li><a href="#intg_used">The '<tt>llvm.used</tt>' Global Variable</a></li>
Chris Lattner58f9bb22009-07-20 06:14:25 +0000[diff] [blame]114 <li><a href="#intg_compiler_used">The '<tt>llvm.compiler.used</tt>'
115 Global Variable</a></li>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]116 <li><a href="#intg_global_ctors">The '<tt>llvm.global_ctors</tt>'
117 Global Variable</a></li>
118 <li><a href="#intg_global_dtors">The '<tt>llvm.global_dtors</tt>'
119 Global Variable</a></li>
120 </ol>
121 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]122 <li><a href="#instref">Instruction Reference</a>
123 <ol>
124 <li><a href="#terminators">Terminator Instructions</a>
125 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]126 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
127 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]128 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]129 <li><a href="#i_indirectbr">'<tt>indirectbr</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]130 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
Bill Wendlingf891bf82011-07-31 06:30:59 +0000[diff] [blame]131 <li><a href="#i_resume">'<tt>resume</tt>' Instruction</a></li>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]132 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]133 </ol>
134 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]135 <li><a href="#binaryops">Binary Operations</a>
136 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]137 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]138 <li><a href="#i_fadd">'<tt>fadd</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]139 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]140 <li><a href="#i_fsub">'<tt>fsub</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]141 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]142 <li><a href="#i_fmul">'<tt>fmul</tt>' Instruction</a></li>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]143 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
144 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
145 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]146 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
147 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
148 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]149 </ol>
150 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]151 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
152 <ol>
Reid Spencer2ab01932007-02-02 13:57:07 +0000[diff] [blame]153 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
154 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
155 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]156 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]157 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]158 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]159 </ol>
160 </li>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]161 <li><a href="#vectorops">Vector Operations</a>
162 <ol>
163 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
164 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
165 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]166 </ol>
167 </li>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]168 <li><a href="#aggregateops">Aggregate Operations</a>
169 <ol>
170 <li><a href="#i_extractvalue">'<tt>extractvalue</tt>' Instruction</a></li>
171 <li><a href="#i_insertvalue">'<tt>insertvalue</tt>' Instruction</a></li>
172 </ol>
173 </li>
Chris Lattner6ab66722006-08-15 00:45:58 +0000[diff] [blame]174 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]175 <ol>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]176 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
177 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
178 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
179 <li><a href="#i_fence">'<tt>fence</tt>' Instruction</a></li>
180 <li><a href="#i_cmpxchg">'<tt>cmpxchg</tt>' Instruction</a></li>
181 <li><a href="#i_atomicrmw">'<tt>atomicrmw</tt>' Instruction</a></li>
Robert Bocchino820bc75b2006-02-17 21:18:08 +0000[diff] [blame]182 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]183 </ol>
184 </li>
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]185 <li><a href="#convertops">Conversion Operations</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]186 <ol>
187 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
188 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
189 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
190 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
191 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]192 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
193 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
194 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
195 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]196 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
197 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]198 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]199 </ol>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]200 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]201 <li><a href="#otherops">Other Operations</a>
202 <ol>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]203 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
204 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]205 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]206 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]207 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Chris Lattner33337472006-01-13 23:26:01 +0000[diff] [blame]208 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]209 <li><a href="#i_landingpad">'<tt>landingpad</tt>' Instruction</a></li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]210 </ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]211 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]212 </ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]213 </li>
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]214 <li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]215 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]216 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
217 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]218 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
219 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
220 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]221 </ol>
222 </li>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]223 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
224 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]225 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
226 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
227 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]228 </ol>
229 </li>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]230 <li><a href="#int_codegen">Code Generator Intrinsics</a>
231 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]232 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
233 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
234 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
235 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
236 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
237 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
Dan Gohmane58f7b32010-05-26 21:56:15 +0000[diff] [blame]238 <li><a href="#int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswellaa1c3c12004-04-09 16:43:20 +0000[diff] [blame]239 </ol>
240 </li>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]241 <li><a href="#int_libc">Standard C Library Intrinsics</a>
242 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]243 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
244 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
245 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
246 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
247 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]248 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
249 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
250 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Dan Gohmane635c522011-05-27 00:36:31 +0000[diff] [blame]251 <li><a href="#int_exp">'<tt>llvm.exp.*</tt>' Intrinsic</a></li>
252 <li><a href="#int_log">'<tt>llvm.log.*</tt>' Intrinsic</a></li>
Cameron Zwarichf03fa182011-07-08 21:39:21 +0000[diff] [blame]253 <li><a href="#int_fma">'<tt>llvm.fma.*</tt>' Intrinsic</a></li>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]254 </ol>
255 </li>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]256 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]257 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]258 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattnerb748c672006-01-16 22:34:14 +0000[diff] [blame]259 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
260 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
261 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]262 </ol>
263 </li>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]264 <li><a href="#int_overflow">Arithmetic with Overflow Intrinsics</a>
265 <ol>
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]266 <li><a href="#int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt> Intrinsics</a></li>
267 <li><a href="#int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt> Intrinsics</a></li>
268 <li><a href="#int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt> Intrinsics</a></li>
269 <li><a href="#int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt> Intrinsics</a></li>
270 <li><a href="#int_smul_overflow">'<tt>llvm.smul.with.overflow.*</tt> Intrinsics</a></li>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]271 <li><a href="#int_umul_overflow">'<tt>llvm.umul.with.overflow.*</tt> Intrinsics</a></li>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]272 </ol>
273 </li>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]274 <li><a href="#int_fp16">Half Precision Floating Point Intrinsics</a>
275 <ol>
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]276 <li><a href="#int_convert_to_fp16">'<tt>llvm.convert.to.fp16</tt>' Intrinsic</a></li>
277 <li><a href="#int_convert_from_fp16">'<tt>llvm.convert.from.fp16</tt>' Intrinsic</a></li>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]278 </ol>
279 </li>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]280 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]281 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sandsa0984362011-09-06 13:37:06 +0000[diff] [blame]282 <li><a href="#int_trampoline">Trampoline Intrinsics</a>
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]283 <ol>
284 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sandsa0984362011-09-06 13:37:06 +0000[diff] [blame]285 <li><a href="#int_at">'<tt>llvm.adjust.trampoline</tt>' Intrinsic</a></li>
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]286 </ol>
287 </li>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]288 <li><a href="#int_memorymarkers">Memory Use Markers</a>
289 <ol>
Jakub Staszak5fd147f2011-12-04 20:44:25 +0000[diff] [blame]290 <li><a href="#int_lifetime_start">'<tt>llvm.lifetime.start</tt>' Intrinsic</a></li>
291 <li><a href="#int_lifetime_end">'<tt>llvm.lifetime.end</tt>' Intrinsic</a></li>
292 <li><a href="#int_invariant_start">'<tt>llvm.invariant.start</tt>' Intrinsic</a></li>
293 <li><a href="#int_invariant_end">'<tt>llvm.invariant.end</tt>' Intrinsic</a></li>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]294 </ol>
295 </li>
Reid Spencer5b2cb0f2007-07-20 19:59:11 +0000[diff] [blame]296 <li><a href="#int_general">General intrinsics</a>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]297 <ol>
Reid Spencer5b2cb0f2007-07-20 19:59:11 +0000[diff] [blame]298 <li><a href="#int_var_annotation">
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]299 '<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]300 <li><a href="#int_annotation">
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]301 '<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]302 <li><a href="#int_trap">
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]303 '<tt>llvm.trap</tt>' Intrinsic</a></li>
304 <li><a href="#int_stackprotector">
305 '<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]306 <li><a href="#int_objectsize">
307 '<tt>llvm.objectsize</tt>' Intrinsic</a></li>
Jakub Staszak5fef7922011-12-04 18:29:26 +0000[diff] [blame]308 <li><a href="#int_expect">
309 '<tt>llvm.expect</tt>' Intrinsic</a></li>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]310 </ol>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]311 </li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]312 </ol>
313 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]314</ol>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]315
316<div class="doc_author">
317 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
318 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]319</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]320
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]321<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]322<h2><a name="abstract">Abstract</a></h2>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]323<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]324
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]325<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]326
327<p>This document is a reference manual for the LLVM assembly language. LLVM is
328 a Static Single Assignment (SSA) based representation that provides type
329 safety, low-level operations, flexibility, and the capability of representing
330 'all' high-level languages cleanly. It is the common code representation
331 used throughout all phases of the LLVM compilation strategy.</p>
332
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]333</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]334
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]335<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]336<h2><a name="introduction">Introduction</a></h2>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]337<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]338
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]339<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]340
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]341<p>The LLVM code representation is designed to be used in three different forms:
342 as an in-memory compiler IR, as an on-disk bitcode representation (suitable
343 for fast loading by a Just-In-Time compiler), and as a human readable
344 assembly language representation. This allows LLVM to provide a powerful
345 intermediate representation for efficient compiler transformations and
346 analysis, while providing a natural means to debug and visualize the
347 transformations. The three different forms of LLVM are all equivalent. This
348 document describes the human readable representation and notation.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]349
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]350<p>The LLVM representation aims to be light-weight and low-level while being
351 expressive, typed, and extensible at the same time. It aims to be a
352 "universal IR" of sorts, by being at a low enough level that high-level ideas
353 may be cleanly mapped to it (similar to how microprocessors are "universal
354 IR's", allowing many source languages to be mapped to them). By providing
355 type information, LLVM can be used as the target of optimizations: for
356 example, through pointer analysis, it can be proven that a C automatic
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]357 variable is never accessed outside of the current function, allowing it to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]358 be promoted to a simple SSA value instead of a memory location.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]359
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]360<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]361<h4>
362 <a name="wellformed">Well-Formedness</a>
363</h4>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]364
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]365<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]366
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]367<p>It is important to note that this document describes 'well formed' LLVM
368 assembly language. There is a difference between what the parser accepts and
369 what is considered 'well formed'. For example, the following instruction is
370 syntactically okay, but not well formed:</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]371
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]372<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]373%x = <a href="#i_add">add</a> i32 1, %x
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]374</pre>
375
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]376<p>because the definition of <tt>%x</tt> does not dominate all of its uses. The
377 LLVM infrastructure provides a verification pass that may be used to verify
378 that an LLVM module is well formed. This pass is automatically run by the
379 parser after parsing input assembly and by the optimizer before it outputs
380 bitcode. The violations pointed out by the verifier pass indicate bugs in
381 transformation passes or input to the parser.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]382
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]383</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]384
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]385</div>
386
Chris Lattner87a3dbe2007-10-03 17:34:29 +0000[diff] [blame]387<!-- Describe the typesetting conventions here. -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]388
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]389<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]390<h2><a name="identifiers">Identifiers</a></h2>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]391<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]392
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]393<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]394
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]395<p>LLVM identifiers come in two basic types: global and local. Global
396 identifiers (functions, global variables) begin with the <tt>'@'</tt>
397 character. Local identifiers (register names, types) begin with
398 the <tt>'%'</tt> character. Additionally, there are three different formats
399 for identifiers, for different purposes:</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]400
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]401<ol>
Reid Spencerb23b65f2007-08-07 14:34:28 +0000[diff] [blame]402 <li>Named values are represented as a string of characters with their prefix.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]403 For example, <tt>%foo</tt>, <tt>@DivisionByZero</tt>,
404 <tt>%a.really.long.identifier</tt>. The actual regular expression used is
405 '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'. Identifiers which require
406 other characters in their names can be surrounded with quotes. Special
407 characters may be escaped using <tt>"\xx"</tt> where <tt>xx</tt> is the
408 ASCII code for the character in hexadecimal. In this way, any character
409 can be used in a name value, even quotes themselves.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]410
Reid Spencerb23b65f2007-08-07 14:34:28 +0000[diff] [blame]411 <li>Unnamed values are represented as an unsigned numeric value with their
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]412 prefix. For example, <tt>%12</tt>, <tt>@2</tt>, <tt>%44</tt>.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]413
Reid Spencer8f08d802004-12-09 18:02:53 +0000[diff] [blame]414 <li>Constants, which are described in a <a href="#constants">section about
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]415 constants</a>, below.</li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]416</ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]417
Reid Spencerb23b65f2007-08-07 14:34:28 +0000[diff] [blame]418<p>LLVM requires that values start with a prefix for two reasons: Compilers
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]419 don't need to worry about name clashes with reserved words, and the set of
420 reserved words may be expanded in the future without penalty. Additionally,
421 unnamed identifiers allow a compiler to quickly come up with a temporary
422 variable without having to avoid symbol table conflicts.</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]423
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]424<p>Reserved words in LLVM are very similar to reserved words in other
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]425 languages. There are keywords for different opcodes
426 ('<tt><a href="#i_add">add</a></tt>',
427 '<tt><a href="#i_bitcast">bitcast</a></tt>',
428 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names
429 ('<tt><a href="#t_void">void</a></tt>',
430 '<tt><a href="#t_primitive">i32</a></tt>', etc...), and others. These
431 reserved words cannot conflict with variable names, because none of them
432 start with a prefix character (<tt>'%'</tt> or <tt>'@'</tt>).</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]433
434<p>Here is an example of LLVM code to multiply the integer variable
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]435 '<tt>%X</tt>' by 8:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]436
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]437<p>The easy way:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]438
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]439<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]440%result = <a href="#i_mul">mul</a> i32 %X, 8
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]441</pre>
442
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]443<p>After strength reduction:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]444
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]445<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]446%result = <a href="#i_shl">shl</a> i32 %X, i8 3
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]447</pre>
448
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]449<p>And the hard way:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]450
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]451<pre class="doc_code">
Gabor Greifbd0328f2009-10-28 13:05:07 +0000[diff] [blame]452%0 = <a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
453%1 = <a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]454%result = <a href="#i_add">add</a> i32 %1, %1
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]455</pre>
456
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]457<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several important
458 lexical features of LLVM:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]459
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]460<ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]461 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]462 line.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]463
464 <li>Unnamed temporaries are created when the result of a computation is not
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]465 assigned to a named value.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]466
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]467 <li>Unnamed temporaries are numbered sequentially</li>
468</ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]469
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]470<p>It also shows a convention that we follow in this document. When
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]471 demonstrating instructions, we will follow an instruction with a comment that
472 defines the type and name of value produced. Comments are shown in italic
473 text.</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]474
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]475</div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]476
477<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]478<h2><a name="highlevel">High Level Structure</a></h2>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]479<!-- *********************************************************************** -->
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]480<div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]481<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]482<h3>
483 <a name="modulestructure">Module Structure</a>
484</h3>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]485
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]486<div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]487
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]488<p>LLVM programs are composed of "Module"s, each of which is a translation unit
489 of the input programs. Each module consists of functions, global variables,
490 and symbol table entries. Modules may be combined together with the LLVM
491 linker, which merges function (and global variable) definitions, resolves
492 forward declarations, and merges symbol table entries. Here is an example of
493 the "hello world" module:</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]494
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]495<pre class="doc_code">
Chris Lattner54a7be72010-08-17 17:13:42 +0000[diff] [blame]496<i>; Declare the string constant as a global constant.</i>&nbsp;
Nick Lewyckyfea7ddc2011-01-29 01:09:53 +0000[diff] [blame]497<a href="#identifiers">@.LC0</a> = <a href="#linkage_internal">internal</a>&nbsp;<a href="#globalvars">constant</a>&nbsp;<a href="#t_array">[13 x i8]</a> c"hello world\0A\00" <i>; [13 x i8]*</i>&nbsp;
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]498
Chris Lattner54a7be72010-08-17 17:13:42 +0000[diff] [blame]499<i>; External declaration of the puts function</i>&nbsp;
500<a href="#functionstructure">declare</a> i32 @puts(i8*) <i>; i32 (i8*)* </i>&nbsp;
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]501
502<i>; Definition of main function</i>
Chris Lattner54a7be72010-08-17 17:13:42 +0000[diff] [blame]503define i32 @main() { <i>; i32()* </i>&nbsp;
504 <i>; Convert [13 x i8]* to i8 *...</i>&nbsp;
505 %cast210 = <a href="#i_getelementptr">getelementptr</a> [13 x i8]* @.LC0, i64 0, i64 0 <i>; i8*</i>&nbsp;
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]506
Chris Lattner54a7be72010-08-17 17:13:42 +0000[diff] [blame]507 <i>; Call puts function to write out the string to stdout.</i>&nbsp;
508 <a href="#i_call">call</a> i32 @puts(i8* %cast210) <i>; i32</i>&nbsp;
509 <a href="#i_ret">ret</a> i32 0&nbsp;
510}
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]511
512<i>; Named metadata</i>
513!1 = metadata !{i32 41}
514!foo = !{!1, null}
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]515</pre>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]516
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]517<p>This example is made up of a <a href="#globalvars">global variable</a> named
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]518 "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>" function,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]519 a <a href="#functionstructure">function definition</a> for
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]520 "<tt>main</tt>" and <a href="#namedmetadatastructure">named metadata</a>
521 "<tt>foo"</tt>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]522
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]523<p>In general, a module is made up of a list of global values, where both
524 functions and global variables are global values. Global values are
525 represented by a pointer to a memory location (in this case, a pointer to an
526 array of char, and a pointer to a function), and have one of the
527 following <a href="#linkage">linkage types</a>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]528
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]529</div>
530
531<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]532<h3>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]533 <a name="linkage">Linkage Types</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]534</h3>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]535
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]536<div>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]537
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]538<p>All Global Variables and Functions have one of the following types of
539 linkage:</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]540
541<dl>
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]542 <dt><tt><b><a name="linkage_private">private</a></b></tt></dt>
Bill Wendling03bcd6e2010-07-01 21:55:59 +0000[diff] [blame]543 <dd>Global values with "<tt>private</tt>" linkage are only directly accessible
544 by objects in the current module. In particular, linking code into a
545 module with an private global value may cause the private to be renamed as
546 necessary to avoid collisions. Because the symbol is private to the
547 module, all references can be updated. This doesn't show up in any symbol
548 table in the object file.</dd>
Rafael Espindola6de96a12009-01-15 20:18:42 +0000[diff] [blame]549
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]550 <dt><tt><b><a name="linkage_linker_private">linker_private</a></b></tt></dt>
Bill Wendling03bcd6e2010-07-01 21:55:59 +0000[diff] [blame]551 <dd>Similar to <tt>private</tt>, but the symbol is passed through the
552 assembler and evaluated by the linker. Unlike normal strong symbols, they
553 are removed by the linker from the final linked image (executable or
554 dynamic library).</dd>
555
556 <dt><tt><b><a name="linkage_linker_private_weak">linker_private_weak</a></b></tt></dt>
557 <dd>Similar to "<tt>linker_private</tt>", but the symbol is weak. Note that
558 <tt>linker_private_weak</tt> symbols are subject to coalescing by the
559 linker. The symbols are removed by the linker from the final linked image
560 (executable or dynamic library).</dd>
Bill Wendlinga3c6f6b2009-07-20 01:03:30 +0000[diff] [blame]561
Bill Wendling578ee402010-08-20 22:05:50 +0000[diff] [blame]562 <dt><tt><b><a name="linkage_linker_private_weak_def_auto">linker_private_weak_def_auto</a></b></tt></dt>
563 <dd>Similar to "<tt>linker_private_weak</tt>", but it's known that the address
564 of the object is not taken. For instance, functions that had an inline
565 definition, but the compiler decided not to inline it. Note,
566 unlike <tt>linker_private</tt> and <tt>linker_private_weak</tt>,
567 <tt>linker_private_weak_def_auto</tt> may have only <tt>default</tt>
568 visibility. The symbols are removed by the linker from the final linked
569 image (executable or dynamic library).</dd>
570
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]571 <dt><tt><b><a name="linkage_internal">internal</a></b></tt></dt>
Bill Wendling36321712010-06-29 22:34:52 +0000[diff] [blame]572 <dd>Similar to private, but the value shows as a local symbol
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]573 (<tt>STB_LOCAL</tt> in the case of ELF) in the object file. This
574 corresponds to the notion of the '<tt>static</tt>' keyword in C.</dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]575
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]576 <dt><tt><b><a name="linkage_available_externally">available_externally</a></b></tt></dt>
Chris Lattner184f1be2009-04-13 05:44:34 +0000[diff] [blame]577 <dd>Globals with "<tt>available_externally</tt>" linkage are never emitted
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]578 into the object file corresponding to the LLVM module. They exist to
579 allow inlining and other optimizations to take place given knowledge of
580 the definition of the global, which is known to be somewhere outside the
581 module. Globals with <tt>available_externally</tt> linkage are allowed to
582 be discarded at will, and are otherwise the same as <tt>linkonce_odr</tt>.
583 This linkage type is only allowed on definitions, not declarations.</dd>
Chris Lattner184f1be2009-04-13 05:44:34 +0000[diff] [blame]584
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]585 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt></dt>
Chris Lattnere20b4702007-01-14 06:51:48 +0000[diff] [blame]586 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
Chris Lattner0de4caa2010-01-09 19:15:14 +0000[diff] [blame]587 the same name when linkage occurs. This can be used to implement
588 some forms of inline functions, templates, or other code which must be
589 generated in each translation unit that uses it, but where the body may
590 be overridden with a more definitive definition later. Unreferenced
591 <tt>linkonce</tt> globals are allowed to be discarded. Note that
592 <tt>linkonce</tt> linkage does not actually allow the optimizer to
593 inline the body of this function into callers because it doesn't know if
594 this definition of the function is the definitive definition within the
595 program or whether it will be overridden by a stronger definition.
596 To enable inlining and other optimizations, use "<tt>linkonce_odr</tt>"
597 linkage.</dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]598
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]599 <dt><tt><b><a name="linkage_weak">weak</a></b></tt></dt>
Chris Lattnerd0554882009-08-05 05:21:07 +0000[diff] [blame]600 <dd>"<tt>weak</tt>" linkage has the same merging semantics as
601 <tt>linkonce</tt> linkage, except that unreferenced globals with
602 <tt>weak</tt> linkage may not be discarded. This is used for globals that
603 are declared "weak" in C source code.</dd>
604
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]605 <dt><tt><b><a name="linkage_common">common</a></b></tt></dt>
Chris Lattnerd0554882009-08-05 05:21:07 +0000[diff] [blame]606 <dd>"<tt>common</tt>" linkage is most similar to "<tt>weak</tt>" linkage, but
607 they are used for tentative definitions in C, such as "<tt>int X;</tt>" at
608 global scope.
609 Symbols with "<tt>common</tt>" linkage are merged in the same way as
610 <tt>weak symbols</tt>, and they may not be deleted if unreferenced.
Chris Lattner0aff0b22009-08-05 05:41:44 +0000[diff] [blame]611 <tt>common</tt> symbols may not have an explicit section,
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]612 must have a zero initializer, and may not be marked '<a
Chris Lattner0aff0b22009-08-05 05:41:44 +0000[diff] [blame]613 href="#globalvars"><tt>constant</tt></a>'. Functions and aliases may not
614 have common linkage.</dd>
Chris Lattnerd0554882009-08-05 05:21:07 +0000[diff] [blame]615
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]616
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]617 <dt><tt><b><a name="linkage_appending">appending</a></b></tt></dt>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]618 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]619 pointer to array type. When two global variables with appending linkage
620 are linked together, the two global arrays are appended together. This is
621 the LLVM, typesafe, equivalent of having the system linker append together
622 "sections" with identical names when .o files are linked.</dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]623
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]624 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]625 <dd>The semantics of this linkage follow the ELF object file model: the symbol
626 is weak until linked, if not linked, the symbol becomes null instead of
627 being an undefined reference.</dd>
Anton Korobeynikova0554d92007-01-12 19:20:47 +0000[diff] [blame]628
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]629 <dt><tt><b><a name="linkage_linkonce_odr">linkonce_odr</a></b></tt></dt>
630 <dt><tt><b><a name="linkage_weak_odr">weak_odr</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]631 <dd>Some languages allow differing globals to be merged, such as two functions
632 with different semantics. Other languages, such as <tt>C++</tt>, ensure
Bill Wendling03bcd6e2010-07-01 21:55:59 +0000[diff] [blame]633 that only equivalent globals are ever merged (the "one definition rule"
634 &mdash; "ODR"). Such languages can use the <tt>linkonce_odr</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]635 and <tt>weak_odr</tt> linkage types to indicate that the global will only
636 be merged with equivalent globals. These linkage types are otherwise the
637 same as their non-<tt>odr</tt> versions.</dd>
Duncan Sands12da8ce2009-03-07 15:45:40 +0000[diff] [blame]638
Bill Wendlingef3cdea2011-11-04 20:40:41 +0000[diff] [blame]639 <dt><tt><b><a name="linkage_external">external</a></b></tt></dt>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]640 <dd>If none of the above identifiers are used, the global is externally
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]641 visible, meaning that it participates in linkage and can be used to
642 resolve external symbol references.</dd>
Reid Spencer7972c472007-04-11 23:49:50 +0000[diff] [blame]643</dl>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]644
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]645<p>The next two types of linkage are targeted for Microsoft Windows platform
646 only. They are designed to support importing (exporting) symbols from (to)
647 DLLs (Dynamic Link Libraries).</p>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]648
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]649<dl>
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]650 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt></dt>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]651 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]652 or variable via a global pointer to a pointer that is set up by the DLL
653 exporting the symbol. On Microsoft Windows targets, the pointer name is
654 formed by combining <code>__imp_</code> and the function or variable
655 name.</dd>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]656
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]657 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt></dt>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]658 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]659 pointer to a pointer in a DLL, so that it can be referenced with the
660 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
661 name is formed by combining <code>__imp_</code> and the function or
662 variable name.</dd>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]663</dl>
664
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]665<p>For example, since the "<tt>.LC0</tt>" variable is defined to be internal, if
666 another module defined a "<tt>.LC0</tt>" variable and was linked with this
667 one, one of the two would be renamed, preventing a collision. Since
668 "<tt>main</tt>" and "<tt>puts</tt>" are external (i.e., lacking any linkage
669 declarations), they are accessible outside of the current module.</p>
670
671<p>It is illegal for a function <i>declaration</i> to have any linkage type
Bill Wendlingb4d076e2011-10-11 06:41:28 +0000[diff] [blame]672 other than <tt>external</tt>, <tt>dllimport</tt>
673 or <tt>extern_weak</tt>.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]674
Duncan Sands12da8ce2009-03-07 15:45:40 +0000[diff] [blame]675<p>Aliases can have only <tt>external</tt>, <tt>internal</tt>, <tt>weak</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]676 or <tt>weak_odr</tt> linkages.</p>
677
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]678</div>
679
680<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]681<h3>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]682 <a name="callingconv">Calling Conventions</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]683</h3>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]684
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]685<div>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]686
687<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]688 and <a href="#i_invoke">invokes</a> can all have an optional calling
689 convention specified for the call. The calling convention of any pair of
690 dynamic caller/callee must match, or the behavior of the program is
691 undefined. The following calling conventions are supported by LLVM, and more
692 may be added in the future:</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]693
694<dl>
695 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]696 <dd>This calling convention (the default if no other calling convention is
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]697 specified) matches the target C calling conventions. This calling
698 convention supports varargs function calls and tolerates some mismatch in
699 the declared prototype and implemented declaration of the function (as
700 does normal C).</dd>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]701
702 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]703 <dd>This calling convention attempts to make calls as fast as possible
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]704 (e.g. by passing things in registers). This calling convention allows the
705 target to use whatever tricks it wants to produce fast code for the
706 target, without having to conform to an externally specified ABI
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]707 (Application Binary Interface).
708 <a href="CodeGenerator.html#tailcallopt">Tail calls can only be optimized
Chris Lattnera179e4d2010-03-11 00:22:57 +0000[diff] [blame]709 when this or the GHC convention is used.</a> This calling convention
710 does not support varargs and requires the prototype of all callees to
711 exactly match the prototype of the function definition.</dd>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]712
713 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]714 <dd>This calling convention attempts to make code in the caller as efficient
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]715 as possible under the assumption that the call is not commonly executed.
716 As such, these calls often preserve all registers so that the call does
717 not break any live ranges in the caller side. This calling convention
718 does not support varargs and requires the prototype of all callees to
719 exactly match the prototype of the function definition.</dd>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]720
Chris Lattnera179e4d2010-03-11 00:22:57 +0000[diff] [blame]721 <dt><b>"<tt>cc <em>10</em></tt>" - GHC convention</b>:</dt>
722 <dd>This calling convention has been implemented specifically for use by the
723 <a href="http://www.haskell.org/ghc">Glasgow Haskell Compiler (GHC)</a>.
724 It passes everything in registers, going to extremes to achieve this by
725 disabling callee save registers. This calling convention should not be
726 used lightly but only for specific situations such as an alternative to
727 the <em>register pinning</em> performance technique often used when
728 implementing functional programming languages.At the moment only X86
729 supports this convention and it has the following limitations:
730 <ul>
731 <li>On <em>X86-32</em> only supports up to 4 bit type parameters. No
732 floating point types are supported.</li>
733 <li>On <em>X86-64</em> only supports up to 10 bit type parameters and
734 6 floating point parameters.</li>
735 </ul>
736 This calling convention supports
737 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> but
738 requires both the caller and callee are using it.
739 </dd>
740
Chris Lattner573f64e2005-05-07 01:46:40 +0000[diff] [blame]741 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]742 <dd>Any calling convention may be specified by number, allowing
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]743 target-specific calling conventions to be used. Target specific calling
744 conventions start at 64.</dd>
Chris Lattner573f64e2005-05-07 01:46:40 +0000[diff] [blame]745</dl>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]746
747<p>More calling conventions can be added/defined on an as-needed basis, to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]748 support Pascal conventions or any other well-known target-independent
749 convention.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]750
751</div>
752
753<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]754<h3>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]755 <a name="visibility">Visibility Styles</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]756</h3>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]757
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]758<div>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]759
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]760<p>All Global Variables and Functions have one of the following visibility
761 styles:</p>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]762
763<dl>
764 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]765 <dd>On targets that use the ELF object file format, default visibility means
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]766 that the declaration is visible to other modules and, in shared libraries,
767 means that the declared entity may be overridden. On Darwin, default
768 visibility means that the declaration is visible to other modules. Default
769 visibility corresponds to "external linkage" in the language.</dd>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]770
771 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]772 <dd>Two declarations of an object with hidden visibility refer to the same
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]773 object if they are in the same shared object. Usually, hidden visibility
774 indicates that the symbol will not be placed into the dynamic symbol
775 table, so no other module (executable or shared library) can reference it
776 directly.</dd>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]777
Anton Korobeynikov39f3cff2007-04-29 18:35:00 +0000[diff] [blame]778 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
Anton Korobeynikov39f3cff2007-04-29 18:35:00 +0000[diff] [blame]779 <dd>On ELF, protected visibility indicates that the symbol will be placed in
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]780 the dynamic symbol table, but that references within the defining module
781 will bind to the local symbol. That is, the symbol cannot be overridden by
782 another module.</dd>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]783</dl>
784
785</div>
786
787<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]788<h3>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]789 <a name="namedtypes">Named Types</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]790</h3>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]791
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]792<div>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]793
794<p>LLVM IR allows you to specify name aliases for certain types. This can make
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]795 it easier to read the IR and make the IR more condensed (particularly when
796 recursive types are involved). An example of a name specification is:</p>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]797
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]798<pre class="doc_code">
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]799%mytype = type { %mytype*, i32 }
800</pre>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]801
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]802<p>You may give a name to any <a href="#typesystem">type</a> except
Chris Lattner249b9762010-08-17 23:26:04 +0000[diff] [blame]803 "<a href="#t_void">void</a>". Type name aliases may be used anywhere a type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]804 is expected with the syntax "%mytype".</p>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]805
806<p>Note that type names are aliases for the structural type that they indicate,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]807 and that you can therefore specify multiple names for the same type. This
808 often leads to confusing behavior when dumping out a .ll file. Since LLVM IR
809 uses structural typing, the name is not part of the type. When printing out
810 LLVM IR, the printer will pick <em>one name</em> to render all types of a
811 particular shape. This means that if you have code where two different
812 source types end up having the same LLVM type, that the dumper will sometimes
813 print the "wrong" or unexpected type. This is an important design point and
814 isn't going to change.</p>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]815
816</div>
817
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]818<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]819<h3>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]820 <a name="globalvars">Global Variables</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]821</h3>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]822
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]823<div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]824
Chris Lattner5d5aede2005-02-12 19:30:21 +0000[diff] [blame]825<p>Global variables define regions of memory allocated at compilation time
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]826 instead of run-time. Global variables may optionally be initialized, may
827 have an explicit section to be placed in, and may have an optional explicit
828 alignment specified. A variable may be defined as "thread_local", which
829 means that it will not be shared by threads (each thread will have a
830 separated copy of the variable). A variable may be defined as a global
831 "constant," which indicates that the contents of the variable
832 will <b>never</b> be modified (enabling better optimization, allowing the
833 global data to be placed in the read-only section of an executable, etc).
834 Note that variables that need runtime initialization cannot be marked
835 "constant" as there is a store to the variable.</p>
Chris Lattner5d5aede2005-02-12 19:30:21 +0000[diff] [blame]836
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]837<p>LLVM explicitly allows <em>declarations</em> of global variables to be marked
838 constant, even if the final definition of the global is not. This capability
839 can be used to enable slightly better optimization of the program, but
840 requires the language definition to guarantee that optimizations based on the
841 'constantness' are valid for the translation units that do not include the
842 definition.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]843
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]844<p>As SSA values, global variables define pointer values that are in scope
845 (i.e. they dominate) all basic blocks in the program. Global variables
846 always define a pointer to their "content" type because they describe a
847 region of memory, and all memory objects in LLVM are accessed through
848 pointers.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]849
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame]850<p>Global variables can be marked with <tt>unnamed_addr</tt> which indicates
851 that the address is not significant, only the content. Constants marked
Rafael Espindolaf1ed7812011-01-15 08:20:57 +0000[diff] [blame]852 like this can be merged with other constants if they have the same
853 initializer. Note that a constant with significant address <em>can</em>
854 be merged with a <tt>unnamed_addr</tt> constant, the result being a
855 constant whose address is significant.</p>
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame]856
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]857<p>A global variable may be declared to reside in a target-specific numbered
858 address space. For targets that support them, address spaces may affect how
859 optimizations are performed and/or what target instructions are used to
860 access the variable. The default address space is zero. The address space
861 qualifier must precede any other attributes.</p>
Christopher Lamb308121c2007-12-11 09:31:00 +0000[diff] [blame]862
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]863<p>LLVM allows an explicit section to be specified for globals. If the target
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]864 supports it, it will emit globals to the section specified.</p>
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]865
Chris Lattner78e00bc2010-04-28 00:13:42 +0000[diff] [blame]866<p>An explicit alignment may be specified for a global, which must be a power
867 of 2. If not present, or if the alignment is set to zero, the alignment of
868 the global is set by the target to whatever it feels convenient. If an
869 explicit alignment is specified, the global is forced to have exactly that
Chris Lattner4bd85e42010-04-28 00:31:12 +0000[diff] [blame]870 alignment. Targets and optimizers are not allowed to over-align the global
871 if the global has an assigned section. In this case, the extra alignment
872 could be observable: for example, code could assume that the globals are
873 densely packed in their section and try to iterate over them as an array,
874 alignment padding would break this iteration.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]875
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]876<p>For example, the following defines a global in a numbered address space with
877 an initializer, section, and alignment:</p>
Chris Lattner5760c502007-01-14 00:27:09 +0000[diff] [blame]878
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]879<pre class="doc_code">
Dan Gohmanaaa679b2009-01-11 00:40:00 +0000[diff] [blame]880@G = addrspace(5) constant float 1.0, section "foo", align 4
Chris Lattner5760c502007-01-14 00:27:09 +0000[diff] [blame]881</pre>
882
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]883</div>
884
885
886<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]887<h3>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]888 <a name="functionstructure">Functions</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]889</h3>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]890
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]891<div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]892
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]893<p>LLVM function definitions consist of the "<tt>define</tt>" keyword, an
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]894 optional <a href="#linkage">linkage type</a>, an optional
895 <a href="#visibility">visibility style</a>, an optional
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame]896 <a href="#callingconv">calling convention</a>,
897 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]898 <a href="#paramattrs">parameter attribute</a> for the return type, a function
899 name, a (possibly empty) argument list (each with optional
900 <a href="#paramattrs">parameter attributes</a>), optional
901 <a href="#fnattrs">function attributes</a>, an optional section, an optional
902 alignment, an optional <a href="#gc">garbage collector name</a>, an opening
903 curly brace, a list of basic blocks, and a closing curly brace.</p>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]904
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]905<p>LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
906 optional <a href="#linkage">linkage type</a>, an optional
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]907 <a href="#visibility">visibility style</a>, an optional
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame]908 <a href="#callingconv">calling convention</a>,
909 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]910 <a href="#paramattrs">parameter attribute</a> for the return type, a function
911 name, a possibly empty list of arguments, an optional alignment, and an
912 optional <a href="#gc">garbage collector name</a>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]913
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]914<p>A function definition contains a list of basic blocks, forming the CFG
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]915 (Control Flow Graph) for the function. Each basic block may optionally start
916 with a label (giving the basic block a symbol table entry), contains a list
917 of instructions, and ends with a <a href="#terminators">terminator</a>
918 instruction (such as a branch or function return).</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]919
Chris Lattnera59fb102007-06-08 16:52:14 +0000[diff] [blame]920<p>The first basic block in a function is special in two ways: it is immediately
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]921 executed on entrance to the function, and it is not allowed to have
922 predecessor basic blocks (i.e. there can not be any branches to the entry
923 block of a function). Because the block can have no predecessors, it also
924 cannot have any <a href="#i_phi">PHI nodes</a>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]925
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]926<p>LLVM allows an explicit section to be specified for functions. If the target
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]927 supports it, it will emit functions to the section specified.</p>
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]928
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]929<p>An explicit alignment may be specified for a function. If not present, or if
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]930 the alignment is set to zero, the alignment of the function is set by the
931 target to whatever it feels convenient. If an explicit alignment is
932 specified, the function is forced to have at least that much alignment. All
933 alignments must be a power of 2.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]934
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame]935<p>If the <tt>unnamed_addr</tt> attribute is given, the address is know to not
Bill Wendlingef3cdea2011-11-04 20:40:41 +0000[diff] [blame]936 be significant and two identical functions can be merged.</p>
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame]937
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]938<h5>Syntax:</h5>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]939<pre class="doc_code">
Chris Lattner0ae02092008-10-13 16:55:18 +0000[diff] [blame]940define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]941 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
942 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
943 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
944 [<a href="#gc">gc</a>] { ... }
945</pre>
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]946
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]947</div>
948
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]949<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]950<h3>
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]951 <a name="aliasstructure">Aliases</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]952</h3>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]953
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]954<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]955
956<p>Aliases act as "second name" for the aliasee value (which can be either
957 function, global variable, another alias or bitcast of global value). Aliases
958 may have an optional <a href="#linkage">linkage type</a>, and an
959 optional <a href="#visibility">visibility style</a>.</p>
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]960
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]961<h5>Syntax:</h5>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]962<pre class="doc_code">
Duncan Sands7e99a942008-09-12 20:48:21 +0000[diff] [blame]963@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Bill Wendling2d8b9a82007-05-29 09:42:13 +0000[diff] [blame]964</pre>
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]965
966</div>
967
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]968<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]969<h3>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]970 <a name="namedmetadatastructure">Named Metadata</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]971</h3>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]972
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]973<div>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]974
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]975<p>Named metadata is a collection of metadata. <a href="#metadata">Metadata
Dan Gohman093cb792010-07-21 18:54:18 +0000[diff] [blame]976 nodes</a> (but not metadata strings) are the only valid operands for
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]977 a named metadata.</p>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]978
979<h5>Syntax:</h5>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]980<pre class="doc_code">
Dan Gohman093cb792010-07-21 18:54:18 +0000[diff] [blame]981; Some unnamed metadata nodes, which are referenced by the named metadata.
982!0 = metadata !{metadata !"zero"}
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]983!1 = metadata !{metadata !"one"}
Dan Gohman093cb792010-07-21 18:54:18 +0000[diff] [blame]984!2 = metadata !{metadata !"two"}
Dan Gohman58cd65f2010-07-13 19:48:13 +0000[diff] [blame]985; A named metadata.
Dan Gohman093cb792010-07-21 18:54:18 +0000[diff] [blame]986!name = !{!0, !1, !2}
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]987</pre>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]988
989</div>
990
991<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]992<h3>
993 <a name="paramattrs">Parameter Attributes</a>
994</h3>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]995
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]996<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]997
998<p>The return type and each parameter of a function type may have a set of
999 <i>parameter attributes</i> associated with them. Parameter attributes are
1000 used to communicate additional information about the result or parameters of
1001 a function. Parameter attributes are considered to be part of the function,
1002 not of the function type, so functions with different parameter attributes
1003 can have the same function type.</p>
1004
1005<p>Parameter attributes are simple keywords that follow the type specified. If
1006 multiple parameter attributes are needed, they are space separated. For
1007 example:</p>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]1008
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1009<pre class="doc_code">
Nick Lewyckydac78d82009-02-15 23:06:14 +0000[diff] [blame]1010declare i32 @printf(i8* noalias nocapture, ...)
Chris Lattnerd2597d72008-10-04 18:33:34 +0000[diff] [blame]1011declare i32 @atoi(i8 zeroext)
1012declare signext i8 @returns_signed_char()
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]1013</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]1014
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1015<p>Note that any attributes for the function result (<tt>nounwind</tt>,
1016 <tt>readonly</tt>) come immediately after the argument list.</p>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1017
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1018<p>Currently, only the following parameter attributes are defined:</p>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000[diff] [blame]1019
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1020<dl>
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1021 <dt><tt><b>zeroext</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1022 <dd>This indicates to the code generator that the parameter or return value
Cameron Zwarichac106272011-03-16 22:20:18 +0000[diff] [blame]1023 should be zero-extended to the extent required by the target's ABI (which
1024 is usually 32-bits, but is 8-bits for a i1 on x86-64) by the caller (for a
1025 parameter) or the callee (for a return value).</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000[diff] [blame]1026
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1027 <dt><tt><b>signext</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1028 <dd>This indicates to the code generator that the parameter or return value
Cameron Zwarich341c36d2011-03-17 14:21:58 +0000[diff] [blame]1029 should be sign-extended to the extent required by the target's ABI (which
1030 is usually 32-bits) by the caller (for a parameter) or the callee (for a
1031 return value).</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000[diff] [blame]1032
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1033 <dt><tt><b>inreg</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1034 <dd>This indicates that this parameter or return value should be treated in a
1035 special target-dependent fashion during while emitting code for a function
1036 call or return (usually, by putting it in a register as opposed to memory,
1037 though some targets use it to distinguish between two different kinds of
1038 registers). Use of this attribute is target-specific.</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000[diff] [blame]1039
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1040 <dt><tt><b><a name="byval">byval</a></b></tt></dt>
Chris Lattnerd78dbee2010-11-20 23:49:06 +0000[diff] [blame]1041 <dd><p>This indicates that the pointer parameter should really be passed by
1042 value to the function. The attribute implies that a hidden copy of the
1043 pointee
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1044 is made between the caller and the callee, so the callee is unable to
1045 modify the value in the callee. This attribute is only valid on LLVM
1046 pointer arguments. It is generally used to pass structs and arrays by
1047 value, but is also valid on pointers to scalars. The copy is considered
1048 to belong to the caller not the callee (for example,
1049 <tt><a href="#readonly">readonly</a></tt> functions should not write to
1050 <tt>byval</tt> parameters). This is not a valid attribute for return
Chris Lattnerd78dbee2010-11-20 23:49:06 +0000[diff] [blame]1051 values.</p>
1052
1053 <p>The byval attribute also supports specifying an alignment with
1054 the align attribute. It indicates the alignment of the stack slot to
1055 form and the known alignment of the pointer specified to the call site. If
1056 the alignment is not specified, then the code generator makes a
1057 target-specific assumption.</p></dd>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1058
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]1059 <dt><tt><b><a name="sret">sret</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1060 <dd>This indicates that the pointer parameter specifies the address of a
1061 structure that is the return value of the function in the source program.
1062 This pointer must be guaranteed by the caller to be valid: loads and
1063 stores to the structure may be assumed by the callee to not to trap. This
1064 may only be applied to the first parameter. This is not a valid attribute
1065 for return values. </dd>
1066
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]1067 <dt><tt><b><a name="noalias">noalias</a></b></tt></dt>
Dan Gohmandf12d082010-07-02 18:41:32 +0000[diff] [blame]1068 <dd>This indicates that pointer values
1069 <a href="#pointeraliasing"><i>based</i></a> on the argument or return
Dan Gohmande256292010-07-02 23:46:54 +0000[diff] [blame]1070 value do not alias pointer values which are not <i>based</i> on it,
1071 ignoring certain "irrelevant" dependencies.
1072 For a call to the parent function, dependencies between memory
1073 references from before or after the call and from those during the call
1074 are "irrelevant" to the <tt>noalias</tt> keyword for the arguments and
1075 return value used in that call.
Dan Gohmandf12d082010-07-02 18:41:32 +0000[diff] [blame]1076 The caller shares the responsibility with the callee for ensuring that
1077 these requirements are met.
1078 For further details, please see the discussion of the NoAlias response in
Dan Gohman6c858db2010-07-06 15:26:33 +0000[diff] [blame]1079 <a href="AliasAnalysis.html#MustMayNo">alias analysis</a>.<br>
1080<br>
John McCall72ed8902010-07-06 21:07:14 +0000[diff] [blame]1081 Note that this definition of <tt>noalias</tt> is intentionally
1082 similar to the definition of <tt>restrict</tt> in C99 for function
Chris Lattner5eff9ca2010-07-06 20:51:35 +0000[diff] [blame]1083 arguments, though it is slightly weaker.
Dan Gohman6c858db2010-07-06 15:26:33 +0000[diff] [blame]1084<br>
1085 For function return values, C99's <tt>restrict</tt> is not meaningful,
1086 while LLVM's <tt>noalias</tt> is.
1087 </dd>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1088
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]1089 <dt><tt><b><a name="nocapture">nocapture</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1090 <dd>This indicates that the callee does not make any copies of the pointer
1091 that outlive the callee itself. This is not a valid attribute for return
1092 values.</dd>
1093
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]1094 <dt><tt><b><a name="nest">nest</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1095 <dd>This indicates that the pointer parameter can be excised using the
1096 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
1097 attribute for return values.</dd>
1098</dl>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1099
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1100</div>
1101
1102<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1103<h3>
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1104 <a name="gc">Garbage Collector Names</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1105</h3>
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1106
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1107<div>
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1108
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1109<p>Each function may specify a garbage collector name, which is simply a
1110 string:</p>
1111
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1112<pre class="doc_code">
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1113define void @f() gc "name" { ... }
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1114</pre>
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1115
1116<p>The compiler declares the supported values of <i>name</i>. Specifying a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1117 collector which will cause the compiler to alter its output in order to
1118 support the named garbage collection algorithm.</p>
1119
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1120</div>
1121
1122<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1123<h3>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1124 <a name="fnattrs">Function Attributes</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1125</h3>
Devang Patelcaacdba2008-09-04 23:05:13 +0000[diff] [blame]1126
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1127<div>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1128
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1129<p>Function attributes are set to communicate additional information about a
1130 function. Function attributes are considered to be part of the function, not
1131 of the function type, so functions with different parameter attributes can
1132 have the same function type.</p>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1133
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1134<p>Function attributes are simple keywords that follow the type specified. If
1135 multiple attributes are needed, they are space separated. For example:</p>
Devang Patelcaacdba2008-09-04 23:05:13 +0000[diff] [blame]1136
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1137<pre class="doc_code">
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1138define void @f() noinline { ... }
1139define void @f() alwaysinline { ... }
1140define void @f() alwaysinline optsize { ... }
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1141define void @f() optsize { ... }
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1142</pre>
Devang Patelcaacdba2008-09-04 23:05:13 +0000[diff] [blame]1143
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1144<dl>
Kostya Serebryanya5054ad2012-01-20 17:56:17 +0000[diff] [blame]1145 <dt><tt><b>address_safety</b></tt></dt>
1146 <dd>This attribute indicates that the address safety analysis
1147 is enabled for this function. </dd>
1148
Charles Davisbe5557e2010-02-12 00:31:15 +0000[diff] [blame]1149 <dt><tt><b>alignstack(&lt;<em>n</em>&gt;)</b></tt></dt>
1150 <dd>This attribute indicates that, when emitting the prologue and epilogue,
1151 the backend should forcibly align the stack pointer. Specify the
1152 desired alignment, which must be a power of two, in parentheses.
1153
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1154 <dt><tt><b>alwaysinline</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1155 <dd>This attribute indicates that the inliner should attempt to inline this
1156 function into callers whenever possible, ignoring any active inlining size
1157 threshold for this caller.</dd>
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1158
Dan Gohman8bd11f12011-06-16 16:03:13 +0000[diff] [blame]1159 <dt><tt><b>nonlazybind</b></tt></dt>
1160 <dd>This attribute suppresses lazy symbol binding for the function. This
1161 may make calls to the function faster, at the cost of extra program
1162 startup time if the function is not called during program startup.</dd>
1163
Jakob Stoklund Olesen74bb06c2010-02-06 01:16:28 +0000[diff] [blame]1164 <dt><tt><b>inlinehint</b></tt></dt>
1165 <dd>This attribute indicates that the source code contained a hint that inlining
1166 this function is desirable (such as the "inline" keyword in C/C++). It
1167 is just a hint; it imposes no requirements on the inliner.</dd>
1168
Nick Lewycky14b58da2010-07-06 18:24:09 +0000[diff] [blame]1169 <dt><tt><b>naked</b></tt></dt>
1170 <dd>This attribute disables prologue / epilogue emission for the function.
1171 This can have very system-specific consequences.</dd>
1172
1173 <dt><tt><b>noimplicitfloat</b></tt></dt>
1174 <dd>This attributes disables implicit floating point instructions.</dd>
1175
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1176 <dt><tt><b>noinline</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1177 <dd>This attribute indicates that the inliner should never inline this
1178 function in any situation. This attribute may not be used together with
1179 the <tt>alwaysinline</tt> attribute.</dd>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1180
Nick Lewycky14b58da2010-07-06 18:24:09 +0000[diff] [blame]1181 <dt><tt><b>noredzone</b></tt></dt>
1182 <dd>This attribute indicates that the code generator should not use a red
1183 zone, even if the target-specific ABI normally permits it.</dd>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1184
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1185 <dt><tt><b>noreturn</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1186 <dd>This function attribute indicates that the function never returns
1187 normally. This produces undefined behavior at runtime if the function
1188 ever does dynamically return.</dd>
Bill Wendlinga8130172008-11-13 01:02:51 +0000[diff] [blame]1189
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1190 <dt><tt><b>nounwind</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1191 <dd>This function attribute indicates that the function never returns with an
1192 unwind or exceptional control flow. If the function does unwind, its
1193 runtime behavior is undefined.</dd>
Bill Wendling0f5541e2008-11-26 19:07:40 +0000[diff] [blame]1194
Nick Lewycky14b58da2010-07-06 18:24:09 +0000[diff] [blame]1195 <dt><tt><b>optsize</b></tt></dt>
1196 <dd>This attribute suggests that optimization passes and code generator passes
1197 make choices that keep the code size of this function low, and otherwise
1198 do optimizations specifically to reduce code size.</dd>
1199
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1200 <dt><tt><b>readnone</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1201 <dd>This attribute indicates that the function computes its result (or decides
1202 to unwind an exception) based strictly on its arguments, without
1203 dereferencing any pointer arguments or otherwise accessing any mutable
1204 state (e.g. memory, control registers, etc) visible to caller functions.
1205 It does not write through any pointer arguments
1206 (including <tt><a href="#byval">byval</a></tt> arguments) and never
1207 changes any state visible to callers. This means that it cannot unwind
Bill Wendling3f6a3a22012-02-06 21:57:33 +0000[diff] [blame^]1208 exceptions by calling the <tt>C++</tt> exception throwing methods.</dd>
Devang Patel310fd4a2009-06-12 19:45:19 +0000[diff] [blame]1209
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1210 <dt><tt><b><a name="readonly">readonly</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1211 <dd>This attribute indicates that the function does not write through any
1212 pointer arguments (including <tt><a href="#byval">byval</a></tt>
1213 arguments) or otherwise modify any state (e.g. memory, control registers,
1214 etc) visible to caller functions. It may dereference pointer arguments
1215 and read state that may be set in the caller. A readonly function always
1216 returns the same value (or unwinds an exception identically) when called
1217 with the same set of arguments and global state. It cannot unwind an
Bill Wendling3f6a3a22012-02-06 21:57:33 +0000[diff] [blame^]1218 exception by calling the <tt>C++</tt> exception throwing methods.</dd>
Anton Korobeynikovc8ce7b082009-07-17 18:07:26 +0000[diff] [blame]1219
Bill Wendlingb437ab82011-12-05 21:27:54 +0000[diff] [blame]1220 <dt><tt><b><a name="returns_twice">returns_twice</a></b></tt></dt>
1221 <dd>This attribute indicates that this function can return twice. The
1222 C <code>setjmp</code> is an example of such a function. The compiler
1223 disables some optimizations (like tail calls) in the caller of these
1224 functions.</dd>
1225
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1226 <dt><tt><b><a name="ssp">ssp</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1227 <dd>This attribute indicates that the function should emit a stack smashing
1228 protector. It is in the form of a "canary"&mdash;a random value placed on
1229 the stack before the local variables that's checked upon return from the
1230 function to see if it has been overwritten. A heuristic is used to
1231 determine if a function needs stack protectors or not.<br>
1232<br>
1233 If a function that has an <tt>ssp</tt> attribute is inlined into a
1234 function that doesn't have an <tt>ssp</tt> attribute, then the resulting
1235 function will have an <tt>ssp</tt> attribute.</dd>
1236
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1237 <dt><tt><b>sspreq</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1238 <dd>This attribute indicates that the function should <em>always</em> emit a
1239 stack smashing protector. This overrides
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1240 the <tt><a href="#ssp">ssp</a></tt> function attribute.<br>
1241<br>
1242 If a function that has an <tt>sspreq</tt> attribute is inlined into a
1243 function that doesn't have an <tt>sspreq</tt> attribute or which has
1244 an <tt>ssp</tt> attribute, then the resulting function will have
1245 an <tt>sspreq</tt> attribute.</dd>
Rafael Espindola163d6752011-07-25 15:27:59 +0000[diff] [blame]1246
1247 <dt><tt><b><a name="uwtable">uwtable</a></b></tt></dt>
1248 <dd>This attribute indicates that the ABI being targeted requires that
1249 an unwind table entry be produce for this function even if we can
1250 show that no exceptions passes by it. This is normally the case for
1251 the ELF x86-64 abi, but it can be disabled for some compilation
1252 units.</dd>
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1253</dl>
1254
Devang Patelcaacdba2008-09-04 23:05:13 +0000[diff] [blame]1255</div>
1256
1257<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1258<h3>
Chris Lattner93564892006-04-08 04:40:53 +0000[diff] [blame]1259 <a name="moduleasm">Module-Level Inline Assembly</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1260</h3>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1261
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1262<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1263
1264<p>Modules may contain "module-level inline asm" blocks, which corresponds to
1265 the GCC "file scope inline asm" blocks. These blocks are internally
1266 concatenated by LLVM and treated as a single unit, but may be separated in
1267 the <tt>.ll</tt> file if desired. The syntax is very simple:</p>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1268
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1269<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]1270module asm "inline asm code goes here"
1271module asm "more can go here"
1272</pre>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1273
1274<p>The strings can contain any character by escaping non-printable characters.
1275 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1276 for the number.</p>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1277
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1278<p>The inline asm code is simply printed to the machine code .s file when
1279 assembly code is generated.</p>
1280
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1281</div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1282
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1283<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1284<h3>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1285 <a name="datalayout">Data Layout</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1286</h3>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1287
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1288<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1289
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1290<p>A module may specify a target specific data layout string that specifies how
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1291 data is to be laid out in memory. The syntax for the data layout is
1292 simply:</p>
1293
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1294<pre class="doc_code">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1295target datalayout = "<i>layout specification</i>"
1296</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1297
1298<p>The <i>layout specification</i> consists of a list of specifications
1299 separated by the minus sign character ('-'). Each specification starts with
1300 a letter and may include other information after the letter to define some
1301 aspect of the data layout. The specifications accepted are as follows:</p>
1302
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1303<dl>
1304 <dt><tt>E</tt></dt>
1305 <dd>Specifies that the target lays out data in big-endian form. That is, the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1306 bits with the most significance have the lowest address location.</dd>
1307
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1308 <dt><tt>e</tt></dt>
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]1309 <dd>Specifies that the target lays out data in little-endian form. That is,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1310 the bits with the least significance have the lowest address
1311 location.</dd>
1312
Lang Hamesde7ab802011-10-10 23:42:08 +0000[diff] [blame]1313 <dt><tt>S<i>size</i></tt></dt>
1314 <dd>Specifies the natural alignment of the stack in bits. Alignment promotion
1315 of stack variables is limited to the natural stack alignment to avoid
1316 dynamic stack realignment. The stack alignment must be a multiple of
Lang Hamesff2c52c2011-10-11 17:50:14 +0000[diff] [blame]1317 8-bits. If omitted, the natural stack alignment defaults to "unspecified",
1318 which does not prevent any alignment promotions.</dd>
Lang Hamesde7ab802011-10-10 23:42:08 +0000[diff] [blame]1319
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1320 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1321 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1322 <i>preferred</i> alignments. All sizes are in bits. Specifying
1323 the <i>pref</i> alignment is optional. If omitted, the
1324 preceding <tt>:</tt> should be omitted too.</dd>
1325
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1326 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1327 <dd>This specifies the alignment for an integer type of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1328 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1329
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1330 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1331 <dd>This specifies the alignment for a vector type of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1332 <i>size</i>.</dd>
1333
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1334 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1335 <dd>This specifies the alignment for a floating point type of a given bit
Dale Johannesence522852010-05-28 18:54:47 +0000[diff] [blame]1336 <i>size</i>. Only values of <i>size</i> that are supported by the target
1337 will work. 32 (float) and 64 (double) are supported on all targets;
1338 80 or 128 (different flavors of long double) are also supported on some
1339 targets.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1340
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1341 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1342 <dd>This specifies the alignment for an aggregate type of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1343 <i>size</i>.</dd>
1344
Daniel Dunbar7921a592009-06-08 22:17:53 +0000[diff] [blame]1345 <dt><tt>s<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1346 <dd>This specifies the alignment for a stack object of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1347 <i>size</i>.</dd>
Chris Lattnera381eff2009-11-07 09:35:34 +0000[diff] [blame]1348
1349 <dt><tt>n<i>size1</i>:<i>size2</i>:<i>size3</i>...</tt></dt>
1350 <dd>This specifies a set of native integer widths for the target CPU
1351 in bits. For example, it might contain "n32" for 32-bit PowerPC,
1352 "n32:64" for PowerPC 64, or "n8:16:32:64" for X86-64. Elements of
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1353 this set are considered to support most general arithmetic
Chris Lattnera381eff2009-11-07 09:35:34 +0000[diff] [blame]1354 operations efficiently.</dd>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1355</dl>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1356
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1357<p>When constructing the data layout for a given target, LLVM starts with a
Dan Gohman61110ae2010-04-28 00:36:01 +0000[diff] [blame]1358 default set of specifications which are then (possibly) overridden by the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1359 specifications in the <tt>datalayout</tt> keyword. The default specifications
1360 are given in this list:</p>
1361
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1362<ul>
1363 <li><tt>E</tt> - big endian</li>
Dan Gohman8ad777d2010-02-23 02:44:03 +0000[diff] [blame]1364 <li><tt>p:64:64:64</tt> - 64-bit pointers with 64-bit alignment</li>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1365 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1366 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1367 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1368 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]1369 <li><tt>i64:32:64</tt> - i64 has ABI alignment of 32-bits but preferred
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1370 alignment of 64-bits</li>
1371 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1372 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1373 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1374 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1375 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
Daniel Dunbar7921a592009-06-08 22:17:53 +0000[diff] [blame]1376 <li><tt>s0:64:64</tt> - stack objects are 64-bit aligned</li>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1377</ul>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1378
1379<p>When LLVM is determining the alignment for a given type, it uses the
1380 following rules:</p>
1381
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1382<ol>
1383 <li>If the type sought is an exact match for one of the specifications, that
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1384 specification is used.</li>
1385
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1386 <li>If no match is found, and the type sought is an integer type, then the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1387 smallest integer type that is larger than the bitwidth of the sought type
1388 is used. If none of the specifications are larger than the bitwidth then
1389 the the largest integer type is used. For example, given the default
1390 specifications above, the i7 type will use the alignment of i8 (next
1391 largest) while both i65 and i256 will use the alignment of i64 (largest
1392 specified).</li>
1393
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1394 <li>If no match is found, and the type sought is a vector type, then the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1395 largest vector type that is smaller than the sought vector type will be
1396 used as a fall back. This happens because &lt;128 x double&gt; can be
1397 implemented in terms of 64 &lt;2 x double&gt;, for example.</li>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1398</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1399
Chris Lattner48797402011-10-11 23:01:39 +0000[diff] [blame]1400<p>The function of the data layout string may not be what you expect. Notably,
1401 this is not a specification from the frontend of what alignment the code
1402 generator should use.</p>
1403
1404<p>Instead, if specified, the target data layout is required to match what the
1405 ultimate <em>code generator</em> expects. This string is used by the
1406 mid-level optimizers to
1407 improve code, and this only works if it matches what the ultimate code
1408 generator uses. If you would like to generate IR that does not embed this
1409 target-specific detail into the IR, then you don't have to specify the
1410 string. This will disable some optimizations that require precise layout
1411 information, but this also prevents those optimizations from introducing
1412 target specificity into the IR.</p>
1413
1414
1415
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1416</div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1417
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1418<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1419<h3>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1420 <a name="pointeraliasing">Pointer Aliasing Rules</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1421</h3>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1422
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1423<div>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1424
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1425<p>Any memory access must be done through a pointer value associated
Andreas Bolkae39f0332009-07-27 20:37:10 +0000[diff] [blame]1426with an address range of the memory access, otherwise the behavior
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1427is undefined. Pointer values are associated with address ranges
1428according to the following rules:</p>
1429
1430<ul>
Dan Gohmandf12d082010-07-02 18:41:32 +0000[diff] [blame]1431 <li>A pointer value is associated with the addresses associated with
1432 any value it is <i>based</i> on.
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1433 <li>An address of a global variable is associated with the address
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1434 range of the variable's storage.</li>
1435 <li>The result value of an allocation instruction is associated with
1436 the address range of the allocated storage.</li>
1437 <li>A null pointer in the default address-space is associated with
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1438 no address.</li>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1439 <li>An integer constant other than zero or a pointer value returned
1440 from a function not defined within LLVM may be associated with address
1441 ranges allocated through mechanisms other than those provided by
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1442 LLVM. Such ranges shall not overlap with any ranges of addresses
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1443 allocated by mechanisms provided by LLVM.</li>
Dan Gohmandf12d082010-07-02 18:41:32 +0000[diff] [blame]1444</ul>
1445
1446<p>A pointer value is <i>based</i> on another pointer value according
1447 to the following rules:</p>
1448
1449<ul>
1450 <li>A pointer value formed from a
1451 <tt><a href="#i_getelementptr">getelementptr</a></tt> operation
1452 is <i>based</i> on the first operand of the <tt>getelementptr</tt>.</li>
1453 <li>The result value of a
1454 <tt><a href="#i_bitcast">bitcast</a></tt> is <i>based</i> on the operand
1455 of the <tt>bitcast</tt>.</li>
1456 <li>A pointer value formed by an
1457 <tt><a href="#i_inttoptr">inttoptr</a></tt> is <i>based</i> on all
1458 pointer values that contribute (directly or indirectly) to the
1459 computation of the pointer's value.</li>
1460 <li>The "<i>based</i> on" relationship is transitive.</li>
1461</ul>
1462
1463<p>Note that this definition of <i>"based"</i> is intentionally
1464 similar to the definition of <i>"based"</i> in C99, though it is
1465 slightly weaker.</p>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1466
1467<p>LLVM IR does not associate types with memory. The result type of a
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1468<tt><a href="#i_load">load</a></tt> merely indicates the size and
1469alignment of the memory from which to load, as well as the
Dan Gohman4eb47192010-06-17 19:23:50 +0000[diff] [blame]1470interpretation of the value. The first operand type of a
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1471<tt><a href="#i_store">store</a></tt> similarly only indicates the size
1472and alignment of the store.</p>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1473
1474<p>Consequently, type-based alias analysis, aka TBAA, aka
1475<tt>-fstrict-aliasing</tt>, is not applicable to general unadorned
1476LLVM IR. <a href="#metadata">Metadata</a> may be used to encode
1477additional information which specialized optimization passes may use
1478to implement type-based alias analysis.</p>
1479
1480</div>
1481
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]1482<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1483<h3>
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]1484 <a name="volatile">Volatile Memory Accesses</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1485</h3>
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]1486
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1487<div>
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]1488
1489<p>Certain memory accesses, such as <a href="#i_load"><tt>load</tt></a>s, <a
1490href="#i_store"><tt>store</tt></a>s, and <a
1491href="#int_memcpy"><tt>llvm.memcpy</tt></a>s may be marked <tt>volatile</tt>.
1492The optimizers must not change the number of volatile operations or change their
1493order of execution relative to other volatile operations. The optimizers
1494<i>may</i> change the order of volatile operations relative to non-volatile
1495operations. This is not Java's "volatile" and has no cross-thread
1496synchronization behavior.</p>
1497
1498</div>
1499
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1500<!-- ======================================================================= -->
1501<h3>
1502 <a name="memmodel">Memory Model for Concurrent Operations</a>
1503</h3>
1504
1505<div>
1506
1507<p>The LLVM IR does not define any way to start parallel threads of execution
1508or to register signal handlers. Nonetheless, there are platform-specific
1509ways to create them, and we define LLVM IR's behavior in their presence. This
1510model is inspired by the C++0x memory model.</p>
1511
Eli Friedman95f69a42011-08-22 21:35:27 +0000[diff] [blame]1512<p>For a more informal introduction to this model, see the
1513<a href="Atomics.html">LLVM Atomic Instructions and Concurrency Guide</a>.
1514
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1515<p>We define a <i>happens-before</i> partial order as the least partial order
1516that</p>
1517<ul>
1518 <li>Is a superset of single-thread program order, and</li>
1519 <li>When a <i>synchronizes-with</i> <tt>b</tt>, includes an edge from
1520 <tt>a</tt> to <tt>b</tt>. <i>Synchronizes-with</i> pairs are introduced
1521 by platform-specific techniques, like pthread locks, thread
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1522 creation, thread joining, etc., and by atomic instructions.
1523 (See also <a href="#ordering">Atomic Memory Ordering Constraints</a>).
1524 </li>
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1525</ul>
1526
1527<p>Note that program order does not introduce <i>happens-before</i> edges
1528between a thread and signals executing inside that thread.</p>
1529
1530<p>Every (defined) read operation (load instructions, memcpy, atomic
1531loads/read-modify-writes, etc.) <var>R</var> reads a series of bytes written by
1532(defined) write operations (store instructions, atomic
Eli Friedmanf12e4e92011-07-22 03:04:45 +0000[diff] [blame]1533stores/read-modify-writes, memcpy, etc.). For the purposes of this section,
1534initialized globals are considered to have a write of the initializer which is
1535atomic and happens before any other read or write of the memory in question.
1536For each byte of a read <var>R</var>, <var>R<sub>byte</sub></var> may see
1537any write to the same byte, except:</p>
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1538
1539<ul>
1540 <li>If <var>write<sub>1</sub></var> happens before
1541 <var>write<sub>2</sub></var>, and <var>write<sub>2</sub></var> happens
1542 before <var>R<sub>byte</sub></var>, then <var>R<sub>byte</sub></var>
Eli Friedmanf12e4e92011-07-22 03:04:45 +0000[diff] [blame]1543 does not see <var>write<sub>1</sub></var>.
Bill Wendling537603b2011-07-31 06:45:03 +0000[diff] [blame]1544 <li>If <var>R<sub>byte</sub></var> happens before
1545 <var>write<sub>3</sub></var>, then <var>R<sub>byte</sub></var> does not
1546 see <var>write<sub>3</sub></var>.
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1547</ul>
1548
1549<p>Given that definition, <var>R<sub>byte</sub></var> is defined as follows:
1550<ul>
Eli Friedman95f69a42011-08-22 21:35:27 +0000[diff] [blame]1551 <li>If <var>R</var> is volatile, the result is target-dependent. (Volatile
1552 is supposed to give guarantees which can support
1553 <code>sig_atomic_t</code> in C/C++, and may be used for accesses to
1554 addresses which do not behave like normal memory. It does not generally
1555 provide cross-thread synchronization.)
1556 <li>Otherwise, if there is no write to the same byte that happens before
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1557 <var>R<sub>byte</sub></var>, <var>R<sub>byte</sub></var> returns
1558 <tt>undef</tt> for that byte.
Eli Friedmanf12e4e92011-07-22 03:04:45 +0000[diff] [blame]1559 <li>Otherwise, if <var>R<sub>byte</sub></var> may see exactly one write,
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1560 <var>R<sub>byte</sub></var> returns the value written by that
1561 write.</li>
Eli Friedmanf12e4e92011-07-22 03:04:45 +0000[diff] [blame]1562 <li>Otherwise, if <var>R</var> is atomic, and all the writes
1563 <var>R<sub>byte</sub></var> may see are atomic, it chooses one of the
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1564 values written. See the <a href="#ordering">Atomic Memory Ordering
1565 Constraints</a> section for additional constraints on how the choice
1566 is made.
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1567 <li>Otherwise <var>R<sub>byte</sub></var> returns <tt>undef</tt>.</li>
1568</ul>
1569
1570<p><var>R</var> returns the value composed of the series of bytes it read.
1571This implies that some bytes within the value may be <tt>undef</tt>
1572<b>without</b> the entire value being <tt>undef</tt>. Note that this only
1573defines the semantics of the operation; it doesn't mean that targets will
1574emit more than one instruction to read the series of bytes.</p>
1575
1576<p>Note that in cases where none of the atomic intrinsics are used, this model
1577places only one restriction on IR transformations on top of what is required
1578for single-threaded execution: introducing a store to a byte which might not
Eli Friedman4bc9f3c2011-08-02 01:15:34 +0000[diff] [blame]1579otherwise be stored is not allowed in general. (Specifically, in the case
1580where another thread might write to and read from an address, introducing a
1581store can change a load that may see exactly one write into a load that may
1582see multiple writes.)</p>
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1583
1584<!-- FIXME: This model assumes all targets where concurrency is relevant have
1585a byte-size store which doesn't affect adjacent bytes. As far as I can tell,
1586none of the backends currently in the tree fall into this category; however,
1587there might be targets which care. If there are, we want a paragraph
1588like the following:
1589
1590Targets may specify that stores narrower than a certain width are not
1591available; on such a target, for the purposes of this model, treat any
1592non-atomic write with an alignment or width less than the minimum width
1593as if it writes to the relevant surrounding bytes.
1594-->
1595
1596</div>
1597
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1598<!-- ======================================================================= -->
NAKAMURA Takumi0300d882011-08-12 06:17:17 +0000[diff] [blame]1599<h3>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1600 <a name="ordering">Atomic Memory Ordering Constraints</a>
NAKAMURA Takumi0300d882011-08-12 06:17:17 +0000[diff] [blame]1601</h3>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1602
NAKAMURA Takumi0300d882011-08-12 06:17:17 +0000[diff] [blame]1603<div>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1604
1605<p>Atomic instructions (<a href="#i_cmpxchg"><code>cmpxchg</code></a>,
Eli Friedman59b66882011-08-09 23:02:53 +0000[diff] [blame]1606<a href="#i_atomicrmw"><code>atomicrmw</code></a>,
1607<a href="#i_fence"><code>fence</code></a>,
1608<a href="#i_load"><code>atomic load</code></a>, and
Eli Friedman75362532011-08-09 23:26:12 +0000[diff] [blame]1609<a href="#i_store"><code>atomic store</code></a>) take an ordering parameter
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1610that determines which other atomic instructions on the same address they
1611<i>synchronize with</i>. These semantics are borrowed from Java and C++0x,
1612but are somewhat more colloquial. If these descriptions aren't precise enough,
Eli Friedman95f69a42011-08-22 21:35:27 +0000[diff] [blame]1613check those specs (see spec references in the
Nick Lewycky75499f52012-01-23 08:47:21 +0000[diff] [blame]1614<a href="Atomics.html#introduction">atomics guide</a>).
Eli Friedman95f69a42011-08-22 21:35:27 +0000[diff] [blame]1615<a href="#i_fence"><code>fence</code></a> instructions
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1616treat these orderings somewhat differently since they don't take an address.
1617See that instruction's documentation for details.</p>
1618
Eli Friedman95f69a42011-08-22 21:35:27 +0000[diff] [blame]1619<p>For a simpler introduction to the ordering constraints, see the
1620<a href="Atomics.html">LLVM Atomic Instructions and Concurrency Guide</a>.</p>
1621
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1622<dl>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1623<dt><code>unordered</code></dt>
1624<dd>The set of values that can be read is governed by the happens-before
1625partial order. A value cannot be read unless some operation wrote it.
1626This is intended to provide a guarantee strong enough to model Java's
1627non-volatile shared variables. This ordering cannot be specified for
1628read-modify-write operations; it is not strong enough to make them atomic
1629in any interesting way.</dd>
1630<dt><code>monotonic</code></dt>
1631<dd>In addition to the guarantees of <code>unordered</code>, there is a single
1632total order for modifications by <code>monotonic</code> operations on each
1633address. All modification orders must be compatible with the happens-before
1634order. There is no guarantee that the modification orders can be combined to
1635a global total order for the whole program (and this often will not be
1636possible). The read in an atomic read-modify-write operation
1637(<a href="#i_cmpxchg"><code>cmpxchg</code></a> and
1638<a href="#i_atomicrmw"><code>atomicrmw</code></a>)
1639reads the value in the modification order immediately before the value it
1640writes. If one atomic read happens before another atomic read of the same
1641address, the later read must see the same value or a later value in the
1642address's modification order. This disallows reordering of
1643<code>monotonic</code> (or stronger) operations on the same address. If an
1644address is written <code>monotonic</code>ally by one thread, and other threads
1645<code>monotonic</code>ally read that address repeatedly, the other threads must
Eli Friedman95f69a42011-08-22 21:35:27 +0000[diff] [blame]1646eventually see the write. This corresponds to the C++0x/C1x
1647<code>memory_order_relaxed</code>.</dd>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1648<dt><code>acquire</code></dt>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1649<dd>In addition to the guarantees of <code>monotonic</code>,
Eli Friedman0cb3b562011-08-24 20:28:39 +0000[diff] [blame]1650a <i>synchronizes-with</i> edge may be formed with a <code>release</code>
1651operation. This is intended to model C++'s <code>memory_order_acquire</code>.</dd>
1652<dt><code>release</code></dt>
1653<dd>In addition to the guarantees of <code>monotonic</code>, if this operation
1654writes a value which is subsequently read by an <code>acquire</code> operation,
1655it <i>synchronizes-with</i> that operation. (This isn't a complete
1656description; see the C++0x definition of a release sequence.) This corresponds
1657to the C++0x/C1x <code>memory_order_release</code>.</dd>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1658<dt><code>acq_rel</code> (acquire+release)</dt><dd>Acts as both an
Eli Friedman95f69a42011-08-22 21:35:27 +0000[diff] [blame]1659<code>acquire</code> and <code>release</code> operation on its address.
1660This corresponds to the C++0x/C1x <code>memory_order_acq_rel</code>.</dd>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1661<dt><code>seq_cst</code> (sequentially consistent)</dt><dd>
1662<dd>In addition to the guarantees of <code>acq_rel</code>
1663(<code>acquire</code> for an operation which only reads, <code>release</code>
1664for an operation which only writes), there is a global total order on all
1665sequentially-consistent operations on all addresses, which is consistent with
1666the <i>happens-before</i> partial order and with the modification orders of
1667all the affected addresses. Each sequentially-consistent read sees the last
Eli Friedman95f69a42011-08-22 21:35:27 +0000[diff] [blame]1668preceding write to the same address in this global order. This corresponds
1669to the C++0x/C1x <code>memory_order_seq_cst</code> and Java volatile.</dd>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1670</dl>
1671
1672<p id="singlethread">If an atomic operation is marked <code>singlethread</code>,
1673it only <i>synchronizes with</i> or participates in modification and seq_cst
1674total orderings with other operations running in the same thread (for example,
1675in signal handlers).</p>
1676
1677</div>
1678
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1679</div>
1680
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1681<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1682<h2><a name="typesystem">Type System</a></h2>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1683<!-- *********************************************************************** -->
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1684
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1685<div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1686
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1687<p>The LLVM type system is one of the most important features of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1688 intermediate representation. Being typed enables a number of optimizations
1689 to be performed on the intermediate representation directly, without having
1690 to do extra analyses on the side before the transformation. A strong type
1691 system makes it easier to read the generated code and enables novel analyses
1692 and transformations that are not feasible to perform on normal three address
1693 code representations.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1694
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1695<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1696<h3>
1697 <a name="t_classifications">Type Classifications</a>
1698</h3>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1699
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1700<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1701
1702<p>The types fall into a few useful classifications:</p>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]1703
1704<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1705 <tbody>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1706 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1707 <tr>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1708 <td><a href="#t_integer">integer</a></td>
Reid Spencer138249b2007-05-16 18:44:01 +0000[diff] [blame]1709 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1710 </tr>
1711 <tr>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1712 <td><a href="#t_floating">floating point</a></td>
Dan Gohman518cda42011-12-17 00:04:22 +0000[diff] [blame]1713 <td><tt>half, float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1714 </tr>
1715 <tr>
1716 <td><a name="t_firstclass">first class</a></td>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1717 <td><a href="#t_integer">integer</a>,
1718 <a href="#t_floating">floating point</a>,
1719 <a href="#t_pointer">pointer</a>,
Dan Gohman08783a882008-06-18 18:42:13 +0000[diff] [blame]1720 <a href="#t_vector">vector</a>,
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]1721 <a href="#t_struct">structure</a>,
1722 <a href="#t_array">array</a>,
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1723 <a href="#t_label">label</a>,
1724 <a href="#t_metadata">metadata</a>.
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1725 </td>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1726 </tr>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1727 <tr>
1728 <td><a href="#t_primitive">primitive</a></td>
1729 <td><a href="#t_label">label</a>,
1730 <a href="#t_void">void</a>,
Tobias Grosser4c8c95b2010-12-28 20:29:31 +0000[diff] [blame]1731 <a href="#t_integer">integer</a>,
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1732 <a href="#t_floating">floating point</a>,
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]1733 <a href="#t_x86mmx">x86mmx</a>,
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1734 <a href="#t_metadata">metadata</a>.</td>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1735 </tr>
1736 <tr>
1737 <td><a href="#t_derived">derived</a></td>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1738 <td><a href="#t_array">array</a>,
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1739 <a href="#t_function">function</a>,
1740 <a href="#t_pointer">pointer</a>,
1741 <a href="#t_struct">structure</a>,
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1742 <a href="#t_vector">vector</a>,
1743 <a href="#t_opaque">opaque</a>.
Dan Gohman93bf60d2008-10-14 16:32:04 +0000[diff] [blame]1744 </td>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1745 </tr>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1746 </tbody>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1747</table>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]1748
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1749<p>The <a href="#t_firstclass">first class</a> types are perhaps the most
1750 important. Values of these types are the only ones which can be produced by
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1751 instructions.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1752
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1753</div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1754
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1755<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1756<h3>
1757 <a name="t_primitive">Primitive Types</a>
1758</h3>
Chris Lattner43542b32008-01-04 04:34:14 +0000[diff] [blame]1759
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1760<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1761
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1762<p>The primitive types are the fundamental building blocks of the LLVM
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1763 system.</p>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1764
1765<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1766<h4>
1767 <a name="t_integer">Integer Type</a>
1768</h4>
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1769
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1770<div>
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1771
1772<h5>Overview:</h5>
1773<p>The integer type is a very simple type that simply specifies an arbitrary
1774 bit width for the integer type desired. Any bit width from 1 bit to
1775 2<sup>23</sup>-1 (about 8 million) can be specified.</p>
1776
1777<h5>Syntax:</h5>
1778<pre>
1779 iN
1780</pre>
1781
1782<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1783 value.</p>
1784
1785<h5>Examples:</h5>
1786<table class="layout">
1787 <tr class="layout">
1788 <td class="left"><tt>i1</tt></td>
1789 <td class="left">a single-bit integer.</td>
1790 </tr>
1791 <tr class="layout">
1792 <td class="left"><tt>i32</tt></td>
1793 <td class="left">a 32-bit integer.</td>
1794 </tr>
1795 <tr class="layout">
1796 <td class="left"><tt>i1942652</tt></td>
1797 <td class="left">a really big integer of over 1 million bits.</td>
1798 </tr>
1799</table>
1800
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1801</div>
1802
1803<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1804<h4>
1805 <a name="t_floating">Floating Point Types</a>
1806</h4>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1807
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1808<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1809
1810<table>
1811 <tbody>
1812 <tr><th>Type</th><th>Description</th></tr>
Dan Gohman518cda42011-12-17 00:04:22 +0000[diff] [blame]1813 <tr><td><tt>half</tt></td><td>16-bit floating point value</td></tr>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1814 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1815 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1816 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1817 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1818 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1819 </tbody>
1820</table>
1821
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1822</div>
1823
1824<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1825<h4>
1826 <a name="t_x86mmx">X86mmx Type</a>
1827</h4>
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]1828
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1829<div>
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]1830
1831<h5>Overview:</h5>
1832<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>
1833
1834<h5>Syntax:</h5>
1835<pre>
Dale Johannesenb1f0ff12010-10-01 01:07:02 +0000[diff] [blame]1836 x86mmx
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]1837</pre>
1838
1839</div>
1840
1841<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1842<h4>
1843 <a name="t_void">Void Type</a>
1844</h4>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1845
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1846<div>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1847
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1848<h5>Overview:</h5>
1849<p>The void type does not represent any value and has no size.</p>
1850
1851<h5>Syntax:</h5>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1852<pre>
1853 void
1854</pre>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1855
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1856</div>
1857
1858<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1859<h4>
1860 <a name="t_label">Label Type</a>
1861</h4>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1862
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1863<div>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1864
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1865<h5>Overview:</h5>
1866<p>The label type represents code labels.</p>
1867
1868<h5>Syntax:</h5>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1869<pre>
1870 label
1871</pre>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1872
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1873</div>
1874
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1875<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1876<h4>
1877 <a name="t_metadata">Metadata Type</a>
1878</h4>
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1879
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1880<div>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1881
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1882<h5>Overview:</h5>
Nick Lewycky93e06a52009-09-27 23:27:42 +0000[diff] [blame]1883<p>The metadata type represents embedded metadata. No derived types may be
1884 created from metadata except for <a href="#t_function">function</a>
1885 arguments.
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1886
1887<h5>Syntax:</h5>
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1888<pre>
1889 metadata
1890</pre>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1891
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1892</div>
1893
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1894</div>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1895
1896<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1897<h3>
1898 <a name="t_derived">Derived Types</a>
1899</h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1900
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1901<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1902
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1903<p>The real power in LLVM comes from the derived types in the system. This is
1904 what allows a programmer to represent arrays, functions, pointers, and other
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1905 useful types. Each of these types contain one or more element types which
1906 may be a primitive type, or another derived type. For example, it is
1907 possible to have a two dimensional array, using an array as the element type
1908 of another array.</p>
Dan Gohman142ccc02009-01-24 15:58:40 +0000[diff] [blame]1909
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1910<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1911<h4>
1912 <a name="t_aggregate">Aggregate Types</a>
1913</h4>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1914
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1915<div>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1916
1917<p>Aggregate Types are a subset of derived types that can contain multiple
Duncan Sands9aaec152011-12-14 15:44:20 +0000[diff] [blame]1918 member types. <a href="#t_array">Arrays</a> and
1919 <a href="#t_struct">structs</a> are aggregate types.
1920 <a href="#t_vector">Vectors</a> are not considered to be aggregate types.</p>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1921
1922</div>
1923
Reid Spencer138249b2007-05-16 18:44:01 +0000[diff] [blame]1924<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1925<h4>
1926 <a name="t_array">Array Type</a>
1927</h4>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1928
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1929<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1930
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1931<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1932<p>The array type is a very simple derived type that arranges elements
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1933 sequentially in memory. The array type requires a size (number of elements)
1934 and an underlying data type.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1935
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]1936<h5>Syntax:</h5>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1937<pre>
1938 [&lt;# elements&gt; x &lt;elementtype&gt;]
1939</pre>
1940
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1941<p>The number of elements is a constant integer value; <tt>elementtype</tt> may
1942 be any type with a size.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1943
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]1944<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1945<table class="layout">
1946 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1947 <td class="left"><tt>[40 x i32]</tt></td>
1948 <td class="left">Array of 40 32-bit integer values.</td>
1949 </tr>
1950 <tr class="layout">
1951 <td class="left"><tt>[41 x i32]</tt></td>
1952 <td class="left">Array of 41 32-bit integer values.</td>
1953 </tr>
1954 <tr class="layout">
1955 <td class="left"><tt>[4 x i8]</tt></td>
1956 <td class="left">Array of 4 8-bit integer values.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1957 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1958</table>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1959<p>Here are some examples of multidimensional arrays:</p>
1960<table class="layout">
1961 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1962 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1963 <td class="left">3x4 array of 32-bit integer values.</td>
1964 </tr>
1965 <tr class="layout">
1966 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1967 <td class="left">12x10 array of single precision floating point values.</td>
1968 </tr>
1969 <tr class="layout">
1970 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1971 <td class="left">2x3x4 array of 16-bit integer values.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1972 </tr>
1973</table>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +0000[diff] [blame]1974
Dan Gohmanc74bc282009-11-09 19:01:53 +0000[diff] [blame]1975<p>There is no restriction on indexing beyond the end of the array implied by
1976 a static type (though there are restrictions on indexing beyond the bounds
1977 of an allocated object in some cases). This means that single-dimension
1978 'variable sized array' addressing can be implemented in LLVM with a zero
1979 length array type. An implementation of 'pascal style arrays' in LLVM could
1980 use the type "<tt>{ i32, [0 x float]}</tt>", for example.</p>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +0000[diff] [blame]1981
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1982</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1983
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1984<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1985<h4>
1986 <a name="t_function">Function Type</a>
1987</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1988
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1989<div>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]1990
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1991<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1992<p>The function type can be thought of as a function signature. It consists of
1993 a return type and a list of formal parameter types. The return type of a
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]1994 function type is a first class type or a void type.</p>
Devang Pateld6cff512008-03-10 20:49:15 +0000[diff] [blame]1995
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1996<h5>Syntax:</h5>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]1997<pre>
Nick Lewycky14d1ccc2009-09-27 07:55:32 +0000[diff] [blame]1998 &lt;returntype&gt; (&lt;parameter list&gt;)
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]1999</pre>
2000
John Criswell4c0cf7f2005-10-24 16:17:18 +0000[diff] [blame]2001<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2002 specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
2003 which indicates that the function takes a variable number of arguments.
2004 Variable argument functions can access their arguments with
2005 the <a href="#int_varargs">variable argument handling intrinsic</a>
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]2006 functions. '<tt>&lt;returntype&gt;</tt>' is any type except
Nick Lewycky93e06a52009-09-27 23:27:42 +0000[diff] [blame]2007 <a href="#t_label">label</a>.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]2008
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2009<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2010<table class="layout">
2011 <tr class="layout">
Reid Spencer58c08712006-12-31 07:18:34 +0000[diff] [blame]2012 <td class="left"><tt>i32 (i32)</tt></td>
2013 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2014 </td>
Reid Spencer58c08712006-12-31 07:18:34 +0000[diff] [blame]2015 </tr><tr class="layout">
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]2016 <td class="left"><tt>float&nbsp;(i16,&nbsp;i32&nbsp;*)&nbsp;*
Reid Spencer655dcc62006-12-31 07:20:23 +0000[diff] [blame]2017 </tt></td>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2018 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]2019 an <tt>i16</tt> and a <a href="#t_pointer">pointer</a> to <tt>i32</tt>,
2020 returning <tt>float</tt>.
Reid Spencer58c08712006-12-31 07:18:34 +0000[diff] [blame]2021 </td>
2022 </tr><tr class="layout">
2023 <td class="left"><tt>i32 (i8*, ...)</tt></td>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2024 <td class="left">A vararg function that takes at least one
2025 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
2026 which returns an integer. This is the signature for <tt>printf</tt> in
Reid Spencer58c08712006-12-31 07:18:34 +0000[diff] [blame]2027 LLVM.
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2028 </td>
Devang Patele3dfc1c2008-03-24 05:35:41 +0000[diff] [blame]2029 </tr><tr class="layout">
2030 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Nick Lewycky14d1ccc2009-09-27 07:55:32 +0000[diff] [blame]2031 <td class="left">A function taking an <tt>i32</tt>, returning a
2032 <a href="#t_struct">structure</a> containing two <tt>i32</tt> values
Devang Patele3dfc1c2008-03-24 05:35:41 +0000[diff] [blame]2033 </td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2034 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2035</table>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]2036
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2037</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2038
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2039<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2040<h4>
2041 <a name="t_struct">Structure Type</a>
2042</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2043
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2044<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2045
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2046<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2047<p>The structure type is used to represent a collection of data members together
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2048 in memory. The elements of a structure may be any type that has a size.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2049
Jeffrey Yasskinf991bbb2010-01-11 19:19:26 +0000[diff] [blame]2050<p>Structures in memory are accessed using '<tt><a href="#i_load">load</a></tt>'
2051 and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field
2052 with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
2053 Structures in registers are accessed using the
2054 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' and
2055 '<tt><a href="#i_insertvalue">insertvalue</a></tt>' instructions.</p>
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2056
2057<p>Structures may optionally be "packed" structures, which indicate that the
2058 alignment of the struct is one byte, and that there is no padding between
Chris Lattner190552d2011-08-12 17:31:02 +0000[diff] [blame]2059 the elements. In non-packed structs, padding between field types is inserted
2060 as defined by the TargetData string in the module, which is required to match
Chris Lattner7bd0ea32011-10-11 23:02:17 +0000[diff] [blame]2061 what the underlying code generator expects.</p>
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2062
Chris Lattner190552d2011-08-12 17:31:02 +0000[diff] [blame]2063<p>Structures can either be "literal" or "identified". A literal structure is
2064 defined inline with other types (e.g. <tt>{i32, i32}*</tt>) whereas identified
2065 types are always defined at the top level with a name. Literal types are
2066 uniqued by their contents and can never be recursive or opaque since there is
Chris Lattner32531732011-08-12 18:12:40 +0000[diff] [blame]2067 no way to write one. Identified types can be recursive, can be opaqued, and are
Chris Lattner190552d2011-08-12 17:31:02 +0000[diff] [blame]2068 never uniqued.
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2069</p>
2070
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2071<h5>Syntax:</h5>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]2072<pre>
Chris Lattner190552d2011-08-12 17:31:02 +0000[diff] [blame]2073 %T1 = type { &lt;type list&gt; } <i>; Identified normal struct type</i>
2074 %T2 = type &lt;{ &lt;type list&gt; }&gt; <i>; Identified packed struct type</i>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]2075</pre>
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2076
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2077<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2078<table class="layout">
2079 <tr class="layout">
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]2080 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
2081 <td class="left">A triple of three <tt>i32</tt> values</td>
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2082 </tr>
2083 <tr class="layout">
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]2084 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
2085 <td class="left">A pair, where the first element is a <tt>float</tt> and the
2086 second element is a <a href="#t_pointer">pointer</a> to a
2087 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
2088 an <tt>i32</tt>.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2089 </tr>
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2090 <tr class="layout">
2091 <td class="left"><tt>&lt;{ i8, i32 }&gt;</tt></td>
2092 <td class="left">A packed struct known to be 5 bytes in size.</td>
2093 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2094</table>
Dan Gohman142ccc02009-01-24 15:58:40 +0000[diff] [blame]2095
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2096</div>
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2097
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2098<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2099<h4>
Chris Lattner2a843822011-07-23 19:59:08 +0000[diff] [blame]2100 <a name="t_opaque">Opaque Structure Types</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2101</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2102
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2103<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2104
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]2105<h5>Overview:</h5>
Chris Lattner2a843822011-07-23 19:59:08 +0000[diff] [blame]2106<p>Opaque structure types are used to represent named structure types that do
2107 not have a body specified. This corresponds (for example) to the C notion of
2108 a forward declared structure.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2109
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]2110<h5>Syntax:</h5>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]2111<pre>
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2112 %X = type opaque
2113 %52 = type opaque
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]2114</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2115
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]2116<h5>Examples:</h5>
2117<table class="layout">
2118 <tr class="layout">
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2119 <td class="left"><tt>opaque</tt></td>
2120 <td class="left">An opaque type.</td>
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]2121 </tr>
2122</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2123
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]2124</div>
2125
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2126
2127
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]2128<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2129<h4>
2130 <a name="t_pointer">Pointer Type</a>
2131</h4>
Chris Lattner4a67c912009-02-08 19:53:29 +0000[diff] [blame]2132
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2133<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2134
2135<h5>Overview:</h5>
Dan Gohman88481112010-02-25 16:50:07 +0000[diff] [blame]2136<p>The pointer type is used to specify memory locations.
2137 Pointers are commonly used to reference objects in memory.</p>
2138
2139<p>Pointer types may have an optional address space attribute defining the
2140 numbered address space where the pointed-to object resides. The default
2141 address space is number zero. The semantics of non-zero address
2142 spaces are target-specific.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2143
2144<p>Note that LLVM does not permit pointers to void (<tt>void*</tt>) nor does it
2145 permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</tt> instead.</p>
Chris Lattner4a67c912009-02-08 19:53:29 +0000[diff] [blame]2146
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]2147<h5>Syntax:</h5>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]2148<pre>
2149 &lt;type&gt; *
2150</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2151
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]2152<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2153<table class="layout">
2154 <tr class="layout">
Dan Gohman623806e2009-01-04 23:44:43 +0000[diff] [blame]2155 <td class="left"><tt>[4 x i32]*</tt></td>
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]2156 <td class="left">A <a href="#t_pointer">pointer</a> to <a
2157 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
2158 </tr>
2159 <tr class="layout">
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]2160 <td class="left"><tt>i32 (i32*) *</tt></td>
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]2161 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]2162 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]2163 <tt>i32</tt>.</td>
2164 </tr>
2165 <tr class="layout">
2166 <td class="left"><tt>i32 addrspace(5)*</tt></td>
2167 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
2168 that resides in address space #5.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2169 </tr>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2170</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2171
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2172</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2173
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]2174<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2175<h4>
2176 <a name="t_vector">Vector Type</a>
2177</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2178
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2179<div>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]2180
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]2181<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2182<p>A vector type is a simple derived type that represents a vector of elements.
2183 Vector types are used when multiple primitive data are operated in parallel
2184 using a single instruction (SIMD). A vector type requires a size (number of
Duncan Sands31c0e0e2009-11-27 13:38:03 +0000[diff] [blame]2185 elements) and an underlying primitive data type. Vector types are considered
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2186 <a href="#t_firstclass">first class</a>.</p>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]2187
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]2188<h5>Syntax:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]2189<pre>
2190 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
2191</pre>
2192
Chris Lattnerf11031a2010-10-10 18:20:35 +0000[diff] [blame]2193<p>The number of elements is a constant integer value larger than 0; elementtype
Nadav Rotem3924cb02011-12-05 06:29:09 +0000[diff] [blame]2194 may be any integer or floating point type, or a pointer to these types.
2195 Vectors of size zero are not allowed. </p>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]2196
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]2197<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2198<table class="layout">
2199 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]2200 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
2201 <td class="left">Vector of 4 32-bit integer values.</td>
2202 </tr>
2203 <tr class="layout">
2204 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
2205 <td class="left">Vector of 8 32-bit floating-point values.</td>
2206 </tr>
2207 <tr class="layout">
2208 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
2209 <td class="left">Vector of 2 64-bit integer values.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2210 </tr>
Nadav Rotem3924cb02011-12-05 06:29:09 +0000[diff] [blame]2211 <tr class="layout">
2212 <td class="left"><tt>&lt;4 x i64*&gt;</tt></td>
2213 <td class="left">Vector of 4 pointers to 64-bit integer values.</td>
2214 </tr>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2215</table>
Dan Gohman142ccc02009-01-24 15:58:40 +0000[diff] [blame]2216
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2217</div>
2218
Bill Wendlingae8b5ea2011-07-31 06:47:33 +0000[diff] [blame]2219</div>
2220
NAKAMURA Takumia35cdd62011-10-31 13:04:26 +0000[diff] [blame]2221</div>
2222
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2223<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2224<h2><a name="constants">Constants</a></h2>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2225<!-- *********************************************************************** -->
2226
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2227<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2228
2229<p>LLVM has several different basic types of constants. This section describes
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2230 them all and their syntax.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2231
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2232<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2233<h3>
2234 <a name="simpleconstants">Simple Constants</a>
2235</h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2236
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2237<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2238
2239<dl>
2240 <dt><b>Boolean constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2241 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]2242 constants of the <tt><a href="#t_integer">i1</a></tt> type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2243
2244 <dt><b>Integer constants</b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2245 <dd>Standard integers (such as '4') are constants of
2246 the <a href="#t_integer">integer</a> type. Negative numbers may be used
2247 with integer types.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2248
2249 <dt><b>Floating point constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2250 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2251 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
2252 notation (see below). The assembler requires the exact decimal value of a
2253 floating-point constant. For example, the assembler accepts 1.25 but
2254 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
2255 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2256
2257 <dt><b>Null pointer constants</b></dt>
John Criswelldfe6a862004-12-10 15:51:16 +0000[diff] [blame]2258 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2259 and must be of <a href="#t_pointer">pointer type</a>.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2260</dl>
2261
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2262<p>The one non-intuitive notation for constants is the hexadecimal form of
2263 floating point constants. For example, the form '<tt>double
2264 0x432ff973cafa8000</tt>' is equivalent to (but harder to read than)
2265 '<tt>double 4.5e+15</tt>'. The only time hexadecimal floating point
2266 constants are required (and the only time that they are generated by the
2267 disassembler) is when a floating point constant must be emitted but it cannot
2268 be represented as a decimal floating point number in a reasonable number of
2269 digits. For example, NaN's, infinities, and other special values are
2270 represented in their IEEE hexadecimal format so that assembly and disassembly
2271 do not cause any bits to change in the constants.</p>
2272
Dan Gohman518cda42011-12-17 00:04:22 +0000[diff] [blame]2273<p>When using the hexadecimal form, constants of types half, float, and double are
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2274 represented using the 16-digit form shown above (which matches the IEEE754
Dan Gohman518cda42011-12-17 00:04:22 +0000[diff] [blame]2275 representation for double); half and float values must, however, be exactly
2276 representable as IEE754 half and single precision, respectively.
2277 Hexadecimal format is always used
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2278 for long double, and there are three forms of long double. The 80-bit format
2279 used by x86 is represented as <tt>0xK</tt> followed by 20 hexadecimal digits.
2280 The 128-bit format used by PowerPC (two adjacent doubles) is represented
2281 by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit format
2282 is represented by <tt>0xL</tt> followed by 32 hexadecimal digits; no
2283 currently supported target uses this format. Long doubles will only work if
2284 they match the long double format on your target. All hexadecimal formats
2285 are big-endian (sign bit at the left).</p>
2286
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]2287<p>There are no constants of type x86mmx.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2288</div>
2289
2290<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2291<h3>
Bill Wendling972b7202009-07-20 02:32:41 +0000[diff] [blame]2292<a name="aggregateconstants"></a> <!-- old anchor -->
2293<a name="complexconstants">Complex Constants</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2294</h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2295
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2296<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2297
Chris Lattner361bfcd2009-02-28 18:32:25 +0000[diff] [blame]2298<p>Complex constants are a (potentially recursive) combination of simple
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2299 constants and smaller complex constants.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2300
2301<dl>
2302 <dt><b>Structure constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2303 <dd>Structure constants are represented with notation similar to structure
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2304 type definitions (a comma separated list of elements, surrounded by braces
2305 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
2306 where "<tt>@G</tt>" is declared as "<tt>@G = external global i32</tt>".
2307 Structure constants must have <a href="#t_struct">structure type</a>, and
2308 the number and types of elements must match those specified by the
2309 type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2310
2311 <dt><b>Array constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2312 <dd>Array constants are represented with notation similar to array type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2313 definitions (a comma separated list of elements, surrounded by square
2314 brackets (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74
2315 ]</tt>". Array constants must have <a href="#t_array">array type</a>, and
2316 the number and types of elements must match those specified by the
2317 type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2318
Reid Spencer404a3252007-02-15 03:07:05 +0000[diff] [blame]2319 <dt><b>Vector constants</b></dt>
Reid Spencer404a3252007-02-15 03:07:05 +0000[diff] [blame]2320 <dd>Vector constants are represented with notation similar to vector type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2321 definitions (a comma separated list of elements, surrounded by
2322 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32
2323 42, i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must
2324 have <a href="#t_vector">vector type</a>, and the number and types of
2325 elements must match those specified by the type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2326
2327 <dt><b>Zero initialization</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2328 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]2329 value to zero of <em>any</em> type, including scalar and
2330 <a href="#t_aggregate">aggregate</a> types.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2331 This is often used to avoid having to print large zero initializers
2332 (e.g. for large arrays) and is always exactly equivalent to using explicit
2333 zero initializers.</dd>
Nick Lewycky49f89192009-04-04 07:22:01 +0000[diff] [blame]2334
2335 <dt><b>Metadata node</b></dt>
Nick Lewycky8e2c4f42009-05-30 16:08:30 +0000[diff] [blame]2336 <dd>A metadata node is a structure-like constant with
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2337 <a href="#t_metadata">metadata type</a>. For example: "<tt>metadata !{
2338 i32 0, metadata !"test" }</tt>". Unlike other constants that are meant to
2339 be interpreted as part of the instruction stream, metadata is a place to
2340 attach additional information such as debug info.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2341</dl>
2342
2343</div>
2344
2345<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2346<h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2347 <a name="globalconstants">Global Variable and Function Addresses</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2348</h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2349
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2350<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2351
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2352<p>The addresses of <a href="#globalvars">global variables</a>
2353 and <a href="#functionstructure">functions</a> are always implicitly valid
2354 (link-time) constants. These constants are explicitly referenced when
2355 the <a href="#identifiers">identifier for the global</a> is used and always
2356 have <a href="#t_pointer">pointer</a> type. For example, the following is a
2357 legal LLVM file:</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2358
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2359<pre class="doc_code">
Chris Lattner00538a12007-06-06 18:28:13 +0000[diff] [blame]2360@X = global i32 17
2361@Y = global i32 42
2362@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2363</pre>
2364
2365</div>
2366
2367<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2368<h3>
2369 <a name="undefvalues">Undefined Values</a>
2370</h3>
2371
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2372<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2373
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2374<p>The string '<tt>undef</tt>' can be used anywhere a constant is expected, and
Benjamin Kramer0f420382009-10-12 14:46:08 +0000[diff] [blame]2375 indicates that the user of the value may receive an unspecified bit-pattern.
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2376 Undefined values may be of any type (other than '<tt>label</tt>'
2377 or '<tt>void</tt>') and be used anywhere a constant is permitted.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2378
Chris Lattner92ada5d2009-09-11 01:49:31 +0000[diff] [blame]2379<p>Undefined values are useful because they indicate to the compiler that the
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2380 program is well defined no matter what value is used. This gives the
2381 compiler more freedom to optimize. Here are some examples of (potentially
2382 surprising) transformations that are valid (in pseudo IR):</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2383
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2384
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2385<pre class="doc_code">
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2386 %A = add %X, undef
2387 %B = sub %X, undef
2388 %C = xor %X, undef
2389Safe:
2390 %A = undef
2391 %B = undef
2392 %C = undef
2393</pre>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2394
2395<p>This is safe because all of the output bits are affected by the undef bits.
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2396 Any output bit can have a zero or one depending on the input bits.</p>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2397
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2398<pre class="doc_code">
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2399 %A = or %X, undef
2400 %B = and %X, undef
2401Safe:
2402 %A = -1
2403 %B = 0
2404Unsafe:
2405 %A = undef
2406 %B = undef
2407</pre>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2408
2409<p>These logical operations have bits that are not always affected by the input.
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2410 For example, if <tt>%X</tt> has a zero bit, then the output of the
2411 '<tt>and</tt>' operation will always be a zero for that bit, no matter what
2412 the corresponding bit from the '<tt>undef</tt>' is. As such, it is unsafe to
2413 optimize or assume that the result of the '<tt>and</tt>' is '<tt>undef</tt>'.
2414 However, it is safe to assume that all bits of the '<tt>undef</tt>' could be
2415 0, and optimize the '<tt>and</tt>' to 0. Likewise, it is safe to assume that
2416 all the bits of the '<tt>undef</tt>' operand to the '<tt>or</tt>' could be
2417 set, allowing the '<tt>or</tt>' to be folded to -1.</p>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2418
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2419<pre class="doc_code">
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2420 %A = select undef, %X, %Y
2421 %B = select undef, 42, %Y
2422 %C = select %X, %Y, undef
2423Safe:
2424 %A = %X (or %Y)
2425 %B = 42 (or %Y)
2426 %C = %Y
2427Unsafe:
2428 %A = undef
2429 %B = undef
2430 %C = undef
2431</pre>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2432
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2433<p>This set of examples shows that undefined '<tt>select</tt>' (and conditional
2434 branch) conditions can go <em>either way</em>, but they have to come from one
2435 of the two operands. In the <tt>%A</tt> example, if <tt>%X</tt> and
2436 <tt>%Y</tt> were both known to have a clear low bit, then <tt>%A</tt> would
2437 have to have a cleared low bit. However, in the <tt>%C</tt> example, the
2438 optimizer is allowed to assume that the '<tt>undef</tt>' operand could be the
2439 same as <tt>%Y</tt>, allowing the whole '<tt>select</tt>' to be
2440 eliminated.</p>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2441
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2442<pre class="doc_code">
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2443 %A = xor undef, undef
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2444
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2445 %B = undef
2446 %C = xor %B, %B
2447
2448 %D = undef
2449 %E = icmp lt %D, 4
2450 %F = icmp gte %D, 4
2451
2452Safe:
2453 %A = undef
2454 %B = undef
2455 %C = undef
2456 %D = undef
2457 %E = undef
2458 %F = undef
2459</pre>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2460
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2461<p>This example points out that two '<tt>undef</tt>' operands are not
2462 necessarily the same. This can be surprising to people (and also matches C
2463 semantics) where they assume that "<tt>X^X</tt>" is always zero, even
2464 if <tt>X</tt> is undefined. This isn't true for a number of reasons, but the
2465 short answer is that an '<tt>undef</tt>' "variable" can arbitrarily change
2466 its value over its "live range". This is true because the variable doesn't
2467 actually <em>have a live range</em>. Instead, the value is logically read
2468 from arbitrary registers that happen to be around when needed, so the value
2469 is not necessarily consistent over time. In fact, <tt>%A</tt> and <tt>%C</tt>
2470 need to have the same semantics or the core LLVM "replace all uses with"
2471 concept would not hold.</p>
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2472
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2473<pre class="doc_code">
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2474 %A = fdiv undef, %X
2475 %B = fdiv %X, undef
2476Safe:
2477 %A = undef
2478b: unreachable
2479</pre>
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2480
2481<p>These examples show the crucial difference between an <em>undefined
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2482 value</em> and <em>undefined behavior</em>. An undefined value (like
2483 '<tt>undef</tt>') is allowed to have an arbitrary bit-pattern. This means that
2484 the <tt>%A</tt> operation can be constant folded to '<tt>undef</tt>', because
2485 the '<tt>undef</tt>' could be an SNaN, and <tt>fdiv</tt> is not (currently)
2486 defined on SNaN's. However, in the second example, we can make a more
2487 aggressive assumption: because the <tt>undef</tt> is allowed to be an
2488 arbitrary value, we are allowed to assume that it could be zero. Since a
2489 divide by zero has <em>undefined behavior</em>, we are allowed to assume that
2490 the operation does not execute at all. This allows us to delete the divide and
2491 all code after it. Because the undefined operation "can't happen", the
2492 optimizer can assume that it occurs in dead code.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2493
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2494<pre class="doc_code">
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2495a: store undef -> %X
2496b: store %X -> undef
2497Safe:
2498a: &lt;deleted&gt;
2499b: unreachable
2500</pre>
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2501
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2502<p>These examples reiterate the <tt>fdiv</tt> example: a store <em>of</em> an
2503 undefined value can be assumed to not have any effect; we can assume that the
2504 value is overwritten with bits that happen to match what was already there.
2505 However, a store <em>to</em> an undefined location could clobber arbitrary
2506 memory, therefore, it has undefined behavior.</p>
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2507
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2508</div>
2509
2510<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2511<h3>
Dan Gohman9a2a0932011-12-06 03:18:47 +0000[diff] [blame]2512 <a name="poisonvalues">Poison Values</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2513</h3>
2514
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2515<div>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2516
Dan Gohman9a2a0932011-12-06 03:18:47 +0000[diff] [blame]2517<p>Poison values are similar to <a href="#undefvalues">undef values</a>, however
Dan Gohman32772f72011-12-06 03:35:58 +0000[diff] [blame]2518 they also represent the fact that an instruction or constant expression which
2519 cannot evoke side effects has nevertheless detected a condition which results
2520 in undefined behavior.</p>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2521
Dan Gohman9a2a0932011-12-06 03:18:47 +0000[diff] [blame]2522<p>There is currently no way of representing a poison value in the IR; they
Dan Gohmanac355aa2010-05-03 14:51:43 +0000[diff] [blame]2523 only exist when produced by operations such as
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2524 <a href="#i_add"><tt>add</tt></a> with the <tt>nsw</tt> flag.</p>
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2525
Dan Gohman9a2a0932011-12-06 03:18:47 +0000[diff] [blame]2526<p>Poison value behavior is defined in terms of value <i>dependence</i>:</p>
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2527
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2528<ul>
2529<li>Values other than <a href="#i_phi"><tt>phi</tt></a> nodes depend on
2530 their operands.</li>
2531
2532<li><a href="#i_phi"><tt>Phi</tt></a> nodes depend on the operand corresponding
2533 to their dynamic predecessor basic block.</li>
2534
2535<li>Function arguments depend on the corresponding actual argument values in
2536 the dynamic callers of their functions.</li>
2537
2538<li><a href="#i_call"><tt>Call</tt></a> instructions depend on the
2539 <a href="#i_ret"><tt>ret</tt></a> instructions that dynamically transfer
2540 control back to them.</li>
2541
Dan Gohman7292a752010-05-03 14:55:22 +0000[diff] [blame]2542<li><a href="#i_invoke"><tt>Invoke</tt></a> instructions depend on the
Bill Wendling3f6a3a22012-02-06 21:57:33 +0000[diff] [blame^]2543 <a href="#i_ret"><tt>ret</tt></a>, <a href="#i_resume"><tt>resume</tt></a>,
Dan Gohman7292a752010-05-03 14:55:22 +0000[diff] [blame]2544 or exception-throwing call instructions that dynamically transfer control
2545 back to them.</li>
2546
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2547<li>Non-volatile loads and stores depend on the most recent stores to all of the
2548 referenced memory addresses, following the order in the IR
2549 (including loads and stores implied by intrinsics such as
2550 <a href="#int_memcpy"><tt>@llvm.memcpy</tt></a>.)</li>
2551
Dan Gohman3513ea52010-05-03 14:59:34 +0000[diff] [blame]2552<!-- TODO: In the case of multiple threads, this only applies if the store
2553 "happens-before" the load or store. -->
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2554
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2555<!-- TODO: floating-point exception state -->
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2556
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2557<li>An instruction with externally visible side effects depends on the most
2558 recent preceding instruction with externally visible side effects, following
Dan Gohman6c858db2010-07-06 15:26:33 +0000[diff] [blame]2559 the order in the IR. (This includes
2560 <a href="#volatile">volatile operations</a>.)</li>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2561
Dan Gohman7292a752010-05-03 14:55:22 +0000[diff] [blame]2562<li>An instruction <i>control-depends</i> on a
2563 <a href="#terminators">terminator instruction</a>
2564 if the terminator instruction has multiple successors and the instruction
2565 is always executed when control transfers to one of the successors, and
Chris Lattner0ab5e2c2011-04-15 05:18:47 +0000[diff] [blame]2566 may not be executed when control is transferred to another.</li>
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2567
Dan Gohmanc8454ee2011-04-12 23:05:59 +0000[diff] [blame]2568<li>Additionally, an instruction also <i>control-depends</i> on a terminator
2569 instruction if the set of instructions it otherwise depends on would be
Chris Lattner0ab5e2c2011-04-15 05:18:47 +0000[diff] [blame]2570 different if the terminator had transferred control to a different
Dan Gohmanc8454ee2011-04-12 23:05:59 +0000[diff] [blame]2571 successor.</li>
2572
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2573<li>Dependence is transitive.</li>
2574
2575</ul>
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2576
Dan Gohman32772f72011-12-06 03:35:58 +0000[diff] [blame]2577<p>Poison Values have the same behavior as <a href="#undefvalues">undef values</a>,
2578 with the additional affect that any instruction which has a <i>dependence</i>
2579 on a poison value has undefined behavior.</p>
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2580
2581<p>Here are some examples:</p>
Dan Gohman48a25882010-04-26 20:54:53 +0000[diff] [blame]2582
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2583<pre class="doc_code">
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2584entry:
Dan Gohman9a2a0932011-12-06 03:18:47 +0000[diff] [blame]2585 %poison = sub nuw i32 0, 1 ; Results in a poison value.
Dan Gohman32772f72011-12-06 03:35:58 +0000[diff] [blame]2586 %still_poison = and i32 %poison, 0 ; 0, but also poison.
Dan Gohman9a2a0932011-12-06 03:18:47 +0000[diff] [blame]2587 %poison_yet_again = getelementptr i32* @h, i32 %still_poison
Dan Gohman32772f72011-12-06 03:35:58 +0000[diff] [blame]2588 store i32 0, i32* %poison_yet_again ; memory at @h[0] is poisoned
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2589
Dan Gohman32772f72011-12-06 03:35:58 +0000[diff] [blame]2590 store i32 %poison, i32* @g ; Poison value stored to memory.
2591 %poison2 = load i32* @g ; Poison value loaded back from memory.
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2592
Dan Gohman9a2a0932011-12-06 03:18:47 +0000[diff] [blame]2593 store volatile i32 %poison, i32* @g ; External observation; undefined behavior.
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2594
2595 %narrowaddr = bitcast i32* @g to i16*
2596 %wideaddr = bitcast i32* @g to i64*
Dan Gohman9a2a0932011-12-06 03:18:47 +0000[diff] [blame]2597 %poison3 = load i16* %narrowaddr ; Returns a poison value.
2598 %poison4 = load i64* %wideaddr ; Returns a poison value.
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2599
Dan Gohman5f115a72011-12-06 03:31:14 +0000[diff] [blame]2600 %cmp = icmp slt i32 %poison, 0 ; Returns a poison value.
2601 br i1 %cmp, label %true, label %end ; Branch to either destination.
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2602
2603true:
Dan Gohman5f115a72011-12-06 03:31:14 +0000[diff] [blame]2604 store volatile i32 0, i32* @g ; This is control-dependent on %cmp, so
2605 ; it has undefined behavior.
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2606 br label %end
2607
2608end:
2609 %p = phi i32 [ 0, %entry ], [ 1, %true ]
Dan Gohman5f115a72011-12-06 03:31:14 +0000[diff] [blame]2610 ; Both edges into this PHI are
2611 ; control-dependent on %cmp, so this
2612 ; always results in a poison value.
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2613
Dan Gohman5f115a72011-12-06 03:31:14 +0000[diff] [blame]2614 store volatile i32 0, i32* @g ; This would depend on the store in %true
2615 ; if %cmp is true, or the store in %entry
2616 ; otherwise, so this is undefined behavior.
Dan Gohmanc8454ee2011-04-12 23:05:59 +0000[diff] [blame]2617
Nick Lewycky9c876bf2011-05-16 19:29:30 +0000[diff] [blame]2618 br i1 %cmp, label %second_true, label %second_end
Dan Gohman5f115a72011-12-06 03:31:14 +0000[diff] [blame]2619 ; The same branch again, but this time the
2620 ; true block doesn't have side effects.
Dan Gohmanc8454ee2011-04-12 23:05:59 +0000[diff] [blame]2621
2622second_true:
2623 ; No side effects!
Nick Lewycky9c876bf2011-05-16 19:29:30 +0000[diff] [blame]2624 ret void
Dan Gohmanc8454ee2011-04-12 23:05:59 +0000[diff] [blame]2625
2626second_end:
Dan Gohman5f115a72011-12-06 03:31:14 +0000[diff] [blame]2627 store volatile i32 0, i32* @g ; This time, the instruction always depends
2628 ; on the store in %end. Also, it is
2629 ; control-equivalent to %end, so this is
Dan Gohman32772f72011-12-06 03:35:58 +0000[diff] [blame]2630 ; well-defined (ignoring earlier undefined
2631 ; behavior in this example).
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2632</pre>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2633
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2634</div>
2635
2636<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2637<h3>
2638 <a name="blockaddress">Addresses of Basic Blocks</a>
2639</h3>
2640
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2641<div>
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2642
Chris Lattneraa99c942009-11-01 01:27:45 +0000[diff] [blame]2643<p><b><tt>blockaddress(@function, %block)</tt></b></p>
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2644
2645<p>The '<tt>blockaddress</tt>' constant computes the address of the specified
Chris Lattner5c5f0ac2009-10-27 21:49:40 +0000[diff] [blame]2646 basic block in the specified function, and always has an i8* type. Taking
Chris Lattneraa99c942009-11-01 01:27:45 +0000[diff] [blame]2647 the address of the entry block is illegal.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2648
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2649<p>This value only has defined behavior when used as an operand to the
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2650 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>' instruction, or for
2651 comparisons against null. Pointer equality tests between labels addresses
2652 results in undefined behavior &mdash; though, again, comparison against null
2653 is ok, and no label is equal to the null pointer. This may be passed around
2654 as an opaque pointer sized value as long as the bits are not inspected. This
2655 allows <tt>ptrtoint</tt> and arithmetic to be performed on these values so
2656 long as the original value is reconstituted before the <tt>indirectbr</tt>
2657 instruction.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2658
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2659<p>Finally, some targets may provide defined semantics when using the value as
2660 the operand to an inline assembly, but that is target specific.</p>
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2661
2662</div>
2663
2664
2665<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2666<h3>
2667 <a name="constantexprs">Constant Expressions</a>
2668</h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2669
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2670<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2671
2672<p>Constant expressions are used to allow expressions involving other constants
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2673 to be used as constants. Constant expressions may be of
2674 any <a href="#t_firstclass">first class</a> type and may involve any LLVM
2675 operation that does not have side effects (e.g. load and call are not
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2676 supported). The following is the syntax for constant expressions:</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2677
2678<dl>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2679 <dt><b><tt>trunc (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2680 <dd>Truncate a constant to another type. The bit size of CST must be larger
2681 than the bit size of TYPE. Both types must be integers.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2682
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2683 <dt><b><tt>zext (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2684 <dd>Zero extend a constant to another type. The bit size of CST must be
Duncan Sandsa522e562010-07-13 12:06:14 +0000[diff] [blame]2685 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2686
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2687 <dt><b><tt>sext (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2688 <dd>Sign extend a constant to another type. The bit size of CST must be
Duncan Sandsa522e562010-07-13 12:06:14 +0000[diff] [blame]2689 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2690
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2691 <dt><b><tt>fptrunc (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2692 <dd>Truncate a floating point constant to another floating point type. The
2693 size of CST must be larger than the size of TYPE. Both types must be
2694 floating point.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2695
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2696 <dt><b><tt>fpext (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2697 <dd>Floating point extend a constant to another type. The size of CST must be
2698 smaller or equal to the size of TYPE. Both types must be floating
2699 point.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2700
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2701 <dt><b><tt>fptoui (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2702 <dd>Convert a floating point constant to the corresponding unsigned integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2703 constant. TYPE must be a scalar or vector integer type. CST must be of
2704 scalar or vector floating point type. Both CST and TYPE must be scalars,
2705 or vectors of the same number of elements. If the value won't fit in the
2706 integer type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2707
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2708 <dt><b><tt>fptosi (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2709 <dd>Convert a floating point constant to the corresponding signed integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2710 constant. TYPE must be a scalar or vector integer type. CST must be of
2711 scalar or vector floating point type. Both CST and TYPE must be scalars,
2712 or vectors of the same number of elements. If the value won't fit in the
2713 integer type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2714
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2715 <dt><b><tt>uitofp (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2716 <dd>Convert an unsigned integer constant to the corresponding floating point
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2717 constant. TYPE must be a scalar or vector floating point type. CST must be
2718 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2719 vectors of the same number of elements. If the value won't fit in the
2720 floating point type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2721
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2722 <dt><b><tt>sitofp (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2723 <dd>Convert a signed integer constant to the corresponding floating point
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2724 constant. TYPE must be a scalar or vector floating point type. CST must be
2725 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2726 vectors of the same number of elements. If the value won't fit in the
2727 floating point type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2728
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2729 <dt><b><tt>ptrtoint (CST to TYPE)</tt></b></dt>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]2730 <dd>Convert a pointer typed constant to the corresponding integer constant
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2731 <tt>TYPE</tt> must be an integer type. <tt>CST</tt> must be of pointer
2732 type. The <tt>CST</tt> value is zero extended, truncated, or unchanged to
2733 make it fit in <tt>TYPE</tt>.</dd>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]2734
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2735 <dt><b><tt>inttoptr (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2736 <dd>Convert a integer constant to a pointer constant. TYPE must be a pointer
2737 type. CST must be of integer type. The CST value is zero extended,
2738 truncated, or unchanged to make it fit in a pointer size. This one is
2739 <i>really</i> dangerous!</dd>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]2740
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2741 <dt><b><tt>bitcast (CST to TYPE)</tt></b></dt>
Chris Lattner789dee32009-02-28 18:27:03 +0000[diff] [blame]2742 <dd>Convert a constant, CST, to another TYPE. The constraints of the operands
2743 are the same as those for the <a href="#i_bitcast">bitcast
2744 instruction</a>.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2745
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2746 <dt><b><tt>getelementptr (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
2747 <dt><b><tt>getelementptr inbounds (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2748 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2749 constants. As with the <a href="#i_getelementptr">getelementptr</a>
2750 instruction, the index list may have zero or more indexes, which are
2751 required to make sense for the type of "CSTPTR".</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2752
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2753 <dt><b><tt>select (COND, VAL1, VAL2)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2754 <dd>Perform the <a href="#i_select">select operation</a> on constants.</dd>
Reid Spencer9965ee72006-12-04 19:23:19 +0000[diff] [blame]2755
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2756 <dt><b><tt>icmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer9965ee72006-12-04 19:23:19 +0000[diff] [blame]2757 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
2758
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2759 <dt><b><tt>fcmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer9965ee72006-12-04 19:23:19 +0000[diff] [blame]2760 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
Robert Bocchino7e97a6d2006-01-10 19:31:34 +0000[diff] [blame]2761
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2762 <dt><b><tt>extractelement (VAL, IDX)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2763 <dd>Perform the <a href="#i_extractelement">extractelement operation</a> on
2764 constants.</dd>
Robert Bocchino7e97a6d2006-01-10 19:31:34 +0000[diff] [blame]2765
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2766 <dt><b><tt>insertelement (VAL, ELT, IDX)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2767 <dd>Perform the <a href="#i_insertelement">insertelement operation</a> on
2768 constants.</dd>
Chris Lattner016a0e52006-04-08 00:13:41 +0000[diff] [blame]2769
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2770 <dt><b><tt>shufflevector (VEC1, VEC2, IDXMASK)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2771 <dd>Perform the <a href="#i_shufflevector">shufflevector operation</a> on
2772 constants.</dd>
Chris Lattner016a0e52006-04-08 00:13:41 +0000[diff] [blame]2773
Nick Lewycky9ab9a7f2010-05-29 06:44:15 +0000[diff] [blame]2774 <dt><b><tt>extractvalue (VAL, IDX0, IDX1, ...)</tt></b></dt>
2775 <dd>Perform the <a href="#i_extractvalue">extractvalue operation</a> on
2776 constants. The index list is interpreted in a similar manner as indices in
2777 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2778 index value must be specified.</dd>
2779
2780 <dt><b><tt>insertvalue (VAL, ELT, IDX0, IDX1, ...)</tt></b></dt>
2781 <dd>Perform the <a href="#i_insertvalue">insertvalue 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
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2786 <dt><b><tt>OPCODE (LHS, RHS)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2787 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
2788 be any of the <a href="#binaryops">binary</a>
2789 or <a href="#bitwiseops">bitwise binary</a> operations. The constraints
2790 on operands are the same as those for the corresponding instruction
2791 (e.g. no bitwise operations on floating point values are allowed).</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2792</dl>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2793
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2794</div>
Chris Lattnerb1652612004-03-08 16:49:10 +0000[diff] [blame]2795
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2796</div>
2797
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2798<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2799<h2><a name="othervalues">Other Values</a></h2>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2800<!-- *********************************************************************** -->
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2801<div>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2802<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2803<h3>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2804<a name="inlineasm">Inline Assembler Expressions</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2805</h3>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2806
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2807<div>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2808
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2809<p>LLVM supports inline assembler expressions (as opposed
Bill Wendlingad8b58b2011-11-30 21:52:43 +0000[diff] [blame]2810 to <a href="#moduleasm">Module-Level Inline Assembly</a>) through the use of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2811 a special value. This value represents the inline assembler as a string
2812 (containing the instructions to emit), a list of operand constraints (stored
Dale Johannesen63c94fe2009-10-13 21:56:55 +0000[diff] [blame]2813 as a string), a flag that indicates whether or not the inline asm
Dale Johannesen1cfb9582009-10-21 23:28:00 +0000[diff] [blame]2814 expression has side effects, and a flag indicating whether the function
2815 containing the asm needs to align its stack conservatively. An example
2816 inline assembler expression is:</p>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2817
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2818<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2819i32 (i32) asm "bswap $0", "=r,r"
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2820</pre>
2821
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2822<p>Inline assembler expressions may <b>only</b> be used as the callee operand of
2823 a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we
2824 have:</p>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2825
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2826<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2827%X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2828</pre>
2829
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2830<p>Inline asms with side effects not visible in the constraint list must be
2831 marked as having side effects. This is done through the use of the
2832 '<tt>sideeffect</tt>' keyword, like so:</p>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2833
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2834<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2835call void asm sideeffect "eieio", ""()
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2836</pre>
2837
Dale Johannesen1cfb9582009-10-21 23:28:00 +0000[diff] [blame]2838<p>In some cases inline asms will contain code that will not work unless the
2839 stack is aligned in some way, such as calls or SSE instructions on x86,
2840 yet will not contain code that does that alignment within the asm.
2841 The compiler should make conservative assumptions about what the asm might
2842 contain and should generate its usual stack alignment code in the prologue
2843 if the '<tt>alignstack</tt>' keyword is present:</p>
Dale Johannesen63c94fe2009-10-13 21:56:55 +0000[diff] [blame]2844
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2845<pre class="doc_code">
Dale Johannesen1cfb9582009-10-21 23:28:00 +0000[diff] [blame]2846call void asm alignstack "eieio", ""()
Dale Johannesen63c94fe2009-10-13 21:56:55 +0000[diff] [blame]2847</pre>
Dale Johannesen63c94fe2009-10-13 21:56:55 +0000[diff] [blame]2848
2849<p>If both keywords appear the '<tt>sideeffect</tt>' keyword must come
2850 first.</p>
2851
Bill Wendlingad8b58b2011-11-30 21:52:43 +0000[diff] [blame]2852<!--
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2853<p>TODO: The format of the asm and constraints string still need to be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2854 documented here. Constraints on what can be done (e.g. duplication, moving,
2855 etc need to be documented). This is probably best done by reference to
2856 another document that covers inline asm from a holistic perspective.</p>
Bill Wendlingad8b58b2011-11-30 21:52:43 +0000[diff] [blame]2857 -->
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2858
Bill Wendlingad8b58b2011-11-30 21:52:43 +0000[diff] [blame]2859<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2860<h4>
Bill Wendlingad8b58b2011-11-30 21:52:43 +0000[diff] [blame]2861 <a name="inlineasm_md">Inline Asm Metadata</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2862</h4>
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2863
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2864<div>
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2865
Bill Wendlingad8b58b2011-11-30 21:52:43 +0000[diff] [blame]2866<p>The call instructions that wrap inline asm nodes may have a
2867 "<tt>!srcloc</tt>" MDNode attached to it that contains a list of constant
2868 integers. If present, the code generator will use the integer as the
2869 location cookie value when report errors through the <tt>LLVMContext</tt>
2870 error reporting mechanisms. This allows a front-end to correlate backend
2871 errors that occur with inline asm back to the source code that produced it.
2872 For example:</p>
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2873
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2874<pre class="doc_code">
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2875call void asm sideeffect "something bad", ""()<b>, !srcloc !42</b>
2876...
2877!42 = !{ i32 1234567 }
2878</pre>
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2879
2880<p>It is up to the front-end to make sense of the magic numbers it places in the
Bill Wendlingad8b58b2011-11-30 21:52:43 +0000[diff] [blame]2881 IR. If the MDNode contains multiple constants, the code generator will use
Chris Lattner79ffdc72010-11-17 08:20:42 +0000[diff] [blame]2882 the one that corresponds to the line of the asm that the error occurs on.</p>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2883
2884</div>
2885
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2886</div>
2887
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]2888<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2889<h3>
2890 <a name="metadata">Metadata Nodes and Metadata Strings</a>
2891</h3>
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]2892
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2893<div>
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]2894
2895<p>LLVM IR allows metadata to be attached to instructions in the program that
2896 can convey extra information about the code to the optimizers and code
2897 generator. One example application of metadata is source-level debug
2898 information. There are two metadata primitives: strings and nodes. All
2899 metadata has the <tt>metadata</tt> type and is identified in syntax by a
2900 preceding exclamation point ('<tt>!</tt>').</p>
2901
2902<p>A metadata string is a string surrounded by double quotes. It can contain
Bill Wendlingb6c22202011-11-30 21:43:43 +0000[diff] [blame]2903 any character by escaping non-printable characters with "<tt>\xx</tt>" where
2904 "<tt>xx</tt>" is the two digit hex code. For example:
2905 "<tt>!"test\00"</tt>".</p>
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]2906
2907<p>Metadata nodes are represented with notation similar to structure constants
2908 (a comma separated list of elements, surrounded by braces and preceded by an
Bill Wendlingb6c22202011-11-30 21:43:43 +0000[diff] [blame]2909 exclamation point). Metadata nodes can have any values as their operand. For
2910 example:</p>
2911
2912<div class="doc_code">
2913<pre>
2914!{ metadata !"test\00", i32 10}
2915</pre>
2916</div>
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]2917
2918<p>A <a href="#namedmetadatastructure">named metadata</a> is a collection of
2919 metadata nodes, which can be looked up in the module symbol table. For
Bill Wendlingb6c22202011-11-30 21:43:43 +0000[diff] [blame]2920 example:</p>
2921
2922<div class="doc_code">
2923<pre>
2924!foo = metadata !{!4, !3}
2925</pre>
2926</div>
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]2927
Devang Patel9984bd62010-03-04 23:44:48 +0000[diff] [blame]2928<p>Metadata can be used as function arguments. Here <tt>llvm.dbg.value</tt>
Bill Wendlingb6c22202011-11-30 21:43:43 +0000[diff] [blame]2929 function is using two metadata arguments:</p>
Devang Patel9984bd62010-03-04 23:44:48 +0000[diff] [blame]2930
Bill Wendlingc0e10672011-03-02 02:17:11 +0000[diff] [blame]2931<div class="doc_code">
2932<pre>
2933call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2934</pre>
2935</div>
Devang Patel9984bd62010-03-04 23:44:48 +0000[diff] [blame]2936
2937<p>Metadata can be attached with an instruction. Here metadata <tt>!21</tt> is
Bill Wendlingb6c22202011-11-30 21:43:43 +0000[diff] [blame]2938 attached to the <tt>add</tt> instruction using the <tt>!dbg</tt>
2939 identifier:</p>
Devang Patel9984bd62010-03-04 23:44:48 +0000[diff] [blame]2940
Bill Wendlingc0e10672011-03-02 02:17:11 +0000[diff] [blame]2941<div class="doc_code">
2942<pre>
2943%indvar.next = add i64 %indvar, 1, !dbg !21
2944</pre>
2945</div>
2946
Peter Collingbourneec9ff672011-10-27 19:19:07 +0000[diff] [blame]2947<p>More information about specific metadata nodes recognized by the optimizers
2948 and code generator is found below.</p>
2949
Bill Wendlingb6c22202011-11-30 21:43:43 +0000[diff] [blame]2950<!-- _______________________________________________________________________ -->
Peter Collingbourneec9ff672011-10-27 19:19:07 +0000[diff] [blame]2951<h4>
2952 <a name="tbaa">'<tt>tbaa</tt>' Metadata</a>
2953</h4>
2954
2955<div>
2956
2957<p>In LLVM IR, memory does not have types, so LLVM's own type system is not
2958 suitable for doing TBAA. Instead, metadata is added to the IR to describe
2959 a type system of a higher level language. This can be used to implement
2960 typical C/C++ TBAA, but it can also be used to implement custom alias
2961 analysis behavior for other languages.</p>
2962
2963<p>The current metadata format is very simple. TBAA metadata nodes have up to
2964 three fields, e.g.:</p>
2965
2966<div class="doc_code">
2967<pre>
2968!0 = metadata !{ metadata !"an example type tree" }
2969!1 = metadata !{ metadata !"int", metadata !0 }
2970!2 = metadata !{ metadata !"float", metadata !0 }
2971!3 = metadata !{ metadata !"const float", metadata !2, i64 1 }
2972</pre>
2973</div>
2974
2975<p>The first field is an identity field. It can be any value, usually
2976 a metadata string, which uniquely identifies the type. The most important
2977 name in the tree is the name of the root node. Two trees with
2978 different root node names are entirely disjoint, even if they
2979 have leaves with common names.</p>
2980
2981<p>The second field identifies the type's parent node in the tree, or
2982 is null or omitted for a root node. A type is considered to alias
2983 all of its descendants and all of its ancestors in the tree. Also,
2984 a type is considered to alias all types in other trees, so that
2985 bitcode produced from multiple front-ends is handled conservatively.</p>
2986
2987<p>If the third field is present, it's an integer which if equal to 1
2988 indicates that the type is "constant" (meaning
2989 <tt>pointsToConstantMemory</tt> should return true; see
2990 <a href="AliasAnalysis.html#OtherItfs">other useful
2991 <tt>AliasAnalysis</tt> methods</a>).</p>
2992
2993</div>
2994
Bill Wendlingb6c22202011-11-30 21:43:43 +0000[diff] [blame]2995<!-- _______________________________________________________________________ -->
Peter Collingbournef7d1e7b2011-10-27 19:19:14 +0000[diff] [blame]2996<h4>
2997 <a name="fpaccuracy">'<tt>fpaccuracy</tt>' Metadata</a>
2998</h4>
2999
3000<div>
3001
3002<p><tt>fpaccuracy</tt> metadata may be attached to any instruction of floating
3003 point type. It expresses the maximum relative error of the result of
3004 that instruction, in ULPs. ULP is defined as follows:</p>
3005
Bill Wendling302d7ce2011-11-09 19:33:56 +0000[diff] [blame]3006<blockquote>
3007
3008<p>If <tt>x</tt> is a real number that lies between two finite consecutive
3009 floating-point numbers <tt>a</tt> and <tt>b</tt>, without being equal to one
3010 of them, then <tt>ulp(x) = |b - a|</tt>, otherwise <tt>ulp(x)</tt> is the
3011 distance between the two non-equal finite floating-point numbers nearest
3012 <tt>x</tt>. Moreover, <tt>ulp(NaN)</tt> is <tt>NaN</tt>.</p>
3013
3014</blockquote>
Peter Collingbournef7d1e7b2011-10-27 19:19:14 +0000[diff] [blame]3015
3016<p>The maximum relative error may be any rational number. The metadata node
3017 shall consist of a pair of unsigned integers respectively representing
3018 the numerator and denominator. For example, 2.5 ULP:</p>
3019
3020<div class="doc_code">
3021<pre>
3022!0 = metadata !{ i32 5, i32 2 }
3023</pre>
3024</div>
3025
3026</div>
3027
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]3028</div>
3029
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3030</div>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3031
3032<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3033<h2>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3034 <a name="intrinsic_globals">Intrinsic Global Variables</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3035</h2>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3036<!-- *********************************************************************** -->
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3037<div>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3038<p>LLVM has a number of "magic" global variables that contain data that affect
3039code generation or other IR semantics. These are documented here. All globals
Chris Lattner58f9bb22009-07-20 06:14:25 +0000[diff] [blame]3040of this sort should have a section specified as "<tt>llvm.metadata</tt>". This
3041section and all globals that start with "<tt>llvm.</tt>" are reserved for use
3042by LLVM.</p>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3043
3044<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3045<h3>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3046<a name="intg_used">The '<tt>llvm.used</tt>' Global Variable</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3047</h3>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3048
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3049<div>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3050
3051<p>The <tt>@llvm.used</tt> global is an array with i8* element type which has <a
3052href="#linkage_appending">appending linkage</a>. This array contains a list of
3053pointers to global variables and functions which may optionally have a pointer
3054cast formed of bitcast or getelementptr. For example, a legal use of it is:</p>
3055
Bill Wendling1654bb22011-11-08 00:32:45 +0000[diff] [blame]3056<div class="doc_code">
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3057<pre>
Bill Wendling1654bb22011-11-08 00:32:45 +0000[diff] [blame]3058@X = global i8 4
3059@Y = global i32 123
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3060
Bill Wendling1654bb22011-11-08 00:32:45 +0000[diff] [blame]3061@llvm.used = appending global [2 x i8*] [
3062 i8* @X,
3063 i8* bitcast (i32* @Y to i8*)
3064], section "llvm.metadata"
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3065</pre>
Bill Wendling1654bb22011-11-08 00:32:45 +0000[diff] [blame]3066</div>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3067
3068<p>If a global variable appears in the <tt>@llvm.used</tt> list, then the
Bill Wendling1654bb22011-11-08 00:32:45 +0000[diff] [blame]3069 compiler, assembler, and linker are required to treat the symbol as if there
3070 is a reference to the global that it cannot see. For example, if a variable
3071 has internal linkage and no references other than that from
3072 the <tt>@llvm.used</tt> list, it cannot be deleted. This is commonly used to
3073 represent references from inline asms and other things the compiler cannot
3074 "see", and corresponds to "<tt>attribute((used))</tt>" in GNU C.</p>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3075
3076<p>On some targets, the code generator must emit a directive to the assembler or
Bill Wendling1654bb22011-11-08 00:32:45 +0000[diff] [blame]3077 object file to prevent the assembler and linker from molesting the
3078 symbol.</p>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3079
3080</div>
3081
3082<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3083<h3>
3084 <a name="intg_compiler_used">
3085 The '<tt>llvm.compiler.used</tt>' Global Variable
3086 </a>
3087</h3>
Chris Lattner58f9bb22009-07-20 06:14:25 +0000[diff] [blame]3088
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3089<div>
Chris Lattner58f9bb22009-07-20 06:14:25 +0000[diff] [blame]3090
3091<p>The <tt>@llvm.compiler.used</tt> directive is the same as the
Bill Wendling1654bb22011-11-08 00:32:45 +0000[diff] [blame]3092 <tt>@llvm.used</tt> directive, except that it only prevents the compiler from
3093 touching the symbol. On targets that support it, this allows an intelligent
3094 linker to optimize references to the symbol without being impeded as it would
3095 be by <tt>@llvm.used</tt>.</p>
Chris Lattner58f9bb22009-07-20 06:14:25 +0000[diff] [blame]3096
3097<p>This is a rare construct that should only be used in rare circumstances, and
Bill Wendling1654bb22011-11-08 00:32:45 +0000[diff] [blame]3098 should not be exposed to source languages.</p>
Chris Lattner58f9bb22009-07-20 06:14:25 +0000[diff] [blame]3099
3100</div>
3101
3102<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3103<h3>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3104<a name="intg_global_ctors">The '<tt>llvm.global_ctors</tt>' Global Variable</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3105</h3>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3106
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3107<div>
Bill Wendling1654bb22011-11-08 00:32:45 +0000[diff] [blame]3108
3109<div class="doc_code">
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]3110<pre>
3111%0 = type { i32, void ()* }
David Chisnallb492b812010-04-30 19:27:35 +0000[diff] [blame]3112@llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor }]
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]3113</pre>
Bill Wendling1654bb22011-11-08 00:32:45 +0000[diff] [blame]3114</div>
3115
3116<p>The <tt>@llvm.global_ctors</tt> array contains a list of constructor
3117 functions and associated priorities. The functions referenced by this array
3118 will be called in ascending order of priority (i.e. lowest first) when the
3119 module is loaded. The order of functions with the same priority is not
3120 defined.</p>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3121
3122</div>
3123
3124<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3125<h3>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3126<a name="intg_global_dtors">The '<tt>llvm.global_dtors</tt>' Global Variable</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3127</h3>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3128
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3129<div>
Bill Wendling1654bb22011-11-08 00:32:45 +0000[diff] [blame]3130
3131<div class="doc_code">
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]3132<pre>
3133%0 = type { i32, void ()* }
David Chisnallb492b812010-04-30 19:27:35 +0000[diff] [blame]3134@llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor }]
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]3135</pre>
Bill Wendling1654bb22011-11-08 00:32:45 +0000[diff] [blame]3136</div>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3137
Bill Wendling1654bb22011-11-08 00:32:45 +0000[diff] [blame]3138<p>The <tt>@llvm.global_dtors</tt> array contains a list of destructor functions
3139 and associated priorities. The functions referenced by this array will be
3140 called in descending order of priority (i.e. highest first) when the module
3141 is loaded. The order of functions with the same priority is not defined.</p>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3142
3143</div>
3144
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3145</div>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3146
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]3147<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3148<h2><a name="instref">Instruction Reference</a></h2>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3149<!-- *********************************************************************** -->
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]3150
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3151<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]3152
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3153<p>The LLVM instruction set consists of several different classifications of
3154 instructions: <a href="#terminators">terminator
3155 instructions</a>, <a href="#binaryops">binary instructions</a>,
3156 <a href="#bitwiseops">bitwise binary instructions</a>,
3157 <a href="#memoryops">memory instructions</a>, and
3158 <a href="#otherops">other instructions</a>.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]3159
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3160<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3161<h3>
3162 <a name="terminators">Terminator Instructions</a>
3163</h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]3164
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3165<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]3166
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3167<p>As mentioned <a href="#functionstructure">previously</a>, every basic block
3168 in a program ends with a "Terminator" instruction, which indicates which
3169 block should be executed after the current block is finished. These
3170 terminator instructions typically yield a '<tt>void</tt>' value: they produce
3171 control flow, not values (the one exception being the
3172 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
3173
Chris Lattnerd3d65ab2011-08-02 20:29:13 +0000[diff] [blame]3174<p>The terminator instructions are:
3175 '<a href="#i_ret"><tt>ret</tt></a>',
3176 '<a href="#i_br"><tt>br</tt></a>',
3177 '<a href="#i_switch"><tt>switch</tt></a>',
3178 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>',
3179 '<a href="#i_invoke"><tt>invoke</tt></a>',
Chris Lattnerd3d65ab2011-08-02 20:29:13 +0000[diff] [blame]3180 '<a href="#i_resume"><tt>resume</tt></a>', and
3181 '<a href="#i_unreachable"><tt>unreachable</tt></a>'.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]3182
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3183<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3184<h4>
3185 <a name="i_ret">'<tt>ret</tt>' Instruction</a>
3186</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3187
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3188<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3189
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3190<h5>Syntax:</h5>
Dan Gohmancc3132e2008-10-04 19:00:07 +0000[diff] [blame]3191<pre>
3192 ret &lt;type&gt; &lt;value&gt; <i>; Return a value from a non-void function</i>
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]3193 ret void <i>; Return from void function</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3194</pre>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]3195
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3196<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3197<p>The '<tt>ret</tt>' instruction is used to return control flow (and optionally
3198 a value) from a function back to the caller.</p>
3199
3200<p>There are two forms of the '<tt>ret</tt>' instruction: one that returns a
3201 value and then causes control flow, and one that just causes control flow to
3202 occur.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]3203
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3204<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3205<p>The '<tt>ret</tt>' instruction optionally accepts a single argument, the
3206 return value. The type of the return value must be a
3207 '<a href="#t_firstclass">first class</a>' type.</p>
Dan Gohmancc3132e2008-10-04 19:00:07 +0000[diff] [blame]3208
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3209<p>A function is not <a href="#wellformed">well formed</a> if it it has a
3210 non-void return type and contains a '<tt>ret</tt>' instruction with no return
3211 value or a return value with a type that does not match its type, or if it
3212 has a void return type and contains a '<tt>ret</tt>' instruction with a
3213 return value.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]3214
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3215<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3216<p>When the '<tt>ret</tt>' instruction is executed, control flow returns back to
3217 the calling function's context. If the caller is a
3218 "<a href="#i_call"><tt>call</tt></a>" instruction, execution continues at the
3219 instruction after the call. If the caller was an
3220 "<a href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues at
3221 the beginning of the "normal" destination block. If the instruction returns
3222 a value, that value shall set the call or invoke instruction's return
3223 value.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]3224
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3225<h5>Example:</h5>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]3226<pre>
3227 ret i32 5 <i>; Return an integer value of 5</i>
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]3228 ret void <i>; Return from a void function</i>
Bill Wendling050ee8f2009-02-28 22:12:54 +0000[diff] [blame]3229 ret { i32, i8 } { i32 4, i8 2 } <i>; Return a struct of values 4 and 2</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3230</pre>
Dan Gohman3065b612009-01-12 23:12:39 +0000[diff] [blame]3231
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3232</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3233<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3234<h4>
3235 <a name="i_br">'<tt>br</tt>' Instruction</a>
3236</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3237
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3238<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3239
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3240<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3241<pre>
Bill Wendling16b86742011-07-26 10:41:15 +0000[diff] [blame]3242 br i1 &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;
3243 br label &lt;dest&gt; <i>; Unconditional branch</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3244</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3245
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3246<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3247<p>The '<tt>br</tt>' instruction is used to cause control flow to transfer to a
3248 different basic block in the current function. There are two forms of this
3249 instruction, corresponding to a conditional branch and an unconditional
3250 branch.</p>
3251
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3252<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3253<p>The conditional branch form of the '<tt>br</tt>' instruction takes a single
3254 '<tt>i1</tt>' value and two '<tt>label</tt>' values. The unconditional form
3255 of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>' value as a
3256 target.</p>
3257
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3258<h5>Semantics:</h5>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]3259<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3260 argument is evaluated. If the value is <tt>true</tt>, control flows to the
3261 '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
3262 control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
3263
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3264<h5>Example:</h5>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]3265<pre>
3266Test:
3267 %cond = <a href="#i_icmp">icmp</a> eq i32 %a, %b
3268 br i1 %cond, label %IfEqual, label %IfUnequal
3269IfEqual:
3270 <a href="#i_ret">ret</a> i32 1
3271IfUnequal:
3272 <a href="#i_ret">ret</a> i32 0
3273</pre>
3274
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3275</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3276
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3277<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3278<h4>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3279 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3280</h4>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3281
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3282<div>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3283
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3284<h5>Syntax:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3285<pre>
3286 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
3287</pre>
3288
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3289<h5>Overview:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3290<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3291 several different places. It is a generalization of the '<tt>br</tt>'
3292 instruction, allowing a branch to occur to one of many possible
3293 destinations.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3294
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3295<h5>Arguments:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3296<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3297 comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination,
3298 and an array of pairs of comparison value constants and '<tt>label</tt>'s.
3299 The table is not allowed to contain duplicate constant entries.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3300
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3301<h5>Semantics:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3302<p>The <tt>switch</tt> instruction specifies a table of values and
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3303 destinations. When the '<tt>switch</tt>' instruction is executed, this table
3304 is searched for the given value. If the value is found, control flow is
Benjamin Kramer0f420382009-10-12 14:46:08 +0000[diff] [blame]3305 transferred to the corresponding destination; otherwise, control flow is
3306 transferred to the default destination.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3307
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3308<h5>Implementation:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3309<p>Depending on properties of the target machine and the particular
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3310 <tt>switch</tt> instruction, this instruction may be code generated in
3311 different ways. For example, it could be generated as a series of chained
3312 conditional branches or with a lookup table.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3313
3314<h5>Example:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3315<pre>
3316 <i>; Emulate a conditional br instruction</i>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]3317 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Dan Gohman623806e2009-01-04 23:44:43 +0000[diff] [blame]3318 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3319
3320 <i>; Emulate an unconditional br instruction</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]3321 switch i32 0, label %dest [ ]
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3322
3323 <i>; Implement a jump table:</i>
Dan Gohman623806e2009-01-04 23:44:43 +0000[diff] [blame]3324 switch i32 %val, label %otherwise [ i32 0, label %onzero
3325 i32 1, label %onone
3326 i32 2, label %ontwo ]
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3327</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3328
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3329</div>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3330
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3331
3332<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3333<h4>
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]3334 <a name="i_indirectbr">'<tt>indirectbr</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3335</h4>
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3336
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3337<div>
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3338
3339<h5>Syntax:</h5>
3340<pre>
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]3341 indirectbr &lt;somety&gt;* &lt;address&gt;, [ label &lt;dest1&gt;, label &lt;dest2&gt;, ... ]
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3342</pre>
3343
3344<h5>Overview:</h5>
3345
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]3346<p>The '<tt>indirectbr</tt>' instruction implements an indirect branch to a label
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3347 within the current function, whose address is specified by
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]3348 "<tt>address</tt>". Address must be derived from a <a
3349 href="#blockaddress">blockaddress</a> constant.</p>
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3350
3351<h5>Arguments:</h5>
3352
3353<p>The '<tt>address</tt>' argument is the address of the label to jump to. The
3354 rest of the arguments indicate the full set of possible destinations that the
3355 address may point to. Blocks are allowed to occur multiple times in the
3356 destination list, though this isn't particularly useful.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3357
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3358<p>This destination list is required so that dataflow analysis has an accurate
3359 understanding of the CFG.</p>
3360
3361<h5>Semantics:</h5>
3362
3363<p>Control transfers to the block specified in the address argument. All
3364 possible destination blocks must be listed in the label list, otherwise this
3365 instruction has undefined behavior. This implies that jumps to labels
3366 defined in other functions have undefined behavior as well.</p>
3367
3368<h5>Implementation:</h5>
3369
3370<p>This is typically implemented with a jump through a register.</p>
3371
3372<h5>Example:</h5>
3373<pre>
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]3374 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3375</pre>
3376
3377</div>
3378
3379
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3380<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3381<h4>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3382 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3383</h4>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3384
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3385<div>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3386
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3387<h5>Syntax:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3388<pre>
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]3389 &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 Lattner6b7a0082006-05-14 18:23:06 +0000[diff] [blame]3390 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3391</pre>
3392
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]3393<h5>Overview:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3394<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3395 function, with the possibility of control flow transfer to either the
3396 '<tt>normal</tt>' label or the '<tt>exception</tt>' label. If the callee
3397 function returns with the "<tt><a href="#i_ret">ret</a></tt>" instruction,
3398 control flow will return to the "normal" label. If the callee (or any
Bill Wendling3f6a3a22012-02-06 21:57:33 +0000[diff] [blame^]3399 indirect callees) returns via the "<a href="#i_resume"><tt>resume</tt></a>"
3400 instruction or other exception handling mechanism, control is interrupted and
3401 continued at the dynamically nearest "exception" label.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3402
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]3403<p>The '<tt>exception</tt>' label is a
3404 <i><a href="ExceptionHandling.html#overview">landing pad</a></i> for the
3405 exception. As such, '<tt>exception</tt>' label is required to have the
3406 "<a href="#i_landingpad"><tt>landingpad</tt></a>" instruction, which contains
Chad Rosierc28f3e92011-12-09 02:00:44 +0000[diff] [blame]3407 the information about the behavior of the program after unwinding
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]3408 happens, as its first non-PHI instruction. The restrictions on the
3409 "<tt>landingpad</tt>" instruction's tightly couples it to the
3410 "<tt>invoke</tt>" instruction, so that the important information contained
3411 within the "<tt>landingpad</tt>" instruction can't be lost through normal
3412 code motion.</p>
3413
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3414<h5>Arguments:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3415<p>This instruction requires several arguments:</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3416
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3417<ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3418 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
3419 convention</a> the call should use. If none is specified, the call
3420 defaults to using C calling conventions.</li>
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]3421
3422 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3423 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
3424 '<tt>inreg</tt>' attributes are valid here.</li>
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]3425
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3426 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3427 function value being invoked. In most cases, this is a direct function
3428 invocation, but indirect <tt>invoke</tt>s are just as possible, branching
3429 off an arbitrary pointer to function value.</li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3430
3431 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3432 function to be invoked. </li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3433
3434 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]3435 signature argument types and parameter attributes. All arguments must be
3436 of <a href="#t_firstclass">first class</a> type. If the function
3437 signature indicates the function accepts a variable number of arguments,
3438 the extra arguments can be specified.</li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3439
3440 <li>'<tt>normal label</tt>': the label reached when the called function
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3441 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3442
Bill Wendling3f6a3a22012-02-06 21:57:33 +0000[diff] [blame^]3443 <li>'<tt>exception label</tt>': the label reached when a callee returns via
3444 the <a href="#i_resume"><tt>resume</tt></a> instruction or other exception
3445 handling mechanism.</li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3446
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]3447 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3448 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
3449 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3450</ol>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3451
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3452<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3453<p>This instruction is designed to operate as a standard
3454 '<tt><a href="#i_call">call</a></tt>' instruction in most regards. The
3455 primary difference is that it establishes an association with a label, which
3456 is used by the runtime library to unwind the stack.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3457
3458<p>This instruction is used in languages with destructors to ensure that proper
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3459 cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
3460 exception. Additionally, this is important for implementation of
3461 '<tt>catch</tt>' clauses in high-level languages that support them.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3462
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3463<p>For the purposes of the SSA form, the definition of the value returned by the
3464 '<tt>invoke</tt>' instruction is deemed to occur on the edge from the current
3465 block to the "normal" label. If the callee unwinds then no return value is
3466 available.</p>
Dan Gohman9069d892009-05-22 21:47:08 +0000[diff] [blame]3467
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3468<h5>Example:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3469<pre>
Nick Lewycky084ab472008-03-16 07:18:12 +0000[diff] [blame]3470 %retval = invoke i32 @Test(i32 15) to label %Continue
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]3471 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewycky084ab472008-03-16 07:18:12 +0000[diff] [blame]3472 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]3473 unwind label %TestCleanup <i>; {i32}:retval set</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3474</pre>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3475
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3476</div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3477
Bill Wendlingf891bf82011-07-31 06:30:59 +0000[diff] [blame]3478 <!-- _______________________________________________________________________ -->
3479
3480<h4>
3481 <a name="i_resume">'<tt>resume</tt>' Instruction</a>
3482</h4>
3483
3484<div>
3485
3486<h5>Syntax:</h5>
3487<pre>
3488 resume &lt;type&gt; &lt;value&gt;
3489</pre>
3490
3491<h5>Overview:</h5>
3492<p>The '<tt>resume</tt>' instruction is a terminator instruction that has no
3493 successors.</p>
3494
3495<h5>Arguments:</h5>
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]3496<p>The '<tt>resume</tt>' instruction requires one argument, which must have the
Bill Wendlingc5a13612011-08-03 18:37:32 +0000[diff] [blame]3497 same type as the result of any '<tt>landingpad</tt>' instruction in the same
3498 function.</p>
Bill Wendlingf891bf82011-07-31 06:30:59 +0000[diff] [blame]3499
3500<h5>Semantics:</h5>
3501<p>The '<tt>resume</tt>' instruction resumes propagation of an existing
3502 (in-flight) exception whose unwinding was interrupted with
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]3503 a <a href="#i_landingpad"><tt>landingpad</tt></a> instruction.</p>
Bill Wendlingf891bf82011-07-31 06:30:59 +0000[diff] [blame]3504
3505<h5>Example:</h5>
3506<pre>
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]3507 resume { i8*, i32 } %exn
Bill Wendlingf891bf82011-07-31 06:30:59 +0000[diff] [blame]3508</pre>
3509
3510</div>
3511
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3512<!-- _______________________________________________________________________ -->
3513
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3514<h4>
3515 <a name="i_unreachable">'<tt>unreachable</tt>' Instruction</a>
3516</h4>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3517
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3518<div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3519
3520<h5>Syntax:</h5>
3521<pre>
3522 unreachable
3523</pre>
3524
3525<h5>Overview:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3526<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3527 instruction is used to inform the optimizer that a particular portion of the
3528 code is not reachable. This can be used to indicate that the code after a
3529 no-return function cannot be reached, and other facts.</p>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3530
3531<h5>Semantics:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3532<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3533
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3534</div>
3535
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3536</div>
3537
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3538<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3539<h3>
3540 <a name="binaryops">Binary Operations</a>
3541</h3>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3542
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3543<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3544
3545<p>Binary operators are used to do most of the computation in a program. They
3546 require two operands of the same type, execute an operation on them, and
3547 produce a single value. The operands might represent multiple data, as is
3548 the case with the <a href="#t_vector">vector</a> data type. The result value
3549 has the same type as its operands.</p>
3550
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3551<p>There are several different binary operators:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3552
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3553<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3554<h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3555 <a name="i_add">'<tt>add</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3556</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3557
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3558<div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3559
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3560<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3561<pre>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3562 &lt;result&gt; = add &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohman957b1312009-09-02 17:31:42 +0000[diff] [blame]3563 &lt;result&gt; = add nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3564 &lt;result&gt; = add nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3565 &lt;result&gt; = add nuw nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3566</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3567
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3568<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3569<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3570
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3571<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3572<p>The two arguments to the '<tt>add</tt>' instruction must
3573 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3574 integer values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3575
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3576<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3577<p>The value produced is the integer sum of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3578
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3579<p>If the sum has unsigned overflow, the result returned is the mathematical
3580 result modulo 2<sup>n</sup>, where n is the bit width of the result.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3581
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3582<p>Because LLVM integers use a two's complement representation, this instruction
3583 is appropriate for both signed and unsigned integers.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3584
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3585<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3586 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3587 <tt>nsw</tt> keywords are present, the result value of the <tt>add</tt>
Dan Gohman9a2a0932011-12-06 03:18:47 +0000[diff] [blame]3588 is a <a href="#poisonvalues">poison value</a> if unsigned and/or signed overflow,
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]3589 respectively, occurs.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3590
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3591<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3592<pre>
3593 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3594</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3595
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3596</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3597
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3598<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3599<h4>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3600 <a name="i_fadd">'<tt>fadd</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3601</h4>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3602
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3603<div>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3604
3605<h5>Syntax:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3606<pre>
3607 &lt;result&gt; = fadd &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3608</pre>
3609
3610<h5>Overview:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3611<p>The '<tt>fadd</tt>' instruction returns the sum of its two operands.</p>
3612
3613<h5>Arguments:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3614<p>The two arguments to the '<tt>fadd</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3615 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3616 floating point values. Both arguments must have identical types.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3617
3618<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3619<p>The value produced is the floating point sum of the two operands.</p>
3620
3621<h5>Example:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3622<pre>
3623 &lt;result&gt; = fadd float 4.0, %var <i>; yields {float}:result = 4.0 + %var</i>
3624</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3625
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3626</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3627
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3628<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3629<h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3630 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3631</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3632
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3633<div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3634
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3635<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3636<pre>
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3637 &lt;result&gt; = sub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohman957b1312009-09-02 17:31:42 +0000[diff] [blame]3638 &lt;result&gt; = sub nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3639 &lt;result&gt; = sub nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3640 &lt;result&gt; = sub nuw nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3641</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3642
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3643<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3644<p>The '<tt>sub</tt>' instruction returns the difference of its two
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3645 operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3646
3647<p>Note that the '<tt>sub</tt>' instruction is used to represent the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3648 '<tt>neg</tt>' instruction present in most other intermediate
3649 representations.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3650
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3651<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3652<p>The two arguments to the '<tt>sub</tt>' instruction must
3653 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3654 integer values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3655
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3656<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3657<p>The value produced is the integer difference of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3658
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3659<p>If the difference has unsigned overflow, the result returned is the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3660 mathematical result modulo 2<sup>n</sup>, where n is the bit width of the
3661 result.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3662
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3663<p>Because LLVM integers use a two's complement representation, this instruction
3664 is appropriate for both signed and unsigned integers.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3665
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3666<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3667 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3668 <tt>nsw</tt> keywords are present, the result value of the <tt>sub</tt>
Dan Gohman9a2a0932011-12-06 03:18:47 +0000[diff] [blame]3669 is a <a href="#poisonvalues">poison value</a> if unsigned and/or signed overflow,
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]3670 respectively, occurs.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3671
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3672<h5>Example:</h5>
Bill Wendling2d8b9a82007-05-29 09:42:13 +0000[diff] [blame]3673<pre>
3674 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]3675 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3676</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3677
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3678</div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3679
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3680<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3681<h4>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3682 <a name="i_fsub">'<tt>fsub</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3683</h4>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3684
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3685<div>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3686
3687<h5>Syntax:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3688<pre>
3689 &lt;result&gt; = fsub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3690</pre>
3691
3692<h5>Overview:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3693<p>The '<tt>fsub</tt>' instruction returns the difference of its two
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3694 operands.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3695
3696<p>Note that the '<tt>fsub</tt>' instruction is used to represent the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3697 '<tt>fneg</tt>' instruction present in most other intermediate
3698 representations.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3699
3700<h5>Arguments:</h5>
Bill Wendling972b7202009-07-20 02:32:41 +0000[diff] [blame]3701<p>The two arguments to the '<tt>fsub</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3702 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3703 floating point values. Both arguments must have identical types.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3704
3705<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3706<p>The value produced is the floating point difference of the two operands.</p>
3707
3708<h5>Example:</h5>
3709<pre>
3710 &lt;result&gt; = fsub float 4.0, %var <i>; yields {float}:result = 4.0 - %var</i>
3711 &lt;result&gt; = fsub float -0.0, %val <i>; yields {float}:result = -%var</i>
3712</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3713
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3714</div>
3715
3716<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3717<h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3718 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3719</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3720
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3721<div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3722
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3723<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3724<pre>
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3725 &lt;result&gt; = mul &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohman957b1312009-09-02 17:31:42 +0000[diff] [blame]3726 &lt;result&gt; = mul nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3727 &lt;result&gt; = mul nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3728 &lt;result&gt; = mul nuw nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3729</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3730
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3731<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3732<p>The '<tt>mul</tt>' instruction returns the product of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3733
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3734<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3735<p>The two arguments to the '<tt>mul</tt>' instruction must
3736 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3737 integer values. Both arguments must have identical types.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3738
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3739<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3740<p>The value produced is the integer product of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3741
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3742<p>If the result of the multiplication has unsigned overflow, the result
3743 returned is the mathematical result modulo 2<sup>n</sup>, where n is the bit
3744 width of the result.</p>
3745
3746<p>Because LLVM integers use a two's complement representation, and the result
3747 is the same width as the operands, this instruction returns the correct
3748 result for both signed and unsigned integers. If a full product
3749 (e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands should
3750 be sign-extended or zero-extended as appropriate to the width of the full
3751 product.</p>
3752
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3753<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3754 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3755 <tt>nsw</tt> keywords are present, the result value of the <tt>mul</tt>
Dan Gohman9a2a0932011-12-06 03:18:47 +0000[diff] [blame]3756 is a <a href="#poisonvalues">poison value</a> if unsigned and/or signed overflow,
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]3757 respectively, occurs.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3758
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3759<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3760<pre>
3761 &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3762</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3763
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3764</div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3765
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3766<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3767<h4>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3768 <a name="i_fmul">'<tt>fmul</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3769</h4>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3770
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3771<div>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3772
3773<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3774<pre>
3775 &lt;result&gt; = fmul &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3776</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3777
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3778<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3779<p>The '<tt>fmul</tt>' instruction returns the product of its two operands.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3780
3781<h5>Arguments:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3782<p>The two arguments to the '<tt>fmul</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3783 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3784 floating point values. Both arguments must have identical types.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3785
3786<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3787<p>The value produced is the floating point product of the two operands.</p>
3788
3789<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3790<pre>
3791 &lt;result&gt; = fmul float 4.0, %var <i>; yields {float}:result = 4.0 * %var</i>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3792</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3793
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3794</div>
3795
3796<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3797<h4>
3798 <a name="i_udiv">'<tt>udiv</tt>' Instruction</a>
3799</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3800
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3801<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3802
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3803<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3804<pre>
Chris Lattner35315d02011-02-06 21:44:57 +0000[diff] [blame]3805 &lt;result&gt; = udiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3806 &lt;result&gt; = udiv exact &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3807</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3808
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3809<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3810<p>The '<tt>udiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3811
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3812<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3813<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3814 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3815 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3816
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3817<h5>Semantics:</h5>
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3818<p>The value produced is the unsigned integer quotient of the two operands.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3819
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3820<p>Note that unsigned integer division and signed integer division are distinct
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3821 operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
3822
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3823<p>Division by zero leads to undefined behavior.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3824
Chris Lattner35315d02011-02-06 21:44:57 +0000[diff] [blame]3825<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohman9a2a0932011-12-06 03:18:47 +0000[diff] [blame]3826 <tt>udiv</tt> is a <a href="#poisonvalues">poison value</a> if %op1 is not a
Chris Lattner35315d02011-02-06 21:44:57 +0000[diff] [blame]3827 multiple of %op2 (as such, "((a udiv exact b) mul b) == a").</p>
3828
3829
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3830<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3831<pre>
3832 &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3833</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3834
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3835</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3836
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3837<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3838<h4>
3839 <a name="i_sdiv">'<tt>sdiv</tt>' Instruction</a>
3840</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3841
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3842<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3843
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3844<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3845<pre>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3846 &lt;result&gt; = sdiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohman957b1312009-09-02 17:31:42 +0000[diff] [blame]3847 &lt;result&gt; = sdiv exact &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3848</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3849
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3850<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3851<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3852
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3853<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3854<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3855 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3856 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3857
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3858<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3859<p>The value produced is the signed integer quotient of the two operands rounded
3860 towards zero.</p>
3861
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3862<p>Note that signed integer division and unsigned integer division are distinct
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3863 operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
3864
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3865<p>Division by zero leads to undefined behavior. Overflow also leads to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3866 undefined behavior; this is a rare case, but can occur, for example, by doing
3867 a 32-bit division of -2147483648 by -1.</p>
3868
Dan Gohman71dfd782009-07-22 00:04:19 +0000[diff] [blame]3869<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohman9a2a0932011-12-06 03:18:47 +0000[diff] [blame]3870 <tt>sdiv</tt> is a <a href="#poisonvalues">poison value</a> if the result would
Dan Gohmane501ff72010-07-11 00:08:34 +0000[diff] [blame]3871 be rounded.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3872
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3873<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3874<pre>
3875 &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3876</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3877
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3878</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3879
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3880<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3881<h4>
3882 <a name="i_fdiv">'<tt>fdiv</tt>' Instruction</a>
3883</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3884
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3885<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3886
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3887<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3888<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +0000[diff] [blame]3889 &lt;result&gt; = fdiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3890</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3891
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3892<h5>Overview:</h5>
3893<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3894
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3895<h5>Arguments:</h5>
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]3896<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3897 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3898 floating point values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3899
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3900<h5>Semantics:</h5>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3901<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3902
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3903<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3904<pre>
3905 &lt;result&gt; = fdiv float 4.0, %var <i>; yields {float}:result = 4.0 / %var</i>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3906</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3907
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3908</div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3909
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3910<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3911<h4>
3912 <a name="i_urem">'<tt>urem</tt>' Instruction</a>
3913</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3914
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3915<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3916
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3917<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3918<pre>
3919 &lt;result&gt; = urem &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3920</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3921
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3922<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3923<p>The '<tt>urem</tt>' instruction returns the remainder from the unsigned
3924 division of its two arguments.</p>
3925
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3926<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3927<p>The two arguments to the '<tt>urem</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3928 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3929 values. Both arguments must have identical types.</p>
3930
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3931<h5>Semantics:</h5>
3932<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3933 This instruction always performs an unsigned division to get the
3934 remainder.</p>
3935
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3936<p>Note that unsigned integer remainder and signed integer remainder are
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3937 distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
3938
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3939<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3940
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3941<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3942<pre>
3943 &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3944</pre>
3945
3946</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3947
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3948<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3949<h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3950 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3951</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3952
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3953<div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3954
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3955<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3956<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +0000[diff] [blame]3957 &lt;result&gt; = srem &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3958</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3959
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3960<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3961<p>The '<tt>srem</tt>' instruction returns the remainder from the signed
3962 division of its two operands. This instruction can also take
3963 <a href="#t_vector">vector</a> versions of the values in which case the
3964 elements must be integers.</p>
Chris Lattnerb8f816e2008-01-04 04:33:49 +0000[diff] [blame]3965
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3966<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3967<p>The two arguments to the '<tt>srem</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3968 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3969 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3970
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3971<h5>Semantics:</h5>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3972<p>This instruction returns the <i>remainder</i> of a division (where the result
Duncan Sands2769c6e2011-03-07 09:12:24 +0000[diff] [blame]3973 is either zero or has the same sign as the dividend, <tt>op1</tt>), not the
3974 <i>modulo</i> operator (where the result is either zero or has the same sign
3975 as the divisor, <tt>op2</tt>) of a value.
3976 For more information about the difference,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3977 see <a href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
3978 Math Forum</a>. For a table of how this is implemented in various languages,
3979 please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
3980 Wikipedia: modulo operation</a>.</p>
3981
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3982<p>Note that signed integer remainder and unsigned integer remainder are
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3983 distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
3984
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3985<p>Taking the remainder of a division by zero leads to undefined behavior.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3986 Overflow also leads to undefined behavior; this is a rare case, but can
3987 occur, for example, by taking the remainder of a 32-bit division of
3988 -2147483648 by -1. (The remainder doesn't actually overflow, but this rule
3989 lets srem be implemented using instructions that return both the result of
3990 the division and the remainder.)</p>
3991
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3992<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3993<pre>
3994 &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3995</pre>
3996
3997</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3998
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3999<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4000<h4>
4001 <a name="i_frem">'<tt>frem</tt>' Instruction</a>
4002</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4003
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4004<div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4005
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]4006<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4007<pre>
4008 &lt;result&gt; = frem &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]4009</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4010
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]4011<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4012<p>The '<tt>frem</tt>' instruction returns the remainder from the division of
4013 its two operands.</p>
4014
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]4015<h5>Arguments:</h5>
4016<p>The two arguments to the '<tt>frem</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4017 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
4018 floating point values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4019
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]4020<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4021<p>This instruction returns the <i>remainder</i> of a division. The remainder
4022 has the same sign as the dividend.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4023
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]4024<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4025<pre>
4026 &lt;result&gt; = frem float 4.0, %var <i>; yields {float}:result = 4.0 % %var</i>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]4027</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4028
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4029</div>
Robert Bocchino820bc75b2006-02-17 21:18:08 +0000[diff] [blame]4030
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4031</div>
4032
Reid Spencer2ab01932007-02-02 13:57:07 +0000[diff] [blame]4033<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4034<h3>
4035 <a name="bitwiseops">Bitwise Binary Operations</a>
4036</h3>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4037
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4038<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4039
4040<p>Bitwise binary operators are used to do various forms of bit-twiddling in a
4041 program. They are generally very efficient instructions and can commonly be
4042 strength reduced from other instructions. They require two operands of the
4043 same type, execute an operation on them, and produce a single value. The
4044 resulting value is the same type as its operands.</p>
4045
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4046<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4047<h4>
4048 <a name="i_shl">'<tt>shl</tt>' Instruction</a>
4049</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4050
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4051<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4052
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4053<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4054<pre>
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]4055 &lt;result&gt; = shl &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4056 &lt;result&gt; = shl nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4057 &lt;result&gt; = shl nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4058 &lt;result&gt; = shl nuw nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4059</pre>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]4060
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4061<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4062<p>The '<tt>shl</tt>' instruction returns the first operand shifted to the left
4063 a specified number of bits.</p>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]4064
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4065<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4066<p>Both arguments to the '<tt>shl</tt>' instruction must be the
4067 same <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
4068 integer type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4069
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4070<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4071<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod
4072 2<sup>n</sup>, where <tt>n</tt> is the width of the result. If <tt>op2</tt>
4073 is (statically or dynamically) negative or equal to or larger than the number
4074 of bits in <tt>op1</tt>, the result is undefined. If the arguments are
4075 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
4076 shift amount in <tt>op2</tt>.</p>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]4077
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]4078<p>If the <tt>nuw</tt> keyword is present, then the shift produces a
Dan Gohman9a2a0932011-12-06 03:18:47 +0000[diff] [blame]4079 <a href="#poisonvalues">poison value</a> if it shifts out any non-zero bits. If
Chris Lattnerf10dfdc2011-02-09 16:44:44 +0000[diff] [blame]4080 the <tt>nsw</tt> keyword is present, then the shift produces a
Dan Gohman9a2a0932011-12-06 03:18:47 +0000[diff] [blame]4081 <a href="#poisonvalues">poison value</a> if it shifts out any bits that disagree
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]4082 with the resultant sign bit. As such, NUW/NSW have the same semantics as
4083 they would if the shift were expressed as a mul instruction with the same
4084 nsw/nuw bits in (mul %op1, (shl 1, %op2)).</p>
4085
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4086<h5>Example:</h5>
4087<pre>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4088 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
4089 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
4090 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]4091 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Mon P Wang4dd832d2008-12-09 05:46:39 +0000[diff] [blame]4092 &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 Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4093</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4094
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4095</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4096
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4097<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4098<h4>
4099 <a name="i_lshr">'<tt>lshr</tt>' Instruction</a>
4100</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4101
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4102<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4103
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4104<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4105<pre>
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]4106 &lt;result&gt; = lshr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4107 &lt;result&gt; = lshr exact &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4108</pre>
4109
4110<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4111<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
4112 operand shifted to the right a specified number of bits with zero fill.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4113
4114<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4115<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4116 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4117 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4118
4119<h5>Semantics:</h5>
4120<p>This instruction always performs a logical shift right operation. The most
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4121 significant bits of the result will be filled with zero bits after the shift.
4122 If <tt>op2</tt> is (statically or dynamically) equal to or larger than the
4123 number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
4124 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
4125 shift amount in <tt>op2</tt>.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4126
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]4127<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohman9a2a0932011-12-06 03:18:47 +0000[diff] [blame]4128 <tt>lshr</tt> is a <a href="#poisonvalues">poison value</a> if any of the bits
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]4129 shifted out are non-zero.</p>
4130
4131
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4132<h5>Example:</h5>
4133<pre>
4134 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
4135 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
4136 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
4137 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]4138 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Mon P Wang4dd832d2008-12-09 05:46:39 +0000[diff] [blame]4139 &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 Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4140</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4141
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4142</div>
4143
Reid Spencer2ab01932007-02-02 13:57:07 +0000[diff] [blame]4144<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4145<h4>
4146 <a name="i_ashr">'<tt>ashr</tt>' Instruction</a>
4147</h4>
4148
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4149<div>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4150
4151<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4152<pre>
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]4153 &lt;result&gt; = ashr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4154 &lt;result&gt; = ashr exact &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4155</pre>
4156
4157<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4158<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
4159 operand shifted to the right a specified number of bits with sign
4160 extension.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4161
4162<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4163<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4164 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4165 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4166
4167<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4168<p>This instruction always performs an arithmetic shift right operation, The
4169 most significant bits of the result will be filled with the sign bit
4170 of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
4171 larger than the number of bits in <tt>op1</tt>, the result is undefined. If
4172 the arguments are vectors, each vector element of <tt>op1</tt> is shifted by
4173 the corresponding shift amount in <tt>op2</tt>.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4174
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]4175<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohman9a2a0932011-12-06 03:18:47 +0000[diff] [blame]4176 <tt>ashr</tt> is a <a href="#poisonvalues">poison value</a> if any of the bits
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]4177 shifted out are non-zero.</p>
4178
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4179<h5>Example:</h5>
4180<pre>
4181 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
4182 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
4183 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
4184 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]4185 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Mon P Wang4dd832d2008-12-09 05:46:39 +0000[diff] [blame]4186 &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 Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4187</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4188
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4189</div>
4190
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4191<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4192<h4>
4193 <a name="i_and">'<tt>and</tt>' Instruction</a>
4194</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4195
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4196<div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4197
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4198<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4199<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +0000[diff] [blame]4200 &lt;result&gt; = and &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4201</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4202
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4203<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4204<p>The '<tt>and</tt>' instruction returns the bitwise logical and of its two
4205 operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4206
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4207<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4208<p>The two arguments to the '<tt>and</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4209 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4210 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4211
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4212<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4213<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4214
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4215<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]4216 <tbody>
4217 <tr>
Bill Wendling4517fe52011-12-09 22:41:40 +0000[diff] [blame]4218 <th>In0</th>
4219 <th>In1</th>
4220 <th>Out</th>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]4221 </tr>
4222 <tr>
4223 <td>0</td>
4224 <td>0</td>
4225 <td>0</td>
4226 </tr>
4227 <tr>
4228 <td>0</td>
4229 <td>1</td>
4230 <td>0</td>
4231 </tr>
4232 <tr>
4233 <td>1</td>
4234 <td>0</td>
4235 <td>0</td>
4236 </tr>
4237 <tr>
4238 <td>1</td>
4239 <td>1</td>
4240 <td>1</td>
4241 </tr>
4242 </tbody>
4243</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4244
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4245<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4246<pre>
4247 &lt;result&gt; = and i32 4, %var <i>; yields {i32}:result = 4 &amp; %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4248 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
4249 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4250</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4251</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4252<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4253<h4>
4254 <a name="i_or">'<tt>or</tt>' Instruction</a>
4255</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4256
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4257<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4258
4259<h5>Syntax:</h5>
4260<pre>
4261 &lt;result&gt; = or &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4262</pre>
4263
4264<h5>Overview:</h5>
4265<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive or of its
4266 two operands.</p>
4267
4268<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4269<p>The two arguments to the '<tt>or</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4270 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4271 values. Both arguments must have identical types.</p>
4272
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4273<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4274<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4275
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]4276<table border="1" cellspacing="0" cellpadding="4">
4277 <tbody>
4278 <tr>
Bill Wendling4517fe52011-12-09 22:41:40 +0000[diff] [blame]4279 <th>In0</th>
4280 <th>In1</th>
4281 <th>Out</th>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]4282 </tr>
4283 <tr>
4284 <td>0</td>
4285 <td>0</td>
4286 <td>0</td>
4287 </tr>
4288 <tr>
4289 <td>0</td>
4290 <td>1</td>
4291 <td>1</td>
4292 </tr>
4293 <tr>
4294 <td>1</td>
4295 <td>0</td>
4296 <td>1</td>
4297 </tr>
4298 <tr>
4299 <td>1</td>
4300 <td>1</td>
4301 <td>1</td>
4302 </tr>
4303 </tbody>
4304</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4305
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4306<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4307<pre>
4308 &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4309 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
4310 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4311</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4312
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4313</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4314
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4315<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4316<h4>
4317 <a name="i_xor">'<tt>xor</tt>' Instruction</a>
4318</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4319
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4320<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4321
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4322<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4323<pre>
4324 &lt;result&gt; = xor &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4325</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4326
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4327<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4328<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive or of
4329 its two operands. The <tt>xor</tt> is used to implement the "one's
4330 complement" operation, which is the "~" operator in C.</p>
4331
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4332<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4333<p>The two arguments to the '<tt>xor</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4334 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4335 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4336
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4337<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4338<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4339
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]4340<table border="1" cellspacing="0" cellpadding="4">
4341 <tbody>
4342 <tr>
Bill Wendling4517fe52011-12-09 22:41:40 +0000[diff] [blame]4343 <th>In0</th>
4344 <th>In1</th>
4345 <th>Out</th>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]4346 </tr>
4347 <tr>
4348 <td>0</td>
4349 <td>0</td>
4350 <td>0</td>
4351 </tr>
4352 <tr>
4353 <td>0</td>
4354 <td>1</td>
4355 <td>1</td>
4356 </tr>
4357 <tr>
4358 <td>1</td>
4359 <td>0</td>
4360 <td>1</td>
4361 </tr>
4362 <tr>
4363 <td>1</td>
4364 <td>1</td>
4365 <td>0</td>
4366 </tr>
4367 </tbody>
4368</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4369
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4370<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4371<pre>
4372 &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4373 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
4374 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
4375 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4376</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4377
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4378</div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4379
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4380</div>
4381
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4382<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4383<h3>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4384 <a name="vectorops">Vector Operations</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4385</h3>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4386
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4387<div>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4388
4389<p>LLVM supports several instructions to represent vector operations in a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4390 target-independent manner. These instructions cover the element-access and
4391 vector-specific operations needed to process vectors effectively. While LLVM
4392 does directly support these vector operations, many sophisticated algorithms
4393 will want to use target-specific intrinsics to take full advantage of a
4394 specific target.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4395
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4396<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4397<h4>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4398 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4399</h4>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4400
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4401<div>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4402
4403<h5>Syntax:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4404<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4405 &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 Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4406</pre>
4407
4408<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4409<p>The '<tt>extractelement</tt>' instruction extracts a single scalar element
4410 from a vector at a specified index.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4411
4412
4413<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4414<p>The first operand of an '<tt>extractelement</tt>' instruction is a value
4415 of <a href="#t_vector">vector</a> type. The second operand is an index
4416 indicating the position from which to extract the element. The index may be
4417 a variable.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4418
4419<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4420<p>The result is a scalar of the same type as the element type of
4421 <tt>val</tt>. Its value is the value at position <tt>idx</tt> of
4422 <tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4423 results are undefined.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4424
4425<h5>Example:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4426<pre>
Gabor Greif03ab4dc2009-10-28 13:14:50 +0000[diff] [blame]4427 &lt;result&gt; = extractelement &lt;4 x i32&gt; %vec, i32 0 <i>; yields i32</i>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4428</pre>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4429
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4430</div>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4431
4432<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4433<h4>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4434 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4435</h4>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4436
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4437<div>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4438
4439<h5>Syntax:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4440<pre>
Dan Gohman43ba0672008-05-12 23:38:42 +0000[diff] [blame]4441 &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 Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4442</pre>
4443
4444<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4445<p>The '<tt>insertelement</tt>' instruction inserts a scalar element into a
4446 vector at a specified index.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4447
4448<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4449<p>The first operand of an '<tt>insertelement</tt>' instruction is a value
4450 of <a href="#t_vector">vector</a> type. The second operand is a scalar value
4451 whose type must equal the element type of the first operand. The third
4452 operand is an index indicating the position at which to insert the value.
4453 The index may be a variable.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4454
4455<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4456<p>The result is a vector of the same type as <tt>val</tt>. Its element values
4457 are those of <tt>val</tt> except at position <tt>idx</tt>, where it gets the
4458 value <tt>elt</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4459 results are undefined.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4460
4461<h5>Example:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4462<pre>
Gabor Greif03ab4dc2009-10-28 13:14:50 +0000[diff] [blame]4463 &lt;result&gt; = insertelement &lt;4 x i32&gt; %vec, i32 1, i32 0 <i>; yields &lt;4 x i32&gt;</i>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4464</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4465
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4466</div>
4467
4468<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4469<h4>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4470 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4471</h4>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4472
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4473<div>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4474
4475<h5>Syntax:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4476<pre>
Mon P Wang25f01062008-11-10 04:46:22 +0000[diff] [blame]4477 &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 Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4478</pre>
4479
4480<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4481<p>The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
4482 from two input vectors, returning a vector with the same element type as the
4483 input and length that is the same as the shuffle mask.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4484
4485<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4486<p>The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
4487 with types that match each other. The third argument is a shuffle mask whose
4488 element type is always 'i32'. The result of the instruction is a vector
4489 whose length is the same as the shuffle mask and whose element type is the
4490 same as the element type of the first two operands.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4491
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4492<p>The shuffle mask operand is required to be a constant vector with either
4493 constant integer or undef values.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4494
4495<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4496<p>The elements of the two input vectors are numbered from left to right across
4497 both of the vectors. The shuffle mask operand specifies, for each element of
4498 the result vector, which element of the two input vectors the result element
4499 gets. The element selector may be undef (meaning "don't care") and the
4500 second operand may be undef if performing a shuffle from only one vector.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4501
4502<h5>Example:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4503<pre>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4504 &lt;result&gt; = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]4505 &lt;4 x i32&gt; &lt;i32 0, i32 4, i32 1, i32 5&gt; <i>; yields &lt;4 x i32&gt;</i>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4506 &lt;result&gt; = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; undef,
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4507 &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 Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4508 &lt;result&gt; = shufflevector &lt;8 x i32&gt; %v1, &lt;8 x i32&gt; undef,
Mon P Wang25f01062008-11-10 04:46:22 +0000[diff] [blame]4509 &lt;4 x i32&gt; &lt;i32 0, i32 1, i32 2, i32 3&gt; <i>; yields &lt;4 x i32&gt;</i>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4510 &lt;result&gt; = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
Mon P Wang25f01062008-11-10 04:46:22 +0000[diff] [blame]4511 &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 Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4512</pre>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4513
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4514</div>
Tanya Lattnerb138bbe2006-04-14 19:24:33 +0000[diff] [blame]4515
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4516</div>
4517
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4518<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4519<h3>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4520 <a name="aggregateops">Aggregate Operations</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4521</h3>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4522
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4523<div>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4524
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4525<p>LLVM supports several instructions for working with
4526 <a href="#t_aggregate">aggregate</a> values.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4527
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4528<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4529<h4>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4530 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4531</h4>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4532
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4533<div>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4534
4535<h5>Syntax:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4536<pre>
4537 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
4538</pre>
4539
4540<h5>Overview:</h5>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4541<p>The '<tt>extractvalue</tt>' instruction extracts the value of a member field
4542 from an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4543
4544<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4545<p>The first operand of an '<tt>extractvalue</tt>' instruction is a value
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]4546 of <a href="#t_struct">struct</a> or
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4547 <a href="#t_array">array</a> type. The operands are constant indices to
4548 specify which value to extract in a similar manner as indices in a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4549 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
Frits van Bommel7cf63ac2010-12-05 20:54:38 +0000[diff] [blame]4550 <p>The major differences to <tt>getelementptr</tt> indexing are:</p>
4551 <ul>
4552 <li>Since the value being indexed is not a pointer, the first index is
4553 omitted and assumed to be zero.</li>
4554 <li>At least one index must be specified.</li>
4555 <li>Not only struct indices but also array indices must be in
4556 bounds.</li>
4557 </ul>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4558
4559<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4560<p>The result is the value at the position in the aggregate specified by the
4561 index operands.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4562
4563<h5>Example:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4564<pre>
Gabor Greif03ab4dc2009-10-28 13:14:50 +0000[diff] [blame]4565 &lt;result&gt; = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4566</pre>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4567
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4568</div>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4569
4570<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4571<h4>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4572 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4573</h4>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4574
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4575<div>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4576
4577<h5>Syntax:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4578<pre>
Bill Wendlingf6a91cf2011-07-26 20:42:28 +0000[diff] [blame]4579 &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 Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4580</pre>
4581
4582<h5>Overview:</h5>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4583<p>The '<tt>insertvalue</tt>' instruction inserts a value into a member field
4584 in an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4585
4586<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4587<p>The first operand of an '<tt>insertvalue</tt>' instruction is a value
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]4588 of <a href="#t_struct">struct</a> or
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4589 <a href="#t_array">array</a> type. The second operand is a first-class
4590 value to insert. The following operands are constant indices indicating
4591 the position at which to insert the value in a similar manner as indices in a
Frits van Bommel7cf63ac2010-12-05 20:54:38 +0000[diff] [blame]4592 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' instruction. The
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4593 value to insert must have the same type as the value identified by the
4594 indices.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4595
4596<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4597<p>The result is an aggregate of the same type as <tt>val</tt>. Its value is
4598 that of <tt>val</tt> except that the value at the position specified by the
4599 indices is that of <tt>elt</tt>.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4600
4601<h5>Example:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4602<pre>
Chris Lattnerc2e85402011-05-22 07:18:08 +0000[diff] [blame]4603 %agg1 = insertvalue {i32, float} undef, i32 1, 0 <i>; yields {i32 1, float undef}</i>
4604 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 <i>; yields {i32 1, float %val}</i>
4605 %agg3 = insertvalue {i32, {float}} %agg1, float %val, 1, 0 <i>; yields {i32 1, float %val}</i>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4606</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4607
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4608</div>
4609
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4610</div>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4611
4612<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4613<h3>
Chris Lattner6ab66722006-08-15 00:45:58 +0000[diff] [blame]4614 <a name="memoryops">Memory Access and Addressing Operations</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4615</h3>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4616
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4617<div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4618
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4619<p>A key design point of an SSA-based representation is how it represents
4620 memory. In LLVM, no memory locations are in SSA form, which makes things
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]4621 very simple. This section describes how to read, write, and allocate
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4622 memory in LLVM.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4623
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4624<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4625<h4>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4626 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4627</h4>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4628
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4629<div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4630
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4631<h5>Syntax:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4632<pre>
Dan Gohman2140a742010-05-28 01:14:11 +0000[diff] [blame]4633 &lt;result&gt; = alloca &lt;type&gt;[, &lt;ty&gt; &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4634</pre>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4635
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4636<h5>Overview:</h5>
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]4637<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4638 currently executing function, to be automatically released when this function
4639 returns to its caller. The object is always allocated in the generic address
4640 space (address space zero).</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4641
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4642<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4643<p>The '<tt>alloca</tt>' instruction
4644 allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt> bytes of memory on the
4645 runtime stack, returning a pointer of the appropriate type to the program.
4646 If "NumElements" is specified, it is the number of elements allocated,
4647 otherwise "NumElements" is defaulted to be one. If a constant alignment is
4648 specified, the value result of the allocation is guaranteed to be aligned to
4649 at least that boundary. If not specified, or if zero, the target can choose
4650 to align the allocation on any convenient boundary compatible with the
4651 type.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4652
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4653<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4654
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4655<h5>Semantics:</h5>
Bill Wendling9ee6a312009-05-08 20:49:29 +0000[diff] [blame]4656<p>Memory is allocated; a pointer is returned. The operation is undefined if
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4657 there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
4658 memory is automatically released when the function returns. The
4659 '<tt>alloca</tt>' instruction is commonly used to represent automatic
4660 variables that must have an address available. When the function returns
4661 (either with the <tt><a href="#i_ret">ret</a></tt>
Bill Wendling3f6a3a22012-02-06 21:57:33 +0000[diff] [blame^]4662 or <tt><a href="#i_resume">resume</a></tt> instructions), the memory is
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4663 reclaimed. Allocating zero bytes is legal, but the result is undefined.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4664
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4665<h5>Example:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4666<pre>
Dan Gohman7a5acb52009-01-04 23:49:44 +0000[diff] [blame]4667 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
4668 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
4669 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
4670 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4671</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4672
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4673</div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4674
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4675<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4676<h4>
4677 <a name="i_load">'<tt>load</tt>' Instruction</a>
4678</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4679
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4680<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4681
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4682<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4683<pre>
Eli Friedman02e737b2011-08-12 22:50:01 +0000[diff] [blame]4684 &lt;result&gt; = load [volatile] &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;]
4685 &lt;result&gt; = load atomic [volatile] &lt;ty&gt;* &lt;pointer&gt; [singlethread] &lt;ordering&gt;, align &lt;alignment&gt;
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4686 !&lt;index&gt; = !{ i32 1 }
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4687</pre>
4688
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4689<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4690<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4691
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4692<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4693<p>The argument to the '<tt>load</tt>' instruction specifies the memory address
4694 from which to load. The pointer must point to
4695 a <a href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
4696 marked as <tt>volatile</tt>, then the optimizer is not allowed to modify the
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]4697 number or order of execution of this <tt>load</tt> with other <a
4698 href="#volatile">volatile operations</a>.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4699
Eli Friedman59b66882011-08-09 23:02:53 +0000[diff] [blame]4700<p>If the <code>load</code> is marked as <code>atomic</code>, it takes an extra
4701 <a href="#ordering">ordering</a> and optional <code>singlethread</code>
4702 argument. The <code>release</code> and <code>acq_rel</code> orderings are
4703 not valid on <code>load</code> instructions. Atomic loads produce <a
4704 href="#memorymodel">defined</a> results when they may see multiple atomic
4705 stores. The type of the pointee must be an integer type whose bit width
4706 is a power of two greater than or equal to eight and less than or equal
4707 to a target-specific size limit. <code>align</code> must be explicitly
4708 specified on atomic loads, and the load has undefined behavior if the
4709 alignment is not set to a value which is at least the size in bytes of
4710 the pointee. <code>!nontemporal</code> does not have any defined semantics
4711 for atomic loads.</p>
4712
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4713<p>The optional constant <tt>align</tt> argument specifies the alignment of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4714 operation (that is, the alignment of the memory address). A value of 0 or an
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4715 omitted <tt>align</tt> argument means that the operation has the preferential
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4716 alignment for the target. It is the responsibility of the code emitter to
4717 ensure that the alignment information is correct. Overestimating the
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4718 alignment results in undefined behavior. Underestimating the alignment may
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4719 produce less efficient code. An alignment of 1 is always safe.</p>
4720
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4721<p>The optional <tt>!nontemporal</tt> metadata must reference a single
4722 metatadata name &lt;index&gt; corresponding to a metadata node with
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]4723 one <tt>i32</tt> entry of value 1. The existence of
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4724 the <tt>!nontemporal</tt> metatadata on the instruction tells the optimizer
4725 and code generator that this load is not expected to be reused in the cache.
4726 The code generator may select special instructions to save cache bandwidth,
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]4727 such as the <tt>MOVNT</tt> instruction on x86.</p>
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4728
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4729<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4730<p>The location of memory pointed to is loaded. If the value being loaded is of
4731 scalar type then the number of bytes read does not exceed the minimum number
4732 of bytes needed to hold all bits of the type. For example, loading an
4733 <tt>i24</tt> reads at most three bytes. When loading a value of a type like
4734 <tt>i20</tt> with a size that is not an integral number of bytes, the result
4735 is undefined if the value was not originally written using a store of the
4736 same type.</p>
4737
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4738<h5>Examples:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4739<pre>
4740 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
4741 <a href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4742 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4743</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4744
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4745</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4746
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4747<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4748<h4>
4749 <a name="i_store">'<tt>store</tt>' Instruction</a>
4750</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4751
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4752<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4753
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4754<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4755<pre>
Bill Wendling4517fe52011-12-09 22:41:40 +0000[diff] [blame]4756 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>
4757 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 Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4758</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4759
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4760<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4761<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4762
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4763<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4764<p>There are two arguments to the '<tt>store</tt>' instruction: a value to store
4765 and an address at which to store it. The type of the
4766 '<tt>&lt;pointer&gt;</tt>' operand must be a pointer to
4767 the <a href="#t_firstclass">first class</a> type of the
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]4768 '<tt>&lt;value&gt;</tt>' operand. If the <tt>store</tt> is marked as
4769 <tt>volatile</tt>, then the optimizer is not allowed to modify the number or
4770 order of execution of this <tt>store</tt> with other <a
4771 href="#volatile">volatile operations</a>.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4772
Eli Friedman59b66882011-08-09 23:02:53 +0000[diff] [blame]4773<p>If the <code>store</code> is marked as <code>atomic</code>, it takes an extra
4774 <a href="#ordering">ordering</a> and optional <code>singlethread</code>
4775 argument. The <code>acquire</code> and <code>acq_rel</code> orderings aren't
4776 valid on <code>store</code> instructions. Atomic loads produce <a
4777 href="#memorymodel">defined</a> results when they may see multiple atomic
4778 stores. The type of the pointee must be an integer type whose bit width
4779 is a power of two greater than or equal to eight and less than or equal
4780 to a target-specific size limit. <code>align</code> must be explicitly
4781 specified on atomic stores, and the store has undefined behavior if the
4782 alignment is not set to a value which is at least the size in bytes of
4783 the pointee. <code>!nontemporal</code> does not have any defined semantics
4784 for atomic stores.</p>
4785
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4786<p>The optional constant "align" argument specifies the alignment of the
4787 operation (that is, the alignment of the memory address). A value of 0 or an
4788 omitted "align" argument means that the operation has the preferential
4789 alignment for the target. It is the responsibility of the code emitter to
4790 ensure that the alignment information is correct. Overestimating the
4791 alignment results in an undefined behavior. Underestimating the alignment may
4792 produce less efficient code. An alignment of 1 is always safe.</p>
4793
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4794<p>The optional !nontemporal metadata must reference a single metatadata
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]4795 name &lt;index&gt; corresponding to a metadata node with one i32 entry of
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]4796 value 1. The existence of the !nontemporal metatadata on the
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4797 instruction tells the optimizer and code generator that this load is
4798 not expected to be reused in the cache. The code generator may
4799 select special instructions to save cache bandwidth, such as the
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]4800 MOVNT instruction on x86.</p>
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4801
4802
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]4803<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4804<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>' at the
4805 location specified by the '<tt>&lt;pointer&gt;</tt>' operand. If
4806 '<tt>&lt;value&gt;</tt>' is of scalar type then the number of bytes written
4807 does not exceed the minimum number of bytes needed to hold all bits of the
4808 type. For example, storing an <tt>i24</tt> writes at most three bytes. When
4809 writing a value of a type like <tt>i20</tt> with a size that is not an
4810 integral number of bytes, it is unspecified what happens to the extra bits
4811 that do not belong to the type, but they will typically be overwritten.</p>
4812
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4813<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4814<pre>
4815 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling8830ffe2007-10-22 05:10:05 +0000[diff] [blame]4816 store i32 3, i32* %ptr <i>; yields {void}</i>
4817 %val = <a href="#i_load">load</a> i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4818</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4819
Reid Spencer443460a2006-11-09 21:15:49 +0000[diff] [blame]4820</div>
4821
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4822<!-- _______________________________________________________________________ -->
NAKAMURA Takumi0300d882011-08-12 06:17:17 +0000[diff] [blame]4823<h4>
4824<a name="i_fence">'<tt>fence</tt>' Instruction</a>
4825</h4>
Eli Friedmanfee02c62011-07-25 23:16:38 +0000[diff] [blame]4826
NAKAMURA Takumi0300d882011-08-12 06:17:17 +0000[diff] [blame]4827<div>
Eli Friedmanfee02c62011-07-25 23:16:38 +0000[diff] [blame]4828
4829<h5>Syntax:</h5>
4830<pre>
4831 fence [singlethread] &lt;ordering&gt; <i>; yields {void}</i>
4832</pre>
4833
4834<h5>Overview:</h5>
4835<p>The '<tt>fence</tt>' instruction is used to introduce happens-before edges
4836between operations.</p>
4837
4838<h5>Arguments:</h5> <p>'<code>fence</code>' instructions take an <a
4839href="#ordering">ordering</a> argument which defines what
4840<i>synchronizes-with</i> edges they add. They can only be given
4841<code>acquire</code>, <code>release</code>, <code>acq_rel</code>, and
4842<code>seq_cst</code> orderings.</p>
4843
4844<h5>Semantics:</h5>
4845<p>A fence <var>A</var> which has (at least) <code>release</code> ordering
4846semantics <i>synchronizes with</i> a fence <var>B</var> with (at least)
4847<code>acquire</code> ordering semantics if and only if there exist atomic
4848operations <var>X</var> and <var>Y</var>, both operating on some atomic object
4849<var>M</var>, such that <var>A</var> is sequenced before <var>X</var>,
4850<var>X</var> modifies <var>M</var> (either directly or through some side effect
4851of a sequence headed by <var>X</var>), <var>Y</var> is sequenced before
4852<var>B</var>, and <var>Y</var> observes <var>M</var>. This provides a
4853<i>happens-before</i> dependency between <var>A</var> and <var>B</var>. Rather
4854than an explicit <code>fence</code>, one (but not both) of the atomic operations
4855<var>X</var> or <var>Y</var> might provide a <code>release</code> or
4856<code>acquire</code> (resp.) ordering constraint and still
4857<i>synchronize-with</i> the explicit <code>fence</code> and establish the
4858<i>happens-before</i> edge.</p>
4859
4860<p>A <code>fence</code> which has <code>seq_cst</code> ordering, in addition to
4861having both <code>acquire</code> and <code>release</code> semantics specified
4862above, participates in the global program order of other <code>seq_cst</code>
4863operations and/or fences.</p>
4864
4865<p>The optional "<a href="#singlethread"><code>singlethread</code></a>" argument
4866specifies that the fence only synchronizes with other fences in the same
4867thread. (This is useful for interacting with signal handlers.)</p>
4868
Eli Friedmanfee02c62011-07-25 23:16:38 +0000[diff] [blame]4869<h5>Example:</h5>
4870<pre>
4871 fence acquire <i>; yields {void}</i>
4872 fence singlethread seq_cst <i>; yields {void}</i>
4873</pre>
4874
4875</div>
4876
4877<!-- _______________________________________________________________________ -->
NAKAMURA Takumi0300d882011-08-12 06:17:17 +0000[diff] [blame]4878<h4>
4879<a name="i_cmpxchg">'<tt>cmpxchg</tt>' Instruction</a>
4880</h4>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]4881
NAKAMURA Takumi0300d882011-08-12 06:17:17 +0000[diff] [blame]4882<div>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]4883
4884<h5>Syntax:</h5>
4885<pre>
Bill Wendling4517fe52011-12-09 22:41:40 +0000[diff] [blame]4886 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 Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]4887</pre>
4888
4889<h5>Overview:</h5>
4890<p>The '<tt>cmpxchg</tt>' instruction is used to atomically modify memory.
4891It loads a value in memory and compares it to a given value. If they are
4892equal, it stores a new value into the memory.</p>
4893
4894<h5>Arguments:</h5>
4895<p>There are three arguments to the '<code>cmpxchg</code>' instruction: an
4896address to operate on, a value to compare to the value currently be at that
4897address, and a new value to place at that address if the compared values are
4898equal. The type of '<var>&lt;cmp&gt;</var>' must be an integer type whose
4899bit width is a power of two greater than or equal to eight and less than
4900or equal to a target-specific size limit. '<var>&lt;cmp&gt;</var>' and
4901'<var>&lt;new&gt;</var>' must have the same type, and the type of
4902'<var>&lt;pointer&gt;</var>' must be a pointer to that type. If the
4903<code>cmpxchg</code> is marked as <code>volatile</code>, then the
4904optimizer is not allowed to modify the number or order of execution
4905of this <code>cmpxchg</code> with other <a href="#volatile">volatile
4906operations</a>.</p>
4907
4908<!-- FIXME: Extend allowed types. -->
4909
4910<p>The <a href="#ordering"><var>ordering</var></a> argument specifies how this
4911<code>cmpxchg</code> synchronizes with other atomic operations.</p>
4912
4913<p>The optional "<code>singlethread</code>" argument declares that the
4914<code>cmpxchg</code> is only atomic with respect to code (usually signal
4915handlers) running in the same thread as the <code>cmpxchg</code>. Otherwise the
4916cmpxchg is atomic with respect to all other code in the system.</p>
4917
4918<p>The pointer passed into cmpxchg must have alignment greater than or equal to
4919the size in memory of the operand.
4920
4921<h5>Semantics:</h5>
4922<p>The contents of memory at the location specified by the
4923'<tt>&lt;pointer&gt;</tt>' operand is read and compared to
4924'<tt>&lt;cmp&gt;</tt>'; if the read value is the equal,
4925'<tt>&lt;new&gt;</tt>' is written. The original value at the location
4926is returned.
4927
4928<p>A successful <code>cmpxchg</code> is a read-modify-write instruction for the
4929purpose of identifying <a href="#release_sequence">release sequences</a>. A
4930failed <code>cmpxchg</code> is equivalent to an atomic load with an ordering
4931parameter determined by dropping any <code>release</code> part of the
4932<code>cmpxchg</code>'s ordering.</p>
4933
4934<!--
4935FIXME: Is compare_exchange_weak() necessary? (Consider after we've done
4936optimization work on ARM.)
4937
4938FIXME: Is a weaker ordering constraint on failure helpful in practice?
4939-->
4940
4941<h5>Example:</h5>
4942<pre>
4943entry:
Bill Wendling4517fe52011-12-09 22:41:40 +0000[diff] [blame]4944 %orig = atomic <a href="#i_load">load</a> i32* %ptr unordered <i>; yields {i32}</i>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]4945 <a href="#i_br">br</a> label %loop
4946
4947loop:
4948 %cmp = <a href="#i_phi">phi</a> i32 [ %orig, %entry ], [%old, %loop]
4949 %squared = <a href="#i_mul">mul</a> i32 %cmp, %cmp
Bill Wendling4517fe52011-12-09 22:41:40 +0000[diff] [blame]4950 %old = cmpxchg i32* %ptr, i32 %cmp, i32 %squared <i>; yields {i32}</i>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]4951 %success = <a href="#i_icmp">icmp</a> eq i32 %cmp, %old
4952 <a href="#i_br">br</a> i1 %success, label %done, label %loop
4953
4954done:
4955 ...
4956</pre>
4957
4958</div>
4959
4960<!-- _______________________________________________________________________ -->
NAKAMURA Takumi0300d882011-08-12 06:17:17 +0000[diff] [blame]4961<h4>
4962<a name="i_atomicrmw">'<tt>atomicrmw</tt>' Instruction</a>
4963</h4>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]4964
NAKAMURA Takumi0300d882011-08-12 06:17:17 +0000[diff] [blame]4965<div>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]4966
4967<h5>Syntax:</h5>
4968<pre>
Eli Friedman02e737b2011-08-12 22:50:01 +0000[diff] [blame]4969 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 Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]4970</pre>
4971
4972<h5>Overview:</h5>
4973<p>The '<tt>atomicrmw</tt>' instruction is used to atomically modify memory.</p>
4974
4975<h5>Arguments:</h5>
4976<p>There are three arguments to the '<code>atomicrmw</code>' instruction: an
4977operation to apply, an address whose value to modify, an argument to the
4978operation. The operation must be one of the following keywords:</p>
4979<ul>
4980 <li>xchg</li>
4981 <li>add</li>
4982 <li>sub</li>
4983 <li>and</li>
4984 <li>nand</li>
4985 <li>or</li>
4986 <li>xor</li>
4987 <li>max</li>
4988 <li>min</li>
4989 <li>umax</li>
4990 <li>umin</li>
4991</ul>
4992
4993<p>The type of '<var>&lt;value&gt;</var>' must be an integer type whose
4994bit width is a power of two greater than or equal to eight and less than
4995or equal to a target-specific size limit. The type of the
4996'<code>&lt;pointer&gt;</code>' operand must be a pointer to that type.
4997If the <code>atomicrmw</code> is marked as <code>volatile</code>, then the
4998optimizer is not allowed to modify the number or order of execution of this
4999<code>atomicrmw</code> with other <a href="#volatile">volatile
5000 operations</a>.</p>
5001
5002<!-- FIXME: Extend allowed types. -->
5003
5004<h5>Semantics:</h5>
5005<p>The contents of memory at the location specified by the
5006'<tt>&lt;pointer&gt;</tt>' operand are atomically read, modified, and written
5007back. The original value at the location is returned. The modification is
5008specified by the <var>operation</var> argument:</p>
5009
5010<ul>
5011 <li>xchg: <code>*ptr = val</code></li>
5012 <li>add: <code>*ptr = *ptr + val</code></li>
5013 <li>sub: <code>*ptr = *ptr - val</code></li>
5014 <li>and: <code>*ptr = *ptr &amp; val</code></li>
5015 <li>nand: <code>*ptr = ~(*ptr &amp; val)</code></li>
5016 <li>or: <code>*ptr = *ptr | val</code></li>
5017 <li>xor: <code>*ptr = *ptr ^ val</code></li>
5018 <li>max: <code>*ptr = *ptr &gt; val ? *ptr : val</code> (using a signed comparison)</li>
5019 <li>min: <code>*ptr = *ptr &lt; val ? *ptr : val</code> (using a signed comparison)</li>
5020 <li>umax: <code>*ptr = *ptr &gt; val ? *ptr : val</code> (using an unsigned comparison)</li>
5021 <li>umin: <code>*ptr = *ptr &lt; val ? *ptr : val</code> (using an unsigned comparison)</li>
5022</ul>
5023
5024<h5>Example:</h5>
5025<pre>
5026 %old = atomicrmw add i32* %ptr, i32 1 acquire <i>; yields {i32}</i>
5027</pre>
5028
5029</div>
5030
5031<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5032<h4>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]5033 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5034</h4>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]5035
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5036<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5037
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]5038<h5>Syntax:</h5>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]5039<pre>
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]5040 &lt;result&gt; = getelementptr &lt;pty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
Dan Gohman1639c392009-07-27 21:53:46 +0000[diff] [blame]5041 &lt;result&gt; = getelementptr inbounds &lt;pty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
Nadav Rotem3924cb02011-12-05 06:29:09 +0000[diff] [blame]5042 &lt;result&gt; = getelementptr &lt;ptr vector&gt; ptrval, &lt;vector index type&gt; idx
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]5043</pre>
5044
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]5045<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5046<p>The '<tt>getelementptr</tt>' instruction is used to get the address of a
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]5047 subelement of an <a href="#t_aggregate">aggregate</a> data structure.
5048 It performs address calculation only and does not access memory.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]5049
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]5050<h5>Arguments:</h5>
Nadav Rotem3924cb02011-12-05 06:29:09 +0000[diff] [blame]5051<p>The first argument is always a pointer or a vector of pointers,
5052 and forms the basis of the
Chris Lattnera40b9122009-07-29 06:44:13 +0000[diff] [blame]5053 calculation. The remaining arguments are indices that indicate which of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5054 elements of the aggregate object are indexed. The interpretation of each
5055 index is dependent on the type being indexed into. The first index always
5056 indexes the pointer value given as the first argument, the second index
5057 indexes a value of the type pointed to (not necessarily the value directly
5058 pointed to, since the first index can be non-zero), etc. The first type
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]5059 indexed into must be a pointer value, subsequent types can be arrays,
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]5060 vectors, and structs. Note that subsequent types being indexed into
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]5061 can never be pointers, since that would require loading the pointer before
5062 continuing calculation.</p>
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]5063
5064<p>The type of each index argument depends on the type it is indexing into.
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]5065 When indexing into a (optionally packed) structure, only <tt>i32</tt>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]5066 integer <b>constants</b> are allowed. When indexing into an array, pointer
5067 or vector, integers of any width are allowed, and they are not required to be
Eli Friedmand8874dc2011-08-12 23:37:55 +0000[diff] [blame]5068 constant. These integers are treated as signed values where relevant.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]5069
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5070<p>For example, let's consider a C code fragment and how it gets compiled to
5071 LLVM:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]5072
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]5073<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]5074struct RT {
5075 char A;
Chris Lattnera446f1b2007-05-29 15:43:56 +0000[diff] [blame]5076 int B[10][20];
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]5077 char C;
5078};
5079struct ST {
Chris Lattnera446f1b2007-05-29 15:43:56 +0000[diff] [blame]5080 int X;
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]5081 double Y;
5082 struct RT Z;
5083};
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]5084
Chris Lattnera446f1b2007-05-29 15:43:56 +0000[diff] [blame]5085int *foo(struct ST *s) {
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]5086 return &amp;s[1].Z.B[5][13];
5087}
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]5088</pre>
5089
Bill Wendling7ad1f362011-12-13 01:07:07 +0000[diff] [blame]5090<p>The LLVM code generated by Clang is:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]5091
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]5092<pre class="doc_code">
Bill Wendling7ad1f362011-12-13 01:07:07 +0000[diff] [blame]5093%struct.RT = <a href="#namedtypes">type</a> { i8, [10 x [20 x i32]], i8 }
5094%struct.ST = <a href="#namedtypes">type</a> { i32, double, %struct.RT }
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]5095
Bill Wendling7ad1f362011-12-13 01:07:07 +0000[diff] [blame]5096define i32* @foo(%struct.ST* %s) nounwind uwtable readnone optsize ssp {
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]5097entry:
Bill Wendling7ad1f362011-12-13 01:07:07 +0000[diff] [blame]5098 %arrayidx = getelementptr inbounds %struct.ST* %s, i64 1, i32 2, i32 1, i64 5, i64 13
5099 ret i32* %arrayidx
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]5100}
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]5101</pre>
5102
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]5103<h5>Semantics:</h5>
Bill Wendling7ad1f362011-12-13 01:07:07 +0000[diff] [blame]5104<p>In the example above, the first index is indexing into the
5105 '<tt>%struct.ST*</tt>' type, which is a pointer, yielding a
5106 '<tt>%struct.ST</tt>' = '<tt>{ i32, double, %struct.RT }</tt>' type, a
5107 structure. The second index indexes into the third element of the structure,
5108 yielding a '<tt>%struct.RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]], i8 }</tt>'
5109 type, another structure. The third index indexes into the second element of
5110 the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an array. The
5111 two dimensions of the array are subscripted into, yielding an '<tt>i32</tt>'
5112 type. The '<tt>getelementptr</tt>' instruction returns a pointer to this
5113 element, thus computing a value of '<tt>i32*</tt>' type.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]5114
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5115<p>Note that it is perfectly legal to index partially through a structure,
5116 returning a pointer to an inner element. Because of this, the LLVM code for
5117 the given testcase is equivalent to:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]5118
Bill Wendling7ad1f362011-12-13 01:07:07 +0000[diff] [blame]5119<pre class="doc_code">
5120define i32* @foo(%struct.ST* %s) {
5121 %t1 = getelementptr %struct.ST* %s, i32 1 <i>; yields %struct.ST*:%t1</i>
5122 %t2 = getelementptr %struct.ST* %t1, i32 0, i32 2 <i>; yields %struct.RT*:%t2</i>
5123 %t3 = getelementptr %struct.RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
5124 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
5125 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
5126 ret i32* %t5
5127}
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]5128</pre>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +0000[diff] [blame]5129
Dan Gohman1639c392009-07-27 21:53:46 +0000[diff] [blame]5130<p>If the <tt>inbounds</tt> keyword is present, the result value of the
Dan Gohman9a2a0932011-12-06 03:18:47 +0000[diff] [blame]5131 <tt>getelementptr</tt> is a <a href="#poisonvalues">poison value</a> if the
Dan Gohman57255802010-04-23 15:23:32 +0000[diff] [blame]5132 base pointer is not an <i>in bounds</i> address of an allocated object,
5133 or if any of the addresses that would be formed by successive addition of
5134 the offsets implied by the indices to the base address with infinitely
Eli Friedmand8874dc2011-08-12 23:37:55 +0000[diff] [blame]5135 precise signed arithmetic are not an <i>in bounds</i> address of that
5136 allocated object. The <i>in bounds</i> addresses for an allocated object
5137 are all the addresses that point into the object, plus the address one
Nadav Rotem3924cb02011-12-05 06:29:09 +0000[diff] [blame]5138 byte past the end.
5139 In cases where the base is a vector of pointers the <tt>inbounds</tt> keyword
5140 applies to each of the computations element-wise. </p>
Dan Gohman1639c392009-07-27 21:53:46 +0000[diff] [blame]5141
5142<p>If the <tt>inbounds</tt> keyword is not present, the offsets are added to
Eli Friedmand8874dc2011-08-12 23:37:55 +0000[diff] [blame]5143 the base address with silently-wrapping two's complement arithmetic. If the
5144 offsets have a different width from the pointer, they are sign-extended or
5145 truncated to the width of the pointer. The result value of the
5146 <tt>getelementptr</tt> may be outside the object pointed to by the base
5147 pointer. The result value may not necessarily be used to access memory
5148 though, even if it happens to point into allocated storage. See the
5149 <a href="#pointeraliasing">Pointer Aliasing Rules</a> section for more
5150 information.</p>
Dan Gohman1639c392009-07-27 21:53:46 +0000[diff] [blame]5151
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5152<p>The getelementptr instruction is often confusing. For some more insight into
5153 how it works, see <a href="GetElementPtr.html">the getelementptr FAQ</a>.</p>
Chris Lattner6ab66722006-08-15 00:45:58 +0000[diff] [blame]5154
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]5155<h5>Example:</h5>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]5156<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5157 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]5158 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
5159 <i>; yields i8*:vptr</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]5160 %vptr = getelementptr {i32, &lt;2 x i8&gt;}* %svptr, i64 0, i32 1, i32 1
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]5161 <i>; yields i8*:eptr</i>
5162 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Sanjiv Gupta0c155e62009-04-25 07:27:44 +0000[diff] [blame]5163 <i>; yields i32*:iptr</i>
Sanjiv Gupta77abea02009-04-24 16:38:13 +0000[diff] [blame]5164 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]5165</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5166
Nadav Rotem3924cb02011-12-05 06:29:09 +0000[diff] [blame]5167<p>In cases where the pointer argument is a vector of pointers, only a
5168 single index may be used, and the number of vector elements has to be
5169 the same. For example: </p>
5170<pre class="doc_code">
5171 %A = getelementptr <4 x i8*> %ptrs, <4 x i64> %offsets,
5172</pre>
5173
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]5174</div>
Reid Spencer443460a2006-11-09 21:15:49 +0000[diff] [blame]5175
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5176</div>
5177
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5178<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5179<h3>
5180 <a name="convertops">Conversion Operations</a>
5181</h3>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5182
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5183<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5184
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5185<p>The instructions in this category are the conversion instructions (casting)
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5186 which all take a single operand and a type. They perform various bit
5187 conversions on the operand.</p>
5188
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]5189<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5190<h4>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5191 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5192</h4>
5193
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5194<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5195
5196<h5>Syntax:</h5>
5197<pre>
5198 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5199</pre>
5200
5201<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5202<p>The '<tt>trunc</tt>' instruction truncates its operand to the
5203 type <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5204
5205<h5>Arguments:</h5>
Nadav Rotem502f1b92011-02-24 21:01:34 +0000[diff] [blame]5206<p>The '<tt>trunc</tt>' instruction takes a value to trunc, and a type to trunc it to.
5207 Both types must be of <a href="#t_integer">integer</a> types, or vectors
5208 of the same number of integers.
5209 The bit size of the <tt>value</tt> must be larger than
5210 the bit size of the destination type, <tt>ty2</tt>.
5211 Equal sized types are not allowed.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5212
5213<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5214<p>The '<tt>trunc</tt>' instruction truncates the high order bits
5215 in <tt>value</tt> and converts the remaining bits to <tt>ty2</tt>. Since the
5216 source size must be larger than the destination size, <tt>trunc</tt> cannot
5217 be a <i>no-op cast</i>. It will always truncate bits.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5218
5219<h5>Example:</h5>
5220<pre>
Nadav Rotem502f1b92011-02-24 21:01:34 +0000[diff] [blame]5221 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
5222 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
5223 %Z = trunc i32 122 to i1 <i>; yields i1:false</i>
5224 %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 Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5225</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5226
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5227</div>
5228
5229<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5230<h4>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5231 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5232</h4>
5233
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5234<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5235
5236<h5>Syntax:</h5>
5237<pre>
5238 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5239</pre>
5240
5241<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5242<p>The '<tt>zext</tt>' instruction zero extends its operand to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5243 <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5244
5245
5246<h5>Arguments:</h5>
Nadav Rotem25f2ac92011-02-20 12:37:50 +0000[diff] [blame]5247<p>The '<tt>zext</tt>' instruction takes a value to cast, and a type to cast it to.
5248 Both types must be of <a href="#t_integer">integer</a> types, or vectors
5249 of the same number of integers.
5250 The bit size of the <tt>value</tt> must be smaller than
5251 the bit size of the destination type,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5252 <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5253
5254<h5>Semantics:</h5>
5255<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5256 bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5257
Reid Spencer07c9c682007-01-12 15:46:11 +0000[diff] [blame]5258<p>When zero extending from i1, the result will always be either 0 or 1.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5259
5260<h5>Example:</h5>
5261<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5262 %X = zext i32 257 to i64 <i>; yields i64:257</i>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]5263 %Y = zext i1 true to i32 <i>; yields i32:1</i>
Nadav Rotem25f2ac92011-02-20 12:37:50 +0000[diff] [blame]5264 %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 Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5265</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5266
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5267</div>
5268
5269<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5270<h4>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5271 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5272</h4>
5273
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5274<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5275
5276<h5>Syntax:</h5>
5277<pre>
5278 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5279</pre>
5280
5281<h5>Overview:</h5>
5282<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
5283
5284<h5>Arguments:</h5>
Nadav Rotem502f1b92011-02-24 21:01:34 +0000[diff] [blame]5285<p>The '<tt>sext</tt>' instruction takes a value to cast, and a type to cast it to.
5286 Both types must be of <a href="#t_integer">integer</a> types, or vectors
5287 of the same number of integers.
5288 The bit size of the <tt>value</tt> must be smaller than
5289 the bit size of the destination type,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5290 <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5291
5292<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5293<p>The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
5294 bit (highest order bit) of the <tt>value</tt> until it reaches the bit size
5295 of the type <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5296
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]5297<p>When sign extending from i1, the extension always results in -1 or 0.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5298
5299<h5>Example:</h5>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5300<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5301 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]5302 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
Nadav Rotem502f1b92011-02-24 21:01:34 +0000[diff] [blame]5303 %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 Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5304</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5305
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5306</div>
5307
5308<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5309<h4>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]5310 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5311</h4>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]5312
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5313<div>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]5314
5315<h5>Syntax:</h5>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]5316<pre>
5317 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5318</pre>
5319
5320<h5>Overview:</h5>
5321<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5322 <tt>ty2</tt>.</p>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]5323
5324<h5>Arguments:</h5>
5325<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5326 point</a> value to cast and a <a href="#t_floating">floating point</a> type
5327 to cast it to. The size of <tt>value</tt> must be larger than the size of
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5328 <tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5329 <i>no-op cast</i>.</p>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]5330
5331<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5332<p>The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5333 <a href="#t_floating">floating point</a> type to a smaller
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5334 <a href="#t_floating">floating point</a> type. If the value cannot fit
5335 within the destination type, <tt>ty2</tt>, then the results are
5336 undefined.</p>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]5337
5338<h5>Example:</h5>
5339<pre>
5340 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
5341 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
5342</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5343
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]5344</div>
5345
5346<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5347<h4>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5348 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5349</h4>
5350
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5351<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5352
5353<h5>Syntax:</h5>
5354<pre>
5355 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5356</pre>
5357
5358<h5>Overview:</h5>
5359<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5360 floating point value.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5361
5362<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5363<p>The '<tt>fpext</tt>' instruction takes a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5364 <a href="#t_floating">floating point</a> <tt>value</tt> to cast, and
5365 a <a href="#t_floating">floating point</a> type to cast it to. The source
5366 type must be smaller than the destination type.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5367
5368<h5>Semantics:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]5369<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5370 <a href="#t_floating">floating point</a> type to a larger
5371 <a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
5372 used to make a <i>no-op cast</i> because it always changes bits. Use
5373 <tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5374
5375<h5>Example:</h5>
5376<pre>
Nick Lewycky9feca672011-03-31 18:20:19 +0000[diff] [blame]5377 %X = fpext float 3.125 to double <i>; yields double:3.125000e+00</i>
5378 %Y = fpext double %X to fp128 <i>; yields fp128:0xL00000000000000004000900000000000</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5379</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5380
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5381</div>
5382
5383<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5384<h4>
Reid Spencer2eadb532007-01-21 00:29:26 +0000[diff] [blame]5385 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5386</h4>
5387
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5388<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5389
5390<h5>Syntax:</h5>
5391<pre>
Reid Spencer753163d2007-07-31 14:40:14 +0000[diff] [blame]5392 &lt;result&gt; = fptoui &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5393</pre>
5394
5395<h5>Overview:</h5>
Reid Spencer753163d2007-07-31 14:40:14 +0000[diff] [blame]5396<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5397 unsigned integer equivalent of type <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5398
5399<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5400<p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
5401 scalar or vector <a href="#t_floating">floating point</a> value, and a type
5402 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
5403 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
5404 vector integer type with the same number of elements as <tt>ty</tt></p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5405
5406<h5>Semantics:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5407<p>The '<tt>fptoui</tt>' instruction converts its
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5408 <a href="#t_floating">floating point</a> operand into the nearest (rounding
5409 towards zero) unsigned integer value. If the value cannot fit
5410 in <tt>ty2</tt>, the results are undefined.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5411
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5412<h5>Example:</h5>
5413<pre>
Reid Spencer753163d2007-07-31 14:40:14 +0000[diff] [blame]5414 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner5b95a172007-09-22 03:17:52 +0000[diff] [blame]5415 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Gabor Greiff50fd572009-10-28 09:21:30 +0000[diff] [blame]5416 %Z = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5417</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5418
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5419</div>
5420
5421<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5422<h4>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]5423 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5424</h4>
5425
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5426<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5427
5428<h5>Syntax:</h5>
5429<pre>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]5430 &lt;result&gt; = fptosi &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5431</pre>
5432
5433<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5434<p>The '<tt>fptosi</tt>' instruction converts
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5435 <a href="#t_floating">floating point</a> <tt>value</tt> to
5436 type <tt>ty2</tt>.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5437
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]5438<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5439<p>The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
5440 scalar or vector <a href="#t_floating">floating point</a> value, and a type
5441 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
5442 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
5443 vector integer type with the same number of elements as <tt>ty</tt></p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5444
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]5445<h5>Semantics:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5446<p>The '<tt>fptosi</tt>' instruction converts its
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5447 <a href="#t_floating">floating point</a> operand into the nearest (rounding
5448 towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
5449 the results are undefined.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5450
Chris Lattner70de6632001-07-09 00:26:23 +0000[diff] [blame]5451<h5>Example:</h5>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5452<pre>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]5453 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner5b95a172007-09-22 03:17:52 +0000[diff] [blame]5454 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Gabor Greiff50fd572009-10-28 09:21:30 +0000[diff] [blame]5455 %Z = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5456</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5457
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5458</div>
5459
5460<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5461<h4>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]5462 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5463</h4>
5464
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5465<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5466
5467<h5>Syntax:</h5>
5468<pre>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]5469 &lt;result&gt; = uitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5470</pre>
5471
5472<h5>Overview:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]5473<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5474 integer and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5475
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5476<h5>Arguments:</h5>
Nate Begemand4d45c22007-11-17 03:58:34 +0000[diff] [blame]5477<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5478 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
5479 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
5480 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
5481 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5482
5483<h5>Semantics:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]5484<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5485 integer quantity and converts it to the corresponding floating point
5486 value. If the value cannot fit in the floating point value, the results are
5487 undefined.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5488
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5489<h5>Example:</h5>
5490<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5491 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]5492 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5493</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5494
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5495</div>
5496
5497<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5498<h4>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]5499 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5500</h4>
5501
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5502<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5503
5504<h5>Syntax:</h5>
5505<pre>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]5506 &lt;result&gt; = sitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5507</pre>
5508
5509<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5510<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed integer
5511 and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5512
5513<h5>Arguments:</h5>
Nate Begemand4d45c22007-11-17 03:58:34 +0000[diff] [blame]5514<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5515 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
5516 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
5517 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
5518 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5519
5520<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5521<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed integer
5522 quantity and converts it to the corresponding floating point value. If the
5523 value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5524
5525<h5>Example:</h5>
5526<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5527 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]5528 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5529</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5530
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5531</div>
5532
5533<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5534<h4>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5535 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5536</h4>
5537
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5538<div>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5539
5540<h5>Syntax:</h5>
5541<pre>
5542 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5543</pre>
5544
5545<h5>Overview:</h5>
Nadav Rotem3924cb02011-12-05 06:29:09 +0000[diff] [blame]5546<p>The '<tt>ptrtoint</tt>' instruction converts the pointer or a vector of
5547 pointers <tt>value</tt> to
5548 the integer (or vector of integers) type <tt>ty2</tt>.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5549
5550<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5551<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
Nadav Rotem3924cb02011-12-05 06:29:09 +0000[diff] [blame]5552 must be a a value of type <a href="#t_pointer">pointer</a> or a vector of
5553 pointers, and a type to cast it to
5554 <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> or a vector
5555 of integers type.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5556
5557<h5>Semantics:</h5>
5558<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5559 <tt>ty2</tt> by interpreting the pointer value as an integer and either
5560 truncating or zero extending that value to the size of the integer type. If
5561 <tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
5562 <tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
5563 are the same size, then nothing is done (<i>no-op cast</i>) other than a type
5564 change.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5565
5566<h5>Example:</h5>
5567<pre>
Nadav Rotem3924cb02011-12-05 06:29:09 +0000[diff] [blame]5568 %X = ptrtoint i32* %P to i8 <i>; yields truncation on 32-bit architecture</i>
5569 %Y = ptrtoint i32* %P to i64 <i>; yields zero extension on 32-bit architecture</i>
5570 %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 Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5571</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5572
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5573</div>
5574
5575<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5576<h4>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5577 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5578</h4>
5579
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5580<div>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5581
5582<h5>Syntax:</h5>
5583<pre>
5584 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5585</pre>
5586
5587<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5588<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to a
5589 pointer type, <tt>ty2</tt>.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5590
5591<h5>Arguments:</h5>
Duncan Sands16f122e2007-03-30 12:22:09 +0000[diff] [blame]5592<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5593 value to cast, and a type to cast it to, which must be a
5594 <a href="#t_pointer">pointer</a> type.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5595
5596<h5>Semantics:</h5>
5597<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5598 <tt>ty2</tt> by applying either a zero extension or a truncation depending on
5599 the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
5600 size of a pointer then a truncation is done. If <tt>value</tt> is smaller
5601 than the size of a pointer then a zero extension is done. If they are the
5602 same size, nothing is done (<i>no-op cast</i>).</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5603
5604<h5>Example:</h5>
5605<pre>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5606 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
Gabor Greiff50fd572009-10-28 09:21:30 +0000[diff] [blame]5607 %Y = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
5608 %Z = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
Nadav Rotem3924cb02011-12-05 06:29:09 +0000[diff] [blame]5609 %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 Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5610</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5611
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5612</div>
5613
5614<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5615<h4>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]5616 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5617</h4>
5618
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5619<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5620
5621<h5>Syntax:</h5>
5622<pre>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]5623 &lt;result&gt; = bitcast &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5624</pre>
5625
5626<h5>Overview:</h5>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]5627<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5628 <tt>ty2</tt> without changing any bits.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5629
5630<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5631<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be a
5632 non-aggregate first class value, and a type to cast it to, which must also be
5633 a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes
5634 of <tt>value</tt> and the destination type, <tt>ty2</tt>, must be
5635 identical. If the source type is a pointer, the destination type must also be
5636 a pointer. This instruction supports bitwise conversion of vectors to
5637 integers and to vectors of other types (as long as they have the same
5638 size).</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5639
5640<h5>Semantics:</h5>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]5641<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5642 <tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
5643 this conversion. The conversion is done as if the <tt>value</tt> had been
Nadav Rotem3924cb02011-12-05 06:29:09 +0000[diff] [blame]5644 stored to memory and read back as type <tt>ty2</tt>.
5645 Pointer (or vector of pointers) types may only be converted to other pointer
5646 (or vector of pointers) types with this instruction. To convert
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5647 pointers to other types, use the <a href="#i_inttoptr">inttoptr</a> or
5648 <a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5649
5650<h5>Example:</h5>
5651<pre>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5652 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5653 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Nadav Rotem3924cb02011-12-05 06:29:09 +0000[diff] [blame]5654 %Z = bitcast &lt;2 x int&gt; %V to i64; <i>; yields i64: %V</i>
5655 %Z = bitcast &lt;2 x i32*&gt; %V to &lt;2 x i64*&gt; <i>; yields &lt;2 x i64*&gt;</i>
Chris Lattner70de6632001-07-09 00:26:23 +0000[diff] [blame]5656</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5657
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5658</div>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5659
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5660</div>
5661
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5662<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5663<h3>
5664 <a name="otherops">Other Operations</a>
5665</h3>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5666
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5667<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5668
5669<p>The instructions in this category are the "miscellaneous" instructions, which
5670 defy better classification.</p>
5671
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5672<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5673<h4>
5674 <a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
5675</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5676
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5677<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5678
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5679<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5680<pre>
5681 &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 Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5682</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5683
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5684<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5685<p>The '<tt>icmp</tt>' instruction returns a boolean value or a vector of
Nadav Rotem3924cb02011-12-05 06:29:09 +0000[diff] [blame]5686 boolean values based on comparison of its two integer, integer vector,
5687 pointer, or pointer vector operands.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5688
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5689<h5>Arguments:</h5>
5690<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5691 the condition code indicating the kind of comparison to perform. It is not a
5692 value, just a keyword. The possible condition code are:</p>
5693
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5694<ol>
5695 <li><tt>eq</tt>: equal</li>
5696 <li><tt>ne</tt>: not equal </li>
5697 <li><tt>ugt</tt>: unsigned greater than</li>
5698 <li><tt>uge</tt>: unsigned greater or equal</li>
5699 <li><tt>ult</tt>: unsigned less than</li>
5700 <li><tt>ule</tt>: unsigned less or equal</li>
5701 <li><tt>sgt</tt>: signed greater than</li>
5702 <li><tt>sge</tt>: signed greater or equal</li>
5703 <li><tt>slt</tt>: signed less than</li>
5704 <li><tt>sle</tt>: signed less or equal</li>
5705</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5706
Chris Lattnerc0f423a2007-01-15 01:54:13 +0000[diff] [blame]5707<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5708 <a href="#t_pointer">pointer</a> or integer <a href="#t_vector">vector</a>
5709 typed. They must also be identical types.</p>
5710
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5711<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5712<p>The '<tt>icmp</tt>' compares <tt>op1</tt> and <tt>op2</tt> according to the
5713 condition code given as <tt>cond</tt>. The comparison performed always yields
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]5714 either an <a href="#t_integer"><tt>i1</tt></a> or vector of <tt>i1</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5715 result, as follows:</p>
5716
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5717<ol>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5718 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5719 <tt>false</tt> otherwise. No sign interpretation is necessary or
5720 performed.</li>
5721
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5722 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5723 <tt>false</tt> otherwise. No sign interpretation is necessary or
5724 performed.</li>
5725
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5726 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5727 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5728
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5729 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5730 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5731 to <tt>op2</tt>.</li>
5732
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5733 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5734 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5735
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5736 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5737 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5738
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5739 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5740 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5741
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5742 <li><tt>sge</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5743 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5744 to <tt>op2</tt>.</li>
5745
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5746 <li><tt>slt</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5747 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5748
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5749 <li><tt>sle</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5750 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5751</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5752
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5753<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5754 values are compared as if they were integers.</p>
5755
5756<p>If the operands are integer vectors, then they are compared element by
5757 element. The result is an <tt>i1</tt> vector with the same number of elements
5758 as the values being compared. Otherwise, the result is an <tt>i1</tt>.</p>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5759
5760<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5761<pre>
5762 &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5763 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
5764 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
5765 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
5766 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
5767 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5768</pre>
Dan Gohmana5127ab2009-01-22 01:39:38 +0000[diff] [blame]5769
5770<p>Note that the code generator does not yet support vector types with
5771 the <tt>icmp</tt> instruction.</p>
5772
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5773</div>
5774
5775<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5776<h4>
5777 <a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
5778</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5779
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5780<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5781
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5782<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5783<pre>
5784 &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 Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5785</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5786
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5787<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5788<p>The '<tt>fcmp</tt>' instruction returns a boolean value or vector of boolean
5789 values based on comparison of its operands.</p>
5790
5791<p>If the operands are floating point scalars, then the result type is a boolean
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]5792(<a href="#t_integer"><tt>i1</tt></a>).</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5793
5794<p>If the operands are floating point vectors, then the result type is a vector
5795 of boolean with the same number of elements as the operands being
5796 compared.</p>
5797
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5798<h5>Arguments:</h5>
5799<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5800 the condition code indicating the kind of comparison to perform. It is not a
5801 value, just a keyword. The possible condition code are:</p>
5802
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5803<ol>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5804 <li><tt>false</tt>: no comparison, always returns false</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5805 <li><tt>oeq</tt>: ordered and equal</li>
5806 <li><tt>ogt</tt>: ordered and greater than </li>
5807 <li><tt>oge</tt>: ordered and greater than or equal</li>
5808 <li><tt>olt</tt>: ordered and less than </li>
5809 <li><tt>ole</tt>: ordered and less than or equal</li>
5810 <li><tt>one</tt>: ordered and not equal</li>
5811 <li><tt>ord</tt>: ordered (no nans)</li>
5812 <li><tt>ueq</tt>: unordered or equal</li>
5813 <li><tt>ugt</tt>: unordered or greater than </li>
5814 <li><tt>uge</tt>: unordered or greater than or equal</li>
5815 <li><tt>ult</tt>: unordered or less than </li>
5816 <li><tt>ule</tt>: unordered or less than or equal</li>
5817 <li><tt>une</tt>: unordered or not equal</li>
5818 <li><tt>uno</tt>: unordered (either nans)</li>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5819 <li><tt>true</tt>: no comparison, always returns true</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5820</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5821
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5822<p><i>Ordered</i> means that neither operand is a QNAN while
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5823 <i>unordered</i> means that either operand may be a QNAN.</p>
5824
5825<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be either
5826 a <a href="#t_floating">floating point</a> type or
5827 a <a href="#t_vector">vector</a> of floating point type. They must have
5828 identical types.</p>
5829
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5830<h5>Semantics:</h5>
Gabor Greif0f75ad02008-08-07 21:46:00 +0000[diff] [blame]5831<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5832 according to the condition code given as <tt>cond</tt>. If the operands are
5833 vectors, then the vectors are compared element by element. Each comparison
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]5834 performed always yields an <a href="#t_integer">i1</a> result, as
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5835 follows:</p>
5836
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5837<ol>
5838 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5839
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5840 <li><tt>oeq</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5841 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5842
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5843 <li><tt>ogt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]5844 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5845
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5846 <li><tt>oge</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5847 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5848
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5849 <li><tt>olt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5850 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5851
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5852 <li><tt>ole</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5853 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5854
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5855 <li><tt>one</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5856 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5857
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5858 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5859
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5860 <li><tt>ueq</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5861 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5862
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5863 <li><tt>ugt</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5864 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5865
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5866 <li><tt>uge</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5867 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5868
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5869 <li><tt>ult</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5870 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5871
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5872 <li><tt>ule</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5873 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5874
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5875 <li><tt>une</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5876 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5877
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5878 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5879
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5880 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
5881</ol>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5882
5883<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5884<pre>
5885 &lt;result&gt; = fcmp oeq float 4.0, 5.0 <i>; yields: result=false</i>
Dan Gohmanc579d972008-09-09 01:02:47 +0000[diff] [blame]5886 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
5887 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
5888 &lt;result&gt; = fcmp ueq double 1.0, 2.0 <i>; yields: result=false</i>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5889</pre>
Dan Gohmana5127ab2009-01-22 01:39:38 +0000[diff] [blame]5890
5891<p>Note that the code generator does not yet support vector types with
5892 the <tt>fcmp</tt> instruction.</p>
5893
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5894</div>
5895
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5896<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5897<h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5898 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5899</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5900
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5901<div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5902
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5903<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5904<pre>
5905 &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...
5906</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5907
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5908<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5909<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in the
5910 SSA graph representing the function.</p>
5911
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5912<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5913<p>The type of the incoming values is specified with the first type field. After
5914 this, the '<tt>phi</tt>' instruction takes a list of pairs as arguments, with
5915 one pair for each predecessor basic block of the current block. Only values
5916 of <a href="#t_firstclass">first class</a> type may be used as the value
5917 arguments to the PHI node. Only labels may be used as the label
5918 arguments.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5919
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5920<p>There must be no non-phi instructions between the start of a basic block and
5921 the PHI instructions: i.e. PHI instructions must be first in a basic
5922 block.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5923
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5924<p>For the purposes of the SSA form, the use of each incoming value is deemed to
5925 occur on the edge from the corresponding predecessor block to the current
5926 block (but after any definition of an '<tt>invoke</tt>' instruction's return
5927 value on the same edge).</p>
Jay Foad1a4eea52009-06-03 10:20:10 +0000[diff] [blame]5928
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5929<h5>Semantics:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5930<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5931 specified by the pair corresponding to the predecessor basic block that
5932 executed just prior to the current block.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5933
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5934<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5935<pre>
5936Loop: ; Infinite loop that counts from 0 on up...
5937 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
5938 %nextindvar = add i32 %indvar, 1
5939 br label %Loop
5940</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5941
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5942</div>
5943
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5944<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5945<h4>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5946 <a name="i_select">'<tt>select</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5947</h4>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5948
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5949<div>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5950
5951<h5>Syntax:</h5>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5952<pre>
Dan Gohmanc579d972008-09-09 01:02:47 +0000[diff] [blame]5953 &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>
5954
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]5955 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5956</pre>
5957
5958<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5959<p>The '<tt>select</tt>' instruction is used to choose one value based on a
5960 condition, without branching.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5961
5962
5963<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5964<p>The '<tt>select</tt>' instruction requires an 'i1' value or a vector of 'i1'
5965 values indicating the condition, and two values of the
5966 same <a href="#t_firstclass">first class</a> type. If the val1/val2 are
5967 vectors and the condition is a scalar, then entire vectors are selected, not
5968 individual elements.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5969
5970<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5971<p>If the condition is an i1 and it evaluates to 1, the instruction returns the
5972 first value argument; otherwise, it returns the second value argument.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5973
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5974<p>If the condition is a vector of i1, then the value arguments must be vectors
5975 of the same size, and the selection is done element by element.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5976
5977<h5>Example:</h5>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5978<pre>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]5979 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5980</pre>
Dan Gohmana5127ab2009-01-22 01:39:38 +0000[diff] [blame]5981
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5982</div>
5983
Robert Bocchinof72fdfe2006-01-15 20:48:27 +0000[diff] [blame]5984<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5985<h4>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5986 <a name="i_call">'<tt>call</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5987</h4>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5988
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5989<div>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5990
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5991<h5>Syntax:</h5>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5992<pre>
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]5993 &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 Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5994</pre>
5995
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5996<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5997<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5998
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5999<h5>Arguments:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]6000<p>This instruction requires several arguments:</p>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]6001
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]6002<ol>
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]6003 <li>The optional "tail" marker indicates that the callee function does not
6004 access any allocas or varargs in the caller. Note that calls may be
6005 marked "tail" even if they do not occur before
6006 a <a href="#i_ret"><tt>ret</tt></a> instruction. If the "tail" marker is
6007 present, the function call is eligible for tail call optimization,
6008 but <a href="CodeGenerator.html#tailcallopt">might not in fact be
Evan Cheng59676492010-03-08 21:05:02 +0000[diff] [blame]6009 optimized into a jump</a>. The code generator may optimize calls marked
6010 "tail" with either 1) automatic <a href="CodeGenerator.html#sibcallopt">
6011 sibling call optimization</a> when the caller and callee have
6012 matching signatures, or 2) forced tail call optimization when the
6013 following extra requirements are met:
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]6014 <ul>
6015 <li>Caller and callee both have the calling
6016 convention <tt>fastcc</tt>.</li>
6017 <li>The call is in tail position (ret immediately follows call and ret
6018 uses value of call or is void).</li>
6019 <li>Option <tt>-tailcallopt</tt> is enabled,
Dan Gohman6232f732010-03-02 01:08:11 +0000[diff] [blame]6020 or <code>llvm::GuaranteedTailCallOpt</code> is <code>true</code>.</li>
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]6021 <li><a href="CodeGenerator.html#tailcallopt">Platform specific
6022 constraints are met.</a></li>
6023 </ul>
6024 </li>
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]6025
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6026 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
6027 convention</a> the call should use. If none is specified, the call
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]6028 defaults to using C calling conventions. The calling convention of the
6029 call must match the calling convention of the target function, or else the
6030 behavior is undefined.</li>
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]6031
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6032 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
6033 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
6034 '<tt>inreg</tt>' attributes are valid here.</li>
6035
6036 <li>'<tt>ty</tt>': the type of the call instruction itself which is also the
6037 type of the return value. Functions that return no value are marked
6038 <tt><a href="#t_void">void</a></tt>.</li>
6039
6040 <li>'<tt>fnty</tt>': shall be the signature of the pointer to function value
6041 being invoked. The argument types must match the types implied by this
6042 signature. This type can be omitted if the function is not varargs and if
6043 the function type does not return a pointer to a function.</li>
6044
6045 <li>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
6046 be invoked. In most cases, this is a direct function invocation, but
6047 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
6048 to function value.</li>
6049
6050 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]6051 signature argument types and parameter attributes. All arguments must be
6052 of <a href="#t_firstclass">first class</a> type. If the function
6053 signature indicates the function accepts a variable number of arguments,
6054 the extra arguments can be specified.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6055
6056 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
6057 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
6058 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]6059</ol>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]6060
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]6061<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6062<p>The '<tt>call</tt>' instruction is used to cause control flow to transfer to
6063 a specified function, with its incoming arguments bound to the specified
6064 values. Upon a '<tt><a href="#i_ret">ret</a></tt>' instruction in the called
6065 function, control flow continues with the instruction after the function
6066 call, and the return value of the function is bound to the result
6067 argument.</p>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]6068
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]6069<h5>Example:</h5>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]6070<pre>
Nick Lewyckya9b13d52007-09-08 13:57:50 +0000[diff] [blame]6071 %retval = call i32 @test(i32 %argc)
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6072 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) <i>; yields i32</i>
Chris Lattnerfb7c88d2008-03-21 17:24:17 +0000[diff] [blame]6073 %X = tail call i32 @foo() <i>; yields i32</i>
6074 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
6075 call void %foo(i8 97 signext)
Devang Pateld6cff512008-03-10 20:49:15 +0000[diff] [blame]6076
6077 %struct.A = type { i32, i8 }
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]6078 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohmancc3132e2008-10-04 19:00:07 +0000[diff] [blame]6079 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
6080 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattner6cbe8e92008-10-08 06:26:11 +0000[diff] [blame]6081 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijmaneefa7df2008-10-07 10:03:45 +0000[diff] [blame]6082 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]6083</pre>
6084
Dale Johannesen68f971b2009-09-24 18:38:21 +0000[diff] [blame]6085<p>llvm treats calls to some functions with names and arguments that match the
Dale Johannesen722212d2009-09-25 17:04:42 +0000[diff] [blame]6086standard C99 library as being the C99 library functions, and may perform
6087optimizations or generate code for them under that assumption. This is
6088something we'd like to change in the future to provide better support for
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]6089freestanding environments and non-C-based languages.</p>
Dale Johannesen68f971b2009-09-24 18:38:21 +0000[diff] [blame]6090
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]6091</div>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]6092
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6093<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6094<h4>
Chris Lattner33337472006-01-13 23:26:01 +0000[diff] [blame]6095 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6096</h4>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]6097
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6098<div>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]6099
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]6100<h5>Syntax:</h5>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]6101<pre>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]6102 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]6103</pre>
6104
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]6105<h5>Overview:</h5>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]6106<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6107 the "variable argument" area of a function call. It is used to implement the
6108 <tt>va_arg</tt> macro in C.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]6109
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]6110<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6111<p>This instruction takes a <tt>va_list*</tt> value and the type of the
6112 argument. It returns a value of the specified argument type and increments
6113 the <tt>va_list</tt> to point to the next argument. The actual type
6114 of <tt>va_list</tt> is target specific.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]6115
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]6116<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6117<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified type
6118 from the specified <tt>va_list</tt> and causes the <tt>va_list</tt> to point
6119 to the next argument. For more information, see the variable argument
6120 handling <a href="#int_varargs">Intrinsic Functions</a>.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]6121
6122<p>It is legal for this instruction to be called in a function which does not
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6123 take a variable number of arguments, for example, the <tt>vfprintf</tt>
6124 function.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]6125
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6126<p><tt>va_arg</tt> is an LLVM instruction instead of
6127 an <a href="#intrinsics">intrinsic function</a> because it takes a type as an
6128 argument.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]6129
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]6130<h5>Example:</h5>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]6131<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
6132
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6133<p>Note that the code generator does not yet fully support va_arg on many
6134 targets. Also, it does not currently support va_arg with aggregate types on
6135 any target.</p>
Dan Gohman3065b612009-01-12 23:12:39 +0000[diff] [blame]6136
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]6137</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6138
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]6139<!-- _______________________________________________________________________ -->
6140<h4>
6141 <a name="i_landingpad">'<tt>landingpad</tt>' Instruction</a>
6142</h4>
6143
6144<div>
6145
6146<h5>Syntax:</h5>
6147<pre>
Duncan Sandsdf9d7812012-01-13 19:59:16 +0000[diff] [blame]6148 &lt;resultval&gt; = landingpad &lt;resultty&gt; personality &lt;type&gt; &lt;pers_fn&gt; &lt;clause&gt;+
6149 &lt;resultval&gt; = landingpad &lt;resultty&gt; personality &lt;type&gt; &lt;pers_fn&gt; cleanup &lt;clause&gt;*
Bill Wendling49bfb122011-08-08 08:06:05 +0000[diff] [blame]6150
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]6151 &lt;clause&gt; := catch &lt;type&gt; &lt;value&gt;
Bill Wendlingfae14752011-08-12 20:24:12 +0000[diff] [blame]6152 &lt;clause&gt; := filter &lt;array constant type&gt; &lt;array constant&gt;
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]6153</pre>
6154
6155<h5>Overview:</h5>
6156<p>The '<tt>landingpad</tt>' instruction is used by
6157 <a href="ExceptionHandling.html#overview">LLVM's exception handling
6158 system</a> to specify that a basic block is a landing pad &mdash; one where
6159 the exception lands, and corresponds to the code found in the
6160 <i><tt>catch</tt></i> portion of a <i><tt>try/catch</tt></i> sequence. It
6161 defines values supplied by the personality function (<tt>pers_fn</tt>) upon
6162 re-entry to the function. The <tt>resultval</tt> has the
Duncan Sandsdf9d7812012-01-13 19:59:16 +0000[diff] [blame]6163 type <tt>resultty</tt>.</p>
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]6164
6165<h5>Arguments:</h5>
6166<p>This instruction takes a <tt>pers_fn</tt> value. This is the personality
6167 function associated with the unwinding mechanism. The optional
6168 <tt>cleanup</tt> flag indicates that the landing pad block is a cleanup.</p>
6169
6170<p>A <tt>clause</tt> begins with the clause type &mdash; <tt>catch</tt>
Bill Wendlingfae14752011-08-12 20:24:12 +0000[diff] [blame]6171 or <tt>filter</tt> &mdash; and contains the global variable representing the
6172 "type" that may be caught or filtered respectively. Unlike the
6173 <tt>catch</tt> clause, the <tt>filter</tt> clause takes an array constant as
6174 its argument. Use "<tt>[0 x i8**] undef</tt>" for a filter which cannot
6175 throw. The '<tt>landingpad</tt>' instruction must contain <em>at least</em>
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]6176 one <tt>clause</tt> or the <tt>cleanup</tt> flag.</p>
6177
6178<h5>Semantics:</h5>
6179<p>The '<tt>landingpad</tt>' instruction defines the values which are set by the
6180 personality function (<tt>pers_fn</tt>) upon re-entry to the function, and
6181 therefore the "result type" of the <tt>landingpad</tt> instruction. As with
6182 calling conventions, how the personality function results are represented in
6183 LLVM IR is target specific.</p>
6184
Bill Wendling0524b8d2011-08-03 17:17:06 +0000[diff] [blame]6185<p>The clauses are applied in order from top to bottom. If two
6186 <tt>landingpad</tt> instructions are merged together through inlining, the
Duncan Sandsdf9d7812012-01-13 19:59:16 +0000[diff] [blame]6187 clauses from the calling function are appended to the list of clauses.
6188 When the call stack is being unwound due to an exception being thrown, the
6189 exception is compared against each <tt>clause</tt> in turn. If it doesn't
6190 match any of the clauses, and the <tt>cleanup</tt> flag is not set, then
6191 unwinding continues further up the call stack.</p>
Bill Wendling0524b8d2011-08-03 17:17:06 +0000[diff] [blame]6192
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]6193<p>The <tt>landingpad</tt> instruction has several restrictions:</p>
6194
6195<ul>
6196 <li>A landing pad block is a basic block which is the unwind destination of an
6197 '<tt>invoke</tt>' instruction.</li>
6198 <li>A landing pad block must have a '<tt>landingpad</tt>' instruction as its
6199 first non-PHI instruction.</li>
6200 <li>There can be only one '<tt>landingpad</tt>' instruction within the landing
6201 pad block.</li>
6202 <li>A basic block that is not a landing pad block may not include a
6203 '<tt>landingpad</tt>' instruction.</li>
6204 <li>All '<tt>landingpad</tt>' instructions in a function must have the same
6205 personality function.</li>
6206</ul>
6207
6208<h5>Example:</h5>
6209<pre>
6210 ;; A landing pad which can catch an integer.
6211 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6212 catch i8** @_ZTIi
6213 ;; A landing pad that is a cleanup.
6214 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
Bill Wendlingfae14752011-08-12 20:24:12 +0000[diff] [blame]6215 cleanup
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]6216 ;; A landing pad which can catch an integer and can only throw a double.
6217 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6218 catch i8** @_ZTIi
Bill Wendlingfae14752011-08-12 20:24:12 +0000[diff] [blame]6219 filter [1 x i8**] [@_ZTId]
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]6220</pre>
6221
6222</div>
6223
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6224</div>
6225
6226</div>
6227
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6228<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6229<h2><a name="intrinsics">Intrinsic Functions</a></h2>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]6230<!-- *********************************************************************** -->
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6231
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6232<div>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6233
6234<p>LLVM supports the notion of an "intrinsic function". These functions have
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6235 well known names and semantics and are required to follow certain
6236 restrictions. Overall, these intrinsics represent an extension mechanism for
6237 the LLVM language that does not require changing all of the transformations
6238 in LLVM when adding to the language (or the bitcode reader/writer, the
6239 parser, etc...).</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6240
John Criswell88190562005-05-16 16:17:45 +0000[diff] [blame]6241<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6242 prefix is reserved in LLVM for intrinsic names; thus, function names may not
6243 begin with this prefix. Intrinsic functions must always be external
6244 functions: you cannot define the body of intrinsic functions. Intrinsic
6245 functions may only be used in call or invoke instructions: it is illegal to
6246 take the address of an intrinsic function. Additionally, because intrinsic
6247 functions are part of the LLVM language, it is required if any are added that
6248 they be documented here.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6249
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6250<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents a
6251 family of functions that perform the same operation but on different data
6252 types. Because LLVM can represent over 8 million different integer types,
6253 overloading is used commonly to allow an intrinsic function to operate on any
6254 integer type. One or more of the argument types or the result type can be
6255 overloaded to accept any integer type. Argument types may also be defined as
6256 exactly matching a previous argument's type or the result type. This allows
6257 an intrinsic function which accepts multiple arguments, but needs all of them
6258 to be of the same type, to only be overloaded with respect to a single
6259 argument or the result.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6260
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6261<p>Overloaded intrinsics will have the names of its overloaded argument types
6262 encoded into its function name, each preceded by a period. Only those types
6263 which are overloaded result in a name suffix. Arguments whose type is matched
6264 against another type do not. For example, the <tt>llvm.ctpop</tt> function
6265 can take an integer of any width and returns an integer of exactly the same
6266 integer width. This leads to a family of functions such as
6267 <tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29
6268 %val)</tt>. Only one type, the return type, is overloaded, and only one type
6269 suffix is required. Because the argument's type is matched against the return
6270 type, it does not require its own name suffix.</p>
Reid Spencer4eefaab2007-04-01 08:04:23 +0000[diff] [blame]6271
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]6272<p>To learn how to add an intrinsic function, please see the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6273 <a href="ExtendingLLVM.html">Extending LLVM Guide</a>.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6274
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6275<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6276<h3>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6277 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6278</h3>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6279
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6280<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6281
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6282<p>Variable argument support is defined in LLVM with
6283 the <a href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
6284 intrinsic functions. These functions are related to the similarly named
6285 macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6286
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6287<p>All of these functions operate on arguments that use a target-specific value
6288 type "<tt>va_list</tt>". The LLVM assembly language reference manual does
6289 not define what this type is, so all transformations should be prepared to
6290 handle these functions regardless of the type used.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6291
Chris Lattner30b868d2006-05-15 17:26:46 +0000[diff] [blame]6292<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6293 instruction and the variable argument handling intrinsic functions are
6294 used.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6295
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]6296<pre class="doc_code">
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]6297define i32 @test(i32 %X, ...) {
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6298 ; Initialize variable argument processing
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]6299 %ap = alloca i8*
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]6300 %ap2 = bitcast i8** %ap to i8*
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]6301 call void @llvm.va_start(i8* %ap2)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6302
6303 ; Read a single integer argument
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]6304 %tmp = va_arg i8** %ap, i32
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6305
6306 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]6307 %aq = alloca i8*
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]6308 %aq2 = bitcast i8** %aq to i8*
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]6309 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]6310 call void @llvm.va_end(i8* %aq2)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6311
6312 ; Stop processing of arguments.
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]6313 call void @llvm.va_end(i8* %ap2)
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]6314 ret i32 %tmp
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6315}
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]6316
6317declare void @llvm.va_start(i8*)
6318declare void @llvm.va_copy(i8*, i8*)
6319declare void @llvm.va_end(i8*)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6320</pre>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6321
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6322<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6323<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6324 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6325</h4>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6326
6327
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6328<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6329
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6330<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6331<pre>
6332 declare void %llvm.va_start(i8* &lt;arglist&gt;)
6333</pre>
6334
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6335<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6336<p>The '<tt>llvm.va_start</tt>' intrinsic initializes <tt>*&lt;arglist&gt;</tt>
6337 for subsequent use by <tt><a href="#i_va_arg">va_arg</a></tt>.</p>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]6338
6339<h5>Arguments:</h5>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]6340<p>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]6341
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6342<h5>Semantics:</h5>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]6343<p>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6344 macro available in C. In a target-dependent way, it initializes
6345 the <tt>va_list</tt> element to which the argument points, so that the next
6346 call to <tt>va_arg</tt> will produce the first variable argument passed to
6347 the function. Unlike the C <tt>va_start</tt> macro, this intrinsic does not
6348 need to know the last argument of the function as the compiler can figure
6349 that out.</p>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]6350
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]6351</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6352
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6353<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6354<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6355 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6356</h4>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6357
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6358<div>
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]6359
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6360<h5>Syntax:</h5>
6361<pre>
6362 declare void @llvm.va_end(i8* &lt;arglist&gt;)
6363</pre>
6364
6365<h5>Overview:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]6366<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6367 which has been initialized previously
6368 with <tt><a href="#int_va_start">llvm.va_start</a></tt>
6369 or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]6370
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6371<h5>Arguments:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]6372<p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]6373
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6374<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]6375<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6376 macro available in C. In a target-dependent way, it destroys
6377 the <tt>va_list</tt> element to which the argument points. Calls
6378 to <a href="#int_va_start"><tt>llvm.va_start</tt></a>
6379 and <a href="#int_va_copy"> <tt>llvm.va_copy</tt></a> must be matched exactly
6380 with calls to <tt>llvm.va_end</tt>.</p>
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]6381
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]6382</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6383
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6384<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6385<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6386 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6387</h4>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6388
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6389<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6390
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6391<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6392<pre>
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]6393 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6394</pre>
6395
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6396<h5>Overview:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]6397<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6398 from the source argument list to the destination argument list.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6399
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6400<h5>Arguments:</h5>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]6401<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6402 The second argument is a pointer to a <tt>va_list</tt> element to copy
6403 from.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6404
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6405<h5>Semantics:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]6406<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6407 macro available in C. In a target-dependent way, it copies the
6408 source <tt>va_list</tt> element into the destination <tt>va_list</tt>
6409 element. This intrinsic is necessary because
6410 the <tt><a href="#int_va_start"> llvm.va_start</a></tt> intrinsic may be
6411 arbitrarily complex and require, for example, memory allocation.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6412
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]6413</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6414
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6415</div>
6416
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6417<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6418<h3>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6419 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6420</h3>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6421
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6422<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6423
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6424<p>LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]6425Collection</a> (GC) requires the implementation and generation of these
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6426intrinsics. These intrinsics allow identification of <a href="#int_gcroot">GC
6427roots on the stack</a>, as well as garbage collector implementations that
6428require <a href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a>
6429barriers. Front-ends for type-safe garbage collected languages should generate
6430these intrinsics to make use of the LLVM garbage collectors. For more details,
6431see <a href="GarbageCollection.html">Accurate Garbage Collection with
6432LLVM</a>.</p>
Christopher Lamb55c6d4f2007-12-17 01:00:21 +0000[diff] [blame]6433
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6434<p>The garbage collection intrinsics only operate on objects in the generic
6435 address space (address space zero).</p>
Christopher Lamb55c6d4f2007-12-17 01:00:21 +0000[diff] [blame]6436
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6437<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6438<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6439 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6440</h4>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6441
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6442<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6443
6444<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6445<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]6446 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6447</pre>
6448
6449<h5>Overview:</h5>
John Criswelldfe6a862004-12-10 15:51:16 +0000[diff] [blame]6450<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6451 the code generator, and allows some metadata to be associated with it.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6452
6453<h5>Arguments:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6454<p>The first argument specifies the address of a stack object that contains the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6455 root pointer. The second pointer (which must be either a constant or a
6456 global value address) contains the meta-data to be associated with the
6457 root.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6458
6459<h5>Semantics:</h5>
Chris Lattner851b7712008-04-24 05:59:56 +0000[diff] [blame]6460<p>At runtime, a call to this intrinsic stores a null pointer into the "ptrloc"
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6461 location. At compile-time, the code generator generates information to allow
6462 the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
6463 intrinsic may only be used in a function which <a href="#gc">specifies a GC
6464 algorithm</a>.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6465
6466</div>
6467
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6468<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6469<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6470 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6471</h4>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6472
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6473<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6474
6475<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6476<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]6477 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6478</pre>
6479
6480<h5>Overview:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6481<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6482 locations, allowing garbage collector implementations that require read
6483 barriers.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6484
6485<h5>Arguments:</h5>
Chris Lattnerf9228072006-03-14 20:02:51 +0000[diff] [blame]6486<p>The second argument is the address to read from, which should be an address
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6487 allocated from the garbage collector. The first object is a pointer to the
6488 start of the referenced object, if needed by the language runtime (otherwise
6489 null).</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6490
6491<h5>Semantics:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6492<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6493 instruction, but may be replaced with substantially more complex code by the
6494 garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
6495 may only be used in a function which <a href="#gc">specifies a GC
6496 algorithm</a>.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6497
6498</div>
6499
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6500<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6501<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6502 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6503</h4>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6504
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6505<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6506
6507<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6508<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]6509 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6510</pre>
6511
6512<h5>Overview:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6513<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6514 locations, allowing garbage collector implementations that require write
6515 barriers (such as generational or reference counting collectors).</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6516
6517<h5>Arguments:</h5>
Chris Lattnerf9228072006-03-14 20:02:51 +0000[diff] [blame]6518<p>The first argument is the reference to store, the second is the start of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6519 object to store it to, and the third is the address of the field of Obj to
6520 store to. If the runtime does not require a pointer to the object, Obj may
6521 be null.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6522
6523<h5>Semantics:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6524<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6525 instruction, but may be replaced with substantially more complex code by the
6526 garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
6527 may only be used in a function which <a href="#gc">specifies a GC
6528 algorithm</a>.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6529
6530</div>
6531
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6532</div>
6533
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6534<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6535<h3>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6536 <a name="int_codegen">Code Generator Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6537</h3>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6538
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6539<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6540
6541<p>These intrinsics are provided by LLVM to expose special features that may
6542 only be implemented with code generator support.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6543
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6544<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6545<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6546 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6547</h4>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6548
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6549<div>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6550
6551<h5>Syntax:</h5>
6552<pre>
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]6553 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6554</pre>
6555
6556<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6557<p>The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
6558 target-specific value indicating the return address of the current function
6559 or one of its callers.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6560
6561<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6562<p>The argument to this intrinsic indicates which function to return the address
6563 for. Zero indicates the calling function, one indicates its caller, etc.
6564 The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6565
6566<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6567<p>The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer
6568 indicating the return address of the specified call frame, or zero if it
6569 cannot be identified. The value returned by this intrinsic is likely to be
6570 incorrect or 0 for arguments other than zero, so it should only be used for
6571 debugging purposes.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6572
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6573<p>Note that calling this intrinsic does not prevent function inlining or other
6574 aggressive transformations, so the value returned may not be that of the
6575 obvious source-language caller.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6576
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6577</div>
6578
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6579<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6580<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6581 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6582</h4>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6583
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6584<div>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6585
6586<h5>Syntax:</h5>
6587<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6588 declare i8* @llvm.frameaddress(i32 &lt;level&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6589</pre>
6590
6591<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6592<p>The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
6593 target-specific frame pointer value for the specified stack frame.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6594
6595<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6596<p>The argument to this intrinsic indicates which function to return the frame
6597 pointer for. Zero indicates the calling function, one indicates its caller,
6598 etc. The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6599
6600<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6601<p>The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer
6602 indicating the frame address of the specified call frame, or zero if it
6603 cannot be identified. The value returned by this intrinsic is likely to be
6604 incorrect or 0 for arguments other than zero, so it should only be used for
6605 debugging purposes.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6606
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6607<p>Note that calling this intrinsic does not prevent function inlining or other
6608 aggressive transformations, so the value returned may not be that of the
6609 obvious source-language caller.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6610
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6611</div>
6612
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6613<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6614<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6615 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6616</h4>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6617
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6618<div>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6619
6620<h5>Syntax:</h5>
6621<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6622 declare i8* @llvm.stacksave()
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6623</pre>
6624
6625<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6626<p>The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state
6627 of the function stack, for use
6628 with <a href="#int_stackrestore"> <tt>llvm.stackrestore</tt></a>. This is
6629 useful for implementing language features like scoped automatic variable
6630 sized arrays in C99.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6631
6632<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6633<p>This intrinsic returns a opaque pointer value that can be passed
6634 to <a href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When
6635 an <tt>llvm.stackrestore</tt> intrinsic is executed with a value saved
6636 from <tt>llvm.stacksave</tt>, it effectively restores the state of the stack
6637 to the state it was in when the <tt>llvm.stacksave</tt> intrinsic executed.
6638 In practice, this pops any <a href="#i_alloca">alloca</a> blocks from the
6639 stack that were allocated after the <tt>llvm.stacksave</tt> was executed.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6640
6641</div>
6642
6643<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6644<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6645 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6646</h4>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6647
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6648<div>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6649
6650<h5>Syntax:</h5>
6651<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6652 declare void @llvm.stackrestore(i8* %ptr)
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6653</pre>
6654
6655<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6656<p>The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
6657 the function stack to the state it was in when the
6658 corresponding <a href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic
6659 executed. This is useful for implementing language features like scoped
6660 automatic variable sized arrays in C99.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6661
6662<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6663<p>See the description
6664 for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6665
6666</div>
6667
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6668<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6669<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6670 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6671</h4>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6672
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6673<div>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6674
6675<h5>Syntax:</h5>
6676<pre>
Bruno Cardoso Lopesdc9ff3a2011-06-14 04:58:37 +0000[diff] [blame]6677 declare void @llvm.prefetch(i8* &lt;address&gt;, i32 &lt;rw&gt;, i32 &lt;locality&gt;, i32 &lt;cache type&gt;)
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6678</pre>
6679
6680<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6681<p>The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to
6682 insert a prefetch instruction if supported; otherwise, it is a noop.
6683 Prefetches have no effect on the behavior of the program but can change its
6684 performance characteristics.</p>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6685
6686<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6687<p><tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the
6688 specifier determining if the fetch should be for a read (0) or write (1),
6689 and <tt>locality</tt> is a temporal locality specifier ranging from (0) - no
Bruno Cardoso Lopesdc9ff3a2011-06-14 04:58:37 +0000[diff] [blame]6690 locality, to (3) - extremely local keep in cache. The <tt>cache type</tt>
6691 specifies whether the prefetch is performed on the data (1) or instruction (0)
6692 cache. The <tt>rw</tt>, <tt>locality</tt> and <tt>cache type</tt> arguments
6693 must be constant integers.</p>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6694
6695<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6696<p>This intrinsic does not modify the behavior of the program. In particular,
6697 prefetches cannot trap and do not produce a value. On targets that support
6698 this intrinsic, the prefetch can provide hints to the processor cache for
6699 better performance.</p>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6700
6701</div>
6702
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6703<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6704<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6705 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6706</h4>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6707
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6708<div>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6709
6710<h5>Syntax:</h5>
6711<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]6712 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6713</pre>
6714
6715<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6716<p>The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program
6717 Counter (PC) in a region of code to simulators and other tools. The method
6718 is target specific, but it is expected that the marker will use exported
6719 symbols to transmit the PC of the marker. The marker makes no guarantees
6720 that it will remain with any specific instruction after optimizations. It is
6721 possible that the presence of a marker will inhibit optimizations. The
6722 intended use is to be inserted after optimizations to allow correlations of
6723 simulation runs.</p>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6724
6725<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6726<p><tt>id</tt> is a numerical id identifying the marker.</p>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6727
6728<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6729<p>This intrinsic does not modify the behavior of the program. Backends that do
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]6730 not support this intrinsic may ignore it.</p>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6731
6732</div>
6733
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6734<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6735<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6736 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6737</h4>
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6738
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6739<div>
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6740
6741<h5>Syntax:</h5>
6742<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]6743 declare i64 @llvm.readcyclecounter()
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6744</pre>
6745
6746<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6747<p>The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
6748 counter register (or similar low latency, high accuracy clocks) on those
6749 targets that support it. On X86, it should map to RDTSC. On Alpha, it
6750 should map to RPCC. As the backing counters overflow quickly (on the order
6751 of 9 seconds on alpha), this should only be used for small timings.</p>
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6752
6753<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6754<p>When directly supported, reading the cycle counter should not modify any
6755 memory. Implementations are allowed to either return a application specific
6756 value or a system wide value. On backends without support, this is lowered
6757 to a constant 0.</p>
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6758
6759</div>
6760
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6761</div>
6762
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6763<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6764<h3>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6765 <a name="int_libc">Standard C Library Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6766</h3>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6767
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6768<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6769
6770<p>LLVM provides intrinsics for a few important standard C library functions.
6771 These intrinsics allow source-language front-ends to pass information about
6772 the alignment of the pointer arguments to the code generator, providing
6773 opportunity for more efficient code generation.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6774
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6775<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6776<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6777 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6778</h4>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6779
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6780<div>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6781
6782<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6783<p>This is an overloaded intrinsic. You can use <tt>llvm.memcpy</tt> on any
Mon P Wang508127b2010-04-07 06:35:53 +0000[diff] [blame]6784 integer bit width and for different address spaces. Not all targets support
6785 all bit widths however.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6786
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6787<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6788 declare void @llvm.memcpy.p0i8.p0i8.i32(i8* &lt;dest&gt;, i8* &lt;src&gt;,
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6789 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6790 declare void @llvm.memcpy.p0i8.p0i8.i64(i8* &lt;dest&gt;, i8* &lt;src&gt;,
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6791 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6792</pre>
6793
6794<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6795<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6796 source location to the destination location.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6797
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6798<p>Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6799 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6800 and the pointers can be in specified address spaces.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6801
6802<h5>Arguments:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6803
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6804<p>The first argument is a pointer to the destination, the second is a pointer
6805 to the source. The third argument is an integer argument specifying the
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6806 number of bytes to copy, the fourth argument is the alignment of the
6807 source and destination locations, and the fifth is a boolean indicating a
6808 volatile access.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6809
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]6810<p>If the call to this intrinsic has an alignment value that is not 0 or 1,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6811 then the caller guarantees that both the source and destination pointers are
6812 aligned to that boundary.</p>
Chris Lattner4c67c482004-02-12 21:18:15 +0000[diff] [blame]6813
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]6814<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6815 <tt>llvm.memcpy</tt> call is a <a href="#volatile">volatile operation</a>.
6816 The detailed access behavior is not very cleanly specified and it is unwise
6817 to depend on it.</p>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6818
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6819<h5>Semantics:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6820
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6821<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6822 source location to the destination location, which are not allowed to
6823 overlap. It copies "len" bytes of memory over. If the argument is known to
6824 be aligned to some boundary, this can be specified as the fourth argument,
6825 otherwise it should be set to 0 or 1.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6826
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6827</div>
6828
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6829<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6830<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6831 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6832</h4>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6833
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6834<div>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6835
6836<h5>Syntax:</h5>
Chris Lattnerdd708342008-11-21 16:42:48 +0000[diff] [blame]6837<p>This is an overloaded intrinsic. You can use llvm.memmove on any integer bit
Mon P Wang508127b2010-04-07 06:35:53 +0000[diff] [blame]6838 width and for different address space. Not all targets support all bit
6839 widths however.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6840
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6841<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6842 declare void @llvm.memmove.p0i8.p0i8.i32(i8* &lt;dest&gt;, i8* &lt;src&gt;,
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6843 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6844 declare void @llvm.memmove.p0i8.p0i8.i64(i8* &lt;dest&gt;, i8* &lt;src&gt;,
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6845 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6846</pre>
6847
6848<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6849<p>The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the
6850 source location to the destination location. It is similar to the
6851 '<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to
6852 overlap.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6853
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6854<p>Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6855 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6856 and the pointers can be in specified address spaces.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6857
6858<h5>Arguments:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6859
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6860<p>The first argument is a pointer to the destination, the second is a pointer
6861 to the source. The third argument is an integer argument specifying the
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6862 number of bytes to copy, the fourth argument is the alignment of the
6863 source and destination locations, and the fifth is a boolean indicating a
6864 volatile access.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6865
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]6866<p>If the call to this intrinsic has an alignment value that is not 0 or 1,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6867 then the caller guarantees that the source and destination pointers are
6868 aligned to that boundary.</p>
Chris Lattner4c67c482004-02-12 21:18:15 +0000[diff] [blame]6869
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]6870<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6871 <tt>llvm.memmove</tt> call is a <a href="#volatile">volatile operation</a>.
6872 The detailed access behavior is not very cleanly specified and it is unwise
6873 to depend on it.</p>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6874
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6875<h5>Semantics:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6876
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6877<p>The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the
6878 source location to the destination location, which may overlap. It copies
6879 "len" bytes of memory over. If the argument is known to be aligned to some
6880 boundary, this can be specified as the fourth argument, otherwise it should
6881 be set to 0 or 1.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6882
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6883</div>
6884
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6885<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6886<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6887 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6888</h4>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6889
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6890<div>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6891
6892<h5>Syntax:</h5>
Chris Lattnerdd708342008-11-21 16:42:48 +0000[diff] [blame]6893<p>This is an overloaded intrinsic. You can use llvm.memset on any integer bit
John Criswellad05ae42010-07-30 16:30:28 +0000[diff] [blame]6894 width and for different address spaces. However, not all targets support all
6895 bit widths.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6896
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6897<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6898 declare void @llvm.memset.p0i8.i32(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattner685db9d2010-04-08 00:54:34 +0000[diff] [blame]6899 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6900 declare void @llvm.memset.p0i8.i64(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattner685db9d2010-04-08 00:54:34 +0000[diff] [blame]6901 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6902</pre>
6903
6904<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6905<p>The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a
6906 particular byte value.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6907
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6908<p>Note that, unlike the standard libc function, the <tt>llvm.memset</tt>
John Criswellad05ae42010-07-30 16:30:28 +0000[diff] [blame]6909 intrinsic does not return a value and takes extra alignment/volatile
6910 arguments. Also, the destination can be in an arbitrary address space.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6911
6912<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6913<p>The first argument is a pointer to the destination to fill, the second is the
John Criswellad05ae42010-07-30 16:30:28 +0000[diff] [blame]6914 byte value with which to fill it, the third argument is an integer argument
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6915 specifying the number of bytes to fill, and the fourth argument is the known
John Criswellad05ae42010-07-30 16:30:28 +0000[diff] [blame]6916 alignment of the destination location.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6917
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]6918<p>If the call to this intrinsic has an alignment value that is not 0 or 1,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6919 then the caller guarantees that the destination pointer is aligned to that
6920 boundary.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6921
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]6922<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6923 <tt>llvm.memset</tt> call is a <a href="#volatile">volatile operation</a>.
6924 The detailed access behavior is not very cleanly specified and it is unwise
6925 to depend on it.</p>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6926
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6927<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6928<p>The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting
6929 at the destination location. If the argument is known to be aligned to some
6930 boundary, this can be specified as the fourth argument, otherwise it should
6931 be set to 0 or 1.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6932
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6933</div>
6934
Chris Lattner3b4f4372004-06-11 02:28:03 +0000[diff] [blame]6935<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6936<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6937 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6938</h4>
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6939
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6940<div>
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6941
6942<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6943<p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
6944 floating point or vector of floating point type. Not all targets support all
6945 types however.</p>
6946
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6947<pre>
Dale Johannesendd89d272007-10-02 17:47:38 +0000[diff] [blame]6948 declare float @llvm.sqrt.f32(float %Val)
6949 declare double @llvm.sqrt.f64(double %Val)
6950 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
6951 declare fp128 @llvm.sqrt.f128(fp128 %Val)
6952 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6953</pre>
6954
6955<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6956<p>The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
6957 returning the same value as the libm '<tt>sqrt</tt>' functions would.
6958 Unlike <tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined
6959 behavior for negative numbers other than -0.0 (which allows for better
6960 optimization, because there is no need to worry about errno being
6961 set). <tt>llvm.sqrt(-0.0)</tt> is defined to return -0.0 like IEEE sqrt.</p>
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6962
6963<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6964<p>The argument and return value are floating point numbers of the same
6965 type.</p>
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6966
6967<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6968<p>This function returns the sqrt of the specified operand if it is a
6969 nonnegative floating point number.</p>
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6970
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6971</div>
6972
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6973<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6974<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6975 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6976</h4>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6977
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6978<div>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6979
6980<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6981<p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
6982 floating point or vector of floating point type. Not all targets support all
6983 types however.</p>
6984
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6985<pre>
Dale Johannesendd89d272007-10-02 17:47:38 +0000[diff] [blame]6986 declare float @llvm.powi.f32(float %Val, i32 %power)
6987 declare double @llvm.powi.f64(double %Val, i32 %power)
6988 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
6989 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
6990 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6991</pre>
6992
6993<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6994<p>The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
6995 specified (positive or negative) power. The order of evaluation of
6996 multiplications is not defined. When a vector of floating point type is
6997 used, the second argument remains a scalar integer value.</p>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6998
6999<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7000<p>The second argument is an integer power, and the first is a value to raise to
7001 that power.</p>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]7002
7003<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7004<p>This function returns the first value raised to the second power with an
7005 unspecified sequence of rounding operations.</p>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]7006
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]7007</div>
7008
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7009<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7010<h4>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7011 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7012</h4>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7013
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7014<div>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7015
7016<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7017<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
7018 floating point or vector of floating point type. Not all targets support all
7019 types however.</p>
7020
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7021<pre>
7022 declare float @llvm.sin.f32(float %Val)
7023 declare double @llvm.sin.f64(double %Val)
7024 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
7025 declare fp128 @llvm.sin.f128(fp128 %Val)
7026 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
7027</pre>
7028
7029<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7030<p>The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7031
7032<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7033<p>The argument and return value are floating point numbers of the same
7034 type.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7035
7036<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7037<p>This function returns the sine of the specified operand, returning the same
7038 values as the libm <tt>sin</tt> functions would, and handles error conditions
7039 in the same way.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7040
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7041</div>
7042
7043<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7044<h4>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7045 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7046</h4>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7047
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7048<div>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7049
7050<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7051<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
7052 floating point or vector of floating point type. Not all targets support all
7053 types however.</p>
7054
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7055<pre>
7056 declare float @llvm.cos.f32(float %Val)
7057 declare double @llvm.cos.f64(double %Val)
7058 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
7059 declare fp128 @llvm.cos.f128(fp128 %Val)
7060 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
7061</pre>
7062
7063<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7064<p>The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7065
7066<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7067<p>The argument and return value are floating point numbers of the same
7068 type.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7069
7070<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7071<p>This function returns the cosine of the specified operand, returning the same
7072 values as the libm <tt>cos</tt> functions would, and handles error conditions
7073 in the same way.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7074
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7075</div>
7076
7077<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7078<h4>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7079 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7080</h4>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7081
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7082<div>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7083
7084<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7085<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
7086 floating point or vector of floating point type. Not all targets support all
7087 types however.</p>
7088
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7089<pre>
7090 declare float @llvm.pow.f32(float %Val, float %Power)
7091 declare double @llvm.pow.f64(double %Val, double %Power)
7092 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
7093 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
7094 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
7095</pre>
7096
7097<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7098<p>The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
7099 specified (positive or negative) power.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7100
7101<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7102<p>The second argument is a floating point power, and the first is a value to
7103 raise to that power.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7104
7105<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7106<p>This function returns the first value raised to the second power, returning
7107 the same values as the libm <tt>pow</tt> functions would, and handles error
7108 conditions in the same way.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7109
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]7110</div>
7111
Dan Gohman911fa902011-05-23 21:13:03 +0000[diff] [blame]7112<!-- _______________________________________________________________________ -->
7113<h4>
7114 <a name="int_exp">'<tt>llvm.exp.*</tt>' Intrinsic</a>
7115</h4>
7116
7117<div>
7118
7119<h5>Syntax:</h5>
7120<p>This is an overloaded intrinsic. You can use <tt>llvm.exp</tt> on any
7121 floating point or vector of floating point type. Not all targets support all
7122 types however.</p>
7123
7124<pre>
7125 declare float @llvm.exp.f32(float %Val)
7126 declare double @llvm.exp.f64(double %Val)
7127 declare x86_fp80 @llvm.exp.f80(x86_fp80 %Val)
7128 declare fp128 @llvm.exp.f128(fp128 %Val)
7129 declare ppc_fp128 @llvm.exp.ppcf128(ppc_fp128 %Val)
7130</pre>
7131
7132<h5>Overview:</h5>
7133<p>The '<tt>llvm.exp.*</tt>' intrinsics perform the exp function.</p>
7134
7135<h5>Arguments:</h5>
7136<p>The argument and return value are floating point numbers of the same
7137 type.</p>
7138
7139<h5>Semantics:</h5>
7140<p>This function returns the same values as the libm <tt>exp</tt> functions
7141 would, and handles error conditions in the same way.</p>
7142
7143</div>
7144
7145<!-- _______________________________________________________________________ -->
7146<h4>
7147 <a name="int_log">'<tt>llvm.log.*</tt>' Intrinsic</a>
7148</h4>
7149
7150<div>
7151
7152<h5>Syntax:</h5>
7153<p>This is an overloaded intrinsic. You can use <tt>llvm.log</tt> on any
7154 floating point or vector of floating point type. Not all targets support all
7155 types however.</p>
7156
7157<pre>
7158 declare float @llvm.log.f32(float %Val)
7159 declare double @llvm.log.f64(double %Val)
7160 declare x86_fp80 @llvm.log.f80(x86_fp80 %Val)
7161 declare fp128 @llvm.log.f128(fp128 %Val)
7162 declare ppc_fp128 @llvm.log.ppcf128(ppc_fp128 %Val)
7163</pre>
7164
7165<h5>Overview:</h5>
7166<p>The '<tt>llvm.log.*</tt>' intrinsics perform the log function.</p>
7167
7168<h5>Arguments:</h5>
7169<p>The argument and return value are floating point numbers of the same
7170 type.</p>
7171
7172<h5>Semantics:</h5>
7173<p>This function returns the same values as the libm <tt>log</tt> functions
7174 would, and handles error conditions in the same way.</p>
7175
Nick Lewyckycd196f62011-10-31 01:32:21 +0000[diff] [blame]7176</div>
7177
7178<!-- _______________________________________________________________________ -->
Cameron Zwarichf03fa182011-07-08 21:39:21 +0000[diff] [blame]7179<h4>
7180 <a name="int_fma">'<tt>llvm.fma.*</tt>' Intrinsic</a>
7181</h4>
7182
7183<div>
7184
7185<h5>Syntax:</h5>
7186<p>This is an overloaded intrinsic. You can use <tt>llvm.fma</tt> on any
7187 floating point or vector of floating point type. Not all targets support all
7188 types however.</p>
7189
7190<pre>
7191 declare float @llvm.fma.f32(float %a, float %b, float %c)
7192 declare double @llvm.fma.f64(double %a, double %b, double %c)
7193 declare x86_fp80 @llvm.fma.f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c)
7194 declare fp128 @llvm.fma.f128(fp128 %a, fp128 %b, fp128 %c)
7195 declare ppc_fp128 @llvm.fma.ppcf128(ppc_fp128 %a, ppc_fp128 %b, ppc_fp128 %c)
7196</pre>
7197
7198<h5>Overview:</h5>
Cameron Zwaricha32fd212011-07-08 22:13:55 +0000[diff] [blame]7199<p>The '<tt>llvm.fma.*</tt>' intrinsics perform the fused multiply-add
Cameron Zwarichf03fa182011-07-08 21:39:21 +0000[diff] [blame]7200 operation.</p>
7201
7202<h5>Arguments:</h5>
7203<p>The argument and return value are floating point numbers of the same
7204 type.</p>
7205
7206<h5>Semantics:</h5>
7207<p>This function returns the same values as the libm <tt>fma</tt> functions
7208 would.</p>
7209
Dan Gohman911fa902011-05-23 21:13:03 +0000[diff] [blame]7210</div>
7211
NAKAMURA Takumia35cdd62011-10-31 13:04:26 +0000[diff] [blame]7212</div>
7213
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7214<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7215<h3>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]7216 <a name="int_manip">Bit Manipulation Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7217</h3>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7218
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7219<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7220
7221<p>LLVM provides intrinsics for a few important bit manipulation operations.
7222 These allow efficient code generation for some algorithms.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7223
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7224<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7225<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]7226 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7227</h4>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]7228
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7229<div>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]7230
7231<h5>Syntax:</h5>
Reid Spencer4eefaab2007-04-01 08:04:23 +0000[diff] [blame]7232<p>This is an overloaded intrinsic function. You can use bswap on any integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7233 type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
7234
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]7235<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]7236 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
7237 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
7238 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]7239</pre>
7240
7241<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7242<p>The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
7243 values with an even number of bytes (positive multiple of 16 bits). These
7244 are useful for performing operations on data that is not in the target's
7245 native byte order.</p>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]7246
7247<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7248<p>The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
7249 and low byte of the input i16 swapped. Similarly,
7250 the <tt>llvm.bswap.i32</tt> intrinsic returns an i32 value that has the four
7251 bytes of the input i32 swapped, so that if the input bytes are numbered 0, 1,
7252 2, 3 then the returned i32 will have its bytes in 3, 2, 1, 0 order.
7253 The <tt>llvm.bswap.i48</tt>, <tt>llvm.bswap.i64</tt> and other intrinsics
7254 extend this concept to additional even-byte lengths (6 bytes, 8 bytes and
7255 more, respectively).</p>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]7256
7257</div>
7258
7259<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7260<h4>
Reid Spencerb4f9a6f2006-01-16 21:12:35 +0000[diff] [blame]7261 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7262</h4>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7263
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7264<div>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7265
7266<h5>Syntax:</h5>
Reid Spencer4eefaab2007-04-01 08:04:23 +0000[diff] [blame]7267<p>This is an overloaded intrinsic. You can use llvm.ctpop on any integer bit
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]7268 width, or on any vector with integer elements. Not all targets support all
7269 bit widths or vector types, however.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7270
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7271<pre>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7272 declare i8 @llvm.ctpop.i8(i8 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]7273 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]7274 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]7275 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
7276 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]7277 declare &lt;2 x i32&gt; @llvm.ctpop.v2i32(&lt;2 x i32&gt; &lt;src&gt;)
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7278</pre>
7279
7280<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7281<p>The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set
7282 in a value.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7283
7284<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7285<p>The only argument is the value to be counted. The argument may be of any
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]7286 integer type, or a vector with integer elements.
7287 The return type must match the argument type.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7288
7289<h5>Semantics:</h5>
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]7290<p>The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable, or within each
7291 element of a vector.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7292
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7293</div>
7294
7295<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7296<h4>
Chris Lattnerb748c672006-01-16 22:34:14 +0000[diff] [blame]7297 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7298</h4>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7299
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7300<div>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7301
7302<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7303<p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]7304 integer bit width, or any vector whose elements are integers. Not all
7305 targets support all bit widths or vector types, however.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7306
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7307<pre>
Chandler Carruthf6bb2782011-12-12 04:36:04 +0000[diff] [blame]7308 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7309 declare i16 @llvm.ctlz.i16 (i16 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7310 declare i32 @llvm.ctlz.i32 (i32 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7311 declare i64 @llvm.ctlz.i64 (i64 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7312 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7313 declase &lt;2 x i32&gt; @llvm.ctlz.v2i32(&lt;2 x i32&gt; &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7314</pre>
7315
7316<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7317<p>The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
7318 leading zeros in a variable.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7319
7320<h5>Arguments:</h5>
Chandler Carruthf6bb2782011-12-12 04:36:04 +0000[diff] [blame]7321<p>The first argument is the value to be counted. This argument may be of any
7322 integer type, or a vectory with integer element type. The return type
7323 must match the first argument type.</p>
7324
7325<p>The second argument must be a constant and is a flag to indicate whether the
7326 intrinsic should ensure that a zero as the first argument produces a defined
7327 result. Historically some architectures did not provide a defined result for
7328 zero values as efficiently, and many algorithms are now predicated on
7329 avoiding zero-value inputs.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7330
7331<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7332<p>The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant)
Chandler Carruthf6bb2782011-12-12 04:36:04 +0000[diff] [blame]7333 zeros in a variable, or within each element of the vector.
7334 If <tt>src == 0</tt> then the result is the size in bits of the type of
7335 <tt>src</tt> if <tt>is_zero_undef == 0</tt> and <tt>undef</tt> otherwise.
7336 For example, <tt>llvm.ctlz(i32 2) = 30</tt>.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7337
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7338</div>
Chris Lattner3b4f4372004-06-11 02:28:03 +0000[diff] [blame]7339
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]7340<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7341<h4>
Chris Lattnerb748c672006-01-16 22:34:14 +0000[diff] [blame]7342 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7343</h4>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]7344
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7345<div>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]7346
7347<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7348<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]7349 integer bit width, or any vector of integer elements. Not all targets
7350 support all bit widths or vector types, however.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7351
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]7352<pre>
Chandler Carruthf6bb2782011-12-12 04:36:04 +0000[diff] [blame]7353 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7354 declare i16 @llvm.cttz.i16 (i16 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7355 declare i32 @llvm.cttz.i32 (i32 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7356 declare i64 @llvm.cttz.i64 (i64 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7357 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7358 declase &lt;2 x i32&gt; @llvm.cttz.v2i32(&lt;2 x i32&gt; &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]7359</pre>
7360
7361<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7362<p>The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
7363 trailing zeros.</p>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]7364
7365<h5>Arguments:</h5>
Chandler Carruthf6bb2782011-12-12 04:36:04 +0000[diff] [blame]7366<p>The first argument is the value to be counted. This argument may be of any
7367 integer type, or a vectory with integer element type. The return type
7368 must match the first argument type.</p>
7369
7370<p>The second argument must be a constant and is a flag to indicate whether the
7371 intrinsic should ensure that a zero as the first argument produces a defined
7372 result. Historically some architectures did not provide a defined result for
7373 zero values as efficiently, and many algorithms are now predicated on
7374 avoiding zero-value inputs.</p>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]7375
7376<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7377<p>The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant)
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]7378 zeros in a variable, or within each element of a vector.
Chandler Carruthf6bb2782011-12-12 04:36:04 +0000[diff] [blame]7379 If <tt>src == 0</tt> then the result is the size in bits of the type of
7380 <tt>src</tt> if <tt>is_zero_undef == 0</tt> and <tt>undef</tt> otherwise.
7381 For example, <tt>llvm.cttz(2) = 1</tt>.</p>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]7382
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]7383</div>
7384
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7385</div>
7386
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]7387<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7388<h3>
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]7389 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7390</h3>
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]7391
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7392<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7393
7394<p>LLVM provides intrinsics for some arithmetic with overflow operations.</p>
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]7395
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7396<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7397<h4>
7398 <a name="int_sadd_overflow">
7399 '<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics
7400 </a>
7401</h4>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7402
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7403<div>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7404
7405<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7406<p>This is an overloaded intrinsic. You can use <tt>llvm.sadd.with.overflow</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7407 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7408
7409<pre>
7410 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
7411 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
7412 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
7413</pre>
7414
7415<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7416<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7417 a signed addition of the two arguments, and indicate whether an overflow
7418 occurred during the signed summation.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7419
7420<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7421<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7422 be of integer types of any bit width, but they must have the same bit
7423 width. The second element of the result structure must be of
7424 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7425 undergo signed addition.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7426
7427<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7428<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7429 a signed addition of the two variables. They return a structure &mdash; the
7430 first element of which is the signed summation, and the second element of
7431 which is a bit specifying if the signed summation resulted in an
7432 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7433
7434<h5>Examples:</h5>
7435<pre>
7436 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
7437 %sum = extractvalue {i32, i1} %res, 0
7438 %obit = extractvalue {i32, i1} %res, 1
7439 br i1 %obit, label %overflow, label %normal
7440</pre>
7441
7442</div>
7443
7444<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7445<h4>
7446 <a name="int_uadd_overflow">
7447 '<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics
7448 </a>
7449</h4>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7450
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7451<div>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7452
7453<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7454<p>This is an overloaded intrinsic. You can use <tt>llvm.uadd.with.overflow</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7455 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7456
7457<pre>
7458 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
7459 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
7460 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
7461</pre>
7462
7463<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7464<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7465 an unsigned addition of the two arguments, and indicate whether a carry
7466 occurred during the unsigned summation.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7467
7468<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7469<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7470 be of integer types of any bit width, but they must have the same bit
7471 width. The second element of the result structure must be of
7472 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7473 undergo unsigned addition.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7474
7475<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7476<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7477 an unsigned addition of the two arguments. They return a structure &mdash;
7478 the first element of which is the sum, and the second element of which is a
7479 bit specifying if the unsigned summation resulted in a carry.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7480
7481<h5>Examples:</h5>
7482<pre>
7483 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
7484 %sum = extractvalue {i32, i1} %res, 0
7485 %obit = extractvalue {i32, i1} %res, 1
7486 br i1 %obit, label %carry, label %normal
7487</pre>
7488
7489</div>
7490
7491<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7492<h4>
7493 <a name="int_ssub_overflow">
7494 '<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics
7495 </a>
7496</h4>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7497
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7498<div>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7499
7500<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7501<p>This is an overloaded intrinsic. You can use <tt>llvm.ssub.with.overflow</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7502 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7503
7504<pre>
7505 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
7506 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
7507 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
7508</pre>
7509
7510<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7511<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7512 a signed subtraction of the two arguments, and indicate whether an overflow
7513 occurred during the signed subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7514
7515<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7516<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7517 be of integer types of any bit width, but they must have the same bit
7518 width. The second element of the result structure must be of
7519 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7520 undergo signed subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7521
7522<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7523<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7524 a signed subtraction of the two arguments. They return a structure &mdash;
7525 the first element of which is the subtraction, and the second element of
7526 which is a bit specifying if the signed subtraction resulted in an
7527 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7528
7529<h5>Examples:</h5>
7530<pre>
7531 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
7532 %sum = extractvalue {i32, i1} %res, 0
7533 %obit = extractvalue {i32, i1} %res, 1
7534 br i1 %obit, label %overflow, label %normal
7535</pre>
7536
7537</div>
7538
7539<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7540<h4>
7541 <a name="int_usub_overflow">
7542 '<tt>llvm.usub.with.overflow.*</tt>' Intrinsics
7543 </a>
7544</h4>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7545
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7546<div>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7547
7548<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7549<p>This is an overloaded intrinsic. You can use <tt>llvm.usub.with.overflow</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7550 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7551
7552<pre>
7553 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
7554 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
7555 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
7556</pre>
7557
7558<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7559<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7560 an unsigned subtraction of the two arguments, and indicate whether an
7561 overflow occurred during the unsigned subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7562
7563<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7564<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7565 be of integer types of any bit width, but they must have the same bit
7566 width. The second element of the result structure must be of
7567 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7568 undergo unsigned subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7569
7570<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7571<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7572 an unsigned subtraction of the two arguments. They return a structure &mdash;
7573 the first element of which is the subtraction, and the second element of
7574 which is a bit specifying if the unsigned subtraction resulted in an
7575 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7576
7577<h5>Examples:</h5>
7578<pre>
7579 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
7580 %sum = extractvalue {i32, i1} %res, 0
7581 %obit = extractvalue {i32, i1} %res, 1
7582 br i1 %obit, label %overflow, label %normal
7583</pre>
7584
7585</div>
7586
7587<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7588<h4>
7589 <a name="int_smul_overflow">
7590 '<tt>llvm.smul.with.overflow.*</tt>' Intrinsics
7591 </a>
7592</h4>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7593
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7594<div>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7595
7596<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7597<p>This is an overloaded intrinsic. You can use <tt>llvm.smul.with.overflow</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7598 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7599
7600<pre>
7601 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
7602 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
7603 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
7604</pre>
7605
7606<h5>Overview:</h5>
7607
7608<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7609 a signed multiplication of the two arguments, and indicate whether an
7610 overflow occurred during the signed multiplication.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7611
7612<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7613<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7614 be of integer types of any bit width, but they must have the same bit
7615 width. The second element of the result structure must be of
7616 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7617 undergo signed multiplication.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7618
7619<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7620<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7621 a signed multiplication of the two arguments. They return a structure &mdash;
7622 the first element of which is the multiplication, and the second element of
7623 which is a bit specifying if the signed multiplication resulted in an
7624 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7625
7626<h5>Examples:</h5>
7627<pre>
7628 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
7629 %sum = extractvalue {i32, i1} %res, 0
7630 %obit = extractvalue {i32, i1} %res, 1
7631 br i1 %obit, label %overflow, label %normal
7632</pre>
7633
Reid Spencer5bf54c82007-04-11 23:23:49 +0000[diff] [blame]7634</div>
7635
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7636<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7637<h4>
7638 <a name="int_umul_overflow">
7639 '<tt>llvm.umul.with.overflow.*</tt>' Intrinsics
7640 </a>
7641</h4>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7642
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7643<div>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7644
7645<h5>Syntax:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7646<p>This is an overloaded intrinsic. You can use <tt>llvm.umul.with.overflow</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7647 on any integer bit width.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7648
7649<pre>
7650 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
7651 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7652 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
7653</pre>
7654
7655<h5>Overview:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7656<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7657 a unsigned multiplication of the two arguments, and indicate whether an
7658 overflow occurred during the unsigned multiplication.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7659
7660<h5>Arguments:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7661<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7662 be of integer types of any bit width, but they must have the same bit
7663 width. The second element of the result structure must be of
7664 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7665 undergo unsigned multiplication.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7666
7667<h5>Semantics:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7668<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7669 an unsigned multiplication of the two arguments. They return a structure
7670 &mdash; the first element of which is the multiplication, and the second
7671 element of which is a bit specifying if the unsigned multiplication resulted
7672 in an overflow.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7673
7674<h5>Examples:</h5>
7675<pre>
7676 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7677 %sum = extractvalue {i32, i1} %res, 0
7678 %obit = extractvalue {i32, i1} %res, 1
7679 br i1 %obit, label %overflow, label %normal
7680</pre>
7681
7682</div>
7683
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7684</div>
7685
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]7686<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7687<h3>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7688 <a name="int_fp16">Half Precision Floating Point Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7689</h3>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7690
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7691<div>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7692
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]7693<p>Half precision floating point is a storage-only format. This means that it is
7694 a dense encoding (in memory) but does not support computation in the
7695 format.</p>
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]7696
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]7697<p>This means that code must first load the half-precision floating point
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]7698 value as an i16, then convert it to float with <a
7699 href="#int_convert_from_fp16"><tt>llvm.convert.from.fp16</tt></a>.
7700 Computation can then be performed on the float value (including extending to
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]7701 double etc). To store the value back to memory, it is first converted to
7702 float if needed, then converted to i16 with
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]7703 <a href="#int_convert_to_fp16"><tt>llvm.convert.to.fp16</tt></a>, then
7704 storing as an i16 value.</p>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7705
7706<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7707<h4>
7708 <a name="int_convert_to_fp16">
7709 '<tt>llvm.convert.to.fp16</tt>' Intrinsic
7710 </a>
7711</h4>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7712
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7713<div>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7714
7715<h5>Syntax:</h5>
7716<pre>
7717 declare i16 @llvm.convert.to.fp16(f32 %a)
7718</pre>
7719
7720<h5>Overview:</h5>
7721<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
7722 a conversion from single precision floating point format to half precision
7723 floating point format.</p>
7724
7725<h5>Arguments:</h5>
7726<p>The intrinsic function contains single argument - the value to be
7727 converted.</p>
7728
7729<h5>Semantics:</h5>
7730<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
7731 a conversion from single precision floating point format to half precision
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]7732 floating point format. The return value is an <tt>i16</tt> which
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]7733 contains the converted number.</p>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7734
7735<h5>Examples:</h5>
7736<pre>
7737 %res = call i16 @llvm.convert.to.fp16(f32 %a)
7738 store i16 %res, i16* @x, align 2
7739</pre>
7740
7741</div>
7742
7743<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7744<h4>
7745 <a name="int_convert_from_fp16">
7746 '<tt>llvm.convert.from.fp16</tt>' Intrinsic
7747 </a>
7748</h4>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7749
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7750<div>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7751
7752<h5>Syntax:</h5>
7753<pre>
7754 declare f32 @llvm.convert.from.fp16(i16 %a)
7755</pre>
7756
7757<h5>Overview:</h5>
7758<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs
7759 a conversion from half precision floating point format to single precision
7760 floating point format.</p>
7761
7762<h5>Arguments:</h5>
7763<p>The intrinsic function contains single argument - the value to be
7764 converted.</p>
7765
7766<h5>Semantics:</h5>
7767<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs a
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]7768 conversion from half single precision floating point format to single
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]7769 precision floating point format. The input half-float value is represented by
7770 an <tt>i16</tt> value.</p>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7771
7772<h5>Examples:</h5>
7773<pre>
7774 %a = load i16* @x, align 2
7775 %res = call f32 @llvm.convert.from.fp16(i16 %a)
7776</pre>
7777
7778</div>
7779
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7780</div>
7781
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7782<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7783<h3>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]7784 <a name="int_debugger">Debugger Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7785</h3>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]7786
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7787<div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]7788
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7789<p>The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt>
7790 prefix), are described in
7791 the <a href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source
7792 Level Debugging</a> document.</p>
7793
7794</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]7795
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]7796<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7797<h3>
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]7798 <a name="int_eh">Exception Handling Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7799</h3>
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]7800
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7801<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7802
7803<p>The LLVM exception handling intrinsics (which all start with
7804 <tt>llvm.eh.</tt> prefix), are described in
7805 the <a href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
7806 Handling</a> document.</p>
7807
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]7808</div>
7809
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7810<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7811<h3>
Duncan Sandsa0984362011-09-06 13:37:06 +0000[diff] [blame]7812 <a name="int_trampoline">Trampoline Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7813</h3>
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7814
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7815<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7816
Duncan Sandsa0984362011-09-06 13:37:06 +0000[diff] [blame]7817<p>These intrinsics make it possible to excise one parameter, marked with
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]7818 the <a href="#nest"><tt>nest</tt></a> attribute, from a function.
7819 The result is a callable
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7820 function pointer lacking the nest parameter - the caller does not need to
7821 provide a value for it. Instead, the value to use is stored in advance in a
7822 "trampoline", a block of memory usually allocated on the stack, which also
7823 contains code to splice the nest value into the argument list. This is used
7824 to implement the GCC nested function address extension.</p>
7825
7826<p>For example, if the function is
7827 <tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
7828 pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as
7829 follows:</p>
7830
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]7831<pre class="doc_code">
Duncan Sands86e01192007-09-11 14:10:23 +0000[diff] [blame]7832 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
7833 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
Duncan Sandsa0984362011-09-06 13:37:06 +0000[diff] [blame]7834 call i8* @llvm.init.trampoline(i8* %tramp1, i8* bitcast (i32 (i8*, i32, i32)* @f to i8*), i8* %nval)
7835 %p = call i8* @llvm.adjust.trampoline(i8* %tramp1)
Duncan Sands86e01192007-09-11 14:10:23 +0000[diff] [blame]7836 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7837</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7838
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7839<p>The call <tt>%val = call i32 %fp(i32 %x, i32 %y)</tt> is then equivalent
7840 to <tt>%val = call i32 %f(i8* %nval, i32 %x, i32 %y)</tt>.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7841
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7842<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7843<h4>
7844 <a name="int_it">
7845 '<tt>llvm.init.trampoline</tt>' Intrinsic
7846 </a>
7847</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7848
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7849<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7850
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7851<h5>Syntax:</h5>
7852<pre>
Duncan Sandsa0984362011-09-06 13:37:06 +0000[diff] [blame]7853 declare void @llvm.init.trampoline(i8* &lt;tramp&gt;, i8* &lt;func&gt;, i8* &lt;nval&gt;)
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7854</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7855
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7856<h5>Overview:</h5>
Duncan Sandsa0984362011-09-06 13:37:06 +0000[diff] [blame]7857<p>This fills the memory pointed to by <tt>tramp</tt> with executable code,
7858 turning it into a trampoline.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7859
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7860<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7861<p>The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
7862 pointers. The <tt>tramp</tt> argument must point to a sufficiently large and
7863 sufficiently aligned block of memory; this memory is written to by the
7864 intrinsic. Note that the size and the alignment are target-specific - LLVM
7865 currently provides no portable way of determining them, so a front-end that
7866 generates this intrinsic needs to have some target-specific knowledge.
7867 The <tt>func</tt> argument must hold a function bitcast to
7868 an <tt>i8*</tt>.</p>
7869
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7870<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7871<p>The block of memory pointed to by <tt>tramp</tt> is filled with target
Duncan Sandsa0984362011-09-06 13:37:06 +0000[diff] [blame]7872 dependent code, turning it into a function. Then <tt>tramp</tt> needs to be
7873 passed to <a href="#int_at">llvm.adjust.trampoline</a> to get a pointer
7874 which can be <a href="#int_trampoline">bitcast (to a new function) and
7875 called</a>. The new function's signature is the same as that of
7876 <tt>func</tt> with any arguments marked with the <tt>nest</tt> attribute
7877 removed. At most one such <tt>nest</tt> argument is allowed, and it must be of
7878 pointer type. Calling the new function is equivalent to calling <tt>func</tt>
7879 with the same argument list, but with <tt>nval</tt> used for the missing
7880 <tt>nest</tt> argument. If, after calling <tt>llvm.init.trampoline</tt>, the
7881 memory pointed to by <tt>tramp</tt> is modified, then the effect of any later call
7882 to the returned function pointer is undefined.</p>
7883</div>
7884
7885<!-- _______________________________________________________________________ -->
7886<h4>
7887 <a name="int_at">
7888 '<tt>llvm.adjust.trampoline</tt>' Intrinsic
7889 </a>
7890</h4>
7891
7892<div>
7893
7894<h5>Syntax:</h5>
7895<pre>
7896 declare i8* @llvm.adjust.trampoline(i8* &lt;tramp&gt;)
7897</pre>
7898
7899<h5>Overview:</h5>
7900<p>This performs any required machine-specific adjustment to the address of a
7901 trampoline (passed as <tt>tramp</tt>).</p>
7902
7903<h5>Arguments:</h5>
7904<p><tt>tramp</tt> must point to a block of memory which already has trampoline code
7905 filled in by a previous call to <a href="#int_it"><tt>llvm.init.trampoline</tt>
7906 </a>.</p>
7907
7908<h5>Semantics:</h5>
7909<p>On some architectures the address of the code to be executed needs to be
7910 different to the address where the trampoline is actually stored. This
7911 intrinsic returns the executable address corresponding to <tt>tramp</tt>
7912 after performing the required machine specific adjustments.
7913 The pointer returned can then be <a href="#int_trampoline"> bitcast and
7914 executed</a>.
7915</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7916
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7917</div>
7918
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7919</div>
7920
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7921<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7922<h3>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7923 <a name="int_memorymarkers">Memory Use Markers</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7924</h3>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7925
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7926<div>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7927
7928<p>This class of intrinsics exists to information about the lifetime of memory
7929 objects and ranges where variables are immutable.</p>
7930
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7931<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7932<h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7933 <a name="int_lifetime_start">'<tt>llvm.lifetime.start</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7934</h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7935
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7936<div>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7937
7938<h5>Syntax:</h5>
7939<pre>
7940 declare void @llvm.lifetime.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7941</pre>
7942
7943<h5>Overview:</h5>
7944<p>The '<tt>llvm.lifetime.start</tt>' intrinsic specifies the start of a memory
7945 object's lifetime.</p>
7946
7947<h5>Arguments:</h5>
Nick Lewycky9bc89042009-10-13 07:57:33 +0000[diff] [blame]7948<p>The first argument is a constant integer representing the size of the
7949 object, or -1 if it is variable sized. The second argument is a pointer to
7950 the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7951
7952<h5>Semantics:</h5>
7953<p>This intrinsic indicates that before this point in the code, the value of the
7954 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
Nick Lewyckyd20fd592009-10-27 16:56:58 +0000[diff] [blame]7955 never be used and has an undefined value. A load from the pointer that
7956 precedes this intrinsic can be replaced with
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7957 <tt>'<a href="#undefvalues">undef</a>'</tt>.</p>
7958
7959</div>
7960
7961<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7962<h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7963 <a name="int_lifetime_end">'<tt>llvm.lifetime.end</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7964</h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7965
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7966<div>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7967
7968<h5>Syntax:</h5>
7969<pre>
7970 declare void @llvm.lifetime.end(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7971</pre>
7972
7973<h5>Overview:</h5>
7974<p>The '<tt>llvm.lifetime.end</tt>' intrinsic specifies the end of a memory
7975 object's lifetime.</p>
7976
7977<h5>Arguments:</h5>
Nick Lewycky9bc89042009-10-13 07:57:33 +0000[diff] [blame]7978<p>The first argument is a constant integer representing the size of the
7979 object, or -1 if it is variable sized. The second argument is a pointer to
7980 the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7981
7982<h5>Semantics:</h5>
7983<p>This intrinsic indicates that after this point in the code, the value of the
7984 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
7985 never be used and has an undefined value. Any stores into the memory object
7986 following this intrinsic may be removed as dead.
7987
7988</div>
7989
7990<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7991<h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7992 <a name="int_invariant_start">'<tt>llvm.invariant.start</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7993</h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7994
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7995<div>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7996
7997<h5>Syntax:</h5>
7998<pre>
Nick Lewycky2965d3e2010-11-30 04:13:41 +0000[diff] [blame]7999 declare {}* @llvm.invariant.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8000</pre>
8001
8002<h5>Overview:</h5>
8003<p>The '<tt>llvm.invariant.start</tt>' intrinsic specifies that the contents of
8004 a memory object will not change.</p>
8005
8006<h5>Arguments:</h5>
Nick Lewycky9bc89042009-10-13 07:57:33 +0000[diff] [blame]8007<p>The first argument is a constant integer representing the size of the
8008 object, or -1 if it is variable sized. The second argument is a pointer to
8009 the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8010
8011<h5>Semantics:</h5>
8012<p>This intrinsic indicates that until an <tt>llvm.invariant.end</tt> that uses
8013 the return value, the referenced memory location is constant and
8014 unchanging.</p>
8015
8016</div>
8017
8018<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8019<h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8020 <a name="int_invariant_end">'<tt>llvm.invariant.end</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8021</h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8022
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8023<div>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8024
8025<h5>Syntax:</h5>
8026<pre>
8027 declare void @llvm.invariant.end({}* &lt;start&gt;, i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
8028</pre>
8029
8030<h5>Overview:</h5>
8031<p>The '<tt>llvm.invariant.end</tt>' intrinsic specifies that the contents of
8032 a memory object are mutable.</p>
8033
8034<h5>Arguments:</h5>
8035<p>The first argument is the matching <tt>llvm.invariant.start</tt> intrinsic.
Nick Lewycky9bc89042009-10-13 07:57:33 +0000[diff] [blame]8036 The second argument is a constant integer representing the size of the
8037 object, or -1 if it is variable sized and the third argument is a pointer
8038 to the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8039
8040<h5>Semantics:</h5>
8041<p>This intrinsic indicates that the memory is mutable again.</p>
8042
8043</div>
8044
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8045</div>
8046
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]8047<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8048<h3>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]8049 <a name="int_general">General Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8050</h3>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]8051
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8052<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8053
8054<p>This class of intrinsics is designed to be generic and has no specific
8055 purpose.</p>
8056
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]8057<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8058<h4>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]8059 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8060</h4>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]8061
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8062<div>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]8063
8064<h5>Syntax:</h5>
8065<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8066 declare void @llvm.var.annotation(i8* &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]8067</pre>
8068
8069<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8070<p>The '<tt>llvm.var.annotation</tt>' intrinsic.</p>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]8071
8072<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8073<p>The first argument is a pointer to a value, the second is a pointer to a
8074 global string, the third is a pointer to a global string which is the source
8075 file name, and the last argument is the line number.</p>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]8076
8077<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8078<p>This intrinsic allows annotation of local variables with arbitrary strings.
8079 This can be useful for special purpose optimizations that want to look for
John Criswellf0d536a2011-08-19 16:57:55 +0000[diff] [blame]8080 these annotations. These have no other defined use; they are ignored by code
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8081 generation and optimization.</p>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]8082
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]8083</div>
8084
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]8085<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8086<h4>
Tanya Lattner0186a652007-09-21 23:57:59 +0000[diff] [blame]8087 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8088</h4>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]8089
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8090<div>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]8091
8092<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8093<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
8094 any integer bit width.</p>
8095
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]8096<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8097 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8098 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8099 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8100 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8101 declare i256 @llvm.annotation.i256(i256 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]8102</pre>
8103
8104<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8105<p>The '<tt>llvm.annotation</tt>' intrinsic.</p>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]8106
8107<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8108<p>The first argument is an integer value (result of some expression), the
8109 second is a pointer to a global string, the third is a pointer to a global
8110 string which is the source file name, and the last argument is the line
8111 number. It returns the value of the first argument.</p>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]8112
8113<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8114<p>This intrinsic allows annotations to be put on arbitrary expressions with
8115 arbitrary strings. This can be useful for special purpose optimizations that
John Criswellf0d536a2011-08-19 16:57:55 +0000[diff] [blame]8116 want to look for these annotations. These have no other defined use; they
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8117 are ignored by code generation and optimization.</p>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]8118
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]8119</div>
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]8120
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]8121<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8122<h4>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]8123 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8124</h4>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]8125
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8126<div>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]8127
8128<h5>Syntax:</h5>
8129<pre>
8130 declare void @llvm.trap()
8131</pre>
8132
8133<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8134<p>The '<tt>llvm.trap</tt>' intrinsic.</p>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]8135
8136<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8137<p>None.</p>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]8138
8139<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8140<p>This intrinsics is lowered to the target dependent trap instruction. If the
8141 target does not have a trap instruction, this intrinsic will be lowered to
8142 the call of the <tt>abort()</tt> function.</p>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]8143
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]8144</div>
8145
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]8146<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8147<h4>
Misha Brukman50de2b22008-11-22 23:55:29 +0000[diff] [blame]8148 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8149</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8150
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8151<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8152
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]8153<h5>Syntax:</h5>
8154<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8155 declare void @llvm.stackprotector(i8* &lt;guard&gt;, i8** &lt;slot&gt;)
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]8156</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8157
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]8158<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8159<p>The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and
8160 stores it onto the stack at <tt>slot</tt>. The stack slot is adjusted to
8161 ensure that it is placed on the stack before local variables.</p>
8162
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]8163<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8164<p>The <tt>llvm.stackprotector</tt> intrinsic requires two pointer
8165 arguments. The first argument is the value loaded from the stack
8166 guard <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt>
8167 that has enough space to hold the value of the guard.</p>
8168
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]8169<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8170<p>This intrinsic causes the prologue/epilogue inserter to force the position of
8171 the <tt>AllocaInst</tt> stack slot to be before local variables on the
8172 stack. This is to ensure that if a local variable on the stack is
8173 overwritten, it will destroy the value of the guard. When the function exits,
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]8174 the guard on the stack is checked against the original guard. If they are
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8175 different, then the program aborts by calling the <tt>__stack_chk_fail()</tt>
8176 function.</p>
8177
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]8178</div>
8179
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8180<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8181<h4>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8182 <a name="int_objectsize">'<tt>llvm.objectsize</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8183</h4>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8184
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8185<div>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8186
8187<h5>Syntax:</h5>
8188<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8189 declare i32 @llvm.objectsize.i32(i8* &lt;object&gt;, i1 &lt;type&gt;)
8190 declare i64 @llvm.objectsize.i64(i8* &lt;object&gt;, i1 &lt;type&gt;)
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8191</pre>
8192
8193<h5>Overview:</h5>
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]8194<p>The <tt>llvm.objectsize</tt> intrinsic is designed to provide information to
8195 the optimizers to determine at compile time whether a) an operation (like
8196 memcpy) will overflow a buffer that corresponds to an object, or b) that a
8197 runtime check for overflow isn't necessary. An object in this context means
8198 an allocation of a specific class, structure, array, or other object.</p>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8199
8200<h5>Arguments:</h5>
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]8201<p>The <tt>llvm.objectsize</tt> intrinsic takes two arguments. The first
Eric Christopher31e39bd2009-12-23 00:29:49 +0000[diff] [blame]8202 argument is a pointer to or into the <tt>object</tt>. The second argument
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]8203 is a boolean 0 or 1. This argument determines whether you want the
8204 maximum (0) or minimum (1) bytes remaining. This needs to be a literal 0 or
Eric Christopher31e39bd2009-12-23 00:29:49 +0000[diff] [blame]8205 1, variables are not allowed.</p>
8206
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8207<h5>Semantics:</h5>
8208<p>The <tt>llvm.objectsize</tt> intrinsic is lowered to either a constant
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]8209 representing the size of the object concerned, or <tt>i32/i64 -1 or 0</tt>,
8210 depending on the <tt>type</tt> argument, if the size cannot be determined at
8211 compile time.</p>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8212
8213</div>
Jakub Staszak5fef7922011-12-04 18:29:26 +0000[diff] [blame]8214<!-- _______________________________________________________________________ -->
8215<h4>
8216 <a name="int_expect">'<tt>llvm.expect</tt>' Intrinsic</a>
8217</h4>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8218
Jakub Staszak5fef7922011-12-04 18:29:26 +0000[diff] [blame]8219<div>
8220
8221<h5>Syntax:</h5>
8222<pre>
8223 declare i32 @llvm.expect.i32(i32 &lt;val&gt;, i32 &lt;expected_val&gt;)
8224 declare i64 @llvm.expect.i64(i64 &lt;val&gt;, i64 &lt;expected_val&gt;)
8225</pre>
8226
8227<h5>Overview:</h5>
8228<p>The <tt>llvm.expect</tt> intrinsic provides information about expected (the
8229 most probable) value of <tt>val</tt>, which can be used by optimizers.</p>
8230
8231<h5>Arguments:</h5>
8232<p>The <tt>llvm.expect</tt> intrinsic takes two arguments. The first
8233 argument is a value. The second argument is an expected value, this needs to
8234 be a constant value, variables are not allowed.</p>
8235
8236<h5>Semantics:</h5>
8237<p>This intrinsic is lowered to the <tt>val</tt>.</p>
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8238</div>
8239
8240</div>
8241
Jakub Staszak5fef7922011-12-04 18:29:26 +0000[diff] [blame]8242</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]8243<!-- *********************************************************************** -->
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]8244<hr>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]8245<address>
8246 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman86242e12008-12-11 17:34:48 +0000[diff] [blame]8247 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]8248 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman86242e12008-12-11 17:34:48 +0000[diff] [blame]8249 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]8250
8251 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
NAKAMURA Takumica46f5a2011-04-09 02:13:37 +0000[diff] [blame]8252 <a href="http://llvm.org/">The LLVM Compiler Infrastructure</a><br>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]8253 Last modified: $Date$
8254</address>
Chris Lattnerb8f816e2008-01-04 04:33:49 +0000[diff] [blame]8255
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]8256</body>
8257</html>