Gitiles
Code Review Sign In
gerrit-public.fairphone.software / toolchain / llvm-project / 95f69a42d279f122d55374020f7c253ece40950d / . / llvm / docs / LangRef.html
blob: fa4d74b67f2b855cba59845a1a95844be3106dc7 [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>
38 <li><a href="#linkage_external">'<tt>externally visible</tt>' Linkage</a></li>
39 <li><a href="#linkage_dllimport">'<tt>dllimport</tt>' Linkage</a></li>
40 <li><a href="#linkage_dllexport">'<tt>dllexport</tt>' Linkage</a></li>
Bill 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 Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]95 <li><a href="#trapvalues">Trap 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>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]103 <li><a href="#metadata">Metadata Nodes and Metadata Strings</a></li>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]104 </ol>
105 </li>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]106 <li><a href="#intrinsic_globals">Intrinsic Global Variables</a>
107 <ol>
108 <li><a href="#intg_used">The '<tt>llvm.used</tt>' Global Variable</a></li>
Chris Lattner58f9bb22009-07-20 06:14:25 +0000[diff] [blame]109 <li><a href="#intg_compiler_used">The '<tt>llvm.compiler.used</tt>'
110 Global Variable</a></li>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]111 <li><a href="#intg_global_ctors">The '<tt>llvm.global_ctors</tt>'
112 Global Variable</a></li>
113 <li><a href="#intg_global_dtors">The '<tt>llvm.global_dtors</tt>'
114 Global Variable</a></li>
115 </ol>
116 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]117 <li><a href="#instref">Instruction Reference</a>
118 <ol>
119 <li><a href="#terminators">Terminator Instructions</a>
120 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]121 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
122 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]123 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]124 <li><a href="#i_indirectbr">'<tt>indirectbr</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]125 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]126 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
Bill Wendlingf891bf82011-07-31 06:30:59 +0000[diff] [blame]127 <li><a href="#i_resume">'<tt>resume</tt>' Instruction</a></li>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]128 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]129 </ol>
130 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]131 <li><a href="#binaryops">Binary Operations</a>
132 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]133 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]134 <li><a href="#i_fadd">'<tt>fadd</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]135 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]136 <li><a href="#i_fsub">'<tt>fsub</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]137 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]138 <li><a href="#i_fmul">'<tt>fmul</tt>' Instruction</a></li>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]139 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
140 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
141 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]142 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
143 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
144 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]145 </ol>
146 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]147 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
148 <ol>
Reid Spencer2ab01932007-02-02 13:57:07 +0000[diff] [blame]149 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
150 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
151 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]152 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]153 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]154 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]155 </ol>
156 </li>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]157 <li><a href="#vectorops">Vector Operations</a>
158 <ol>
159 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
160 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
161 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]162 </ol>
163 </li>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]164 <li><a href="#aggregateops">Aggregate Operations</a>
165 <ol>
166 <li><a href="#i_extractvalue">'<tt>extractvalue</tt>' Instruction</a></li>
167 <li><a href="#i_insertvalue">'<tt>insertvalue</tt>' Instruction</a></li>
168 </ol>
169 </li>
Chris Lattner6ab66722006-08-15 00:45:58 +0000[diff] [blame]170 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]171 <ol>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]172 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
173 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
174 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
175 <li><a href="#i_fence">'<tt>fence</tt>' Instruction</a></li>
176 <li><a href="#i_cmpxchg">'<tt>cmpxchg</tt>' Instruction</a></li>
177 <li><a href="#i_atomicrmw">'<tt>atomicrmw</tt>' Instruction</a></li>
Robert Bocchino820bc75b2006-02-17 21:18:08 +0000[diff] [blame]178 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]179 </ol>
180 </li>
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]181 <li><a href="#convertops">Conversion Operations</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]182 <ol>
183 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
184 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
185 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
186 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
187 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]188 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
189 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
190 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
191 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]192 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
193 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]194 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]195 </ol>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]196 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]197 <li><a href="#otherops">Other Operations</a>
198 <ol>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]199 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
200 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]201 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]202 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]203 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Chris Lattner33337472006-01-13 23:26:01 +0000[diff] [blame]204 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]205 <li><a href="#i_landingpad">'<tt>landingpad</tt>' Instruction</a></li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]206 </ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]207 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]208 </ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]209 </li>
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]210 <li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]211 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]212 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
213 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]214 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
215 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
216 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]217 </ol>
218 </li>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]219 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
220 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]221 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
222 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
223 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]224 </ol>
225 </li>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]226 <li><a href="#int_codegen">Code Generator Intrinsics</a>
227 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]228 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
229 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
230 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
231 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
232 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
233 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
Dan Gohmane58f7b32010-05-26 21:56:15 +0000[diff] [blame]234 <li><a href="#int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswellaa1c3c12004-04-09 16:43:20 +0000[diff] [blame]235 </ol>
236 </li>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]237 <li><a href="#int_libc">Standard C Library Intrinsics</a>
238 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]239 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
240 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
241 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
242 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
243 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]244 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
245 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
246 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Dan Gohmane635c522011-05-27 00:36:31 +0000[diff] [blame]247 <li><a href="#int_exp">'<tt>llvm.exp.*</tt>' Intrinsic</a></li>
248 <li><a href="#int_log">'<tt>llvm.log.*</tt>' Intrinsic</a></li>
Cameron Zwarichf03fa182011-07-08 21:39:21 +0000[diff] [blame]249 <li><a href="#int_fma">'<tt>llvm.fma.*</tt>' Intrinsic</a></li>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]250 </ol>
251 </li>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]252 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]253 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]254 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattnerb748c672006-01-16 22:34:14 +0000[diff] [blame]255 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
256 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
257 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]258 </ol>
259 </li>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]260 <li><a href="#int_overflow">Arithmetic with Overflow Intrinsics</a>
261 <ol>
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]262 <li><a href="#int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt> Intrinsics</a></li>
263 <li><a href="#int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt> Intrinsics</a></li>
264 <li><a href="#int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt> Intrinsics</a></li>
265 <li><a href="#int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt> Intrinsics</a></li>
266 <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]267 <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]268 </ol>
269 </li>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]270 <li><a href="#int_fp16">Half Precision Floating Point Intrinsics</a>
271 <ol>
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]272 <li><a href="#int_convert_to_fp16">'<tt>llvm.convert.to.fp16</tt>' Intrinsic</a></li>
273 <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]274 </ol>
275 </li>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]276 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]277 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sands86e01192007-09-11 14:10:23 +0000[diff] [blame]278 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]279 <ol>
280 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]281 </ol>
282 </li>
Bill Wendlingf85850f2008-11-18 22:10:53 +0000[diff] [blame]283 <li><a href="#int_atomics">Atomic intrinsics</a>
284 <ol>
285 <li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
286 <li><a href="#int_atomic_cmp_swap"><tt>llvm.atomic.cmp.swap</tt></a></li>
287 <li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
288 <li><a href="#int_atomic_load_add"><tt>llvm.atomic.load.add</tt></a></li>
289 <li><a href="#int_atomic_load_sub"><tt>llvm.atomic.load.sub</tt></a></li>
290 <li><a href="#int_atomic_load_and"><tt>llvm.atomic.load.and</tt></a></li>
291 <li><a href="#int_atomic_load_nand"><tt>llvm.atomic.load.nand</tt></a></li>
292 <li><a href="#int_atomic_load_or"><tt>llvm.atomic.load.or</tt></a></li>
293 <li><a href="#int_atomic_load_xor"><tt>llvm.atomic.load.xor</tt></a></li>
294 <li><a href="#int_atomic_load_max"><tt>llvm.atomic.load.max</tt></a></li>
295 <li><a href="#int_atomic_load_min"><tt>llvm.atomic.load.min</tt></a></li>
296 <li><a href="#int_atomic_load_umax"><tt>llvm.atomic.load.umax</tt></a></li>
297 <li><a href="#int_atomic_load_umin"><tt>llvm.atomic.load.umin</tt></a></li>
298 </ol>
299 </li>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]300 <li><a href="#int_memorymarkers">Memory Use Markers</a>
301 <ol>
302 <li><a href="#int_lifetime_start"><tt>llvm.lifetime.start</tt></a></li>
303 <li><a href="#int_lifetime_end"><tt>llvm.lifetime.end</tt></a></li>
304 <li><a href="#int_invariant_start"><tt>llvm.invariant.start</tt></a></li>
305 <li><a href="#int_invariant_end"><tt>llvm.invariant.end</tt></a></li>
306 </ol>
307 </li>
Reid Spencer5b2cb0f2007-07-20 19:59:11 +0000[diff] [blame]308 <li><a href="#int_general">General intrinsics</a>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]309 <ol>
Reid Spencer5b2cb0f2007-07-20 19:59:11 +0000[diff] [blame]310 <li><a href="#int_var_annotation">
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]311 '<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]312 <li><a href="#int_annotation">
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]313 '<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]314 <li><a href="#int_trap">
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]315 '<tt>llvm.trap</tt>' Intrinsic</a></li>
316 <li><a href="#int_stackprotector">
317 '<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]318 <li><a href="#int_objectsize">
319 '<tt>llvm.objectsize</tt>' Intrinsic</a></li>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]320 </ol>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]321 </li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]322 </ol>
323 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]324</ol>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]325
326<div class="doc_author">
327 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
328 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]329</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]330
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]331<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]332<h2><a name="abstract">Abstract</a></h2>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]333<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]334
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]335<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]336
337<p>This document is a reference manual for the LLVM assembly language. LLVM is
338 a Static Single Assignment (SSA) based representation that provides type
339 safety, low-level operations, flexibility, and the capability of representing
340 'all' high-level languages cleanly. It is the common code representation
341 used throughout all phases of the LLVM compilation strategy.</p>
342
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]343</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]344
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]345<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]346<h2><a name="introduction">Introduction</a></h2>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]347<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]348
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]349<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]350
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]351<p>The LLVM code representation is designed to be used in three different forms:
352 as an in-memory compiler IR, as an on-disk bitcode representation (suitable
353 for fast loading by a Just-In-Time compiler), and as a human readable
354 assembly language representation. This allows LLVM to provide a powerful
355 intermediate representation for efficient compiler transformations and
356 analysis, while providing a natural means to debug and visualize the
357 transformations. The three different forms of LLVM are all equivalent. This
358 document describes the human readable representation and notation.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]359
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]360<p>The LLVM representation aims to be light-weight and low-level while being
361 expressive, typed, and extensible at the same time. It aims to be a
362 "universal IR" of sorts, by being at a low enough level that high-level ideas
363 may be cleanly mapped to it (similar to how microprocessors are "universal
364 IR's", allowing many source languages to be mapped to them). By providing
365 type information, LLVM can be used as the target of optimizations: for
366 example, through pointer analysis, it can be proven that a C automatic
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]367 variable is never accessed outside of the current function, allowing it to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]368 be promoted to a simple SSA value instead of a memory location.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]369
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]370<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]371<h4>
372 <a name="wellformed">Well-Formedness</a>
373</h4>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]374
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]375<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]376
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]377<p>It is important to note that this document describes 'well formed' LLVM
378 assembly language. There is a difference between what the parser accepts and
379 what is considered 'well formed'. For example, the following instruction is
380 syntactically okay, but not well formed:</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]381
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]382<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]383%x = <a href="#i_add">add</a> i32 1, %x
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]384</pre>
385
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]386<p>because the definition of <tt>%x</tt> does not dominate all of its uses. The
387 LLVM infrastructure provides a verification pass that may be used to verify
388 that an LLVM module is well formed. This pass is automatically run by the
389 parser after parsing input assembly and by the optimizer before it outputs
390 bitcode. The violations pointed out by the verifier pass indicate bugs in
391 transformation passes or input to the parser.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]392
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]393</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]394
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]395</div>
396
Chris Lattner87a3dbe2007-10-03 17:34:29 +0000[diff] [blame]397<!-- Describe the typesetting conventions here. -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]398
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]399<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]400<h2><a name="identifiers">Identifiers</a></h2>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]401<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]402
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]403<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]404
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]405<p>LLVM identifiers come in two basic types: global and local. Global
406 identifiers (functions, global variables) begin with the <tt>'@'</tt>
407 character. Local identifiers (register names, types) begin with
408 the <tt>'%'</tt> character. Additionally, there are three different formats
409 for identifiers, for different purposes:</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]410
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]411<ol>
Reid Spencerb23b65f2007-08-07 14:34:28 +0000[diff] [blame]412 <li>Named values are represented as a string of characters with their prefix.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]413 For example, <tt>%foo</tt>, <tt>@DivisionByZero</tt>,
414 <tt>%a.really.long.identifier</tt>. The actual regular expression used is
415 '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'. Identifiers which require
416 other characters in their names can be surrounded with quotes. Special
417 characters may be escaped using <tt>"\xx"</tt> where <tt>xx</tt> is the
418 ASCII code for the character in hexadecimal. In this way, any character
419 can be used in a name value, even quotes themselves.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]420
Reid Spencerb23b65f2007-08-07 14:34:28 +0000[diff] [blame]421 <li>Unnamed values are represented as an unsigned numeric value with their
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]422 prefix. For example, <tt>%12</tt>, <tt>@2</tt>, <tt>%44</tt>.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]423
Reid Spencer8f08d802004-12-09 18:02:53 +0000[diff] [blame]424 <li>Constants, which are described in a <a href="#constants">section about
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]425 constants</a>, below.</li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]426</ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]427
Reid Spencerb23b65f2007-08-07 14:34:28 +0000[diff] [blame]428<p>LLVM requires that values start with a prefix for two reasons: Compilers
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]429 don't need to worry about name clashes with reserved words, and the set of
430 reserved words may be expanded in the future without penalty. Additionally,
431 unnamed identifiers allow a compiler to quickly come up with a temporary
432 variable without having to avoid symbol table conflicts.</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]433
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]434<p>Reserved words in LLVM are very similar to reserved words in other
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]435 languages. There are keywords for different opcodes
436 ('<tt><a href="#i_add">add</a></tt>',
437 '<tt><a href="#i_bitcast">bitcast</a></tt>',
438 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names
439 ('<tt><a href="#t_void">void</a></tt>',
440 '<tt><a href="#t_primitive">i32</a></tt>', etc...), and others. These
441 reserved words cannot conflict with variable names, because none of them
442 start with a prefix character (<tt>'%'</tt> or <tt>'@'</tt>).</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]443
444<p>Here is an example of LLVM code to multiply the integer variable
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]445 '<tt>%X</tt>' by 8:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]446
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]447<p>The easy way:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]448
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]449<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]450%result = <a href="#i_mul">mul</a> i32 %X, 8
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]451</pre>
452
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]453<p>After strength reduction:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]454
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]455<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]456%result = <a href="#i_shl">shl</a> i32 %X, i8 3
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]457</pre>
458
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]459<p>And the hard way:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]460
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]461<pre class="doc_code">
Gabor Greifbd0328f2009-10-28 13:05:07 +0000[diff] [blame]462%0 = <a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
463%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]464%result = <a href="#i_add">add</a> i32 %1, %1
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]465</pre>
466
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]467<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several important
468 lexical features of LLVM:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]469
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]470<ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]471 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]472 line.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]473
474 <li>Unnamed temporaries are created when the result of a computation is not
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]475 assigned to a named value.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]476
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]477 <li>Unnamed temporaries are numbered sequentially</li>
478</ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]479
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]480<p>It also shows a convention that we follow in this document. When
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]481 demonstrating instructions, we will follow an instruction with a comment that
482 defines the type and name of value produced. Comments are shown in italic
483 text.</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]484
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]485</div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]486
487<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]488<h2><a name="highlevel">High Level Structure</a></h2>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]489<!-- *********************************************************************** -->
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]490<div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]491<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]492<h3>
493 <a name="modulestructure">Module Structure</a>
494</h3>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]495
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]496<div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]497
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]498<p>LLVM programs are composed of "Module"s, each of which is a translation unit
499 of the input programs. Each module consists of functions, global variables,
500 and symbol table entries. Modules may be combined together with the LLVM
501 linker, which merges function (and global variable) definitions, resolves
502 forward declarations, and merges symbol table entries. Here is an example of
503 the "hello world" module:</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]504
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]505<pre class="doc_code">
Chris Lattner54a7be72010-08-17 17:13:42 +0000[diff] [blame]506<i>; Declare the string constant as a global constant.</i>&nbsp;
Nick Lewyckyfea7ddc2011-01-29 01:09:53 +0000[diff] [blame]507<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]508
Chris Lattner54a7be72010-08-17 17:13:42 +0000[diff] [blame]509<i>; External declaration of the puts function</i>&nbsp;
510<a href="#functionstructure">declare</a> i32 @puts(i8*) <i>; i32 (i8*)* </i>&nbsp;
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]511
512<i>; Definition of main function</i>
Chris Lattner54a7be72010-08-17 17:13:42 +0000[diff] [blame]513define i32 @main() { <i>; i32()* </i>&nbsp;
514 <i>; Convert [13 x i8]* to i8 *...</i>&nbsp;
515 %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]516
Chris Lattner54a7be72010-08-17 17:13:42 +0000[diff] [blame]517 <i>; Call puts function to write out the string to stdout.</i>&nbsp;
518 <a href="#i_call">call</a> i32 @puts(i8* %cast210) <i>; i32</i>&nbsp;
519 <a href="#i_ret">ret</a> i32 0&nbsp;
520}
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]521
522<i>; Named metadata</i>
523!1 = metadata !{i32 41}
524!foo = !{!1, null}
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]525</pre>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]526
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]527<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]528 "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>" function,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]529 a <a href="#functionstructure">function definition</a> for
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]530 "<tt>main</tt>" and <a href="#namedmetadatastructure">named metadata</a>
531 "<tt>foo"</tt>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]532
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]533<p>In general, a module is made up of a list of global values, where both
534 functions and global variables are global values. Global values are
535 represented by a pointer to a memory location (in this case, a pointer to an
536 array of char, and a pointer to a function), and have one of the
537 following <a href="#linkage">linkage types</a>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]538
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]539</div>
540
541<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]542<h3>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]543 <a name="linkage">Linkage Types</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]544</h3>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]545
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]546<div>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]547
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]548<p>All Global Variables and Functions have one of the following types of
549 linkage:</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]550
551<dl>
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]552 <dt><tt><b><a name="linkage_private">private</a></b></tt></dt>
Bill Wendling03bcd6e2010-07-01 21:55:59 +0000[diff] [blame]553 <dd>Global values with "<tt>private</tt>" linkage are only directly accessible
554 by objects in the current module. In particular, linking code into a
555 module with an private global value may cause the private to be renamed as
556 necessary to avoid collisions. Because the symbol is private to the
557 module, all references can be updated. This doesn't show up in any symbol
558 table in the object file.</dd>
Rafael Espindola6de96a12009-01-15 20:18:42 +0000[diff] [blame]559
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]560 <dt><tt><b><a name="linkage_linker_private">linker_private</a></b></tt></dt>
Bill Wendling03bcd6e2010-07-01 21:55:59 +0000[diff] [blame]561 <dd>Similar to <tt>private</tt>, but the symbol is passed through the
562 assembler and evaluated by the linker. Unlike normal strong symbols, they
563 are removed by the linker from the final linked image (executable or
564 dynamic library).</dd>
565
566 <dt><tt><b><a name="linkage_linker_private_weak">linker_private_weak</a></b></tt></dt>
567 <dd>Similar to "<tt>linker_private</tt>", but the symbol is weak. Note that
568 <tt>linker_private_weak</tt> symbols are subject to coalescing by the
569 linker. The symbols are removed by the linker from the final linked image
570 (executable or dynamic library).</dd>
Bill Wendlinga3c6f6b2009-07-20 01:03:30 +0000[diff] [blame]571
Bill Wendling578ee402010-08-20 22:05:50 +0000[diff] [blame]572 <dt><tt><b><a name="linkage_linker_private_weak_def_auto">linker_private_weak_def_auto</a></b></tt></dt>
573 <dd>Similar to "<tt>linker_private_weak</tt>", but it's known that the address
574 of the object is not taken. For instance, functions that had an inline
575 definition, but the compiler decided not to inline it. Note,
576 unlike <tt>linker_private</tt> and <tt>linker_private_weak</tt>,
577 <tt>linker_private_weak_def_auto</tt> may have only <tt>default</tt>
578 visibility. The symbols are removed by the linker from the final linked
579 image (executable or dynamic library).</dd>
580
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]581 <dt><tt><b><a name="linkage_internal">internal</a></b></tt></dt>
Bill Wendling36321712010-06-29 22:34:52 +0000[diff] [blame]582 <dd>Similar to private, but the value shows as a local symbol
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]583 (<tt>STB_LOCAL</tt> in the case of ELF) in the object file. This
584 corresponds to the notion of the '<tt>static</tt>' keyword in C.</dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]585
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]586 <dt><tt><b><a name="linkage_available_externally">available_externally</a></b></tt></dt>
Chris Lattner184f1be2009-04-13 05:44:34 +0000[diff] [blame]587 <dd>Globals with "<tt>available_externally</tt>" linkage are never emitted
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]588 into the object file corresponding to the LLVM module. They exist to
589 allow inlining and other optimizations to take place given knowledge of
590 the definition of the global, which is known to be somewhere outside the
591 module. Globals with <tt>available_externally</tt> linkage are allowed to
592 be discarded at will, and are otherwise the same as <tt>linkonce_odr</tt>.
593 This linkage type is only allowed on definitions, not declarations.</dd>
Chris Lattner184f1be2009-04-13 05:44:34 +0000[diff] [blame]594
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]595 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt></dt>
Chris Lattnere20b4702007-01-14 06:51:48 +0000[diff] [blame]596 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
Chris Lattner0de4caa2010-01-09 19:15:14 +0000[diff] [blame]597 the same name when linkage occurs. This can be used to implement
598 some forms of inline functions, templates, or other code which must be
599 generated in each translation unit that uses it, but where the body may
600 be overridden with a more definitive definition later. Unreferenced
601 <tt>linkonce</tt> globals are allowed to be discarded. Note that
602 <tt>linkonce</tt> linkage does not actually allow the optimizer to
603 inline the body of this function into callers because it doesn't know if
604 this definition of the function is the definitive definition within the
605 program or whether it will be overridden by a stronger definition.
606 To enable inlining and other optimizations, use "<tt>linkonce_odr</tt>"
607 linkage.</dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]608
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]609 <dt><tt><b><a name="linkage_weak">weak</a></b></tt></dt>
Chris Lattnerd0554882009-08-05 05:21:07 +0000[diff] [blame]610 <dd>"<tt>weak</tt>" linkage has the same merging semantics as
611 <tt>linkonce</tt> linkage, except that unreferenced globals with
612 <tt>weak</tt> linkage may not be discarded. This is used for globals that
613 are declared "weak" in C source code.</dd>
614
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]615 <dt><tt><b><a name="linkage_common">common</a></b></tt></dt>
Chris Lattnerd0554882009-08-05 05:21:07 +0000[diff] [blame]616 <dd>"<tt>common</tt>" linkage is most similar to "<tt>weak</tt>" linkage, but
617 they are used for tentative definitions in C, such as "<tt>int X;</tt>" at
618 global scope.
619 Symbols with "<tt>common</tt>" linkage are merged in the same way as
620 <tt>weak symbols</tt>, and they may not be deleted if unreferenced.
Chris Lattner0aff0b22009-08-05 05:41:44 +0000[diff] [blame]621 <tt>common</tt> symbols may not have an explicit section,
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]622 must have a zero initializer, and may not be marked '<a
Chris Lattner0aff0b22009-08-05 05:41:44 +0000[diff] [blame]623 href="#globalvars"><tt>constant</tt></a>'. Functions and aliases may not
624 have common linkage.</dd>
Chris Lattnerd0554882009-08-05 05:21:07 +0000[diff] [blame]625
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]626
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]627 <dt><tt><b><a name="linkage_appending">appending</a></b></tt></dt>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]628 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]629 pointer to array type. When two global variables with appending linkage
630 are linked together, the two global arrays are appended together. This is
631 the LLVM, typesafe, equivalent of having the system linker append together
632 "sections" with identical names when .o files are linked.</dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]633
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]634 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]635 <dd>The semantics of this linkage follow the ELF object file model: the symbol
636 is weak until linked, if not linked, the symbol becomes null instead of
637 being an undefined reference.</dd>
Anton Korobeynikova0554d92007-01-12 19:20:47 +0000[diff] [blame]638
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]639 <dt><tt><b><a name="linkage_linkonce_odr">linkonce_odr</a></b></tt></dt>
640 <dt><tt><b><a name="linkage_weak_odr">weak_odr</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]641 <dd>Some languages allow differing globals to be merged, such as two functions
642 with different semantics. Other languages, such as <tt>C++</tt>, ensure
Bill Wendling03bcd6e2010-07-01 21:55:59 +0000[diff] [blame]643 that only equivalent globals are ever merged (the "one definition rule"
644 &mdash; "ODR"). Such languages can use the <tt>linkonce_odr</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]645 and <tt>weak_odr</tt> linkage types to indicate that the global will only
646 be merged with equivalent globals. These linkage types are otherwise the
647 same as their non-<tt>odr</tt> versions.</dd>
Duncan Sands12da8ce2009-03-07 15:45:40 +0000[diff] [blame]648
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]649 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]650 <dd>If none of the above identifiers are used, the global is externally
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]651 visible, meaning that it participates in linkage and can be used to
652 resolve external symbol references.</dd>
Reid Spencer7972c472007-04-11 23:49:50 +0000[diff] [blame]653</dl>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]654
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]655<p>The next two types of linkage are targeted for Microsoft Windows platform
656 only. They are designed to support importing (exporting) symbols from (to)
657 DLLs (Dynamic Link Libraries).</p>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]658
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]659<dl>
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]660 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt></dt>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]661 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]662 or variable via a global pointer to a pointer that is set up by the DLL
663 exporting the symbol. On Microsoft Windows targets, the pointer name is
664 formed by combining <code>__imp_</code> and the function or variable
665 name.</dd>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]666
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]667 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt></dt>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]668 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]669 pointer to a pointer in a DLL, so that it can be referenced with the
670 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
671 name is formed by combining <code>__imp_</code> and the function or
672 variable name.</dd>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]673</dl>
674
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]675<p>For example, since the "<tt>.LC0</tt>" variable is defined to be internal, if
676 another module defined a "<tt>.LC0</tt>" variable and was linked with this
677 one, one of the two would be renamed, preventing a collision. Since
678 "<tt>main</tt>" and "<tt>puts</tt>" are external (i.e., lacking any linkage
679 declarations), they are accessible outside of the current module.</p>
680
681<p>It is illegal for a function <i>declaration</i> to have any linkage type
682 other than "externally visible", <tt>dllimport</tt>
683 or <tt>extern_weak</tt>.</p>
684
Duncan Sands12da8ce2009-03-07 15:45:40 +0000[diff] [blame]685<p>Aliases can have only <tt>external</tt>, <tt>internal</tt>, <tt>weak</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]686 or <tt>weak_odr</tt> linkages.</p>
687
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]688</div>
689
690<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]691<h3>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]692 <a name="callingconv">Calling Conventions</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]693</h3>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]694
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]695<div>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]696
697<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]698 and <a href="#i_invoke">invokes</a> can all have an optional calling
699 convention specified for the call. The calling convention of any pair of
700 dynamic caller/callee must match, or the behavior of the program is
701 undefined. The following calling conventions are supported by LLVM, and more
702 may be added in the future:</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]703
704<dl>
705 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]706 <dd>This calling convention (the default if no other calling convention is
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]707 specified) matches the target C calling conventions. This calling
708 convention supports varargs function calls and tolerates some mismatch in
709 the declared prototype and implemented declaration of the function (as
710 does normal C).</dd>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]711
712 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]713 <dd>This calling convention attempts to make calls as fast as possible
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]714 (e.g. by passing things in registers). This calling convention allows the
715 target to use whatever tricks it wants to produce fast code for the
716 target, without having to conform to an externally specified ABI
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]717 (Application Binary Interface).
718 <a href="CodeGenerator.html#tailcallopt">Tail calls can only be optimized
Chris Lattnera179e4d2010-03-11 00:22:57 +0000[diff] [blame]719 when this or the GHC convention is used.</a> This calling convention
720 does not support varargs and requires the prototype of all callees to
721 exactly match the prototype of the function definition.</dd>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]722
723 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]724 <dd>This calling convention attempts to make code in the caller as efficient
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]725 as possible under the assumption that the call is not commonly executed.
726 As such, these calls often preserve all registers so that the call does
727 not break any live ranges in the caller side. This calling convention
728 does not support varargs and requires the prototype of all callees to
729 exactly match the prototype of the function definition.</dd>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]730
Chris Lattnera179e4d2010-03-11 00:22:57 +0000[diff] [blame]731 <dt><b>"<tt>cc <em>10</em></tt>" - GHC convention</b>:</dt>
732 <dd>This calling convention has been implemented specifically for use by the
733 <a href="http://www.haskell.org/ghc">Glasgow Haskell Compiler (GHC)</a>.
734 It passes everything in registers, going to extremes to achieve this by
735 disabling callee save registers. This calling convention should not be
736 used lightly but only for specific situations such as an alternative to
737 the <em>register pinning</em> performance technique often used when
738 implementing functional programming languages.At the moment only X86
739 supports this convention and it has the following limitations:
740 <ul>
741 <li>On <em>X86-32</em> only supports up to 4 bit type parameters. No
742 floating point types are supported.</li>
743 <li>On <em>X86-64</em> only supports up to 10 bit type parameters and
744 6 floating point parameters.</li>
745 </ul>
746 This calling convention supports
747 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> but
748 requires both the caller and callee are using it.
749 </dd>
750
Chris Lattner573f64e2005-05-07 01:46:40 +0000[diff] [blame]751 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]752 <dd>Any calling convention may be specified by number, allowing
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]753 target-specific calling conventions to be used. Target specific calling
754 conventions start at 64.</dd>
Chris Lattner573f64e2005-05-07 01:46:40 +0000[diff] [blame]755</dl>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]756
757<p>More calling conventions can be added/defined on an as-needed basis, to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]758 support Pascal conventions or any other well-known target-independent
759 convention.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]760
761</div>
762
763<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]764<h3>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]765 <a name="visibility">Visibility Styles</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]766</h3>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]767
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]768<div>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]769
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]770<p>All Global Variables and Functions have one of the following visibility
771 styles:</p>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]772
773<dl>
774 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]775 <dd>On targets that use the ELF object file format, default visibility means
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]776 that the declaration is visible to other modules and, in shared libraries,
777 means that the declared entity may be overridden. On Darwin, default
778 visibility means that the declaration is visible to other modules. Default
779 visibility corresponds to "external linkage" in the language.</dd>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]780
781 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]782 <dd>Two declarations of an object with hidden visibility refer to the same
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]783 object if they are in the same shared object. Usually, hidden visibility
784 indicates that the symbol will not be placed into the dynamic symbol
785 table, so no other module (executable or shared library) can reference it
786 directly.</dd>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]787
Anton Korobeynikov39f3cff2007-04-29 18:35:00 +0000[diff] [blame]788 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
Anton Korobeynikov39f3cff2007-04-29 18:35:00 +0000[diff] [blame]789 <dd>On ELF, protected visibility indicates that the symbol will be placed in
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]790 the dynamic symbol table, but that references within the defining module
791 will bind to the local symbol. That is, the symbol cannot be overridden by
792 another module.</dd>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]793</dl>
794
795</div>
796
797<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]798<h3>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]799 <a name="namedtypes">Named Types</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]800</h3>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]801
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]802<div>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]803
804<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]805 it easier to read the IR and make the IR more condensed (particularly when
806 recursive types are involved). An example of a name specification is:</p>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]807
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]808<pre class="doc_code">
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]809%mytype = type { %mytype*, i32 }
810</pre>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]811
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]812<p>You may give a name to any <a href="#typesystem">type</a> except
Chris Lattner249b9762010-08-17 23:26:04 +0000[diff] [blame]813 "<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]814 is expected with the syntax "%mytype".</p>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]815
816<p>Note that type names are aliases for the structural type that they indicate,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]817 and that you can therefore specify multiple names for the same type. This
818 often leads to confusing behavior when dumping out a .ll file. Since LLVM IR
819 uses structural typing, the name is not part of the type. When printing out
820 LLVM IR, the printer will pick <em>one name</em> to render all types of a
821 particular shape. This means that if you have code where two different
822 source types end up having the same LLVM type, that the dumper will sometimes
823 print the "wrong" or unexpected type. This is an important design point and
824 isn't going to change.</p>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]825
826</div>
827
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]828<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]829<h3>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]830 <a name="globalvars">Global Variables</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]831</h3>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]832
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]833<div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]834
Chris Lattner5d5aede2005-02-12 19:30:21 +0000[diff] [blame]835<p>Global variables define regions of memory allocated at compilation time
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]836 instead of run-time. Global variables may optionally be initialized, may
837 have an explicit section to be placed in, and may have an optional explicit
838 alignment specified. A variable may be defined as "thread_local", which
839 means that it will not be shared by threads (each thread will have a
840 separated copy of the variable). A variable may be defined as a global
841 "constant," which indicates that the contents of the variable
842 will <b>never</b> be modified (enabling better optimization, allowing the
843 global data to be placed in the read-only section of an executable, etc).
844 Note that variables that need runtime initialization cannot be marked
845 "constant" as there is a store to the variable.</p>
Chris Lattner5d5aede2005-02-12 19:30:21 +0000[diff] [blame]846
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]847<p>LLVM explicitly allows <em>declarations</em> of global variables to be marked
848 constant, even if the final definition of the global is not. This capability
849 can be used to enable slightly better optimization of the program, but
850 requires the language definition to guarantee that optimizations based on the
851 'constantness' are valid for the translation units that do not include the
852 definition.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]853
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]854<p>As SSA values, global variables define pointer values that are in scope
855 (i.e. they dominate) all basic blocks in the program. Global variables
856 always define a pointer to their "content" type because they describe a
857 region of memory, and all memory objects in LLVM are accessed through
858 pointers.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]859
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame]860<p>Global variables can be marked with <tt>unnamed_addr</tt> which indicates
861 that the address is not significant, only the content. Constants marked
Rafael Espindolaf1ed7812011-01-15 08:20:57 +0000[diff] [blame]862 like this can be merged with other constants if they have the same
863 initializer. Note that a constant with significant address <em>can</em>
864 be merged with a <tt>unnamed_addr</tt> constant, the result being a
865 constant whose address is significant.</p>
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame]866
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]867<p>A global variable may be declared to reside in a target-specific numbered
868 address space. For targets that support them, address spaces may affect how
869 optimizations are performed and/or what target instructions are used to
870 access the variable. The default address space is zero. The address space
871 qualifier must precede any other attributes.</p>
Christopher Lamb308121c2007-12-11 09:31:00 +0000[diff] [blame]872
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]873<p>LLVM allows an explicit section to be specified for globals. If the target
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]874 supports it, it will emit globals to the section specified.</p>
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]875
Chris Lattner78e00bc2010-04-28 00:13:42 +0000[diff] [blame]876<p>An explicit alignment may be specified for a global, which must be a power
877 of 2. If not present, or if the alignment is set to zero, the alignment of
878 the global is set by the target to whatever it feels convenient. If an
879 explicit alignment is specified, the global is forced to have exactly that
Chris Lattner4bd85e42010-04-28 00:31:12 +0000[diff] [blame]880 alignment. Targets and optimizers are not allowed to over-align the global
881 if the global has an assigned section. In this case, the extra alignment
882 could be observable: for example, code could assume that the globals are
883 densely packed in their section and try to iterate over them as an array,
884 alignment padding would break this iteration.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]885
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]886<p>For example, the following defines a global in a numbered address space with
887 an initializer, section, and alignment:</p>
Chris Lattner5760c502007-01-14 00:27:09 +0000[diff] [blame]888
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]889<pre class="doc_code">
Dan Gohmanaaa679b2009-01-11 00:40:00 +0000[diff] [blame]890@G = addrspace(5) constant float 1.0, section "foo", align 4
Chris Lattner5760c502007-01-14 00:27:09 +0000[diff] [blame]891</pre>
892
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]893</div>
894
895
896<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]897<h3>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]898 <a name="functionstructure">Functions</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]899</h3>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]900
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]901<div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]902
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]903<p>LLVM function definitions consist of the "<tt>define</tt>" keyword, an
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]904 optional <a href="#linkage">linkage type</a>, an optional
905 <a href="#visibility">visibility style</a>, an optional
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame]906 <a href="#callingconv">calling convention</a>,
907 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]908 <a href="#paramattrs">parameter attribute</a> for the return type, a function
909 name, a (possibly empty) argument list (each with optional
910 <a href="#paramattrs">parameter attributes</a>), optional
911 <a href="#fnattrs">function attributes</a>, an optional section, an optional
912 alignment, an optional <a href="#gc">garbage collector name</a>, an opening
913 curly brace, a list of basic blocks, and a closing curly brace.</p>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]914
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]915<p>LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
916 optional <a href="#linkage">linkage type</a>, an optional
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]917 <a href="#visibility">visibility style</a>, an optional
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame]918 <a href="#callingconv">calling convention</a>,
919 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]920 <a href="#paramattrs">parameter attribute</a> for the return type, a function
921 name, a possibly empty list of arguments, an optional alignment, and an
922 optional <a href="#gc">garbage collector name</a>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]923
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]924<p>A function definition contains a list of basic blocks, forming the CFG
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]925 (Control Flow Graph) for the function. Each basic block may optionally start
926 with a label (giving the basic block a symbol table entry), contains a list
927 of instructions, and ends with a <a href="#terminators">terminator</a>
928 instruction (such as a branch or function return).</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]929
Chris Lattnera59fb102007-06-08 16:52:14 +0000[diff] [blame]930<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]931 executed on entrance to the function, and it is not allowed to have
932 predecessor basic blocks (i.e. there can not be any branches to the entry
933 block of a function). Because the block can have no predecessors, it also
934 cannot have any <a href="#i_phi">PHI nodes</a>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]935
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]936<p>LLVM allows an explicit section to be specified for functions. If the target
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]937 supports it, it will emit functions to the section specified.</p>
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]938
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]939<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]940 the alignment is set to zero, the alignment of the function is set by the
941 target to whatever it feels convenient. If an explicit alignment is
942 specified, the function is forced to have at least that much alignment. All
943 alignments must be a power of 2.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]944
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame]945<p>If the <tt>unnamed_addr</tt> attribute is given, the address is know to not
946 be significant and two identical functions can be merged</p>.
947
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]948<h5>Syntax:</h5>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]949<pre class="doc_code">
Chris Lattner0ae02092008-10-13 16:55:18 +0000[diff] [blame]950define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]951 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
952 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
953 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
954 [<a href="#gc">gc</a>] { ... }
955</pre>
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]956
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]957</div>
958
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]959<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]960<h3>
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]961 <a name="aliasstructure">Aliases</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]962</h3>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]963
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]964<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]965
966<p>Aliases act as "second name" for the aliasee value (which can be either
967 function, global variable, another alias or bitcast of global value). Aliases
968 may have an optional <a href="#linkage">linkage type</a>, and an
969 optional <a href="#visibility">visibility style</a>.</p>
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]970
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]971<h5>Syntax:</h5>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]972<pre class="doc_code">
Duncan Sands7e99a942008-09-12 20:48:21 +0000[diff] [blame]973@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Bill Wendling2d8b9a82007-05-29 09:42:13 +0000[diff] [blame]974</pre>
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]975
976</div>
977
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]978<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]979<h3>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]980 <a name="namedmetadatastructure">Named Metadata</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]981</h3>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]982
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]983<div>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]984
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]985<p>Named metadata is a collection of metadata. <a href="#metadata">Metadata
Dan Gohman093cb792010-07-21 18:54:18 +0000[diff] [blame]986 nodes</a> (but not metadata strings) are the only valid operands for
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]987 a named metadata.</p>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]988
989<h5>Syntax:</h5>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]990<pre class="doc_code">
Dan Gohman093cb792010-07-21 18:54:18 +0000[diff] [blame]991; Some unnamed metadata nodes, which are referenced by the named metadata.
992!0 = metadata !{metadata !"zero"}
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]993!1 = metadata !{metadata !"one"}
Dan Gohman093cb792010-07-21 18:54:18 +0000[diff] [blame]994!2 = metadata !{metadata !"two"}
Dan Gohman58cd65f2010-07-13 19:48:13 +0000[diff] [blame]995; A named metadata.
Dan Gohman093cb792010-07-21 18:54:18 +0000[diff] [blame]996!name = !{!0, !1, !2}
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]997</pre>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]998
999</div>
1000
1001<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1002<h3>
1003 <a name="paramattrs">Parameter Attributes</a>
1004</h3>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1005
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1006<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1007
1008<p>The return type and each parameter of a function type may have a set of
1009 <i>parameter attributes</i> associated with them. Parameter attributes are
1010 used to communicate additional information about the result or parameters of
1011 a function. Parameter attributes are considered to be part of the function,
1012 not of the function type, so functions with different parameter attributes
1013 can have the same function type.</p>
1014
1015<p>Parameter attributes are simple keywords that follow the type specified. If
1016 multiple parameter attributes are needed, they are space separated. For
1017 example:</p>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]1018
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1019<pre class="doc_code">
Nick Lewyckydac78d82009-02-15 23:06:14 +0000[diff] [blame]1020declare i32 @printf(i8* noalias nocapture, ...)
Chris Lattnerd2597d72008-10-04 18:33:34 +0000[diff] [blame]1021declare i32 @atoi(i8 zeroext)
1022declare signext i8 @returns_signed_char()
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]1023</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]1024
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1025<p>Note that any attributes for the function result (<tt>nounwind</tt>,
1026 <tt>readonly</tt>) come immediately after the argument list.</p>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1027
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1028<p>Currently, only the following parameter attributes are defined:</p>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000[diff] [blame]1029
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1030<dl>
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1031 <dt><tt><b>zeroext</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1032 <dd>This indicates to the code generator that the parameter or return value
Cameron Zwarichac106272011-03-16 22:20:18 +0000[diff] [blame]1033 should be zero-extended to the extent required by the target's ABI (which
1034 is usually 32-bits, but is 8-bits for a i1 on x86-64) by the caller (for a
1035 parameter) or the callee (for a return value).</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000[diff] [blame]1036
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1037 <dt><tt><b>signext</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1038 <dd>This indicates to the code generator that the parameter or return value
Cameron Zwarich341c36d2011-03-17 14:21:58 +0000[diff] [blame]1039 should be sign-extended to the extent required by the target's ABI (which
1040 is usually 32-bits) by the caller (for a parameter) or the callee (for a
1041 return value).</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000[diff] [blame]1042
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1043 <dt><tt><b>inreg</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1044 <dd>This indicates that this parameter or return value should be treated in a
1045 special target-dependent fashion during while emitting code for a function
1046 call or return (usually, by putting it in a register as opposed to memory,
1047 though some targets use it to distinguish between two different kinds of
1048 registers). Use of this attribute is target-specific.</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000[diff] [blame]1049
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1050 <dt><tt><b><a name="byval">byval</a></b></tt></dt>
Chris Lattnerd78dbee2010-11-20 23:49:06 +0000[diff] [blame]1051 <dd><p>This indicates that the pointer parameter should really be passed by
1052 value to the function. The attribute implies that a hidden copy of the
1053 pointee
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1054 is made between the caller and the callee, so the callee is unable to
1055 modify the value in the callee. This attribute is only valid on LLVM
1056 pointer arguments. It is generally used to pass structs and arrays by
1057 value, but is also valid on pointers to scalars. The copy is considered
1058 to belong to the caller not the callee (for example,
1059 <tt><a href="#readonly">readonly</a></tt> functions should not write to
1060 <tt>byval</tt> parameters). This is not a valid attribute for return
Chris Lattnerd78dbee2010-11-20 23:49:06 +0000[diff] [blame]1061 values.</p>
1062
1063 <p>The byval attribute also supports specifying an alignment with
1064 the align attribute. It indicates the alignment of the stack slot to
1065 form and the known alignment of the pointer specified to the call site. If
1066 the alignment is not specified, then the code generator makes a
1067 target-specific assumption.</p></dd>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1068
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]1069 <dt><tt><b><a name="sret">sret</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1070 <dd>This indicates that the pointer parameter specifies the address of a
1071 structure that is the return value of the function in the source program.
1072 This pointer must be guaranteed by the caller to be valid: loads and
1073 stores to the structure may be assumed by the callee to not to trap. This
1074 may only be applied to the first parameter. This is not a valid attribute
1075 for return values. </dd>
1076
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]1077 <dt><tt><b><a name="noalias">noalias</a></b></tt></dt>
Dan Gohmandf12d082010-07-02 18:41:32 +0000[diff] [blame]1078 <dd>This indicates that pointer values
1079 <a href="#pointeraliasing"><i>based</i></a> on the argument or return
Dan Gohmande256292010-07-02 23:46:54 +0000[diff] [blame]1080 value do not alias pointer values which are not <i>based</i> on it,
1081 ignoring certain "irrelevant" dependencies.
1082 For a call to the parent function, dependencies between memory
1083 references from before or after the call and from those during the call
1084 are "irrelevant" to the <tt>noalias</tt> keyword for the arguments and
1085 return value used in that call.
Dan Gohmandf12d082010-07-02 18:41:32 +0000[diff] [blame]1086 The caller shares the responsibility with the callee for ensuring that
1087 these requirements are met.
1088 For further details, please see the discussion of the NoAlias response in
Dan Gohman6c858db2010-07-06 15:26:33 +0000[diff] [blame]1089 <a href="AliasAnalysis.html#MustMayNo">alias analysis</a>.<br>
1090<br>
John McCall72ed8902010-07-06 21:07:14 +0000[diff] [blame]1091 Note that this definition of <tt>noalias</tt> is intentionally
1092 similar to the definition of <tt>restrict</tt> in C99 for function
Chris Lattner5eff9ca2010-07-06 20:51:35 +0000[diff] [blame]1093 arguments, though it is slightly weaker.
Dan Gohman6c858db2010-07-06 15:26:33 +0000[diff] [blame]1094<br>
1095 For function return values, C99's <tt>restrict</tt> is not meaningful,
1096 while LLVM's <tt>noalias</tt> is.
1097 </dd>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1098
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]1099 <dt><tt><b><a name="nocapture">nocapture</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1100 <dd>This indicates that the callee does not make any copies of the pointer
1101 that outlive the callee itself. This is not a valid attribute for return
1102 values.</dd>
1103
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]1104 <dt><tt><b><a name="nest">nest</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1105 <dd>This indicates that the pointer parameter can be excised using the
1106 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
1107 attribute for return values.</dd>
1108</dl>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1109
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1110</div>
1111
1112<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1113<h3>
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1114 <a name="gc">Garbage Collector Names</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1115</h3>
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1116
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1117<div>
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1118
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1119<p>Each function may specify a garbage collector name, which is simply a
1120 string:</p>
1121
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1122<pre class="doc_code">
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1123define void @f() gc "name" { ... }
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1124</pre>
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1125
1126<p>The compiler declares the supported values of <i>name</i>. Specifying a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1127 collector which will cause the compiler to alter its output in order to
1128 support the named garbage collection algorithm.</p>
1129
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1130</div>
1131
1132<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1133<h3>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1134 <a name="fnattrs">Function Attributes</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1135</h3>
Devang Patelcaacdba2008-09-04 23:05:13 +0000[diff] [blame]1136
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1137<div>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1138
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1139<p>Function attributes are set to communicate additional information about a
1140 function. Function attributes are considered to be part of the function, not
1141 of the function type, so functions with different parameter attributes can
1142 have the same function type.</p>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1143
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1144<p>Function attributes are simple keywords that follow the type specified. If
1145 multiple attributes are needed, they are space separated. For example:</p>
Devang Patelcaacdba2008-09-04 23:05:13 +0000[diff] [blame]1146
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1147<pre class="doc_code">
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1148define void @f() noinline { ... }
1149define void @f() alwaysinline { ... }
1150define void @f() alwaysinline optsize { ... }
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1151define void @f() optsize { ... }
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1152</pre>
Devang Patelcaacdba2008-09-04 23:05:13 +0000[diff] [blame]1153
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1154<dl>
Charles Davisbe5557e2010-02-12 00:31:15 +0000[diff] [blame]1155 <dt><tt><b>alignstack(&lt;<em>n</em>&gt;)</b></tt></dt>
1156 <dd>This attribute indicates that, when emitting the prologue and epilogue,
1157 the backend should forcibly align the stack pointer. Specify the
1158 desired alignment, which must be a power of two, in parentheses.
1159
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1160 <dt><tt><b>alwaysinline</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1161 <dd>This attribute indicates that the inliner should attempt to inline this
1162 function into callers whenever possible, ignoring any active inlining size
1163 threshold for this caller.</dd>
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1164
Charles Davis22fe1862010-10-25 15:37:09 +0000[diff] [blame]1165 <dt><tt><b>hotpatch</b></tt></dt>
Charles Davis1b2d3722010-10-25 16:29:03 +0000[diff] [blame]1166 <dd>This attribute indicates that the function should be 'hotpatchable',
Charles Davis74205252010-10-25 19:07:39 +0000[diff] [blame]1167 meaning the function can be patched and/or hooked even while it is
1168 loaded into memory. On x86, the function prologue will be preceded
1169 by six bytes of padding and will begin with a two-byte instruction.
1170 Most of the functions in the Windows system DLLs in Windows XP SP2 or
1171 higher were compiled in this fashion.</dd>
Charles Davis22fe1862010-10-25 15:37:09 +0000[diff] [blame]1172
Dan Gohman8bd11f12011-06-16 16:03:13 +0000[diff] [blame]1173 <dt><tt><b>nonlazybind</b></tt></dt>
1174 <dd>This attribute suppresses lazy symbol binding for the function. This
1175 may make calls to the function faster, at the cost of extra program
1176 startup time if the function is not called during program startup.</dd>
1177
Jakob Stoklund Olesen74bb06c2010-02-06 01:16:28 +0000[diff] [blame]1178 <dt><tt><b>inlinehint</b></tt></dt>
1179 <dd>This attribute indicates that the source code contained a hint that inlining
1180 this function is desirable (such as the "inline" keyword in C/C++). It
1181 is just a hint; it imposes no requirements on the inliner.</dd>
1182
Nick Lewycky14b58da2010-07-06 18:24:09 +0000[diff] [blame]1183 <dt><tt><b>naked</b></tt></dt>
1184 <dd>This attribute disables prologue / epilogue emission for the function.
1185 This can have very system-specific consequences.</dd>
1186
1187 <dt><tt><b>noimplicitfloat</b></tt></dt>
1188 <dd>This attributes disables implicit floating point instructions.</dd>
1189
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1190 <dt><tt><b>noinline</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1191 <dd>This attribute indicates that the inliner should never inline this
1192 function in any situation. This attribute may not be used together with
1193 the <tt>alwaysinline</tt> attribute.</dd>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1194
Nick Lewycky14b58da2010-07-06 18:24:09 +0000[diff] [blame]1195 <dt><tt><b>noredzone</b></tt></dt>
1196 <dd>This attribute indicates that the code generator should not use a red
1197 zone, even if the target-specific ABI normally permits it.</dd>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1198
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1199 <dt><tt><b>noreturn</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1200 <dd>This function attribute indicates that the function never returns
1201 normally. This produces undefined behavior at runtime if the function
1202 ever does dynamically return.</dd>
Bill Wendlinga8130172008-11-13 01:02:51 +0000[diff] [blame]1203
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1204 <dt><tt><b>nounwind</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1205 <dd>This function attribute indicates that the function never returns with an
1206 unwind or exceptional control flow. If the function does unwind, its
1207 runtime behavior is undefined.</dd>
Bill Wendling0f5541e2008-11-26 19:07:40 +0000[diff] [blame]1208
Nick Lewycky14b58da2010-07-06 18:24:09 +0000[diff] [blame]1209 <dt><tt><b>optsize</b></tt></dt>
1210 <dd>This attribute suggests that optimization passes and code generator passes
1211 make choices that keep the code size of this function low, and otherwise
1212 do optimizations specifically to reduce code size.</dd>
1213
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1214 <dt><tt><b>readnone</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1215 <dd>This attribute indicates that the function computes its result (or decides
1216 to unwind an exception) based strictly on its arguments, without
1217 dereferencing any pointer arguments or otherwise accessing any mutable
1218 state (e.g. memory, control registers, etc) visible to caller functions.
1219 It does not write through any pointer arguments
1220 (including <tt><a href="#byval">byval</a></tt> arguments) and never
1221 changes any state visible to callers. This means that it cannot unwind
1222 exceptions by calling the <tt>C++</tt> exception throwing methods, but
1223 could use the <tt>unwind</tt> instruction.</dd>
Devang Patel310fd4a2009-06-12 19:45:19 +0000[diff] [blame]1224
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1225 <dt><tt><b><a name="readonly">readonly</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1226 <dd>This attribute indicates that the function does not write through any
1227 pointer arguments (including <tt><a href="#byval">byval</a></tt>
1228 arguments) or otherwise modify any state (e.g. memory, control registers,
1229 etc) visible to caller functions. It may dereference pointer arguments
1230 and read state that may be set in the caller. A readonly function always
1231 returns the same value (or unwinds an exception identically) when called
1232 with the same set of arguments and global state. It cannot unwind an
1233 exception by calling the <tt>C++</tt> exception throwing methods, but may
1234 use the <tt>unwind</tt> instruction.</dd>
Anton Korobeynikovc8ce7b082009-07-17 18:07:26 +0000[diff] [blame]1235
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1236 <dt><tt><b><a name="ssp">ssp</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1237 <dd>This attribute indicates that the function should emit a stack smashing
1238 protector. It is in the form of a "canary"&mdash;a random value placed on
1239 the stack before the local variables that's checked upon return from the
1240 function to see if it has been overwritten. A heuristic is used to
1241 determine if a function needs stack protectors or not.<br>
1242<br>
1243 If a function that has an <tt>ssp</tt> attribute is inlined into a
1244 function that doesn't have an <tt>ssp</tt> attribute, then the resulting
1245 function will have an <tt>ssp</tt> attribute.</dd>
1246
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1247 <dt><tt><b>sspreq</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1248 <dd>This attribute indicates that the function should <em>always</em> emit a
1249 stack smashing protector. This overrides
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1250 the <tt><a href="#ssp">ssp</a></tt> function attribute.<br>
1251<br>
1252 If a function that has an <tt>sspreq</tt> attribute is inlined into a
1253 function that doesn't have an <tt>sspreq</tt> attribute or which has
1254 an <tt>ssp</tt> attribute, then the resulting function will have
1255 an <tt>sspreq</tt> attribute.</dd>
Rafael Espindola163d6752011-07-25 15:27:59 +0000[diff] [blame]1256
1257 <dt><tt><b><a name="uwtable">uwtable</a></b></tt></dt>
1258 <dd>This attribute indicates that the ABI being targeted requires that
1259 an unwind table entry be produce for this function even if we can
1260 show that no exceptions passes by it. This is normally the case for
1261 the ELF x86-64 abi, but it can be disabled for some compilation
1262 units.</dd>
1263
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1264</dl>
1265
Devang Patelcaacdba2008-09-04 23:05:13 +0000[diff] [blame]1266</div>
1267
1268<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1269<h3>
Chris Lattner93564892006-04-08 04:40:53 +0000[diff] [blame]1270 <a name="moduleasm">Module-Level Inline Assembly</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1271</h3>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1272
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1273<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1274
1275<p>Modules may contain "module-level inline asm" blocks, which corresponds to
1276 the GCC "file scope inline asm" blocks. These blocks are internally
1277 concatenated by LLVM and treated as a single unit, but may be separated in
1278 the <tt>.ll</tt> file if desired. The syntax is very simple:</p>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1279
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1280<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]1281module asm "inline asm code goes here"
1282module asm "more can go here"
1283</pre>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1284
1285<p>The strings can contain any character by escaping non-printable characters.
1286 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]1287 for the number.</p>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1288
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1289<p>The inline asm code is simply printed to the machine code .s file when
1290 assembly code is generated.</p>
1291
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1292</div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1293
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1294<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1295<h3>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1296 <a name="datalayout">Data Layout</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1297</h3>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1298
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1299<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1300
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1301<p>A module may specify a target specific data layout string that specifies how
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1302 data is to be laid out in memory. The syntax for the data layout is
1303 simply:</p>
1304
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1305<pre class="doc_code">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1306target datalayout = "<i>layout specification</i>"
1307</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1308
1309<p>The <i>layout specification</i> consists of a list of specifications
1310 separated by the minus sign character ('-'). Each specification starts with
1311 a letter and may include other information after the letter to define some
1312 aspect of the data layout. The specifications accepted are as follows:</p>
1313
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1314<dl>
1315 <dt><tt>E</tt></dt>
1316 <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]1317 bits with the most significance have the lowest address location.</dd>
1318
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1319 <dt><tt>e</tt></dt>
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]1320 <dd>Specifies that the target lays out data in little-endian form. That is,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1321 the bits with the least significance have the lowest address
1322 location.</dd>
1323
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1324 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1325 <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]1326 <i>preferred</i> alignments. All sizes are in bits. Specifying
1327 the <i>pref</i> alignment is optional. If omitted, the
1328 preceding <tt>:</tt> should be omitted too.</dd>
1329
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1330 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1331 <dd>This specifies the alignment for an integer type of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1332 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1333
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1334 <dt><tt>v<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 vector type of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1336 <i>size</i>.</dd>
1337
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1338 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1339 <dd>This specifies the alignment for a floating point type of a given bit
Dale Johannesence522852010-05-28 18:54:47 +0000[diff] [blame]1340 <i>size</i>. Only values of <i>size</i> that are supported by the target
1341 will work. 32 (float) and 64 (double) are supported on all targets;
1342 80 or 128 (different flavors of long double) are also supported on some
1343 targets.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1344
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1345 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1346 <dd>This specifies the alignment for an aggregate type of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1347 <i>size</i>.</dd>
1348
Daniel Dunbar7921a592009-06-08 22:17:53 +0000[diff] [blame]1349 <dt><tt>s<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1350 <dd>This specifies the alignment for a stack object of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1351 <i>size</i>.</dd>
Chris Lattnera381eff2009-11-07 09:35:34 +0000[diff] [blame]1352
1353 <dt><tt>n<i>size1</i>:<i>size2</i>:<i>size3</i>...</tt></dt>
1354 <dd>This specifies a set of native integer widths for the target CPU
1355 in bits. For example, it might contain "n32" for 32-bit PowerPC,
1356 "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]1357 this set are considered to support most general arithmetic
Chris Lattnera381eff2009-11-07 09:35:34 +0000[diff] [blame]1358 operations efficiently.</dd>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1359</dl>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1360
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1361<p>When constructing the data layout for a given target, LLVM starts with a
Dan Gohman61110ae2010-04-28 00:36:01 +0000[diff] [blame]1362 default set of specifications which are then (possibly) overridden by the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1363 specifications in the <tt>datalayout</tt> keyword. The default specifications
1364 are given in this list:</p>
1365
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1366<ul>
1367 <li><tt>E</tt> - big endian</li>
Dan Gohman8ad777d2010-02-23 02:44:03 +0000[diff] [blame]1368 <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]1369 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1370 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1371 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1372 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]1373 <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]1374 alignment of 64-bits</li>
1375 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1376 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1377 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1378 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1379 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
Daniel Dunbar7921a592009-06-08 22:17:53 +0000[diff] [blame]1380 <li><tt>s0:64:64</tt> - stack objects are 64-bit aligned</li>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1381</ul>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1382
1383<p>When LLVM is determining the alignment for a given type, it uses the
1384 following rules:</p>
1385
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1386<ol>
1387 <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]1388 specification is used.</li>
1389
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1390 <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]1391 smallest integer type that is larger than the bitwidth of the sought type
1392 is used. If none of the specifications are larger than the bitwidth then
1393 the the largest integer type is used. For example, given the default
1394 specifications above, the i7 type will use the alignment of i8 (next
1395 largest) while both i65 and i256 will use the alignment of i64 (largest
1396 specified).</li>
1397
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1398 <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]1399 largest vector type that is smaller than the sought vector type will be
1400 used as a fall back. This happens because &lt;128 x double&gt; can be
1401 implemented in terms of 64 &lt;2 x double&gt;, for example.</li>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1402</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1403
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1404</div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1405
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1406<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1407<h3>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1408 <a name="pointeraliasing">Pointer Aliasing Rules</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1409</h3>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1410
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1411<div>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1412
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1413<p>Any memory access must be done through a pointer value associated
Andreas Bolkae39f0332009-07-27 20:37:10 +0000[diff] [blame]1414with an address range of the memory access, otherwise the behavior
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1415is undefined. Pointer values are associated with address ranges
1416according to the following rules:</p>
1417
1418<ul>
Dan Gohmandf12d082010-07-02 18:41:32 +0000[diff] [blame]1419 <li>A pointer value is associated with the addresses associated with
1420 any value it is <i>based</i> on.
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1421 <li>An address of a global variable is associated with the address
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1422 range of the variable's storage.</li>
1423 <li>The result value of an allocation instruction is associated with
1424 the address range of the allocated storage.</li>
1425 <li>A null pointer in the default address-space is associated with
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1426 no address.</li>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1427 <li>An integer constant other than zero or a pointer value returned
1428 from a function not defined within LLVM may be associated with address
1429 ranges allocated through mechanisms other than those provided by
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1430 LLVM. Such ranges shall not overlap with any ranges of addresses
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1431 allocated by mechanisms provided by LLVM.</li>
Dan Gohmandf12d082010-07-02 18:41:32 +0000[diff] [blame]1432</ul>
1433
1434<p>A pointer value is <i>based</i> on another pointer value according
1435 to the following rules:</p>
1436
1437<ul>
1438 <li>A pointer value formed from a
1439 <tt><a href="#i_getelementptr">getelementptr</a></tt> operation
1440 is <i>based</i> on the first operand of the <tt>getelementptr</tt>.</li>
1441 <li>The result value of a
1442 <tt><a href="#i_bitcast">bitcast</a></tt> is <i>based</i> on the operand
1443 of the <tt>bitcast</tt>.</li>
1444 <li>A pointer value formed by an
1445 <tt><a href="#i_inttoptr">inttoptr</a></tt> is <i>based</i> on all
1446 pointer values that contribute (directly or indirectly) to the
1447 computation of the pointer's value.</li>
1448 <li>The "<i>based</i> on" relationship is transitive.</li>
1449</ul>
1450
1451<p>Note that this definition of <i>"based"</i> is intentionally
1452 similar to the definition of <i>"based"</i> in C99, though it is
1453 slightly weaker.</p>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1454
1455<p>LLVM IR does not associate types with memory. The result type of a
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1456<tt><a href="#i_load">load</a></tt> merely indicates the size and
1457alignment of the memory from which to load, as well as the
Dan Gohman4eb47192010-06-17 19:23:50 +0000[diff] [blame]1458interpretation of the value. The first operand type of a
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1459<tt><a href="#i_store">store</a></tt> similarly only indicates the size
1460and alignment of the store.</p>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1461
1462<p>Consequently, type-based alias analysis, aka TBAA, aka
1463<tt>-fstrict-aliasing</tt>, is not applicable to general unadorned
1464LLVM IR. <a href="#metadata">Metadata</a> may be used to encode
1465additional information which specialized optimization passes may use
1466to implement type-based alias analysis.</p>
1467
1468</div>
1469
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]1470<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1471<h3>
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]1472 <a name="volatile">Volatile Memory Accesses</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1473</h3>
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]1474
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1475<div>
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]1476
1477<p>Certain memory accesses, such as <a href="#i_load"><tt>load</tt></a>s, <a
1478href="#i_store"><tt>store</tt></a>s, and <a
1479href="#int_memcpy"><tt>llvm.memcpy</tt></a>s may be marked <tt>volatile</tt>.
1480The optimizers must not change the number of volatile operations or change their
1481order of execution relative to other volatile operations. The optimizers
1482<i>may</i> change the order of volatile operations relative to non-volatile
1483operations. This is not Java's "volatile" and has no cross-thread
1484synchronization behavior.</p>
1485
1486</div>
1487
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1488<!-- ======================================================================= -->
1489<h3>
1490 <a name="memmodel">Memory Model for Concurrent Operations</a>
1491</h3>
1492
1493<div>
1494
1495<p>The LLVM IR does not define any way to start parallel threads of execution
1496or to register signal handlers. Nonetheless, there are platform-specific
1497ways to create them, and we define LLVM IR's behavior in their presence. This
1498model is inspired by the C++0x memory model.</p>
1499
Eli Friedman95f69a42011-08-22 21:35:27 +0000[diff] [blame^]1500<p>For a more informal introduction to this model, see the
1501<a href="Atomics.html">LLVM Atomic Instructions and Concurrency Guide</a>.
1502
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1503<p>We define a <i>happens-before</i> partial order as the least partial order
1504that</p>
1505<ul>
1506 <li>Is a superset of single-thread program order, and</li>
1507 <li>When a <i>synchronizes-with</i> <tt>b</tt>, includes an edge from
1508 <tt>a</tt> to <tt>b</tt>. <i>Synchronizes-with</i> pairs are introduced
1509 by platform-specific techniques, like pthread locks, thread
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1510 creation, thread joining, etc., and by atomic instructions.
1511 (See also <a href="#ordering">Atomic Memory Ordering Constraints</a>).
1512 </li>
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1513</ul>
1514
1515<p>Note that program order does not introduce <i>happens-before</i> edges
1516between a thread and signals executing inside that thread.</p>
1517
1518<p>Every (defined) read operation (load instructions, memcpy, atomic
1519loads/read-modify-writes, etc.) <var>R</var> reads a series of bytes written by
1520(defined) write operations (store instructions, atomic
Eli Friedmanf12e4e92011-07-22 03:04:45 +0000[diff] [blame]1521stores/read-modify-writes, memcpy, etc.). For the purposes of this section,
1522initialized globals are considered to have a write of the initializer which is
1523atomic and happens before any other read or write of the memory in question.
1524For each byte of a read <var>R</var>, <var>R<sub>byte</sub></var> may see
1525any write to the same byte, except:</p>
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1526
1527<ul>
1528 <li>If <var>write<sub>1</sub></var> happens before
1529 <var>write<sub>2</sub></var>, and <var>write<sub>2</sub></var> happens
1530 before <var>R<sub>byte</sub></var>, then <var>R<sub>byte</sub></var>
Eli Friedmanf12e4e92011-07-22 03:04:45 +0000[diff] [blame]1531 does not see <var>write<sub>1</sub></var>.
Bill Wendling537603b2011-07-31 06:45:03 +0000[diff] [blame]1532 <li>If <var>R<sub>byte</sub></var> happens before
1533 <var>write<sub>3</sub></var>, then <var>R<sub>byte</sub></var> does not
1534 see <var>write<sub>3</sub></var>.
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1535</ul>
1536
1537<p>Given that definition, <var>R<sub>byte</sub></var> is defined as follows:
1538<ul>
Eli Friedman95f69a42011-08-22 21:35:27 +0000[diff] [blame^]1539 <li>If <var>R</var> is volatile, the result is target-dependent. (Volatile
1540 is supposed to give guarantees which can support
1541 <code>sig_atomic_t</code> in C/C++, and may be used for accesses to
1542 addresses which do not behave like normal memory. It does not generally
1543 provide cross-thread synchronization.)
1544 <li>Otherwise, if there is no write to the same byte that happens before
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1545 <var>R<sub>byte</sub></var>, <var>R<sub>byte</sub></var> returns
1546 <tt>undef</tt> for that byte.
Eli Friedmanf12e4e92011-07-22 03:04:45 +0000[diff] [blame]1547 <li>Otherwise, if <var>R<sub>byte</sub></var> may see exactly one write,
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1548 <var>R<sub>byte</sub></var> returns the value written by that
1549 write.</li>
Eli Friedmanf12e4e92011-07-22 03:04:45 +0000[diff] [blame]1550 <li>Otherwise, if <var>R</var> is atomic, and all the writes
1551 <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]1552 values written. See the <a href="#ordering">Atomic Memory Ordering
1553 Constraints</a> section for additional constraints on how the choice
1554 is made.
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1555 <li>Otherwise <var>R<sub>byte</sub></var> returns <tt>undef</tt>.</li>
1556</ul>
1557
1558<p><var>R</var> returns the value composed of the series of bytes it read.
1559This implies that some bytes within the value may be <tt>undef</tt>
1560<b>without</b> the entire value being <tt>undef</tt>. Note that this only
1561defines the semantics of the operation; it doesn't mean that targets will
1562emit more than one instruction to read the series of bytes.</p>
1563
1564<p>Note that in cases where none of the atomic intrinsics are used, this model
1565places only one restriction on IR transformations on top of what is required
1566for single-threaded execution: introducing a store to a byte which might not
Eli Friedman4bc9f3c2011-08-02 01:15:34 +0000[diff] [blame]1567otherwise be stored is not allowed in general. (Specifically, in the case
1568where another thread might write to and read from an address, introducing a
1569store can change a load that may see exactly one write into a load that may
1570see multiple writes.)</p>
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1571
1572<!-- FIXME: This model assumes all targets where concurrency is relevant have
1573a byte-size store which doesn't affect adjacent bytes. As far as I can tell,
1574none of the backends currently in the tree fall into this category; however,
1575there might be targets which care. If there are, we want a paragraph
1576like the following:
1577
1578Targets may specify that stores narrower than a certain width are not
1579available; on such a target, for the purposes of this model, treat any
1580non-atomic write with an alignment or width less than the minimum width
1581as if it writes to the relevant surrounding bytes.
1582-->
1583
1584</div>
1585
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1586<!-- ======================================================================= -->
NAKAMURA Takumi0300d882011-08-12 06:17:17 +0000[diff] [blame]1587<h3>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1588 <a name="ordering">Atomic Memory Ordering Constraints</a>
NAKAMURA Takumi0300d882011-08-12 06:17:17 +0000[diff] [blame]1589</h3>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1590
NAKAMURA Takumi0300d882011-08-12 06:17:17 +0000[diff] [blame]1591<div>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1592
1593<p>Atomic instructions (<a href="#i_cmpxchg"><code>cmpxchg</code></a>,
Eli Friedman59b66882011-08-09 23:02:53 +0000[diff] [blame]1594<a href="#i_atomicrmw"><code>atomicrmw</code></a>,
1595<a href="#i_fence"><code>fence</code></a>,
1596<a href="#i_load"><code>atomic load</code></a>, and
Eli Friedman75362532011-08-09 23:26:12 +0000[diff] [blame]1597<a href="#i_store"><code>atomic store</code></a>) take an ordering parameter
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1598that determines which other atomic instructions on the same address they
1599<i>synchronize with</i>. These semantics are borrowed from Java and C++0x,
1600but are somewhat more colloquial. If these descriptions aren't precise enough,
Eli Friedman95f69a42011-08-22 21:35:27 +0000[diff] [blame^]1601check those specs (see spec references in the
1602<a href="Atomic.html#introduction">atomics guide</a>).
1603<a href="#i_fence"><code>fence</code></a> instructions
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1604treat these orderings somewhat differently since they don't take an address.
1605See that instruction's documentation for details.</p>
1606
Eli Friedman95f69a42011-08-22 21:35:27 +0000[diff] [blame^]1607<p>For a simpler introduction to the ordering constraints, see the
1608<a href="Atomics.html">LLVM Atomic Instructions and Concurrency Guide</a>.</p>
1609
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1610<dl>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1611<dt><code>unordered</code></dt>
1612<dd>The set of values that can be read is governed by the happens-before
1613partial order. A value cannot be read unless some operation wrote it.
1614This is intended to provide a guarantee strong enough to model Java's
1615non-volatile shared variables. This ordering cannot be specified for
1616read-modify-write operations; it is not strong enough to make them atomic
1617in any interesting way.</dd>
1618<dt><code>monotonic</code></dt>
1619<dd>In addition to the guarantees of <code>unordered</code>, there is a single
1620total order for modifications by <code>monotonic</code> operations on each
1621address. All modification orders must be compatible with the happens-before
1622order. There is no guarantee that the modification orders can be combined to
1623a global total order for the whole program (and this often will not be
1624possible). The read in an atomic read-modify-write operation
1625(<a href="#i_cmpxchg"><code>cmpxchg</code></a> and
1626<a href="#i_atomicrmw"><code>atomicrmw</code></a>)
1627reads the value in the modification order immediately before the value it
1628writes. If one atomic read happens before another atomic read of the same
1629address, the later read must see the same value or a later value in the
1630address's modification order. This disallows reordering of
1631<code>monotonic</code> (or stronger) operations on the same address. If an
1632address is written <code>monotonic</code>ally by one thread, and other threads
1633<code>monotonic</code>ally read that address repeatedly, the other threads must
Eli Friedman95f69a42011-08-22 21:35:27 +0000[diff] [blame^]1634eventually see the write. This corresponds to the C++0x/C1x
1635<code>memory_order_relaxed</code>.</dd>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1636<dt><code>acquire</code></dt>
1637<dd>In addition to the guarantees of <code>monotonic</code>, if this operation
1638reads a value written by a <code>release</code> atomic operation, it
Eli Friedman95f69a42011-08-22 21:35:27 +0000[diff] [blame^]1639<i>synchronizes-with</i> that operation. This corresponds to the C++0x/C1x
1640<code>memory_order_acquire</code>.</dd>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1641<dt><code>release</code></dt>
1642<dd>In addition to the guarantees of <code>monotonic</code>,
Eli Friedman95f69a42011-08-22 21:35:27 +0000[diff] [blame^]1643a <i>synchronizes-with</i> edge may be formed with an <code>acquire</code>
1644operation. This is intended to model C++'s <code>memory_order_release</code>.</dd>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1645<dt><code>acq_rel</code> (acquire+release)</dt><dd>Acts as both an
Eli Friedman95f69a42011-08-22 21:35:27 +0000[diff] [blame^]1646<code>acquire</code> and <code>release</code> operation on its address.
1647This corresponds to the C++0x/C1x <code>memory_order_acq_rel</code>.</dd>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1648<dt><code>seq_cst</code> (sequentially consistent)</dt><dd>
1649<dd>In addition to the guarantees of <code>acq_rel</code>
1650(<code>acquire</code> for an operation which only reads, <code>release</code>
1651for an operation which only writes), there is a global total order on all
1652sequentially-consistent operations on all addresses, which is consistent with
1653the <i>happens-before</i> partial order and with the modification orders of
1654all the affected addresses. Each sequentially-consistent read sees the last
Eli Friedman95f69a42011-08-22 21:35:27 +0000[diff] [blame^]1655preceding write to the same address in this global order. This corresponds
1656to the C++0x/C1x <code>memory_order_seq_cst</code> and Java volatile.</dd>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]1657</dl>
1658
1659<p id="singlethread">If an atomic operation is marked <code>singlethread</code>,
1660it only <i>synchronizes with</i> or participates in modification and seq_cst
1661total orderings with other operations running in the same thread (for example,
1662in signal handlers).</p>
1663
1664</div>
1665
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1666</div>
1667
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1668<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1669<h2><a name="typesystem">Type System</a></h2>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1670<!-- *********************************************************************** -->
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1671
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1672<div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1673
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1674<p>The LLVM type system is one of the most important features of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1675 intermediate representation. Being typed enables a number of optimizations
1676 to be performed on the intermediate representation directly, without having
1677 to do extra analyses on the side before the transformation. A strong type
1678 system makes it easier to read the generated code and enables novel analyses
1679 and transformations that are not feasible to perform on normal three address
1680 code representations.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1681
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1682<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1683<h3>
1684 <a name="t_classifications">Type Classifications</a>
1685</h3>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1686
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1687<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1688
1689<p>The types fall into a few useful classifications:</p>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]1690
1691<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1692 <tbody>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1693 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1694 <tr>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1695 <td><a href="#t_integer">integer</a></td>
Reid Spencer138249b2007-05-16 18:44:01 +0000[diff] [blame]1696 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1697 </tr>
1698 <tr>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1699 <td><a href="#t_floating">floating point</a></td>
1700 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1701 </tr>
1702 <tr>
1703 <td><a name="t_firstclass">first class</a></td>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1704 <td><a href="#t_integer">integer</a>,
1705 <a href="#t_floating">floating point</a>,
1706 <a href="#t_pointer">pointer</a>,
Dan Gohman08783a882008-06-18 18:42:13 +0000[diff] [blame]1707 <a href="#t_vector">vector</a>,
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]1708 <a href="#t_struct">structure</a>,
1709 <a href="#t_array">array</a>,
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1710 <a href="#t_label">label</a>,
1711 <a href="#t_metadata">metadata</a>.
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1712 </td>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1713 </tr>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1714 <tr>
1715 <td><a href="#t_primitive">primitive</a></td>
1716 <td><a href="#t_label">label</a>,
1717 <a href="#t_void">void</a>,
Tobias Grosser4c8c95b2010-12-28 20:29:31 +0000[diff] [blame]1718 <a href="#t_integer">integer</a>,
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1719 <a href="#t_floating">floating point</a>,
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]1720 <a href="#t_x86mmx">x86mmx</a>,
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1721 <a href="#t_metadata">metadata</a>.</td>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1722 </tr>
1723 <tr>
1724 <td><a href="#t_derived">derived</a></td>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1725 <td><a href="#t_array">array</a>,
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1726 <a href="#t_function">function</a>,
1727 <a href="#t_pointer">pointer</a>,
1728 <a href="#t_struct">structure</a>,
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1729 <a href="#t_vector">vector</a>,
1730 <a href="#t_opaque">opaque</a>.
Dan Gohman93bf60d2008-10-14 16:32:04 +0000[diff] [blame]1731 </td>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1732 </tr>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1733 </tbody>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1734</table>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]1735
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1736<p>The <a href="#t_firstclass">first class</a> types are perhaps the most
1737 important. Values of these types are the only ones which can be produced by
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1738 instructions.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1739
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1740</div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1741
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1742<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1743<h3>
1744 <a name="t_primitive">Primitive Types</a>
1745</h3>
Chris Lattner43542b32008-01-04 04:34:14 +0000[diff] [blame]1746
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1747<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1748
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1749<p>The primitive types are the fundamental building blocks of the LLVM
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1750 system.</p>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1751
1752<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1753<h4>
1754 <a name="t_integer">Integer Type</a>
1755</h4>
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1756
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1757<div>
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1758
1759<h5>Overview:</h5>
1760<p>The integer type is a very simple type that simply specifies an arbitrary
1761 bit width for the integer type desired. Any bit width from 1 bit to
1762 2<sup>23</sup>-1 (about 8 million) can be specified.</p>
1763
1764<h5>Syntax:</h5>
1765<pre>
1766 iN
1767</pre>
1768
1769<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1770 value.</p>
1771
1772<h5>Examples:</h5>
1773<table class="layout">
1774 <tr class="layout">
1775 <td class="left"><tt>i1</tt></td>
1776 <td class="left">a single-bit integer.</td>
1777 </tr>
1778 <tr class="layout">
1779 <td class="left"><tt>i32</tt></td>
1780 <td class="left">a 32-bit integer.</td>
1781 </tr>
1782 <tr class="layout">
1783 <td class="left"><tt>i1942652</tt></td>
1784 <td class="left">a really big integer of over 1 million bits.</td>
1785 </tr>
1786</table>
1787
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1788</div>
1789
1790<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1791<h4>
1792 <a name="t_floating">Floating Point Types</a>
1793</h4>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1794
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1795<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1796
1797<table>
1798 <tbody>
1799 <tr><th>Type</th><th>Description</th></tr>
1800 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1801 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1802 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1803 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1804 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1805 </tbody>
1806</table>
1807
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1808</div>
1809
1810<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1811<h4>
1812 <a name="t_x86mmx">X86mmx Type</a>
1813</h4>
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]1814
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1815<div>
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]1816
1817<h5>Overview:</h5>
1818<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>
1819
1820<h5>Syntax:</h5>
1821<pre>
Dale Johannesenb1f0ff12010-10-01 01:07:02 +0000[diff] [blame]1822 x86mmx
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]1823</pre>
1824
1825</div>
1826
1827<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1828<h4>
1829 <a name="t_void">Void Type</a>
1830</h4>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1831
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1832<div>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1833
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1834<h5>Overview:</h5>
1835<p>The void type does not represent any value and has no size.</p>
1836
1837<h5>Syntax:</h5>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1838<pre>
1839 void
1840</pre>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1841
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1842</div>
1843
1844<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1845<h4>
1846 <a name="t_label">Label Type</a>
1847</h4>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1848
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1849<div>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1850
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1851<h5>Overview:</h5>
1852<p>The label type represents code labels.</p>
1853
1854<h5>Syntax:</h5>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1855<pre>
1856 label
1857</pre>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1858
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1859</div>
1860
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1861<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1862<h4>
1863 <a name="t_metadata">Metadata Type</a>
1864</h4>
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1865
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1866<div>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1867
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1868<h5>Overview:</h5>
Nick Lewycky93e06a52009-09-27 23:27:42 +0000[diff] [blame]1869<p>The metadata type represents embedded metadata. No derived types may be
1870 created from metadata except for <a href="#t_function">function</a>
1871 arguments.
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1872
1873<h5>Syntax:</h5>
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1874<pre>
1875 metadata
1876</pre>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1877
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1878</div>
1879
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1880</div>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1881
1882<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1883<h3>
1884 <a name="t_derived">Derived Types</a>
1885</h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1886
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1887<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1888
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1889<p>The real power in LLVM comes from the derived types in the system. This is
1890 what allows a programmer to represent arrays, functions, pointers, and other
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1891 useful types. Each of these types contain one or more element types which
1892 may be a primitive type, or another derived type. For example, it is
1893 possible to have a two dimensional array, using an array as the element type
1894 of another array.</p>
Dan Gohman142ccc02009-01-24 15:58:40 +0000[diff] [blame]1895
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]1896</div>
1897
1898
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1899<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1900<h4>
1901 <a name="t_aggregate">Aggregate Types</a>
1902</h4>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1903
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1904<div>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1905
1906<p>Aggregate Types are a subset of derived types that can contain multiple
1907 member types. <a href="#t_array">Arrays</a>,
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]1908 <a href="#t_struct">structs</a>, and <a href="#t_vector">vectors</a> are
1909 aggregate types.</p>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1910
1911</div>
1912
Reid Spencer138249b2007-05-16 18:44:01 +0000[diff] [blame]1913<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1914<h4>
1915 <a name="t_array">Array Type</a>
1916</h4>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1917
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1918<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1919
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1920<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1921<p>The array type is a very simple derived type that arranges elements
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1922 sequentially in memory. The array type requires a size (number of elements)
1923 and an underlying data type.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1924
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]1925<h5>Syntax:</h5>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1926<pre>
1927 [&lt;# elements&gt; x &lt;elementtype&gt;]
1928</pre>
1929
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1930<p>The number of elements is a constant integer value; <tt>elementtype</tt> may
1931 be any type with a size.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1932
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]1933<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1934<table class="layout">
1935 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1936 <td class="left"><tt>[40 x i32]</tt></td>
1937 <td class="left">Array of 40 32-bit integer values.</td>
1938 </tr>
1939 <tr class="layout">
1940 <td class="left"><tt>[41 x i32]</tt></td>
1941 <td class="left">Array of 41 32-bit integer values.</td>
1942 </tr>
1943 <tr class="layout">
1944 <td class="left"><tt>[4 x i8]</tt></td>
1945 <td class="left">Array of 4 8-bit integer values.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1946 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1947</table>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1948<p>Here are some examples of multidimensional arrays:</p>
1949<table class="layout">
1950 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1951 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1952 <td class="left">3x4 array of 32-bit integer values.</td>
1953 </tr>
1954 <tr class="layout">
1955 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1956 <td class="left">12x10 array of single precision floating point values.</td>
1957 </tr>
1958 <tr class="layout">
1959 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1960 <td class="left">2x3x4 array of 16-bit integer values.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1961 </tr>
1962</table>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +0000[diff] [blame]1963
Dan Gohmanc74bc282009-11-09 19:01:53 +0000[diff] [blame]1964<p>There is no restriction on indexing beyond the end of the array implied by
1965 a static type (though there are restrictions on indexing beyond the bounds
1966 of an allocated object in some cases). This means that single-dimension
1967 'variable sized array' addressing can be implemented in LLVM with a zero
1968 length array type. An implementation of 'pascal style arrays' in LLVM could
1969 use the type "<tt>{ i32, [0 x float]}</tt>", for example.</p>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +0000[diff] [blame]1970
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1971</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1972
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1973<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1974<h4>
1975 <a name="t_function">Function Type</a>
1976</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1977
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1978<div>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]1979
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1980<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1981<p>The function type can be thought of as a function signature. It consists of
1982 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]1983 function type is a first class type or a void type.</p>
Devang Pateld6cff512008-03-10 20:49:15 +0000[diff] [blame]1984
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1985<h5>Syntax:</h5>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]1986<pre>
Nick Lewycky14d1ccc2009-09-27 07:55:32 +0000[diff] [blame]1987 &lt;returntype&gt; (&lt;parameter list&gt;)
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]1988</pre>
1989
John Criswell4c0cf7f2005-10-24 16:17:18 +0000[diff] [blame]1990<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]1991 specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1992 which indicates that the function takes a variable number of arguments.
1993 Variable argument functions can access their arguments with
1994 the <a href="#int_varargs">variable argument handling intrinsic</a>
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]1995 functions. '<tt>&lt;returntype&gt;</tt>' is any type except
Nick Lewycky93e06a52009-09-27 23:27:42 +0000[diff] [blame]1996 <a href="#t_label">label</a>.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]1997
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1998<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1999<table class="layout">
2000 <tr class="layout">
Reid Spencer58c08712006-12-31 07:18:34 +0000[diff] [blame]2001 <td class="left"><tt>i32 (i32)</tt></td>
2002 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2003 </td>
Reid Spencer58c08712006-12-31 07:18:34 +0000[diff] [blame]2004 </tr><tr class="layout">
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]2005 <td class="left"><tt>float&nbsp;(i16,&nbsp;i32&nbsp;*)&nbsp;*
Reid Spencer655dcc62006-12-31 07:20:23 +0000[diff] [blame]2006 </tt></td>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2007 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]2008 an <tt>i16</tt> and a <a href="#t_pointer">pointer</a> to <tt>i32</tt>,
2009 returning <tt>float</tt>.
Reid Spencer58c08712006-12-31 07:18:34 +0000[diff] [blame]2010 </td>
2011 </tr><tr class="layout">
2012 <td class="left"><tt>i32 (i8*, ...)</tt></td>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2013 <td class="left">A vararg function that takes at least one
2014 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
2015 which returns an integer. This is the signature for <tt>printf</tt> in
Reid Spencer58c08712006-12-31 07:18:34 +0000[diff] [blame]2016 LLVM.
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2017 </td>
Devang Patele3dfc1c2008-03-24 05:35:41 +0000[diff] [blame]2018 </tr><tr class="layout">
2019 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Nick Lewycky14d1ccc2009-09-27 07:55:32 +0000[diff] [blame]2020 <td class="left">A function taking an <tt>i32</tt>, returning a
2021 <a href="#t_struct">structure</a> containing two <tt>i32</tt> values
Devang Patele3dfc1c2008-03-24 05:35:41 +0000[diff] [blame]2022 </td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2023 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2024</table>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]2025
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2026</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2027
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2028<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2029<h4>
2030 <a name="t_struct">Structure Type</a>
2031</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2032
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2033<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2034
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2035<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2036<p>The structure type is used to represent a collection of data members together
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2037 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]2038
Jeffrey Yasskinf991bbb2010-01-11 19:19:26 +0000[diff] [blame]2039<p>Structures in memory are accessed using '<tt><a href="#i_load">load</a></tt>'
2040 and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field
2041 with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
2042 Structures in registers are accessed using the
2043 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' and
2044 '<tt><a href="#i_insertvalue">insertvalue</a></tt>' instructions.</p>
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2045
2046<p>Structures may optionally be "packed" structures, which indicate that the
2047 alignment of the struct is one byte, and that there is no padding between
Chris Lattner190552d2011-08-12 17:31:02 +0000[diff] [blame]2048 the elements. In non-packed structs, padding between field types is inserted
2049 as defined by the TargetData string in the module, which is required to match
2050 what the underlying processor expects.</p>
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2051
Chris Lattner190552d2011-08-12 17:31:02 +0000[diff] [blame]2052<p>Structures can either be "literal" or "identified". A literal structure is
2053 defined inline with other types (e.g. <tt>{i32, i32}*</tt>) whereas identified
2054 types are always defined at the top level with a name. Literal types are
2055 uniqued by their contents and can never be recursive or opaque since there is
Chris Lattner32531732011-08-12 18:12:40 +0000[diff] [blame]2056 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]2057 never uniqued.
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2058</p>
2059
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2060<h5>Syntax:</h5>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]2061<pre>
Chris Lattner190552d2011-08-12 17:31:02 +0000[diff] [blame]2062 %T1 = type { &lt;type list&gt; } <i>; Identified normal struct type</i>
2063 %T2 = type &lt;{ &lt;type list&gt; }&gt; <i>; Identified packed struct type</i>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]2064</pre>
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2065
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2066<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2067<table class="layout">
2068 <tr class="layout">
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]2069 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
2070 <td class="left">A triple of three <tt>i32</tt> values</td>
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2071 </tr>
2072 <tr class="layout">
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]2073 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
2074 <td class="left">A pair, where the first element is a <tt>float</tt> and the
2075 second element is a <a href="#t_pointer">pointer</a> to a
2076 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
2077 an <tt>i32</tt>.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2078 </tr>
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2079 <tr class="layout">
2080 <td class="left"><tt>&lt;{ i8, i32 }&gt;</tt></td>
2081 <td class="left">A packed struct known to be 5 bytes in size.</td>
2082 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2083</table>
Dan Gohman142ccc02009-01-24 15:58:40 +0000[diff] [blame]2084
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2085</div>
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2086
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2087<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2088<h4>
Chris Lattner2a843822011-07-23 19:59:08 +0000[diff] [blame]2089 <a name="t_opaque">Opaque Structure Types</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2090</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2091
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2092<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2093
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]2094<h5>Overview:</h5>
Chris Lattner2a843822011-07-23 19:59:08 +0000[diff] [blame]2095<p>Opaque structure types are used to represent named structure types that do
2096 not have a body specified. This corresponds (for example) to the C notion of
2097 a forward declared structure.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2098
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]2099<h5>Syntax:</h5>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]2100<pre>
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2101 %X = type opaque
2102 %52 = type opaque
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]2103</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2104
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]2105<h5>Examples:</h5>
2106<table class="layout">
2107 <tr class="layout">
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2108 <td class="left"><tt>opaque</tt></td>
2109 <td class="left">An opaque type.</td>
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]2110 </tr>
2111</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2112
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]2113</div>
2114
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2115
2116
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]2117<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2118<h4>
2119 <a name="t_pointer">Pointer Type</a>
2120</h4>
Chris Lattner4a67c912009-02-08 19:53:29 +0000[diff] [blame]2121
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2122<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2123
2124<h5>Overview:</h5>
Dan Gohman88481112010-02-25 16:50:07 +0000[diff] [blame]2125<p>The pointer type is used to specify memory locations.
2126 Pointers are commonly used to reference objects in memory.</p>
2127
2128<p>Pointer types may have an optional address space attribute defining the
2129 numbered address space where the pointed-to object resides. The default
2130 address space is number zero. The semantics of non-zero address
2131 spaces are target-specific.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2132
2133<p>Note that LLVM does not permit pointers to void (<tt>void*</tt>) nor does it
2134 permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</tt> instead.</p>
Chris Lattner4a67c912009-02-08 19:53:29 +0000[diff] [blame]2135
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]2136<h5>Syntax:</h5>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]2137<pre>
2138 &lt;type&gt; *
2139</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2140
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]2141<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2142<table class="layout">
2143 <tr class="layout">
Dan Gohman623806e2009-01-04 23:44:43 +0000[diff] [blame]2144 <td class="left"><tt>[4 x i32]*</tt></td>
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]2145 <td class="left">A <a href="#t_pointer">pointer</a> to <a
2146 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
2147 </tr>
2148 <tr class="layout">
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]2149 <td class="left"><tt>i32 (i32*) *</tt></td>
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]2150 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]2151 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]2152 <tt>i32</tt>.</td>
2153 </tr>
2154 <tr class="layout">
2155 <td class="left"><tt>i32 addrspace(5)*</tt></td>
2156 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
2157 that resides in address space #5.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2158 </tr>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2159</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2160
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2161</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2162
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]2163<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2164<h4>
2165 <a name="t_vector">Vector Type</a>
2166</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2167
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2168<div>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]2169
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]2170<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2171<p>A vector type is a simple derived type that represents a vector of elements.
2172 Vector types are used when multiple primitive data are operated in parallel
2173 using a single instruction (SIMD). A vector type requires a size (number of
Duncan Sands31c0e0e2009-11-27 13:38:03 +0000[diff] [blame]2174 elements) and an underlying primitive data type. Vector types are considered
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2175 <a href="#t_firstclass">first class</a>.</p>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]2176
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]2177<h5>Syntax:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]2178<pre>
2179 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
2180</pre>
2181
Chris Lattnerf11031a2010-10-10 18:20:35 +0000[diff] [blame]2182<p>The number of elements is a constant integer value larger than 0; elementtype
2183 may be any integer or floating point type. Vectors of size zero are not
2184 allowed, and pointers are not allowed as the element type.</p>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]2185
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]2186<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2187<table class="layout">
2188 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]2189 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
2190 <td class="left">Vector of 4 32-bit integer values.</td>
2191 </tr>
2192 <tr class="layout">
2193 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
2194 <td class="left">Vector of 8 32-bit floating-point values.</td>
2195 </tr>
2196 <tr class="layout">
2197 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
2198 <td class="left">Vector of 2 64-bit integer values.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2199 </tr>
2200</table>
Dan Gohman142ccc02009-01-24 15:58:40 +0000[diff] [blame]2201
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2202</div>
2203
Bill Wendlingae8b5ea2011-07-31 06:47:33 +0000[diff] [blame]2204</div>
2205
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2206<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2207<h2><a name="constants">Constants</a></h2>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2208<!-- *********************************************************************** -->
2209
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2210<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2211
2212<p>LLVM has several different basic types of constants. This section describes
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2213 them all and their syntax.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2214
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2215<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2216<h3>
2217 <a name="simpleconstants">Simple Constants</a>
2218</h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2219
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2220<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2221
2222<dl>
2223 <dt><b>Boolean constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2224 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]2225 constants of the <tt><a href="#t_integer">i1</a></tt> type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2226
2227 <dt><b>Integer constants</b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2228 <dd>Standard integers (such as '4') are constants of
2229 the <a href="#t_integer">integer</a> type. Negative numbers may be used
2230 with integer types.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2231
2232 <dt><b>Floating point constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2233 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2234 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
2235 notation (see below). The assembler requires the exact decimal value of a
2236 floating-point constant. For example, the assembler accepts 1.25 but
2237 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
2238 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2239
2240 <dt><b>Null pointer constants</b></dt>
John Criswelldfe6a862004-12-10 15:51:16 +0000[diff] [blame]2241 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2242 and must be of <a href="#t_pointer">pointer type</a>.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2243</dl>
2244
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2245<p>The one non-intuitive notation for constants is the hexadecimal form of
2246 floating point constants. For example, the form '<tt>double
2247 0x432ff973cafa8000</tt>' is equivalent to (but harder to read than)
2248 '<tt>double 4.5e+15</tt>'. The only time hexadecimal floating point
2249 constants are required (and the only time that they are generated by the
2250 disassembler) is when a floating point constant must be emitted but it cannot
2251 be represented as a decimal floating point number in a reasonable number of
2252 digits. For example, NaN's, infinities, and other special values are
2253 represented in their IEEE hexadecimal format so that assembly and disassembly
2254 do not cause any bits to change in the constants.</p>
2255
Dale Johannesencd4a3012009-02-11 22:14:51 +0000[diff] [blame]2256<p>When using the hexadecimal form, constants of types float and double are
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2257 represented using the 16-digit form shown above (which matches the IEEE754
2258 representation for double); float values must, however, be exactly
2259 representable as IEE754 single precision. Hexadecimal format is always used
2260 for long double, and there are three forms of long double. The 80-bit format
2261 used by x86 is represented as <tt>0xK</tt> followed by 20 hexadecimal digits.
2262 The 128-bit format used by PowerPC (two adjacent doubles) is represented
2263 by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit format
2264 is represented by <tt>0xL</tt> followed by 32 hexadecimal digits; no
2265 currently supported target uses this format. Long doubles will only work if
2266 they match the long double format on your target. All hexadecimal formats
2267 are big-endian (sign bit at the left).</p>
2268
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]2269<p>There are no constants of type x86mmx.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2270</div>
2271
2272<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2273<h3>
Bill Wendling972b7202009-07-20 02:32:41 +0000[diff] [blame]2274<a name="aggregateconstants"></a> <!-- old anchor -->
2275<a name="complexconstants">Complex Constants</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2276</h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2277
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2278<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2279
Chris Lattner361bfcd2009-02-28 18:32:25 +0000[diff] [blame]2280<p>Complex constants are a (potentially recursive) combination of simple
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2281 constants and smaller complex constants.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2282
2283<dl>
2284 <dt><b>Structure constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2285 <dd>Structure constants are represented with notation similar to structure
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2286 type definitions (a comma separated list of elements, surrounded by braces
2287 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
2288 where "<tt>@G</tt>" is declared as "<tt>@G = external global i32</tt>".
2289 Structure constants must have <a href="#t_struct">structure type</a>, and
2290 the number and types of elements must match those specified by the
2291 type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2292
2293 <dt><b>Array constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2294 <dd>Array constants are represented with notation similar to array type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2295 definitions (a comma separated list of elements, surrounded by square
2296 brackets (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74
2297 ]</tt>". Array constants must have <a href="#t_array">array type</a>, and
2298 the number and types of elements must match those specified by the
2299 type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2300
Reid Spencer404a3252007-02-15 03:07:05 +0000[diff] [blame]2301 <dt><b>Vector constants</b></dt>
Reid Spencer404a3252007-02-15 03:07:05 +0000[diff] [blame]2302 <dd>Vector constants are represented with notation similar to vector type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2303 definitions (a comma separated list of elements, surrounded by
2304 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32
2305 42, i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must
2306 have <a href="#t_vector">vector type</a>, and the number and types of
2307 elements must match those specified by the type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2308
2309 <dt><b>Zero initialization</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2310 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]2311 value to zero of <em>any</em> type, including scalar and
2312 <a href="#t_aggregate">aggregate</a> types.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2313 This is often used to avoid having to print large zero initializers
2314 (e.g. for large arrays) and is always exactly equivalent to using explicit
2315 zero initializers.</dd>
Nick Lewycky49f89192009-04-04 07:22:01 +0000[diff] [blame]2316
2317 <dt><b>Metadata node</b></dt>
Nick Lewycky8e2c4f42009-05-30 16:08:30 +0000[diff] [blame]2318 <dd>A metadata node is a structure-like constant with
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2319 <a href="#t_metadata">metadata type</a>. For example: "<tt>metadata !{
2320 i32 0, metadata !"test" }</tt>". Unlike other constants that are meant to
2321 be interpreted as part of the instruction stream, metadata is a place to
2322 attach additional information such as debug info.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2323</dl>
2324
2325</div>
2326
2327<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2328<h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2329 <a name="globalconstants">Global Variable and Function Addresses</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2330</h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2331
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2332<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2333
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2334<p>The addresses of <a href="#globalvars">global variables</a>
2335 and <a href="#functionstructure">functions</a> are always implicitly valid
2336 (link-time) constants. These constants are explicitly referenced when
2337 the <a href="#identifiers">identifier for the global</a> is used and always
2338 have <a href="#t_pointer">pointer</a> type. For example, the following is a
2339 legal LLVM file:</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2340
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2341<pre class="doc_code">
Chris Lattner00538a12007-06-06 18:28:13 +0000[diff] [blame]2342@X = global i32 17
2343@Y = global i32 42
2344@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2345</pre>
2346
2347</div>
2348
2349<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2350<h3>
2351 <a name="undefvalues">Undefined Values</a>
2352</h3>
2353
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2354<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2355
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2356<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]2357 indicates that the user of the value may receive an unspecified bit-pattern.
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2358 Undefined values may be of any type (other than '<tt>label</tt>'
2359 or '<tt>void</tt>') and be used anywhere a constant is permitted.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2360
Chris Lattner92ada5d2009-09-11 01:49:31 +0000[diff] [blame]2361<p>Undefined values are useful because they indicate to the compiler that the
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2362 program is well defined no matter what value is used. This gives the
2363 compiler more freedom to optimize. Here are some examples of (potentially
2364 surprising) transformations that are valid (in pseudo IR):</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2365
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2366
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2367<pre class="doc_code">
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2368 %A = add %X, undef
2369 %B = sub %X, undef
2370 %C = xor %X, undef
2371Safe:
2372 %A = undef
2373 %B = undef
2374 %C = undef
2375</pre>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2376
2377<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]2378 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]2379
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2380<pre class="doc_code">
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2381 %A = or %X, undef
2382 %B = and %X, undef
2383Safe:
2384 %A = -1
2385 %B = 0
2386Unsafe:
2387 %A = undef
2388 %B = undef
2389</pre>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2390
2391<p>These logical operations have bits that are not always affected by the input.
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2392 For example, if <tt>%X</tt> has a zero bit, then the output of the
2393 '<tt>and</tt>' operation will always be a zero for that bit, no matter what
2394 the corresponding bit from the '<tt>undef</tt>' is. As such, it is unsafe to
2395 optimize or assume that the result of the '<tt>and</tt>' is '<tt>undef</tt>'.
2396 However, it is safe to assume that all bits of the '<tt>undef</tt>' could be
2397 0, and optimize the '<tt>and</tt>' to 0. Likewise, it is safe to assume that
2398 all the bits of the '<tt>undef</tt>' operand to the '<tt>or</tt>' could be
2399 set, allowing the '<tt>or</tt>' to be folded to -1.</p>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2400
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2401<pre class="doc_code">
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2402 %A = select undef, %X, %Y
2403 %B = select undef, 42, %Y
2404 %C = select %X, %Y, undef
2405Safe:
2406 %A = %X (or %Y)
2407 %B = 42 (or %Y)
2408 %C = %Y
2409Unsafe:
2410 %A = undef
2411 %B = undef
2412 %C = undef
2413</pre>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2414
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2415<p>This set of examples shows that undefined '<tt>select</tt>' (and conditional
2416 branch) conditions can go <em>either way</em>, but they have to come from one
2417 of the two operands. In the <tt>%A</tt> example, if <tt>%X</tt> and
2418 <tt>%Y</tt> were both known to have a clear low bit, then <tt>%A</tt> would
2419 have to have a cleared low bit. However, in the <tt>%C</tt> example, the
2420 optimizer is allowed to assume that the '<tt>undef</tt>' operand could be the
2421 same as <tt>%Y</tt>, allowing the whole '<tt>select</tt>' to be
2422 eliminated.</p>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2423
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2424<pre class="doc_code">
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2425 %A = xor undef, undef
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2426
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2427 %B = undef
2428 %C = xor %B, %B
2429
2430 %D = undef
2431 %E = icmp lt %D, 4
2432 %F = icmp gte %D, 4
2433
2434Safe:
2435 %A = undef
2436 %B = undef
2437 %C = undef
2438 %D = undef
2439 %E = undef
2440 %F = undef
2441</pre>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2442
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2443<p>This example points out that two '<tt>undef</tt>' operands are not
2444 necessarily the same. This can be surprising to people (and also matches C
2445 semantics) where they assume that "<tt>X^X</tt>" is always zero, even
2446 if <tt>X</tt> is undefined. This isn't true for a number of reasons, but the
2447 short answer is that an '<tt>undef</tt>' "variable" can arbitrarily change
2448 its value over its "live range". This is true because the variable doesn't
2449 actually <em>have a live range</em>. Instead, the value is logically read
2450 from arbitrary registers that happen to be around when needed, so the value
2451 is not necessarily consistent over time. In fact, <tt>%A</tt> and <tt>%C</tt>
2452 need to have the same semantics or the core LLVM "replace all uses with"
2453 concept would not hold.</p>
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2454
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2455<pre class="doc_code">
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2456 %A = fdiv undef, %X
2457 %B = fdiv %X, undef
2458Safe:
2459 %A = undef
2460b: unreachable
2461</pre>
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2462
2463<p>These examples show the crucial difference between an <em>undefined
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2464 value</em> and <em>undefined behavior</em>. An undefined value (like
2465 '<tt>undef</tt>') is allowed to have an arbitrary bit-pattern. This means that
2466 the <tt>%A</tt> operation can be constant folded to '<tt>undef</tt>', because
2467 the '<tt>undef</tt>' could be an SNaN, and <tt>fdiv</tt> is not (currently)
2468 defined on SNaN's. However, in the second example, we can make a more
2469 aggressive assumption: because the <tt>undef</tt> is allowed to be an
2470 arbitrary value, we are allowed to assume that it could be zero. Since a
2471 divide by zero has <em>undefined behavior</em>, we are allowed to assume that
2472 the operation does not execute at all. This allows us to delete the divide and
2473 all code after it. Because the undefined operation "can't happen", the
2474 optimizer can assume that it occurs in dead code.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2475
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2476<pre class="doc_code">
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2477a: store undef -> %X
2478b: store %X -> undef
2479Safe:
2480a: &lt;deleted&gt;
2481b: unreachable
2482</pre>
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2483
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2484<p>These examples reiterate the <tt>fdiv</tt> example: a store <em>of</em> an
2485 undefined value can be assumed to not have any effect; we can assume that the
2486 value is overwritten with bits that happen to match what was already there.
2487 However, a store <em>to</em> an undefined location could clobber arbitrary
2488 memory, therefore, it has undefined behavior.</p>
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2489
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2490</div>
2491
2492<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2493<h3>
2494 <a name="trapvalues">Trap Values</a>
2495</h3>
2496
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2497<div>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2498
Dan Gohmanb2a709b2010-04-26 20:21:21 +0000[diff] [blame]2499<p>Trap values are similar to <a href="#undefvalues">undef values</a>, however
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2500 instead of representing an unspecified bit pattern, they represent the
2501 fact that an instruction or constant expression which cannot evoke side
2502 effects has nevertheless detected a condition which results in undefined
Dan Gohmanb2a709b2010-04-26 20:21:21 +0000[diff] [blame]2503 behavior.</p>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2504
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2505<p>There is currently no way of representing a trap value in the IR; they
Dan Gohmanac355aa2010-05-03 14:51:43 +0000[diff] [blame]2506 only exist when produced by operations such as
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2507 <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]2508
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2509<p>Trap value behavior is defined in terms of value <i>dependence</i>:</p>
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2510
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2511<ul>
2512<li>Values other than <a href="#i_phi"><tt>phi</tt></a> nodes depend on
2513 their operands.</li>
2514
2515<li><a href="#i_phi"><tt>Phi</tt></a> nodes depend on the operand corresponding
2516 to their dynamic predecessor basic block.</li>
2517
2518<li>Function arguments depend on the corresponding actual argument values in
2519 the dynamic callers of their functions.</li>
2520
2521<li><a href="#i_call"><tt>Call</tt></a> instructions depend on the
2522 <a href="#i_ret"><tt>ret</tt></a> instructions that dynamically transfer
2523 control back to them.</li>
2524
Dan Gohman7292a752010-05-03 14:55:22 +0000[diff] [blame]2525<li><a href="#i_invoke"><tt>Invoke</tt></a> instructions depend on the
2526 <a href="#i_ret"><tt>ret</tt></a>, <a href="#i_unwind"><tt>unwind</tt></a>,
2527 or exception-throwing call instructions that dynamically transfer control
2528 back to them.</li>
2529
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2530<li>Non-volatile loads and stores depend on the most recent stores to all of the
2531 referenced memory addresses, following the order in the IR
2532 (including loads and stores implied by intrinsics such as
2533 <a href="#int_memcpy"><tt>@llvm.memcpy</tt></a>.)</li>
2534
Dan Gohman3513ea52010-05-03 14:59:34 +0000[diff] [blame]2535<!-- TODO: In the case of multiple threads, this only applies if the store
2536 "happens-before" the load or store. -->
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2537
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2538<!-- TODO: floating-point exception state -->
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2539
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2540<li>An instruction with externally visible side effects depends on the most
2541 recent preceding instruction with externally visible side effects, following
Dan Gohman6c858db2010-07-06 15:26:33 +0000[diff] [blame]2542 the order in the IR. (This includes
2543 <a href="#volatile">volatile operations</a>.)</li>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2544
Dan Gohman7292a752010-05-03 14:55:22 +0000[diff] [blame]2545<li>An instruction <i>control-depends</i> on a
2546 <a href="#terminators">terminator instruction</a>
2547 if the terminator instruction has multiple successors and the instruction
2548 is always executed when control transfers to one of the successors, and
Chris Lattner0ab5e2c2011-04-15 05:18:47 +0000[diff] [blame]2549 may not be executed when control is transferred to another.</li>
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2550
Dan Gohmanc8454ee2011-04-12 23:05:59 +0000[diff] [blame]2551<li>Additionally, an instruction also <i>control-depends</i> on a terminator
2552 instruction if the set of instructions it otherwise depends on would be
Chris Lattner0ab5e2c2011-04-15 05:18:47 +0000[diff] [blame]2553 different if the terminator had transferred control to a different
Dan Gohmanc8454ee2011-04-12 23:05:59 +0000[diff] [blame]2554 successor.</li>
2555
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2556<li>Dependence is transitive.</li>
2557
2558</ul>
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2559
2560<p>Whenever a trap value is generated, all values which depend on it evaluate
2561 to trap. If they have side effects, the evoke their side effects as if each
2562 operand with a trap value were undef. If they have externally-visible side
2563 effects, the behavior is undefined.</p>
2564
2565<p>Here are some examples:</p>
Dan Gohman48a25882010-04-26 20:54:53 +0000[diff] [blame]2566
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2567<pre class="doc_code">
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2568entry:
2569 %trap = sub nuw i32 0, 1 ; Results in a trap value.
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2570 %still_trap = and i32 %trap, 0 ; Whereas (and i32 undef, 0) would return 0.
2571 %trap_yet_again = getelementptr i32* @h, i32 %still_trap
2572 store i32 0, i32* %trap_yet_again ; undefined behavior
2573
2574 store i32 %trap, i32* @g ; Trap value conceptually stored to memory.
2575 %trap2 = load i32* @g ; Returns a trap value, not just undef.
2576
2577 volatile store i32 %trap, i32* @g ; External observation; undefined behavior.
2578
2579 %narrowaddr = bitcast i32* @g to i16*
2580 %wideaddr = bitcast i32* @g to i64*
Nick Lewycky9c876bf2011-05-16 19:29:30 +0000[diff] [blame]2581 %trap3 = load i16* %narrowaddr ; Returns a trap value.
2582 %trap4 = load i64* %wideaddr ; Returns a trap value.
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2583
Nick Lewycky9c876bf2011-05-16 19:29:30 +0000[diff] [blame]2584 %cmp = icmp slt i32 %trap, 0 ; Returns a trap value.
2585 br i1 %cmp, label %true, label %end ; Branch to either destination.
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2586
2587true:
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2588 volatile store i32 0, i32* @g ; This is control-dependent on %cmp, so
2589 ; it has undefined behavior.
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2590 br label %end
2591
2592end:
2593 %p = phi i32 [ 0, %entry ], [ 1, %true ]
2594 ; Both edges into this PHI are
2595 ; control-dependent on %cmp, so this
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2596 ; always results in a trap value.
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2597
Dan Gohmanc8454ee2011-04-12 23:05:59 +0000[diff] [blame]2598 volatile store i32 0, i32* @g ; This would depend on the store in %true
2599 ; if %cmp is true, or the store in %entry
2600 ; otherwise, so this is undefined behavior.
2601
Nick Lewycky9c876bf2011-05-16 19:29:30 +0000[diff] [blame]2602 br i1 %cmp, label %second_true, label %second_end
Dan Gohmanc8454ee2011-04-12 23:05:59 +0000[diff] [blame]2603 ; The same branch again, but this time the
2604 ; true block doesn't have side effects.
2605
2606second_true:
2607 ; No side effects!
Nick Lewycky9c876bf2011-05-16 19:29:30 +0000[diff] [blame]2608 ret void
Dan Gohmanc8454ee2011-04-12 23:05:59 +0000[diff] [blame]2609
2610second_end:
2611 volatile store i32 0, i32* @g ; This time, the instruction always depends
2612 ; on the store in %end. Also, it is
2613 ; control-equivalent to %end, so this is
Nick Lewycky9c876bf2011-05-16 19:29:30 +0000[diff] [blame]2614 ; well-defined (again, ignoring earlier
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2615 ; undefined behavior in this example).
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2616</pre>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2617
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2618</div>
2619
2620<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2621<h3>
2622 <a name="blockaddress">Addresses of Basic Blocks</a>
2623</h3>
2624
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2625<div>
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2626
Chris Lattneraa99c942009-11-01 01:27:45 +0000[diff] [blame]2627<p><b><tt>blockaddress(@function, %block)</tt></b></p>
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2628
2629<p>The '<tt>blockaddress</tt>' constant computes the address of the specified
Chris Lattner5c5f0ac2009-10-27 21:49:40 +0000[diff] [blame]2630 basic block in the specified function, and always has an i8* type. Taking
Chris Lattneraa99c942009-11-01 01:27:45 +0000[diff] [blame]2631 the address of the entry block is illegal.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2632
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2633<p>This value only has defined behavior when used as an operand to the
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2634 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>' instruction, or for
2635 comparisons against null. Pointer equality tests between labels addresses
2636 results in undefined behavior &mdash; though, again, comparison against null
2637 is ok, and no label is equal to the null pointer. This may be passed around
2638 as an opaque pointer sized value as long as the bits are not inspected. This
2639 allows <tt>ptrtoint</tt> and arithmetic to be performed on these values so
2640 long as the original value is reconstituted before the <tt>indirectbr</tt>
2641 instruction.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2642
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2643<p>Finally, some targets may provide defined semantics when using the value as
2644 the operand to an inline assembly, but that is target specific.</p>
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2645
2646</div>
2647
2648
2649<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2650<h3>
2651 <a name="constantexprs">Constant Expressions</a>
2652</h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2653
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2654<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2655
2656<p>Constant expressions are used to allow expressions involving other constants
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2657 to be used as constants. Constant expressions may be of
2658 any <a href="#t_firstclass">first class</a> type and may involve any LLVM
2659 operation that does not have side effects (e.g. load and call are not
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2660 supported). The following is the syntax for constant expressions:</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2661
2662<dl>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2663 <dt><b><tt>trunc (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2664 <dd>Truncate a constant to another type. The bit size of CST must be larger
2665 than the bit size of TYPE. Both types must be integers.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2666
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2667 <dt><b><tt>zext (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2668 <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]2669 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2670
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2671 <dt><b><tt>sext (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2672 <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]2673 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2674
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2675 <dt><b><tt>fptrunc (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2676 <dd>Truncate a floating point constant to another floating point type. The
2677 size of CST must be larger than the size of TYPE. Both types must be
2678 floating point.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2679
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2680 <dt><b><tt>fpext (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2681 <dd>Floating point extend a constant to another type. The size of CST must be
2682 smaller or equal to the size of TYPE. Both types must be floating
2683 point.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2684
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2685 <dt><b><tt>fptoui (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2686 <dd>Convert a floating point constant to the corresponding unsigned integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2687 constant. TYPE must be a scalar or vector integer type. CST must be of
2688 scalar or vector floating point type. Both CST and TYPE must be scalars,
2689 or vectors of the same number of elements. If the value won't fit in the
2690 integer type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2691
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2692 <dt><b><tt>fptosi (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2693 <dd>Convert a floating point constant to the corresponding signed integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2694 constant. TYPE must be a scalar or vector integer type. CST must be of
2695 scalar or vector floating point type. Both CST and TYPE must be scalars,
2696 or vectors of the same number of elements. If the value won't fit in the
2697 integer type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2698
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2699 <dt><b><tt>uitofp (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2700 <dd>Convert an unsigned integer constant to the corresponding floating point
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2701 constant. TYPE must be a scalar or vector floating point type. CST must be
2702 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2703 vectors of the same number of elements. If the value won't fit in the
2704 floating point type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2705
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2706 <dt><b><tt>sitofp (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2707 <dd>Convert a signed integer constant to the corresponding floating point
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2708 constant. TYPE must be a scalar or vector floating point type. CST must be
2709 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2710 vectors of the same number of elements. If the value won't fit in the
2711 floating point type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2712
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2713 <dt><b><tt>ptrtoint (CST to TYPE)</tt></b></dt>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]2714 <dd>Convert a pointer typed constant to the corresponding integer constant
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2715 <tt>TYPE</tt> must be an integer type. <tt>CST</tt> must be of pointer
2716 type. The <tt>CST</tt> value is zero extended, truncated, or unchanged to
2717 make it fit in <tt>TYPE</tt>.</dd>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]2718
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2719 <dt><b><tt>inttoptr (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2720 <dd>Convert a integer constant to a pointer constant. TYPE must be a pointer
2721 type. CST must be of integer type. The CST value is zero extended,
2722 truncated, or unchanged to make it fit in a pointer size. This one is
2723 <i>really</i> dangerous!</dd>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]2724
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2725 <dt><b><tt>bitcast (CST to TYPE)</tt></b></dt>
Chris Lattner789dee32009-02-28 18:27:03 +0000[diff] [blame]2726 <dd>Convert a constant, CST, to another TYPE. The constraints of the operands
2727 are the same as those for the <a href="#i_bitcast">bitcast
2728 instruction</a>.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2729
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2730 <dt><b><tt>getelementptr (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
2731 <dt><b><tt>getelementptr inbounds (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2732 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2733 constants. As with the <a href="#i_getelementptr">getelementptr</a>
2734 instruction, the index list may have zero or more indexes, which are
2735 required to make sense for the type of "CSTPTR".</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2736
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2737 <dt><b><tt>select (COND, VAL1, VAL2)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2738 <dd>Perform the <a href="#i_select">select operation</a> on constants.</dd>
Reid Spencer9965ee72006-12-04 19:23:19 +0000[diff] [blame]2739
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2740 <dt><b><tt>icmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer9965ee72006-12-04 19:23:19 +0000[diff] [blame]2741 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
2742
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2743 <dt><b><tt>fcmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer9965ee72006-12-04 19:23:19 +0000[diff] [blame]2744 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
Robert Bocchino7e97a6d2006-01-10 19:31:34 +0000[diff] [blame]2745
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2746 <dt><b><tt>extractelement (VAL, IDX)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2747 <dd>Perform the <a href="#i_extractelement">extractelement operation</a> on
2748 constants.</dd>
Robert Bocchino7e97a6d2006-01-10 19:31:34 +0000[diff] [blame]2749
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2750 <dt><b><tt>insertelement (VAL, ELT, IDX)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2751 <dd>Perform the <a href="#i_insertelement">insertelement operation</a> on
2752 constants.</dd>
Chris Lattner016a0e52006-04-08 00:13:41 +0000[diff] [blame]2753
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2754 <dt><b><tt>shufflevector (VEC1, VEC2, IDXMASK)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2755 <dd>Perform the <a href="#i_shufflevector">shufflevector operation</a> on
2756 constants.</dd>
Chris Lattner016a0e52006-04-08 00:13:41 +0000[diff] [blame]2757
Nick Lewycky9ab9a7f2010-05-29 06:44:15 +0000[diff] [blame]2758 <dt><b><tt>extractvalue (VAL, IDX0, IDX1, ...)</tt></b></dt>
2759 <dd>Perform the <a href="#i_extractvalue">extractvalue operation</a> on
2760 constants. The index list is interpreted in a similar manner as indices in
2761 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2762 index value must be specified.</dd>
2763
2764 <dt><b><tt>insertvalue (VAL, ELT, IDX0, IDX1, ...)</tt></b></dt>
2765 <dd>Perform the <a href="#i_insertvalue">insertvalue operation</a> on
2766 constants. The index list is interpreted in a similar manner as indices in
2767 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2768 index value must be specified.</dd>
2769
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2770 <dt><b><tt>OPCODE (LHS, RHS)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2771 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
2772 be any of the <a href="#binaryops">binary</a>
2773 or <a href="#bitwiseops">bitwise binary</a> operations. The constraints
2774 on operands are the same as those for the corresponding instruction
2775 (e.g. no bitwise operations on floating point values are allowed).</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2776</dl>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2777
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2778</div>
Chris Lattnerb1652612004-03-08 16:49:10 +0000[diff] [blame]2779
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2780</div>
2781
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2782<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2783<h2><a name="othervalues">Other Values</a></h2>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2784<!-- *********************************************************************** -->
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2785<div>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2786<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2787<h3>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2788<a name="inlineasm">Inline Assembler Expressions</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2789</h3>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2790
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2791<div>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2792
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2793<p>LLVM supports inline assembler expressions (as opposed
2794 to <a href="#moduleasm"> Module-Level Inline Assembly</a>) through the use of
2795 a special value. This value represents the inline assembler as a string
2796 (containing the instructions to emit), a list of operand constraints (stored
Dale Johannesen63c94fe2009-10-13 21:56:55 +0000[diff] [blame]2797 as a string), a flag that indicates whether or not the inline asm
Dale Johannesen1cfb9582009-10-21 23:28:00 +0000[diff] [blame]2798 expression has side effects, and a flag indicating whether the function
2799 containing the asm needs to align its stack conservatively. An example
2800 inline assembler expression is:</p>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2801
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2802<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2803i32 (i32) asm "bswap $0", "=r,r"
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2804</pre>
2805
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2806<p>Inline assembler expressions may <b>only</b> be used as the callee operand of
2807 a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we
2808 have:</p>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2809
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2810<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2811%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]2812</pre>
2813
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2814<p>Inline asms with side effects not visible in the constraint list must be
2815 marked as having side effects. This is done through the use of the
2816 '<tt>sideeffect</tt>' keyword, like so:</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]2819call void asm sideeffect "eieio", ""()
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2820</pre>
2821
Dale Johannesen1cfb9582009-10-21 23:28:00 +0000[diff] [blame]2822<p>In some cases inline asms will contain code that will not work unless the
2823 stack is aligned in some way, such as calls or SSE instructions on x86,
2824 yet will not contain code that does that alignment within the asm.
2825 The compiler should make conservative assumptions about what the asm might
2826 contain and should generate its usual stack alignment code in the prologue
2827 if the '<tt>alignstack</tt>' keyword is present:</p>
Dale Johannesen63c94fe2009-10-13 21:56:55 +0000[diff] [blame]2828
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2829<pre class="doc_code">
Dale Johannesen1cfb9582009-10-21 23:28:00 +0000[diff] [blame]2830call void asm alignstack "eieio", ""()
Dale Johannesen63c94fe2009-10-13 21:56:55 +0000[diff] [blame]2831</pre>
Dale Johannesen63c94fe2009-10-13 21:56:55 +0000[diff] [blame]2832
2833<p>If both keywords appear the '<tt>sideeffect</tt>' keyword must come
2834 first.</p>
2835
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2836<p>TODO: The format of the asm and constraints string still need to be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2837 documented here. Constraints on what can be done (e.g. duplication, moving,
2838 etc need to be documented). This is probably best done by reference to
2839 another document that covers inline asm from a holistic perspective.</p>
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2840
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2841<h4>
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2842<a name="inlineasm_md">Inline Asm Metadata</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2843</h4>
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2844
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2845<div>
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2846
2847<p>The call instructions that wrap inline asm nodes may have a "!srcloc" MDNode
Chris Lattner79ffdc72010-11-17 08:20:42 +0000[diff] [blame]2848 attached to it that contains a list of constant integers. If present, the
2849 code generator will use the integer as the location cookie value when report
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2850 errors through the LLVMContext error reporting mechanisms. This allows a
Dan Gohman61110ae2010-04-28 00:36:01 +0000[diff] [blame]2851 front-end to correlate backend errors that occur with inline asm back to the
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2852 source code that produced it. For example:</p>
2853
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2854<pre class="doc_code">
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2855call void asm sideeffect "something bad", ""()<b>, !srcloc !42</b>
2856...
2857!42 = !{ i32 1234567 }
2858</pre>
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2859
2860<p>It is up to the front-end to make sense of the magic numbers it places in the
Chris Lattner79ffdc72010-11-17 08:20:42 +0000[diff] [blame]2861 IR. If the MDNode contains multiple constants, the code generator will use
2862 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]2863
2864</div>
2865
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2866</div>
2867
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]2868<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2869<h3>
2870 <a name="metadata">Metadata Nodes and Metadata Strings</a>
2871</h3>
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]2872
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2873<div>
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]2874
2875<p>LLVM IR allows metadata to be attached to instructions in the program that
2876 can convey extra information about the code to the optimizers and code
2877 generator. One example application of metadata is source-level debug
2878 information. There are two metadata primitives: strings and nodes. All
2879 metadata has the <tt>metadata</tt> type and is identified in syntax by a
2880 preceding exclamation point ('<tt>!</tt>').</p>
2881
2882<p>A metadata string is a string surrounded by double quotes. It can contain
2883 any character by escaping non-printable characters with "\xx" where "xx" is
2884 the two digit hex code. For example: "<tt>!"test\00"</tt>".</p>
2885
2886<p>Metadata nodes are represented with notation similar to structure constants
2887 (a comma separated list of elements, surrounded by braces and preceded by an
2888 exclamation point). For example: "<tt>!{ metadata !"test\00", i32
2889 10}</tt>". Metadata nodes can have any values as their operand.</p>
2890
2891<p>A <a href="#namedmetadatastructure">named metadata</a> is a collection of
2892 metadata nodes, which can be looked up in the module symbol table. For
2893 example: "<tt>!foo = metadata !{!4, !3}</tt>".
2894
Devang Patel9984bd62010-03-04 23:44:48 +0000[diff] [blame]2895<p>Metadata can be used as function arguments. Here <tt>llvm.dbg.value</tt>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2896 function is using two metadata arguments.</p>
Devang Patel9984bd62010-03-04 23:44:48 +0000[diff] [blame]2897
Bill Wendlingc0e10672011-03-02 02:17:11 +0000[diff] [blame]2898<div class="doc_code">
2899<pre>
2900call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2901</pre>
2902</div>
Devang Patel9984bd62010-03-04 23:44:48 +0000[diff] [blame]2903
2904<p>Metadata can be attached with an instruction. Here metadata <tt>!21</tt> is
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2905 attached with <tt>add</tt> instruction using <tt>!dbg</tt> identifier.</p>
Devang Patel9984bd62010-03-04 23:44:48 +0000[diff] [blame]2906
Bill Wendlingc0e10672011-03-02 02:17:11 +0000[diff] [blame]2907<div class="doc_code">
2908<pre>
2909%indvar.next = add i64 %indvar, 1, !dbg !21
2910</pre>
2911</div>
2912
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]2913</div>
2914
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2915</div>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2916
2917<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2918<h2>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2919 <a name="intrinsic_globals">Intrinsic Global Variables</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2920</h2>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2921<!-- *********************************************************************** -->
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2922<div>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2923<p>LLVM has a number of "magic" global variables that contain data that affect
2924code generation or other IR semantics. These are documented here. All globals
Chris Lattner58f9bb22009-07-20 06:14:25 +0000[diff] [blame]2925of this sort should have a section specified as "<tt>llvm.metadata</tt>". This
2926section and all globals that start with "<tt>llvm.</tt>" are reserved for use
2927by LLVM.</p>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2928
2929<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2930<h3>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2931<a name="intg_used">The '<tt>llvm.used</tt>' Global Variable</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2932</h3>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2933
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2934<div>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2935
2936<p>The <tt>@llvm.used</tt> global is an array with i8* element type which has <a
2937href="#linkage_appending">appending linkage</a>. This array contains a list of
2938pointers to global variables and functions which may optionally have a pointer
2939cast formed of bitcast or getelementptr. For example, a legal use of it is:</p>
2940
2941<pre>
2942 @X = global i8 4
2943 @Y = global i32 123
2944
2945 @llvm.used = appending global [2 x i8*] [
2946 i8* @X,
2947 i8* bitcast (i32* @Y to i8*)
2948 ], section "llvm.metadata"
2949</pre>
2950
2951<p>If a global variable appears in the <tt>@llvm.used</tt> list, then the
2952compiler, assembler, and linker are required to treat the symbol as if there is
2953a reference to the global that it cannot see. For example, if a variable has
2954internal linkage and no references other than that from the <tt>@llvm.used</tt>
2955list, it cannot be deleted. This is commonly used to represent references from
2956inline asms and other things the compiler cannot "see", and corresponds to
2957"attribute((used))" in GNU C.</p>
2958
2959<p>On some targets, the code generator must emit a directive to the assembler or
2960object file to prevent the assembler and linker from molesting the symbol.</p>
2961
2962</div>
2963
2964<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2965<h3>
2966 <a name="intg_compiler_used">
2967 The '<tt>llvm.compiler.used</tt>' Global Variable
2968 </a>
2969</h3>
Chris Lattner58f9bb22009-07-20 06:14:25 +0000[diff] [blame]2970
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2971<div>
Chris Lattner58f9bb22009-07-20 06:14:25 +0000[diff] [blame]2972
2973<p>The <tt>@llvm.compiler.used</tt> directive is the same as the
2974<tt>@llvm.used</tt> directive, except that it only prevents the compiler from
2975touching the symbol. On targets that support it, this allows an intelligent
2976linker to optimize references to the symbol without being impeded as it would be
2977by <tt>@llvm.used</tt>.</p>
2978
2979<p>This is a rare construct that should only be used in rare circumstances, and
2980should not be exposed to source languages.</p>
2981
2982</div>
2983
2984<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2985<h3>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2986<a name="intg_global_ctors">The '<tt>llvm.global_ctors</tt>' Global Variable</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2987</h3>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2988
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2989<div>
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]2990<pre>
2991%0 = type { i32, void ()* }
David Chisnallb492b812010-04-30 19:27:35 +0000[diff] [blame]2992@llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor }]
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]2993</pre>
2994<p>The <tt>@llvm.global_ctors</tt> array contains a list of constructor functions and associated priorities. The functions referenced by this array will be called in ascending order of priority (i.e. lowest first) when the module is loaded. The order of functions with the same priority is not defined.
2995</p>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2996
2997</div>
2998
2999<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3000<h3>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3001<a name="intg_global_dtors">The '<tt>llvm.global_dtors</tt>' Global Variable</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3002</h3>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3003
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3004<div>
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]3005<pre>
3006%0 = type { i32, void ()* }
David Chisnallb492b812010-04-30 19:27:35 +0000[diff] [blame]3007@llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor }]
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]3008</pre>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3009
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]3010<p>The <tt>@llvm.global_dtors</tt> array contains a list of destructor functions and associated priorities. The functions referenced by this array will be called in descending order of priority (i.e. highest first) when the module is loaded. The order of functions with the same priority is not defined.
3011</p>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3012
3013</div>
3014
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3015</div>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]3016
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]3017<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3018<h2><a name="instref">Instruction Reference</a></h2>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3019<!-- *********************************************************************** -->
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]3020
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3021<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]3022
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3023<p>The LLVM instruction set consists of several different classifications of
3024 instructions: <a href="#terminators">terminator
3025 instructions</a>, <a href="#binaryops">binary instructions</a>,
3026 <a href="#bitwiseops">bitwise binary instructions</a>,
3027 <a href="#memoryops">memory instructions</a>, and
3028 <a href="#otherops">other instructions</a>.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]3029
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3030<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3031<h3>
3032 <a name="terminators">Terminator Instructions</a>
3033</h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]3034
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3035<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]3036
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3037<p>As mentioned <a href="#functionstructure">previously</a>, every basic block
3038 in a program ends with a "Terminator" instruction, which indicates which
3039 block should be executed after the current block is finished. These
3040 terminator instructions typically yield a '<tt>void</tt>' value: they produce
3041 control flow, not values (the one exception being the
3042 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
3043
Chris Lattnerd3d65ab2011-08-02 20:29:13 +0000[diff] [blame]3044<p>The terminator instructions are:
3045 '<a href="#i_ret"><tt>ret</tt></a>',
3046 '<a href="#i_br"><tt>br</tt></a>',
3047 '<a href="#i_switch"><tt>switch</tt></a>',
3048 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>',
3049 '<a href="#i_invoke"><tt>invoke</tt></a>',
3050 '<a href="#i_unwind"><tt>unwind</tt></a>',
3051 '<a href="#i_resume"><tt>resume</tt></a>', and
3052 '<a href="#i_unreachable"><tt>unreachable</tt></a>'.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]3053
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3054<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3055<h4>
3056 <a name="i_ret">'<tt>ret</tt>' Instruction</a>
3057</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3058
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3059<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3060
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3061<h5>Syntax:</h5>
Dan Gohmancc3132e2008-10-04 19:00:07 +0000[diff] [blame]3062<pre>
3063 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]3064 ret void <i>; Return from void function</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3065</pre>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]3066
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3067<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3068<p>The '<tt>ret</tt>' instruction is used to return control flow (and optionally
3069 a value) from a function back to the caller.</p>
3070
3071<p>There are two forms of the '<tt>ret</tt>' instruction: one that returns a
3072 value and then causes control flow, and one that just causes control flow to
3073 occur.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]3074
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3075<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3076<p>The '<tt>ret</tt>' instruction optionally accepts a single argument, the
3077 return value. The type of the return value must be a
3078 '<a href="#t_firstclass">first class</a>' type.</p>
Dan Gohmancc3132e2008-10-04 19:00:07 +0000[diff] [blame]3079
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3080<p>A function is not <a href="#wellformed">well formed</a> if it it has a
3081 non-void return type and contains a '<tt>ret</tt>' instruction with no return
3082 value or a return value with a type that does not match its type, or if it
3083 has a void return type and contains a '<tt>ret</tt>' instruction with a
3084 return value.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]3085
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3086<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3087<p>When the '<tt>ret</tt>' instruction is executed, control flow returns back to
3088 the calling function's context. If the caller is a
3089 "<a href="#i_call"><tt>call</tt></a>" instruction, execution continues at the
3090 instruction after the call. If the caller was an
3091 "<a href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues at
3092 the beginning of the "normal" destination block. If the instruction returns
3093 a value, that value shall set the call or invoke instruction's return
3094 value.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]3095
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3096<h5>Example:</h5>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]3097<pre>
3098 ret i32 5 <i>; Return an integer value of 5</i>
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]3099 ret void <i>; Return from a void function</i>
Bill Wendling050ee8f2009-02-28 22:12:54 +0000[diff] [blame]3100 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]3101</pre>
Dan Gohman3065b612009-01-12 23:12:39 +0000[diff] [blame]3102
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3103</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3104<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3105<h4>
3106 <a name="i_br">'<tt>br</tt>' Instruction</a>
3107</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3108
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3109<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3110
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3111<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3112<pre>
Bill Wendling16b86742011-07-26 10:41:15 +0000[diff] [blame]3113 br i1 &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;
3114 br label &lt;dest&gt; <i>; Unconditional branch</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3115</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3116
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3117<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3118<p>The '<tt>br</tt>' instruction is used to cause control flow to transfer to a
3119 different basic block in the current function. There are two forms of this
3120 instruction, corresponding to a conditional branch and an unconditional
3121 branch.</p>
3122
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3123<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3124<p>The conditional branch form of the '<tt>br</tt>' instruction takes a single
3125 '<tt>i1</tt>' value and two '<tt>label</tt>' values. The unconditional form
3126 of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>' value as a
3127 target.</p>
3128
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3129<h5>Semantics:</h5>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]3130<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3131 argument is evaluated. If the value is <tt>true</tt>, control flows to the
3132 '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
3133 control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
3134
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3135<h5>Example:</h5>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]3136<pre>
3137Test:
3138 %cond = <a href="#i_icmp">icmp</a> eq i32 %a, %b
3139 br i1 %cond, label %IfEqual, label %IfUnequal
3140IfEqual:
3141 <a href="#i_ret">ret</a> i32 1
3142IfUnequal:
3143 <a href="#i_ret">ret</a> i32 0
3144</pre>
3145
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3146</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3147
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3148<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3149<h4>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3150 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3151</h4>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3152
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3153<div>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3154
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3155<h5>Syntax:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3156<pre>
3157 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
3158</pre>
3159
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3160<h5>Overview:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3161<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]3162 several different places. It is a generalization of the '<tt>br</tt>'
3163 instruction, allowing a branch to occur to one of many possible
3164 destinations.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3165
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3166<h5>Arguments:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3167<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3168 comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination,
3169 and an array of pairs of comparison value constants and '<tt>label</tt>'s.
3170 The table is not allowed to contain duplicate constant entries.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3171
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3172<h5>Semantics:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3173<p>The <tt>switch</tt> instruction specifies a table of values and
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3174 destinations. When the '<tt>switch</tt>' instruction is executed, this table
3175 is searched for the given value. If the value is found, control flow is
Benjamin Kramer0f420382009-10-12 14:46:08 +0000[diff] [blame]3176 transferred to the corresponding destination; otherwise, control flow is
3177 transferred to the default destination.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3178
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3179<h5>Implementation:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3180<p>Depending on properties of the target machine and the particular
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3181 <tt>switch</tt> instruction, this instruction may be code generated in
3182 different ways. For example, it could be generated as a series of chained
3183 conditional branches or with a lookup table.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3184
3185<h5>Example:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3186<pre>
3187 <i>; Emulate a conditional br instruction</i>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]3188 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Dan Gohman623806e2009-01-04 23:44:43 +0000[diff] [blame]3189 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3190
3191 <i>; Emulate an unconditional br instruction</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]3192 switch i32 0, label %dest [ ]
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3193
3194 <i>; Implement a jump table:</i>
Dan Gohman623806e2009-01-04 23:44:43 +0000[diff] [blame]3195 switch i32 %val, label %otherwise [ i32 0, label %onzero
3196 i32 1, label %onone
3197 i32 2, label %ontwo ]
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3198</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3199
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3200</div>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3201
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3202
3203<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3204<h4>
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]3205 <a name="i_indirectbr">'<tt>indirectbr</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3206</h4>
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3207
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3208<div>
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3209
3210<h5>Syntax:</h5>
3211<pre>
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]3212 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]3213</pre>
3214
3215<h5>Overview:</h5>
3216
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]3217<p>The '<tt>indirectbr</tt>' instruction implements an indirect branch to a label
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3218 within the current function, whose address is specified by
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]3219 "<tt>address</tt>". Address must be derived from a <a
3220 href="#blockaddress">blockaddress</a> constant.</p>
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3221
3222<h5>Arguments:</h5>
3223
3224<p>The '<tt>address</tt>' argument is the address of the label to jump to. The
3225 rest of the arguments indicate the full set of possible destinations that the
3226 address may point to. Blocks are allowed to occur multiple times in the
3227 destination list, though this isn't particularly useful.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3228
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3229<p>This destination list is required so that dataflow analysis has an accurate
3230 understanding of the CFG.</p>
3231
3232<h5>Semantics:</h5>
3233
3234<p>Control transfers to the block specified in the address argument. All
3235 possible destination blocks must be listed in the label list, otherwise this
3236 instruction has undefined behavior. This implies that jumps to labels
3237 defined in other functions have undefined behavior as well.</p>
3238
3239<h5>Implementation:</h5>
3240
3241<p>This is typically implemented with a jump through a register.</p>
3242
3243<h5>Example:</h5>
3244<pre>
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]3245 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3246</pre>
3247
3248</div>
3249
3250
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3251<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3252<h4>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3253 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3254</h4>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3255
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3256<div>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3257
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3258<h5>Syntax:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3259<pre>
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]3260 &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]3261 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3262</pre>
3263
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]3264<h5>Overview:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3265<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3266 function, with the possibility of control flow transfer to either the
3267 '<tt>normal</tt>' label or the '<tt>exception</tt>' label. If the callee
3268 function returns with the "<tt><a href="#i_ret">ret</a></tt>" instruction,
3269 control flow will return to the "normal" label. If the callee (or any
3270 indirect callees) returns with the "<a href="#i_unwind"><tt>unwind</tt></a>"
3271 instruction, control is interrupted and continued at the dynamically nearest
3272 "exception" label.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3273
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]3274<p>The '<tt>exception</tt>' label is a
3275 <i><a href="ExceptionHandling.html#overview">landing pad</a></i> for the
3276 exception. As such, '<tt>exception</tt>' label is required to have the
3277 "<a href="#i_landingpad"><tt>landingpad</tt></a>" instruction, which contains
3278 the information about about the behavior of the program after unwinding
3279 happens, as its first non-PHI instruction. The restrictions on the
3280 "<tt>landingpad</tt>" instruction's tightly couples it to the
3281 "<tt>invoke</tt>" instruction, so that the important information contained
3282 within the "<tt>landingpad</tt>" instruction can't be lost through normal
3283 code motion.</p>
3284
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3285<h5>Arguments:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3286<p>This instruction requires several arguments:</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3287
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3288<ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3289 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
3290 convention</a> the call should use. If none is specified, the call
3291 defaults to using C calling conventions.</li>
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]3292
3293 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3294 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
3295 '<tt>inreg</tt>' attributes are valid here.</li>
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]3296
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3297 <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]3298 function value being invoked. In most cases, this is a direct function
3299 invocation, but indirect <tt>invoke</tt>s are just as possible, branching
3300 off an arbitrary pointer to function value.</li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3301
3302 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3303 function to be invoked. </li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3304
3305 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]3306 signature argument types and parameter attributes. All arguments must be
3307 of <a href="#t_firstclass">first class</a> type. If the function
3308 signature indicates the function accepts a variable number of arguments,
3309 the extra arguments can be specified.</li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3310
3311 <li>'<tt>normal label</tt>': the label reached when the called function
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3312 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3313
3314 <li>'<tt>exception label</tt>': the label reached when a callee returns with
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3315 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3316
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]3317 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3318 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
3319 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3320</ol>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3321
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3322<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3323<p>This instruction is designed to operate as a standard
3324 '<tt><a href="#i_call">call</a></tt>' instruction in most regards. The
3325 primary difference is that it establishes an association with a label, which
3326 is used by the runtime library to unwind the stack.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3327
3328<p>This instruction is used in languages with destructors to ensure that proper
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3329 cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
3330 exception. Additionally, this is important for implementation of
3331 '<tt>catch</tt>' clauses in high-level languages that support them.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3332
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3333<p>For the purposes of the SSA form, the definition of the value returned by the
3334 '<tt>invoke</tt>' instruction is deemed to occur on the edge from the current
3335 block to the "normal" label. If the callee unwinds then no return value is
3336 available.</p>
Dan Gohman9069d892009-05-22 21:47:08 +0000[diff] [blame]3337
Chris Lattner97257f82010-01-15 18:08:37 +0000[diff] [blame]3338<p>Note that the code generator does not yet completely support unwind, and
3339that the invoke/unwind semantics are likely to change in future versions.</p>
3340
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3341<h5>Example:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3342<pre>
Nick Lewycky084ab472008-03-16 07:18:12 +0000[diff] [blame]3343 %retval = invoke i32 @Test(i32 15) to label %Continue
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]3344 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewycky084ab472008-03-16 07:18:12 +0000[diff] [blame]3345 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]3346 unwind label %TestCleanup <i>; {i32}:retval set</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3347</pre>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3348
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3349</div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3350
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]3351<!-- _______________________________________________________________________ -->
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3352
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3353<h4>
3354 <a name="i_unwind">'<tt>unwind</tt>' Instruction</a>
3355</h4>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3356
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3357<div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3358
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]3359<h5>Syntax:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3360<pre>
3361 unwind
3362</pre>
3363
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]3364<h5>Overview:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3365<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3366 at the first callee in the dynamic call stack which used
3367 an <a href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call.
3368 This is primarily used to implement exception handling.</p>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3369
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]3370<h5>Semantics:</h5>
Chris Lattnerfe8519c2008-04-19 21:01:16 +0000[diff] [blame]3371<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3372 immediately halt. The dynamic call stack is then searched for the
3373 first <a href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack.
3374 Once found, execution continues at the "exceptional" destination block
3375 specified by the <tt>invoke</tt> instruction. If there is no <tt>invoke</tt>
3376 instruction in the dynamic call chain, undefined behavior results.</p>
3377
Chris Lattner97257f82010-01-15 18:08:37 +0000[diff] [blame]3378<p>Note that the code generator does not yet completely support unwind, and
3379that the invoke/unwind semantics are likely to change in future versions.</p>
3380
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3381</div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3382
Bill Wendlingf891bf82011-07-31 06:30:59 +0000[diff] [blame]3383 <!-- _______________________________________________________________________ -->
3384
3385<h4>
3386 <a name="i_resume">'<tt>resume</tt>' Instruction</a>
3387</h4>
3388
3389<div>
3390
3391<h5>Syntax:</h5>
3392<pre>
3393 resume &lt;type&gt; &lt;value&gt;
3394</pre>
3395
3396<h5>Overview:</h5>
3397<p>The '<tt>resume</tt>' instruction is a terminator instruction that has no
3398 successors.</p>
3399
3400<h5>Arguments:</h5>
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]3401<p>The '<tt>resume</tt>' instruction requires one argument, which must have the
Bill Wendlingc5a13612011-08-03 18:37:32 +0000[diff] [blame]3402 same type as the result of any '<tt>landingpad</tt>' instruction in the same
3403 function.</p>
Bill Wendlingf891bf82011-07-31 06:30:59 +0000[diff] [blame]3404
3405<h5>Semantics:</h5>
3406<p>The '<tt>resume</tt>' instruction resumes propagation of an existing
3407 (in-flight) exception whose unwinding was interrupted with
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]3408 a <a href="#i_landingpad"><tt>landingpad</tt></a> instruction.</p>
Bill Wendlingf891bf82011-07-31 06:30:59 +0000[diff] [blame]3409
3410<h5>Example:</h5>
3411<pre>
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]3412 resume { i8*, i32 } %exn
Bill Wendlingf891bf82011-07-31 06:30:59 +0000[diff] [blame]3413</pre>
3414
3415</div>
3416
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3417<!-- _______________________________________________________________________ -->
3418
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3419<h4>
3420 <a name="i_unreachable">'<tt>unreachable</tt>' Instruction</a>
3421</h4>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3422
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3423<div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3424
3425<h5>Syntax:</h5>
3426<pre>
3427 unreachable
3428</pre>
3429
3430<h5>Overview:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3431<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3432 instruction is used to inform the optimizer that a particular portion of the
3433 code is not reachable. This can be used to indicate that the code after a
3434 no-return function cannot be reached, and other facts.</p>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3435
3436<h5>Semantics:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3437<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3438
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3439</div>
3440
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3441</div>
3442
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3443<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3444<h3>
3445 <a name="binaryops">Binary Operations</a>
3446</h3>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3447
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3448<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3449
3450<p>Binary operators are used to do most of the computation in a program. They
3451 require two operands of the same type, execute an operation on them, and
3452 produce a single value. The operands might represent multiple data, as is
3453 the case with the <a href="#t_vector">vector</a> data type. The result value
3454 has the same type as its operands.</p>
3455
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3456<p>There are several different binary operators:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3457
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3458<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3459<h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3460 <a name="i_add">'<tt>add</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3461</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3462
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3463<div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3464
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3465<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3466<pre>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3467 &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]3468 &lt;result&gt; = add nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3469 &lt;result&gt; = add nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3470 &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]3471</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3472
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3473<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3474<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3475
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3476<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3477<p>The two arguments to the '<tt>add</tt>' instruction must
3478 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3479 integer values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3480
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3481<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3482<p>The value produced is the integer sum of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3483
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3484<p>If the sum has unsigned overflow, the result returned is the mathematical
3485 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]3486
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3487<p>Because LLVM integers use a two's complement representation, this instruction
3488 is appropriate for both signed and unsigned integers.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3489
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3490<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3491 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3492 <tt>nsw</tt> keywords are present, the result value of the <tt>add</tt>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]3493 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3494 respectively, occurs.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3495
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3496<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3497<pre>
3498 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3499</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3500
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3501</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3502
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3503<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3504<h4>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3505 <a name="i_fadd">'<tt>fadd</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3506</h4>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3507
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3508<div>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3509
3510<h5>Syntax:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3511<pre>
3512 &lt;result&gt; = fadd &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3513</pre>
3514
3515<h5>Overview:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3516<p>The '<tt>fadd</tt>' instruction returns the sum of its two operands.</p>
3517
3518<h5>Arguments:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3519<p>The two arguments to the '<tt>fadd</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3520 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3521 floating point values. Both arguments must have identical types.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3522
3523<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3524<p>The value produced is the floating point sum of the two operands.</p>
3525
3526<h5>Example:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3527<pre>
3528 &lt;result&gt; = fadd float 4.0, %var <i>; yields {float}:result = 4.0 + %var</i>
3529</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3530
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3531</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3532
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3533<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3534<h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3535 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3536</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3537
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3538<div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3539
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3540<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3541<pre>
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3542 &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]3543 &lt;result&gt; = sub nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3544 &lt;result&gt; = sub nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3545 &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]3546</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3547
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3548<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3549<p>The '<tt>sub</tt>' instruction returns the difference of its two
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3550 operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3551
3552<p>Note that the '<tt>sub</tt>' instruction is used to represent the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3553 '<tt>neg</tt>' instruction present in most other intermediate
3554 representations.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3555
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3556<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3557<p>The two arguments to the '<tt>sub</tt>' instruction must
3558 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3559 integer values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3560
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3561<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3562<p>The value produced is the integer difference of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3563
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3564<p>If the difference has unsigned overflow, the result returned is the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3565 mathematical result modulo 2<sup>n</sup>, where n is the bit width of the
3566 result.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3567
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3568<p>Because LLVM integers use a two's complement representation, this instruction
3569 is appropriate for both signed and unsigned integers.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3570
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3571<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3572 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3573 <tt>nsw</tt> keywords are present, the result value of the <tt>sub</tt>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]3574 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3575 respectively, occurs.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3576
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3577<h5>Example:</h5>
Bill Wendling2d8b9a82007-05-29 09:42:13 +0000[diff] [blame]3578<pre>
3579 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]3580 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3581</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3582
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3583</div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3584
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3585<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3586<h4>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3587 <a name="i_fsub">'<tt>fsub</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3588</h4>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3589
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3590<div>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3591
3592<h5>Syntax:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3593<pre>
3594 &lt;result&gt; = fsub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3595</pre>
3596
3597<h5>Overview:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3598<p>The '<tt>fsub</tt>' instruction returns the difference of its two
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3599 operands.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3600
3601<p>Note that the '<tt>fsub</tt>' instruction is used to represent the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3602 '<tt>fneg</tt>' instruction present in most other intermediate
3603 representations.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3604
3605<h5>Arguments:</h5>
Bill Wendling972b7202009-07-20 02:32:41 +0000[diff] [blame]3606<p>The two arguments to the '<tt>fsub</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3607 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3608 floating point values. Both arguments must have identical types.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3609
3610<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3611<p>The value produced is the floating point difference of the two operands.</p>
3612
3613<h5>Example:</h5>
3614<pre>
3615 &lt;result&gt; = fsub float 4.0, %var <i>; yields {float}:result = 4.0 - %var</i>
3616 &lt;result&gt; = fsub float -0.0, %val <i>; yields {float}:result = -%var</i>
3617</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3618
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3619</div>
3620
3621<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3622<h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3623 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3624</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3625
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3626<div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3627
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3628<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3629<pre>
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3630 &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]3631 &lt;result&gt; = mul nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3632 &lt;result&gt; = mul nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3633 &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]3634</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3635
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3636<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3637<p>The '<tt>mul</tt>' instruction returns the product of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3638
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3639<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3640<p>The two arguments to the '<tt>mul</tt>' instruction must
3641 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3642 integer values. Both arguments must have identical types.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3643
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3644<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3645<p>The value produced is the integer product of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3646
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3647<p>If the result of the multiplication has unsigned overflow, the result
3648 returned is the mathematical result modulo 2<sup>n</sup>, where n is the bit
3649 width of the result.</p>
3650
3651<p>Because LLVM integers use a two's complement representation, and the result
3652 is the same width as the operands, this instruction returns the correct
3653 result for both signed and unsigned integers. If a full product
3654 (e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands should
3655 be sign-extended or zero-extended as appropriate to the width of the full
3656 product.</p>
3657
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3658<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3659 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3660 <tt>nsw</tt> keywords are present, the result value of the <tt>mul</tt>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]3661 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3662 respectively, occurs.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3663
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3664<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3665<pre>
3666 &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3667</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3668
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3669</div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3670
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3671<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3672<h4>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3673 <a name="i_fmul">'<tt>fmul</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3674</h4>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3675
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3676<div>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3677
3678<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3679<pre>
3680 &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]3681</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3682
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3683<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3684<p>The '<tt>fmul</tt>' instruction returns the product of its two operands.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3685
3686<h5>Arguments:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3687<p>The two arguments to the '<tt>fmul</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3688 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3689 floating point values. Both arguments must have identical types.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3690
3691<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3692<p>The value produced is the floating point product of the two operands.</p>
3693
3694<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3695<pre>
3696 &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]3697</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3698
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3699</div>
3700
3701<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3702<h4>
3703 <a name="i_udiv">'<tt>udiv</tt>' Instruction</a>
3704</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3705
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3706<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3707
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3708<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3709<pre>
Chris Lattner35315d02011-02-06 21:44:57 +0000[diff] [blame]3710 &lt;result&gt; = udiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3711 &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]3712</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3713
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3714<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3715<p>The '<tt>udiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3716
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3717<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3718<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3719 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3720 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3721
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3722<h5>Semantics:</h5>
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3723<p>The value produced is the unsigned integer quotient of the two operands.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3724
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3725<p>Note that unsigned integer division and signed integer division are distinct
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3726 operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
3727
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3728<p>Division by zero leads to undefined behavior.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3729
Chris Lattner35315d02011-02-06 21:44:57 +0000[diff] [blame]3730<p>If the <tt>exact</tt> keyword is present, the result value of the
3731 <tt>udiv</tt> is a <a href="#trapvalues">trap value</a> if %op1 is not a
3732 multiple of %op2 (as such, "((a udiv exact b) mul b) == a").</p>
3733
3734
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3735<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3736<pre>
3737 &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3738</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3739
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3740</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3741
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3742<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3743<h4>
3744 <a name="i_sdiv">'<tt>sdiv</tt>' Instruction</a>
3745</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3746
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3747<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3748
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3749<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3750<pre>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3751 &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]3752 &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]3753</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3754
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3755<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3756<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3757
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3758<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3759<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3760 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3761 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3762
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3763<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3764<p>The value produced is the signed integer quotient of the two operands rounded
3765 towards zero.</p>
3766
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3767<p>Note that signed integer division and unsigned integer division are distinct
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3768 operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
3769
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3770<p>Division by zero leads to undefined behavior. Overflow also leads to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3771 undefined behavior; this is a rare case, but can occur, for example, by doing
3772 a 32-bit division of -2147483648 by -1.</p>
3773
Dan Gohman71dfd782009-07-22 00:04:19 +0000[diff] [blame]3774<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohman57255802010-04-23 15:23:32 +0000[diff] [blame]3775 <tt>sdiv</tt> is a <a href="#trapvalues">trap value</a> if the result would
Dan Gohmane501ff72010-07-11 00:08:34 +0000[diff] [blame]3776 be rounded.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3777
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3778<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3779<pre>
3780 &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3781</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3782
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3783</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3784
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3785<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3786<h4>
3787 <a name="i_fdiv">'<tt>fdiv</tt>' Instruction</a>
3788</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3789
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3790<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3791
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3792<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3793<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +0000[diff] [blame]3794 &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]3795</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3796
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3797<h5>Overview:</h5>
3798<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3799
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3800<h5>Arguments:</h5>
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]3801<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3802 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3803 floating point values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3804
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3805<h5>Semantics:</h5>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3806<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3807
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3808<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3809<pre>
3810 &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]3811</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3812
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3813</div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3814
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3815<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3816<h4>
3817 <a name="i_urem">'<tt>urem</tt>' Instruction</a>
3818</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3819
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3820<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3821
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3822<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3823<pre>
3824 &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]3825</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3826
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3827<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3828<p>The '<tt>urem</tt>' instruction returns the remainder from the unsigned
3829 division of its two arguments.</p>
3830
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3831<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3832<p>The two arguments to the '<tt>urem</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3833 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3834 values. Both arguments must have identical types.</p>
3835
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3836<h5>Semantics:</h5>
3837<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3838 This instruction always performs an unsigned division to get the
3839 remainder.</p>
3840
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3841<p>Note that unsigned integer remainder and signed integer remainder are
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3842 distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
3843
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3844<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3845
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3846<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3847<pre>
3848 &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3849</pre>
3850
3851</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3852
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3853<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3854<h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3855 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3856</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3857
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3858<div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3859
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3860<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3861<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +0000[diff] [blame]3862 &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]3863</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3864
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3865<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3866<p>The '<tt>srem</tt>' instruction returns the remainder from the signed
3867 division of its two operands. This instruction can also take
3868 <a href="#t_vector">vector</a> versions of the values in which case the
3869 elements must be integers.</p>
Chris Lattnerb8f816e2008-01-04 04:33:49 +0000[diff] [blame]3870
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3871<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3872<p>The two arguments to the '<tt>srem</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3873 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3874 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3875
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3876<h5>Semantics:</h5>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3877<p>This instruction returns the <i>remainder</i> of a division (where the result
Duncan Sands2769c6e2011-03-07 09:12:24 +0000[diff] [blame]3878 is either zero or has the same sign as the dividend, <tt>op1</tt>), not the
3879 <i>modulo</i> operator (where the result is either zero or has the same sign
3880 as the divisor, <tt>op2</tt>) of a value.
3881 For more information about the difference,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3882 see <a href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
3883 Math Forum</a>. For a table of how this is implemented in various languages,
3884 please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
3885 Wikipedia: modulo operation</a>.</p>
3886
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3887<p>Note that signed integer remainder and unsigned integer remainder are
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3888 distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
3889
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3890<p>Taking the remainder of a division by zero leads to undefined behavior.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3891 Overflow also leads to undefined behavior; this is a rare case, but can
3892 occur, for example, by taking the remainder of a 32-bit division of
3893 -2147483648 by -1. (The remainder doesn't actually overflow, but this rule
3894 lets srem be implemented using instructions that return both the result of
3895 the division and the remainder.)</p>
3896
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3897<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3898<pre>
3899 &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3900</pre>
3901
3902</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3903
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3904<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3905<h4>
3906 <a name="i_frem">'<tt>frem</tt>' Instruction</a>
3907</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3908
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3909<div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3910
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3911<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3912<pre>
3913 &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]3914</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3915
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3916<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3917<p>The '<tt>frem</tt>' instruction returns the remainder from the division of
3918 its two operands.</p>
3919
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3920<h5>Arguments:</h5>
3921<p>The two arguments to the '<tt>frem</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3922 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3923 floating point values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3924
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3925<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3926<p>This instruction returns the <i>remainder</i> of a division. The remainder
3927 has the same sign as the dividend.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3928
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3929<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3930<pre>
3931 &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]3932</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3933
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3934</div>
Robert Bocchino820bc75b2006-02-17 21:18:08 +0000[diff] [blame]3935
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3936</div>
3937
Reid Spencer2ab01932007-02-02 13:57:07 +0000[diff] [blame]3938<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3939<h3>
3940 <a name="bitwiseops">Bitwise Binary Operations</a>
3941</h3>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3942
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3943<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3944
3945<p>Bitwise binary operators are used to do various forms of bit-twiddling in a
3946 program. They are generally very efficient instructions and can commonly be
3947 strength reduced from other instructions. They require two operands of the
3948 same type, execute an operation on them, and produce a single value. The
3949 resulting value is the same type as its operands.</p>
3950
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3951<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3952<h4>
3953 <a name="i_shl">'<tt>shl</tt>' Instruction</a>
3954</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3955
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3956<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3957
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3958<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3959<pre>
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]3960 &lt;result&gt; = shl &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3961 &lt;result&gt; = shl nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3962 &lt;result&gt; = shl nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3963 &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]3964</pre>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3965
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3966<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3967<p>The '<tt>shl</tt>' instruction returns the first operand shifted to the left
3968 a specified number of bits.</p>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3969
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3970<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3971<p>Both arguments to the '<tt>shl</tt>' instruction must be the
3972 same <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3973 integer type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3974
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3975<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3976<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod
3977 2<sup>n</sup>, where <tt>n</tt> is the width of the result. If <tt>op2</tt>
3978 is (statically or dynamically) negative or equal to or larger than the number
3979 of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3980 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3981 shift amount in <tt>op2</tt>.</p>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3982
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]3983<p>If the <tt>nuw</tt> keyword is present, then the shift produces a
3984 <a href="#trapvalues">trap value</a> if it shifts out any non-zero bits. If
Chris Lattnerf10dfdc2011-02-09 16:44:44 +0000[diff] [blame]3985 the <tt>nsw</tt> keyword is present, then the shift produces a
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]3986 <a href="#trapvalues">trap value</a> if it shifts out any bits that disagree
3987 with the resultant sign bit. As such, NUW/NSW have the same semantics as
3988 they would if the shift were expressed as a mul instruction with the same
3989 nsw/nuw bits in (mul %op1, (shl 1, %op2)).</p>
3990
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3991<h5>Example:</h5>
3992<pre>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3993 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
3994 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
3995 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3996 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Mon P Wang4dd832d2008-12-09 05:46:39 +0000[diff] [blame]3997 &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]3998</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3999
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4000</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4001
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4002<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4003<h4>
4004 <a name="i_lshr">'<tt>lshr</tt>' Instruction</a>
4005</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4006
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4007<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4008
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4009<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4010<pre>
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]4011 &lt;result&gt; = lshr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4012 &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]4013</pre>
4014
4015<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4016<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
4017 operand shifted to the right a specified number of bits with zero fill.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4018
4019<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4020<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4021 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4022 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4023
4024<h5>Semantics:</h5>
4025<p>This instruction always performs a logical shift right operation. The most
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4026 significant bits of the result will be filled with zero bits after the shift.
4027 If <tt>op2</tt> is (statically or dynamically) equal to or larger than the
4028 number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
4029 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
4030 shift amount in <tt>op2</tt>.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4031
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]4032<p>If the <tt>exact</tt> keyword is present, the result value of the
4033 <tt>lshr</tt> is a <a href="#trapvalues">trap value</a> if any of the bits
4034 shifted out are non-zero.</p>
4035
4036
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4037<h5>Example:</h5>
4038<pre>
4039 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
4040 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
4041 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
4042 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]4043 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Mon P Wang4dd832d2008-12-09 05:46:39 +0000[diff] [blame]4044 &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]4045</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4046
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4047</div>
4048
Reid Spencer2ab01932007-02-02 13:57:07 +0000[diff] [blame]4049<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4050<h4>
4051 <a name="i_ashr">'<tt>ashr</tt>' Instruction</a>
4052</h4>
4053
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4054<div>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4055
4056<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4057<pre>
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]4058 &lt;result&gt; = ashr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4059 &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]4060</pre>
4061
4062<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4063<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
4064 operand shifted to the right a specified number of bits with sign
4065 extension.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4066
4067<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4068<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4069 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4070 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4071
4072<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4073<p>This instruction always performs an arithmetic shift right operation, The
4074 most significant bits of the result will be filled with the sign bit
4075 of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
4076 larger than the number of bits in <tt>op1</tt>, the result is undefined. If
4077 the arguments are vectors, each vector element of <tt>op1</tt> is shifted by
4078 the corresponding shift amount in <tt>op2</tt>.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4079
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]4080<p>If the <tt>exact</tt> keyword is present, the result value of the
4081 <tt>ashr</tt> is a <a href="#trapvalues">trap value</a> if any of the bits
4082 shifted out are non-zero.</p>
4083
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4084<h5>Example:</h5>
4085<pre>
4086 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
4087 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
4088 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
4089 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]4090 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Mon P Wang4dd832d2008-12-09 05:46:39 +0000[diff] [blame]4091 &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]4092</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4093
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]4094</div>
4095
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4096<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4097<h4>
4098 <a name="i_and">'<tt>and</tt>' Instruction</a>
4099</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4100
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4101<div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4102
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4103<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4104<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +0000[diff] [blame]4105 &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]4106</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4107
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4108<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4109<p>The '<tt>and</tt>' instruction returns the bitwise logical and of its two
4110 operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4111
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4112<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4113<p>The two arguments to the '<tt>and</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4114 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4115 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4116
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4117<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4118<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4119
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4120<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]4121 <tbody>
4122 <tr>
4123 <td>In0</td>
4124 <td>In1</td>
4125 <td>Out</td>
4126 </tr>
4127 <tr>
4128 <td>0</td>
4129 <td>0</td>
4130 <td>0</td>
4131 </tr>
4132 <tr>
4133 <td>0</td>
4134 <td>1</td>
4135 <td>0</td>
4136 </tr>
4137 <tr>
4138 <td>1</td>
4139 <td>0</td>
4140 <td>0</td>
4141 </tr>
4142 <tr>
4143 <td>1</td>
4144 <td>1</td>
4145 <td>1</td>
4146 </tr>
4147 </tbody>
4148</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4149
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4150<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4151<pre>
4152 &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]4153 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
4154 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4155</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4156</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4157<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4158<h4>
4159 <a name="i_or">'<tt>or</tt>' Instruction</a>
4160</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4161
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4162<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4163
4164<h5>Syntax:</h5>
4165<pre>
4166 &lt;result&gt; = or &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4167</pre>
4168
4169<h5>Overview:</h5>
4170<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive or of its
4171 two operands.</p>
4172
4173<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4174<p>The two arguments to the '<tt>or</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4175 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4176 values. Both arguments must have identical types.</p>
4177
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4178<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4179<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4180
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]4181<table border="1" cellspacing="0" cellpadding="4">
4182 <tbody>
4183 <tr>
4184 <td>In0</td>
4185 <td>In1</td>
4186 <td>Out</td>
4187 </tr>
4188 <tr>
4189 <td>0</td>
4190 <td>0</td>
4191 <td>0</td>
4192 </tr>
4193 <tr>
4194 <td>0</td>
4195 <td>1</td>
4196 <td>1</td>
4197 </tr>
4198 <tr>
4199 <td>1</td>
4200 <td>0</td>
4201 <td>1</td>
4202 </tr>
4203 <tr>
4204 <td>1</td>
4205 <td>1</td>
4206 <td>1</td>
4207 </tr>
4208 </tbody>
4209</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4210
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4211<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4212<pre>
4213 &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4214 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
4215 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4216</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4217
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4218</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4219
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4220<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4221<h4>
4222 <a name="i_xor">'<tt>xor</tt>' Instruction</a>
4223</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4224
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4225<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4226
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4227<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4228<pre>
4229 &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]4230</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4231
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4232<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4233<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive or of
4234 its two operands. The <tt>xor</tt> is used to implement the "one's
4235 complement" operation, which is the "~" operator in C.</p>
4236
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4237<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4238<p>The two arguments to the '<tt>xor</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4239 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4240 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4241
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4242<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4243<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4244
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]4245<table border="1" cellspacing="0" cellpadding="4">
4246 <tbody>
4247 <tr>
4248 <td>In0</td>
4249 <td>In1</td>
4250 <td>Out</td>
4251 </tr>
4252 <tr>
4253 <td>0</td>
4254 <td>0</td>
4255 <td>0</td>
4256 </tr>
4257 <tr>
4258 <td>0</td>
4259 <td>1</td>
4260 <td>1</td>
4261 </tr>
4262 <tr>
4263 <td>1</td>
4264 <td>0</td>
4265 <td>1</td>
4266 </tr>
4267 <tr>
4268 <td>1</td>
4269 <td>1</td>
4270 <td>0</td>
4271 </tr>
4272 </tbody>
4273</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4274
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4275<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4276<pre>
4277 &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4278 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
4279 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
4280 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4281</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4282
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4283</div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4284
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4285</div>
4286
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4287<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4288<h3>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4289 <a name="vectorops">Vector Operations</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4290</h3>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4291
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4292<div>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4293
4294<p>LLVM supports several instructions to represent vector operations in a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4295 target-independent manner. These instructions cover the element-access and
4296 vector-specific operations needed to process vectors effectively. While LLVM
4297 does directly support these vector operations, many sophisticated algorithms
4298 will want to use target-specific intrinsics to take full advantage of a
4299 specific target.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4300
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4301<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4302<h4>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4303 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4304</h4>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4305
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4306<div>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4307
4308<h5>Syntax:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4309<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4310 &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]4311</pre>
4312
4313<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4314<p>The '<tt>extractelement</tt>' instruction extracts a single scalar element
4315 from a vector at a specified index.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4316
4317
4318<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4319<p>The first operand of an '<tt>extractelement</tt>' instruction is a value
4320 of <a href="#t_vector">vector</a> type. The second operand is an index
4321 indicating the position from which to extract the element. The index may be
4322 a variable.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4323
4324<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4325<p>The result is a scalar of the same type as the element type of
4326 <tt>val</tt>. Its value is the value at position <tt>idx</tt> of
4327 <tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4328 results are undefined.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4329
4330<h5>Example:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4331<pre>
Gabor Greif03ab4dc2009-10-28 13:14:50 +0000[diff] [blame]4332 &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]4333</pre>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4334
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4335</div>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4336
4337<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4338<h4>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4339 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4340</h4>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4341
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4342<div>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4343
4344<h5>Syntax:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4345<pre>
Dan Gohman43ba0672008-05-12 23:38:42 +0000[diff] [blame]4346 &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]4347</pre>
4348
4349<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4350<p>The '<tt>insertelement</tt>' instruction inserts a scalar element into a
4351 vector at a specified index.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4352
4353<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4354<p>The first operand of an '<tt>insertelement</tt>' instruction is a value
4355 of <a href="#t_vector">vector</a> type. The second operand is a scalar value
4356 whose type must equal the element type of the first operand. The third
4357 operand is an index indicating the position at which to insert the value.
4358 The index may be a variable.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4359
4360<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4361<p>The result is a vector of the same type as <tt>val</tt>. Its element values
4362 are those of <tt>val</tt> except at position <tt>idx</tt>, where it gets the
4363 value <tt>elt</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4364 results are undefined.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4365
4366<h5>Example:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4367<pre>
Gabor Greif03ab4dc2009-10-28 13:14:50 +0000[diff] [blame]4368 &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]4369</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4370
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4371</div>
4372
4373<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4374<h4>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4375 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4376</h4>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4377
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4378<div>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4379
4380<h5>Syntax:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4381<pre>
Mon P Wang25f01062008-11-10 04:46:22 +0000[diff] [blame]4382 &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]4383</pre>
4384
4385<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4386<p>The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
4387 from two input vectors, returning a vector with the same element type as the
4388 input and length that is the same as the shuffle mask.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4389
4390<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4391<p>The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
4392 with types that match each other. The third argument is a shuffle mask whose
4393 element type is always 'i32'. The result of the instruction is a vector
4394 whose length is the same as the shuffle mask and whose element type is the
4395 same as the element type of the first two operands.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4396
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4397<p>The shuffle mask operand is required to be a constant vector with either
4398 constant integer or undef values.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4399
4400<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4401<p>The elements of the two input vectors are numbered from left to right across
4402 both of the vectors. The shuffle mask operand specifies, for each element of
4403 the result vector, which element of the two input vectors the result element
4404 gets. The element selector may be undef (meaning "don't care") and the
4405 second operand may be undef if performing a shuffle from only one vector.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4406
4407<h5>Example:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4408<pre>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4409 &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]4410 &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]4411 &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]4412 &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]4413 &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]4414 &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]4415 &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]4416 &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]4417</pre>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4418
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4419</div>
Tanya Lattnerb138bbe2006-04-14 19:24:33 +0000[diff] [blame]4420
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4421</div>
4422
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4423<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4424<h3>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4425 <a name="aggregateops">Aggregate Operations</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4426</h3>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4427
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4428<div>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4429
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4430<p>LLVM supports several instructions for working with
4431 <a href="#t_aggregate">aggregate</a> values.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4432
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4433<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4434<h4>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4435 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4436</h4>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4437
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4438<div>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4439
4440<h5>Syntax:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4441<pre>
4442 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
4443</pre>
4444
4445<h5>Overview:</h5>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4446<p>The '<tt>extractvalue</tt>' instruction extracts the value of a member field
4447 from an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4448
4449<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4450<p>The first operand of an '<tt>extractvalue</tt>' instruction is a value
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]4451 of <a href="#t_struct">struct</a> or
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4452 <a href="#t_array">array</a> type. The operands are constant indices to
4453 specify which value to extract in a similar manner as indices in a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4454 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
Frits van Bommel7cf63ac2010-12-05 20:54:38 +0000[diff] [blame]4455 <p>The major differences to <tt>getelementptr</tt> indexing are:</p>
4456 <ul>
4457 <li>Since the value being indexed is not a pointer, the first index is
4458 omitted and assumed to be zero.</li>
4459 <li>At least one index must be specified.</li>
4460 <li>Not only struct indices but also array indices must be in
4461 bounds.</li>
4462 </ul>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4463
4464<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4465<p>The result is the value at the position in the aggregate specified by the
4466 index operands.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4467
4468<h5>Example:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4469<pre>
Gabor Greif03ab4dc2009-10-28 13:14:50 +0000[diff] [blame]4470 &lt;result&gt; = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4471</pre>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4472
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4473</div>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4474
4475<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4476<h4>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4477 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4478</h4>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4479
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4480<div>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4481
4482<h5>Syntax:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4483<pre>
Bill Wendlingf6a91cf2011-07-26 20:42:28 +0000[diff] [blame]4484 &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]4485</pre>
4486
4487<h5>Overview:</h5>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4488<p>The '<tt>insertvalue</tt>' instruction inserts a value into a member field
4489 in an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4490
4491<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4492<p>The first operand of an '<tt>insertvalue</tt>' instruction is a value
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]4493 of <a href="#t_struct">struct</a> or
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4494 <a href="#t_array">array</a> type. The second operand is a first-class
4495 value to insert. The following operands are constant indices indicating
4496 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]4497 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' instruction. The
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4498 value to insert must have the same type as the value identified by the
4499 indices.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4500
4501<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4502<p>The result is an aggregate of the same type as <tt>val</tt>. Its value is
4503 that of <tt>val</tt> except that the value at the position specified by the
4504 indices is that of <tt>elt</tt>.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4505
4506<h5>Example:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4507<pre>
Chris Lattnerc2e85402011-05-22 07:18:08 +0000[diff] [blame]4508 %agg1 = insertvalue {i32, float} undef, i32 1, 0 <i>; yields {i32 1, float undef}</i>
4509 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 <i>; yields {i32 1, float %val}</i>
4510 %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]4511</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4512
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4513</div>
4514
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4515</div>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4516
4517<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4518<h3>
Chris Lattner6ab66722006-08-15 00:45:58 +0000[diff] [blame]4519 <a name="memoryops">Memory Access and Addressing Operations</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4520</h3>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4521
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4522<div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4523
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4524<p>A key design point of an SSA-based representation is how it represents
4525 memory. In LLVM, no memory locations are in SSA form, which makes things
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]4526 very simple. This section describes how to read, write, and allocate
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4527 memory in LLVM.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4528
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4529<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4530<h4>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4531 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4532</h4>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4533
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4534<div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4535
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4536<h5>Syntax:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4537<pre>
Dan Gohman2140a742010-05-28 01:14:11 +0000[diff] [blame]4538 &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]4539</pre>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4540
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4541<h5>Overview:</h5>
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]4542<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4543 currently executing function, to be automatically released when this function
4544 returns to its caller. The object is always allocated in the generic address
4545 space (address space zero).</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4546
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4547<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4548<p>The '<tt>alloca</tt>' instruction
4549 allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt> bytes of memory on the
4550 runtime stack, returning a pointer of the appropriate type to the program.
4551 If "NumElements" is specified, it is the number of elements allocated,
4552 otherwise "NumElements" is defaulted to be one. If a constant alignment is
4553 specified, the value result of the allocation is guaranteed to be aligned to
4554 at least that boundary. If not specified, or if zero, the target can choose
4555 to align the allocation on any convenient boundary compatible with the
4556 type.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4557
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4558<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4559
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4560<h5>Semantics:</h5>
Bill Wendling9ee6a312009-05-08 20:49:29 +0000[diff] [blame]4561<p>Memory is allocated; a pointer is returned. The operation is undefined if
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4562 there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
4563 memory is automatically released when the function returns. The
4564 '<tt>alloca</tt>' instruction is commonly used to represent automatic
4565 variables that must have an address available. When the function returns
4566 (either with the <tt><a href="#i_ret">ret</a></tt>
4567 or <tt><a href="#i_unwind">unwind</a></tt> instructions), the memory is
4568 reclaimed. Allocating zero bytes is legal, but the result is undefined.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4569
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4570<h5>Example:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4571<pre>
Dan Gohman7a5acb52009-01-04 23:49:44 +0000[diff] [blame]4572 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
4573 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
4574 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
4575 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4576</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4577
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4578</div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4579
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4580<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4581<h4>
4582 <a name="i_load">'<tt>load</tt>' Instruction</a>
4583</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4584
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4585<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4586
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4587<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4588<pre>
Eli Friedman02e737b2011-08-12 22:50:01 +0000[diff] [blame]4589 &lt;result&gt; = load [volatile] &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;]
4590 &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]4591 !&lt;index&gt; = !{ i32 1 }
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4592</pre>
4593
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4594<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4595<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4596
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4597<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4598<p>The argument to the '<tt>load</tt>' instruction specifies the memory address
4599 from which to load. The pointer must point to
4600 a <a href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
4601 marked as <tt>volatile</tt>, then the optimizer is not allowed to modify the
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]4602 number or order of execution of this <tt>load</tt> with other <a
4603 href="#volatile">volatile operations</a>.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4604
Eli Friedman59b66882011-08-09 23:02:53 +0000[diff] [blame]4605<p>If the <code>load</code> is marked as <code>atomic</code>, it takes an extra
4606 <a href="#ordering">ordering</a> and optional <code>singlethread</code>
4607 argument. The <code>release</code> and <code>acq_rel</code> orderings are
4608 not valid on <code>load</code> instructions. Atomic loads produce <a
4609 href="#memorymodel">defined</a> results when they may see multiple atomic
4610 stores. The type of the pointee must be an integer type whose bit width
4611 is a power of two greater than or equal to eight and less than or equal
4612 to a target-specific size limit. <code>align</code> must be explicitly
4613 specified on atomic loads, and the load has undefined behavior if the
4614 alignment is not set to a value which is at least the size in bytes of
4615 the pointee. <code>!nontemporal</code> does not have any defined semantics
4616 for atomic loads.</p>
4617
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4618<p>The optional constant <tt>align</tt> argument specifies the alignment of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4619 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]4620 omitted <tt>align</tt> argument means that the operation has the preferential
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4621 alignment for the target. It is the responsibility of the code emitter to
4622 ensure that the alignment information is correct. Overestimating the
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4623 alignment results in undefined behavior. Underestimating the alignment may
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4624 produce less efficient code. An alignment of 1 is always safe.</p>
4625
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4626<p>The optional <tt>!nontemporal</tt> metadata must reference a single
4627 metatadata name &lt;index&gt; corresponding to a metadata node with
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]4628 one <tt>i32</tt> entry of value 1. The existence of
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4629 the <tt>!nontemporal</tt> metatadata on the instruction tells the optimizer
4630 and code generator that this load is not expected to be reused in the cache.
4631 The code generator may select special instructions to save cache bandwidth,
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]4632 such as the <tt>MOVNT</tt> instruction on x86.</p>
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4633
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4634<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4635<p>The location of memory pointed to is loaded. If the value being loaded is of
4636 scalar type then the number of bytes read does not exceed the minimum number
4637 of bytes needed to hold all bits of the type. For example, loading an
4638 <tt>i24</tt> reads at most three bytes. When loading a value of a type like
4639 <tt>i20</tt> with a size that is not an integral number of bytes, the result
4640 is undefined if the value was not originally written using a store of the
4641 same type.</p>
4642
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4643<h5>Examples:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4644<pre>
4645 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
4646 <a href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4647 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4648</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4649
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4650</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4651
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4652<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4653<h4>
4654 <a name="i_store">'<tt>store</tt>' Instruction</a>
4655</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4656
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4657<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4658
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4659<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4660<pre>
Eli Friedman02e737b2011-08-12 22:50:01 +0000[diff] [blame]4661 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>
4662 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]4663</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4664
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4665<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4666<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4667
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4668<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4669<p>There are two arguments to the '<tt>store</tt>' instruction: a value to store
4670 and an address at which to store it. The type of the
4671 '<tt>&lt;pointer&gt;</tt>' operand must be a pointer to
4672 the <a href="#t_firstclass">first class</a> type of the
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]4673 '<tt>&lt;value&gt;</tt>' operand. If the <tt>store</tt> is marked as
4674 <tt>volatile</tt>, then the optimizer is not allowed to modify the number or
4675 order of execution of this <tt>store</tt> with other <a
4676 href="#volatile">volatile operations</a>.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4677
Eli Friedman59b66882011-08-09 23:02:53 +0000[diff] [blame]4678<p>If the <code>store</code> is marked as <code>atomic</code>, it takes an extra
4679 <a href="#ordering">ordering</a> and optional <code>singlethread</code>
4680 argument. The <code>acquire</code> and <code>acq_rel</code> orderings aren't
4681 valid on <code>store</code> instructions. Atomic loads produce <a
4682 href="#memorymodel">defined</a> results when they may see multiple atomic
4683 stores. The type of the pointee must be an integer type whose bit width
4684 is a power of two greater than or equal to eight and less than or equal
4685 to a target-specific size limit. <code>align</code> must be explicitly
4686 specified on atomic stores, and the store has undefined behavior if the
4687 alignment is not set to a value which is at least the size in bytes of
4688 the pointee. <code>!nontemporal</code> does not have any defined semantics
4689 for atomic stores.</p>
4690
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4691<p>The optional constant "align" argument specifies the alignment of the
4692 operation (that is, the alignment of the memory address). A value of 0 or an
4693 omitted "align" argument means that the operation has the preferential
4694 alignment for the target. It is the responsibility of the code emitter to
4695 ensure that the alignment information is correct. Overestimating the
4696 alignment results in an undefined behavior. Underestimating the alignment may
4697 produce less efficient code. An alignment of 1 is always safe.</p>
4698
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4699<p>The optional !nontemporal metadata must reference a single metatadata
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]4700 name &lt;index&gt; corresponding to a metadata node with one i32 entry of
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]4701 value 1. The existence of the !nontemporal metatadata on the
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4702 instruction tells the optimizer and code generator that this load is
4703 not expected to be reused in the cache. The code generator may
4704 select special instructions to save cache bandwidth, such as the
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]4705 MOVNT instruction on x86.</p>
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4706
4707
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]4708<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4709<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>' at the
4710 location specified by the '<tt>&lt;pointer&gt;</tt>' operand. If
4711 '<tt>&lt;value&gt;</tt>' is of scalar type then the number of bytes written
4712 does not exceed the minimum number of bytes needed to hold all bits of the
4713 type. For example, storing an <tt>i24</tt> writes at most three bytes. When
4714 writing a value of a type like <tt>i20</tt> with a size that is not an
4715 integral number of bytes, it is unspecified what happens to the extra bits
4716 that do not belong to the type, but they will typically be overwritten.</p>
4717
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4718<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4719<pre>
4720 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling8830ffe2007-10-22 05:10:05 +0000[diff] [blame]4721 store i32 3, i32* %ptr <i>; yields {void}</i>
4722 %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]4723</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4724
Reid Spencer443460a2006-11-09 21:15:49 +0000[diff] [blame]4725</div>
4726
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4727<!-- _______________________________________________________________________ -->
NAKAMURA Takumi0300d882011-08-12 06:17:17 +0000[diff] [blame]4728<h4>
4729<a name="i_fence">'<tt>fence</tt>' Instruction</a>
4730</h4>
Eli Friedmanfee02c62011-07-25 23:16:38 +0000[diff] [blame]4731
NAKAMURA Takumi0300d882011-08-12 06:17:17 +0000[diff] [blame]4732<div>
Eli Friedmanfee02c62011-07-25 23:16:38 +0000[diff] [blame]4733
4734<h5>Syntax:</h5>
4735<pre>
4736 fence [singlethread] &lt;ordering&gt; <i>; yields {void}</i>
4737</pre>
4738
4739<h5>Overview:</h5>
4740<p>The '<tt>fence</tt>' instruction is used to introduce happens-before edges
4741between operations.</p>
4742
4743<h5>Arguments:</h5> <p>'<code>fence</code>' instructions take an <a
4744href="#ordering">ordering</a> argument which defines what
4745<i>synchronizes-with</i> edges they add. They can only be given
4746<code>acquire</code>, <code>release</code>, <code>acq_rel</code>, and
4747<code>seq_cst</code> orderings.</p>
4748
4749<h5>Semantics:</h5>
4750<p>A fence <var>A</var> which has (at least) <code>release</code> ordering
4751semantics <i>synchronizes with</i> a fence <var>B</var> with (at least)
4752<code>acquire</code> ordering semantics if and only if there exist atomic
4753operations <var>X</var> and <var>Y</var>, both operating on some atomic object
4754<var>M</var>, such that <var>A</var> is sequenced before <var>X</var>,
4755<var>X</var> modifies <var>M</var> (either directly or through some side effect
4756of a sequence headed by <var>X</var>), <var>Y</var> is sequenced before
4757<var>B</var>, and <var>Y</var> observes <var>M</var>. This provides a
4758<i>happens-before</i> dependency between <var>A</var> and <var>B</var>. Rather
4759than an explicit <code>fence</code>, one (but not both) of the atomic operations
4760<var>X</var> or <var>Y</var> might provide a <code>release</code> or
4761<code>acquire</code> (resp.) ordering constraint and still
4762<i>synchronize-with</i> the explicit <code>fence</code> and establish the
4763<i>happens-before</i> edge.</p>
4764
4765<p>A <code>fence</code> which has <code>seq_cst</code> ordering, in addition to
4766having both <code>acquire</code> and <code>release</code> semantics specified
4767above, participates in the global program order of other <code>seq_cst</code>
4768operations and/or fences.</p>
4769
4770<p>The optional "<a href="#singlethread"><code>singlethread</code></a>" argument
4771specifies that the fence only synchronizes with other fences in the same
4772thread. (This is useful for interacting with signal handlers.)</p>
4773
Eli Friedmanfee02c62011-07-25 23:16:38 +0000[diff] [blame]4774<h5>Example:</h5>
4775<pre>
4776 fence acquire <i>; yields {void}</i>
4777 fence singlethread seq_cst <i>; yields {void}</i>
4778</pre>
4779
4780</div>
4781
4782<!-- _______________________________________________________________________ -->
NAKAMURA Takumi0300d882011-08-12 06:17:17 +0000[diff] [blame]4783<h4>
4784<a name="i_cmpxchg">'<tt>cmpxchg</tt>' Instruction</a>
4785</h4>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]4786
NAKAMURA Takumi0300d882011-08-12 06:17:17 +0000[diff] [blame]4787<div>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]4788
4789<h5>Syntax:</h5>
4790<pre>
Eli Friedman02e737b2011-08-12 22:50:01 +0000[diff] [blame]4791 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]4792</pre>
4793
4794<h5>Overview:</h5>
4795<p>The '<tt>cmpxchg</tt>' instruction is used to atomically modify memory.
4796It loads a value in memory and compares it to a given value. If they are
4797equal, it stores a new value into the memory.</p>
4798
4799<h5>Arguments:</h5>
4800<p>There are three arguments to the '<code>cmpxchg</code>' instruction: an
4801address to operate on, a value to compare to the value currently be at that
4802address, and a new value to place at that address if the compared values are
4803equal. The type of '<var>&lt;cmp&gt;</var>' must be an integer type whose
4804bit width is a power of two greater than or equal to eight and less than
4805or equal to a target-specific size limit. '<var>&lt;cmp&gt;</var>' and
4806'<var>&lt;new&gt;</var>' must have the same type, and the type of
4807'<var>&lt;pointer&gt;</var>' must be a pointer to that type. If the
4808<code>cmpxchg</code> is marked as <code>volatile</code>, then the
4809optimizer is not allowed to modify the number or order of execution
4810of this <code>cmpxchg</code> with other <a href="#volatile">volatile
4811operations</a>.</p>
4812
4813<!-- FIXME: Extend allowed types. -->
4814
4815<p>The <a href="#ordering"><var>ordering</var></a> argument specifies how this
4816<code>cmpxchg</code> synchronizes with other atomic operations.</p>
4817
4818<p>The optional "<code>singlethread</code>" argument declares that the
4819<code>cmpxchg</code> is only atomic with respect to code (usually signal
4820handlers) running in the same thread as the <code>cmpxchg</code>. Otherwise the
4821cmpxchg is atomic with respect to all other code in the system.</p>
4822
4823<p>The pointer passed into cmpxchg must have alignment greater than or equal to
4824the size in memory of the operand.
4825
4826<h5>Semantics:</h5>
4827<p>The contents of memory at the location specified by the
4828'<tt>&lt;pointer&gt;</tt>' operand is read and compared to
4829'<tt>&lt;cmp&gt;</tt>'; if the read value is the equal,
4830'<tt>&lt;new&gt;</tt>' is written. The original value at the location
4831is returned.
4832
4833<p>A successful <code>cmpxchg</code> is a read-modify-write instruction for the
4834purpose of identifying <a href="#release_sequence">release sequences</a>. A
4835failed <code>cmpxchg</code> is equivalent to an atomic load with an ordering
4836parameter determined by dropping any <code>release</code> part of the
4837<code>cmpxchg</code>'s ordering.</p>
4838
4839<!--
4840FIXME: Is compare_exchange_weak() necessary? (Consider after we've done
4841optimization work on ARM.)
4842
4843FIXME: Is a weaker ordering constraint on failure helpful in practice?
4844-->
4845
4846<h5>Example:</h5>
4847<pre>
4848entry:
4849 %orig = atomic <a href="#i_load">load</a> i32* %ptr unordered <i>; yields {i32}</i>
4850 <a href="#i_br">br</a> label %loop
4851
4852loop:
4853 %cmp = <a href="#i_phi">phi</a> i32 [ %orig, %entry ], [%old, %loop]
4854 %squared = <a href="#i_mul">mul</a> i32 %cmp, %cmp
4855 %old = cmpxchg i32* %ptr, i32 %cmp, i32 %squared <i>; yields {i32}</i>
4856 %success = <a href="#i_icmp">icmp</a> eq i32 %cmp, %old
4857 <a href="#i_br">br</a> i1 %success, label %done, label %loop
4858
4859done:
4860 ...
4861</pre>
4862
4863</div>
4864
4865<!-- _______________________________________________________________________ -->
NAKAMURA Takumi0300d882011-08-12 06:17:17 +0000[diff] [blame]4866<h4>
4867<a name="i_atomicrmw">'<tt>atomicrmw</tt>' Instruction</a>
4868</h4>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]4869
NAKAMURA Takumi0300d882011-08-12 06:17:17 +0000[diff] [blame]4870<div>
Eli Friedmanc9a551e2011-07-28 21:48:00 +0000[diff] [blame]4871
4872<h5>Syntax:</h5>
4873<pre>
Eli Friedman02e737b2011-08-12 22:50:01 +0000[diff] [blame]4874 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]4875</pre>
4876
4877<h5>Overview:</h5>
4878<p>The '<tt>atomicrmw</tt>' instruction is used to atomically modify memory.</p>
4879
4880<h5>Arguments:</h5>
4881<p>There are three arguments to the '<code>atomicrmw</code>' instruction: an
4882operation to apply, an address whose value to modify, an argument to the
4883operation. The operation must be one of the following keywords:</p>
4884<ul>
4885 <li>xchg</li>
4886 <li>add</li>
4887 <li>sub</li>
4888 <li>and</li>
4889 <li>nand</li>
4890 <li>or</li>
4891 <li>xor</li>
4892 <li>max</li>
4893 <li>min</li>
4894 <li>umax</li>
4895 <li>umin</li>
4896</ul>
4897
4898<p>The type of '<var>&lt;value&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. The type of the
4901'<code>&lt;pointer&gt;</code>' operand must be a pointer to that type.
4902If the <code>atomicrmw</code> is marked as <code>volatile</code>, then the
4903optimizer is not allowed to modify the number or order of execution of this
4904<code>atomicrmw</code> with other <a href="#volatile">volatile
4905 operations</a>.</p>
4906
4907<!-- FIXME: Extend allowed types. -->
4908
4909<h5>Semantics:</h5>
4910<p>The contents of memory at the location specified by the
4911'<tt>&lt;pointer&gt;</tt>' operand are atomically read, modified, and written
4912back. The original value at the location is returned. The modification is
4913specified by the <var>operation</var> argument:</p>
4914
4915<ul>
4916 <li>xchg: <code>*ptr = val</code></li>
4917 <li>add: <code>*ptr = *ptr + val</code></li>
4918 <li>sub: <code>*ptr = *ptr - val</code></li>
4919 <li>and: <code>*ptr = *ptr &amp; val</code></li>
4920 <li>nand: <code>*ptr = ~(*ptr &amp; val)</code></li>
4921 <li>or: <code>*ptr = *ptr | val</code></li>
4922 <li>xor: <code>*ptr = *ptr ^ val</code></li>
4923 <li>max: <code>*ptr = *ptr &gt; val ? *ptr : val</code> (using a signed comparison)</li>
4924 <li>min: <code>*ptr = *ptr &lt; val ? *ptr : val</code> (using a signed comparison)</li>
4925 <li>umax: <code>*ptr = *ptr &gt; val ? *ptr : val</code> (using an unsigned comparison)</li>
4926 <li>umin: <code>*ptr = *ptr &lt; val ? *ptr : val</code> (using an unsigned comparison)</li>
4927</ul>
4928
4929<h5>Example:</h5>
4930<pre>
4931 %old = atomicrmw add i32* %ptr, i32 1 acquire <i>; yields {i32}</i>
4932</pre>
4933
4934</div>
4935
4936<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4937<h4>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4938 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4939</h4>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4940
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4941<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4942
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]4943<h5>Syntax:</h5>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4944<pre>
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]4945 &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]4946 &lt;result&gt; = getelementptr inbounds &lt;pty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4947</pre>
4948
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]4949<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4950<p>The '<tt>getelementptr</tt>' instruction is used to get the address of a
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4951 subelement of an <a href="#t_aggregate">aggregate</a> data structure.
4952 It performs address calculation only and does not access memory.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4953
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]4954<h5>Arguments:</h5>
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]4955<p>The first argument is always a pointer, and forms the basis of the
Chris Lattnera40b9122009-07-29 06:44:13 +0000[diff] [blame]4956 calculation. The remaining arguments are indices that indicate which of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4957 elements of the aggregate object are indexed. The interpretation of each
4958 index is dependent on the type being indexed into. The first index always
4959 indexes the pointer value given as the first argument, the second index
4960 indexes a value of the type pointed to (not necessarily the value directly
4961 pointed to, since the first index can be non-zero), etc. The first type
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4962 indexed into must be a pointer value, subsequent types can be arrays,
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]4963 vectors, and structs. Note that subsequent types being indexed into
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4964 can never be pointers, since that would require loading the pointer before
4965 continuing calculation.</p>
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]4966
4967<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]4968 When indexing into a (optionally packed) structure, only <tt>i32</tt>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4969 integer <b>constants</b> are allowed. When indexing into an array, pointer
4970 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]4971 constant. These integers are treated as signed values where relevant.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4972
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4973<p>For example, let's consider a C code fragment and how it gets compiled to
4974 LLVM:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4975
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]4976<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4977struct RT {
4978 char A;
Chris Lattnera446f1b2007-05-29 15:43:56 +0000[diff] [blame]4979 int B[10][20];
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4980 char C;
4981};
4982struct ST {
Chris Lattnera446f1b2007-05-29 15:43:56 +0000[diff] [blame]4983 int X;
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4984 double Y;
4985 struct RT Z;
4986};
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4987
Chris Lattnera446f1b2007-05-29 15:43:56 +0000[diff] [blame]4988int *foo(struct ST *s) {
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4989 return &amp;s[1].Z.B[5][13];
4990}
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4991</pre>
4992
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4993<p>The LLVM code generated by the GCC frontend is:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4994
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]4995<pre class="doc_code">
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]4996%RT = <a href="#namedtypes">type</a> { i8 , [10 x [20 x i32]], i8 }
4997%ST = <a href="#namedtypes">type</a> { i32, double, %RT }
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4998
Dan Gohman6b867702009-07-25 02:23:48 +0000[diff] [blame]4999define i32* @foo(%ST* %s) {
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]5000entry:
5001 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
5002 ret i32* %reg
5003}
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]5004</pre>
5005
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]5006<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5007<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5008 type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
5009 }</tt>' type, a structure. The second index indexes into the third element
5010 of the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
5011 i8 }</tt>' type, another structure. The third index indexes into the second
5012 element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
5013 array. The two dimensions of the array are subscripted into, yielding an
5014 '<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a
5015 pointer to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]5016
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5017<p>Note that it is perfectly legal to index partially through a structure,
5018 returning a pointer to an inner element. Because of this, the LLVM code for
5019 the given testcase is equivalent to:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]5020
5021<pre>
Dan Gohman6b867702009-07-25 02:23:48 +0000[diff] [blame]5022 define i32* @foo(%ST* %s) {
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5023 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]5024 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
5025 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5026 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
5027 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
5028 ret i32* %t5
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]5029 }
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]5030</pre>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +0000[diff] [blame]5031
Dan Gohman1639c392009-07-27 21:53:46 +0000[diff] [blame]5032<p>If the <tt>inbounds</tt> keyword is present, the result value of the
Dan Gohman57255802010-04-23 15:23:32 +0000[diff] [blame]5033 <tt>getelementptr</tt> is a <a href="#trapvalues">trap value</a> if the
5034 base pointer is not an <i>in bounds</i> address of an allocated object,
5035 or if any of the addresses that would be formed by successive addition of
5036 the offsets implied by the indices to the base address with infinitely
Eli Friedmand8874dc2011-08-12 23:37:55 +0000[diff] [blame]5037 precise signed arithmetic are not an <i>in bounds</i> address of that
5038 allocated object. The <i>in bounds</i> addresses for an allocated object
5039 are all the addresses that point into the object, plus the address one
5040 byte past the end.</p>
Dan Gohman1639c392009-07-27 21:53:46 +0000[diff] [blame]5041
5042<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]5043 the base address with silently-wrapping two's complement arithmetic. If the
5044 offsets have a different width from the pointer, they are sign-extended or
5045 truncated to the width of the pointer. The result value of the
5046 <tt>getelementptr</tt> may be outside the object pointed to by the base
5047 pointer. The result value may not necessarily be used to access memory
5048 though, even if it happens to point into allocated storage. See the
5049 <a href="#pointeraliasing">Pointer Aliasing Rules</a> section for more
5050 information.</p>
Dan Gohman1639c392009-07-27 21:53:46 +0000[diff] [blame]5051
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5052<p>The getelementptr instruction is often confusing. For some more insight into
5053 how it works, see <a href="GetElementPtr.html">the getelementptr FAQ</a>.</p>
Chris Lattner6ab66722006-08-15 00:45:58 +0000[diff] [blame]5054
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]5055<h5>Example:</h5>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]5056<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5057 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]5058 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
5059 <i>; yields i8*:vptr</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]5060 %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]5061 <i>; yields i8*:eptr</i>
5062 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Sanjiv Gupta0c155e62009-04-25 07:27:44 +0000[diff] [blame]5063 <i>; yields i32*:iptr</i>
Sanjiv Gupta77abea02009-04-24 16:38:13 +0000[diff] [blame]5064 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]5065</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5066
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]5067</div>
Reid Spencer443460a2006-11-09 21:15:49 +0000[diff] [blame]5068
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5069</div>
5070
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5071<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5072<h3>
5073 <a name="convertops">Conversion Operations</a>
5074</h3>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5075
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5076<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5077
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5078<p>The instructions in this category are the conversion instructions (casting)
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5079 which all take a single operand and a type. They perform various bit
5080 conversions on the operand.</p>
5081
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]5082<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5083<h4>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5084 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5085</h4>
5086
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5087<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5088
5089<h5>Syntax:</h5>
5090<pre>
5091 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5092</pre>
5093
5094<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5095<p>The '<tt>trunc</tt>' instruction truncates its operand to the
5096 type <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5097
5098<h5>Arguments:</h5>
Nadav Rotem502f1b92011-02-24 21:01:34 +0000[diff] [blame]5099<p>The '<tt>trunc</tt>' instruction takes a value to trunc, and a type to trunc it to.
5100 Both types must be of <a href="#t_integer">integer</a> types, or vectors
5101 of the same number of integers.
5102 The bit size of the <tt>value</tt> must be larger than
5103 the bit size of the destination type, <tt>ty2</tt>.
5104 Equal sized types are not allowed.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5105
5106<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5107<p>The '<tt>trunc</tt>' instruction truncates the high order bits
5108 in <tt>value</tt> and converts the remaining bits to <tt>ty2</tt>. Since the
5109 source size must be larger than the destination size, <tt>trunc</tt> cannot
5110 be a <i>no-op cast</i>. It will always truncate bits.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5111
5112<h5>Example:</h5>
5113<pre>
Nadav Rotem502f1b92011-02-24 21:01:34 +0000[diff] [blame]5114 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
5115 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
5116 %Z = trunc i32 122 to i1 <i>; yields i1:false</i>
5117 %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]5118</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5119
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5120</div>
5121
5122<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5123<h4>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5124 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5125</h4>
5126
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5127<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5128
5129<h5>Syntax:</h5>
5130<pre>
5131 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5132</pre>
5133
5134<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5135<p>The '<tt>zext</tt>' instruction zero extends its operand to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5136 <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5137
5138
5139<h5>Arguments:</h5>
Nadav Rotem25f2ac92011-02-20 12:37:50 +0000[diff] [blame]5140<p>The '<tt>zext</tt>' instruction takes a value to cast, and a type to cast it to.
5141 Both types must be of <a href="#t_integer">integer</a> types, or vectors
5142 of the same number of integers.
5143 The bit size of the <tt>value</tt> must be smaller than
5144 the bit size of the destination type,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5145 <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5146
5147<h5>Semantics:</h5>
5148<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]5149 bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5150
Reid Spencer07c9c682007-01-12 15:46:11 +0000[diff] [blame]5151<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]5152
5153<h5>Example:</h5>
5154<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5155 %X = zext i32 257 to i64 <i>; yields i64:257</i>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]5156 %Y = zext i1 true to i32 <i>; yields i32:1</i>
Nadav Rotem25f2ac92011-02-20 12:37:50 +0000[diff] [blame]5157 %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]5158</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5159
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5160</div>
5161
5162<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5163<h4>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5164 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5165</h4>
5166
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5167<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5168
5169<h5>Syntax:</h5>
5170<pre>
5171 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5172</pre>
5173
5174<h5>Overview:</h5>
5175<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
5176
5177<h5>Arguments:</h5>
Nadav Rotem502f1b92011-02-24 21:01:34 +0000[diff] [blame]5178<p>The '<tt>sext</tt>' instruction takes a value to cast, and a type to cast it to.
5179 Both types must be of <a href="#t_integer">integer</a> types, or vectors
5180 of the same number of integers.
5181 The bit size of the <tt>value</tt> must be smaller than
5182 the bit size of the destination type,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5183 <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5184
5185<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5186<p>The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
5187 bit (highest order bit) of the <tt>value</tt> until it reaches the bit size
5188 of the type <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5189
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]5190<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]5191
5192<h5>Example:</h5>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5193<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5194 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]5195 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
Nadav Rotem502f1b92011-02-24 21:01:34 +0000[diff] [blame]5196 %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]5197</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5198
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5199</div>
5200
5201<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5202<h4>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]5203 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5204</h4>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]5205
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5206<div>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]5207
5208<h5>Syntax:</h5>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]5209<pre>
5210 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5211</pre>
5212
5213<h5>Overview:</h5>
5214<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5215 <tt>ty2</tt>.</p>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]5216
5217<h5>Arguments:</h5>
5218<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5219 point</a> value to cast and a <a href="#t_floating">floating point</a> type
5220 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]5221 <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]5222 <i>no-op cast</i>.</p>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]5223
5224<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5225<p>The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5226 <a href="#t_floating">floating point</a> type to a smaller
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5227 <a href="#t_floating">floating point</a> type. If the value cannot fit
5228 within the destination type, <tt>ty2</tt>, then the results are
5229 undefined.</p>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]5230
5231<h5>Example:</h5>
5232<pre>
5233 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
5234 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
5235</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5236
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]5237</div>
5238
5239<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5240<h4>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5241 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5242</h4>
5243
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5244<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5245
5246<h5>Syntax:</h5>
5247<pre>
5248 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5249</pre>
5250
5251<h5>Overview:</h5>
5252<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]5253 floating point value.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5254
5255<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5256<p>The '<tt>fpext</tt>' instruction takes a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5257 <a href="#t_floating">floating point</a> <tt>value</tt> to cast, and
5258 a <a href="#t_floating">floating point</a> type to cast it to. The source
5259 type must be smaller than the destination type.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5260
5261<h5>Semantics:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]5262<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5263 <a href="#t_floating">floating point</a> type to a larger
5264 <a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
5265 used to make a <i>no-op cast</i> because it always changes bits. Use
5266 <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]5267
5268<h5>Example:</h5>
5269<pre>
Nick Lewycky9feca672011-03-31 18:20:19 +0000[diff] [blame]5270 %X = fpext float 3.125 to double <i>; yields double:3.125000e+00</i>
5271 %Y = fpext double %X to fp128 <i>; yields fp128:0xL00000000000000004000900000000000</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5272</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5273
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5274</div>
5275
5276<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5277<h4>
Reid Spencer2eadb532007-01-21 00:29:26 +0000[diff] [blame]5278 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5279</h4>
5280
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5281<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5282
5283<h5>Syntax:</h5>
5284<pre>
Reid Spencer753163d2007-07-31 14:40:14 +0000[diff] [blame]5285 &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]5286</pre>
5287
5288<h5>Overview:</h5>
Reid Spencer753163d2007-07-31 14:40:14 +0000[diff] [blame]5289<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5290 unsigned integer equivalent of type <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5291
5292<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5293<p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
5294 scalar or vector <a href="#t_floating">floating point</a> value, and a type
5295 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
5296 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
5297 vector integer type with the same number of elements as <tt>ty</tt></p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5298
5299<h5>Semantics:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5300<p>The '<tt>fptoui</tt>' instruction converts its
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5301 <a href="#t_floating">floating point</a> operand into the nearest (rounding
5302 towards zero) unsigned integer value. If the value cannot fit
5303 in <tt>ty2</tt>, the results are undefined.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5304
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5305<h5>Example:</h5>
5306<pre>
Reid Spencer753163d2007-07-31 14:40:14 +0000[diff] [blame]5307 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner5b95a172007-09-22 03:17:52 +0000[diff] [blame]5308 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Gabor Greiff50fd572009-10-28 09:21:30 +0000[diff] [blame]5309 %Z = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5310</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5311
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5312</div>
5313
5314<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5315<h4>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]5316 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5317</h4>
5318
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5319<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5320
5321<h5>Syntax:</h5>
5322<pre>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]5323 &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]5324</pre>
5325
5326<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5327<p>The '<tt>fptosi</tt>' instruction converts
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5328 <a href="#t_floating">floating point</a> <tt>value</tt> to
5329 type <tt>ty2</tt>.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5330
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]5331<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5332<p>The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
5333 scalar or vector <a href="#t_floating">floating point</a> value, and a type
5334 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
5335 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
5336 vector integer type with the same number of elements as <tt>ty</tt></p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5337
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]5338<h5>Semantics:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5339<p>The '<tt>fptosi</tt>' instruction converts its
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5340 <a href="#t_floating">floating point</a> operand into the nearest (rounding
5341 towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
5342 the results are undefined.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5343
Chris Lattner70de6632001-07-09 00:26:23 +0000[diff] [blame]5344<h5>Example:</h5>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5345<pre>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]5346 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner5b95a172007-09-22 03:17:52 +0000[diff] [blame]5347 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Gabor Greiff50fd572009-10-28 09:21:30 +0000[diff] [blame]5348 %Z = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5349</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5350
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5351</div>
5352
5353<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5354<h4>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]5355 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5356</h4>
5357
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5358<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5359
5360<h5>Syntax:</h5>
5361<pre>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]5362 &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]5363</pre>
5364
5365<h5>Overview:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]5366<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5367 integer and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5368
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5369<h5>Arguments:</h5>
Nate Begemand4d45c22007-11-17 03:58:34 +0000[diff] [blame]5370<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]5371 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
5372 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
5373 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
5374 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5375
5376<h5>Semantics:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]5377<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5378 integer quantity and converts it to the corresponding floating point
5379 value. If the value cannot fit in the floating point value, the results are
5380 undefined.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5381
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5382<h5>Example:</h5>
5383<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5384 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]5385 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5386</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5387
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5388</div>
5389
5390<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5391<h4>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]5392 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5393</h4>
5394
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5395<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5396
5397<h5>Syntax:</h5>
5398<pre>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]5399 &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]5400</pre>
5401
5402<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5403<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed integer
5404 and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5405
5406<h5>Arguments:</h5>
Nate Begemand4d45c22007-11-17 03:58:34 +0000[diff] [blame]5407<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]5408 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
5409 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
5410 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
5411 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5412
5413<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5414<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed integer
5415 quantity and converts it to the corresponding floating point value. If the
5416 value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5417
5418<h5>Example:</h5>
5419<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5420 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]5421 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5422</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5423
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5424</div>
5425
5426<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5427<h4>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5428 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5429</h4>
5430
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5431<div>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5432
5433<h5>Syntax:</h5>
5434<pre>
5435 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5436</pre>
5437
5438<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5439<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
5440 the integer type <tt>ty2</tt>.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5441
5442<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5443<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
5444 must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
5445 <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> type.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5446
5447<h5>Semantics:</h5>
5448<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5449 <tt>ty2</tt> by interpreting the pointer value as an integer and either
5450 truncating or zero extending that value to the size of the integer type. If
5451 <tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
5452 <tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
5453 are the same size, then nothing is done (<i>no-op cast</i>) other than a type
5454 change.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5455
5456<h5>Example:</h5>
5457<pre>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5458 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
5459 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5460</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5461
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5462</div>
5463
5464<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5465<h4>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5466 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5467</h4>
5468
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5469<div>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5470
5471<h5>Syntax:</h5>
5472<pre>
5473 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5474</pre>
5475
5476<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5477<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to a
5478 pointer type, <tt>ty2</tt>.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5479
5480<h5>Arguments:</h5>
Duncan Sands16f122e2007-03-30 12:22:09 +0000[diff] [blame]5481<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5482 value to cast, and a type to cast it to, which must be a
5483 <a href="#t_pointer">pointer</a> type.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5484
5485<h5>Semantics:</h5>
5486<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5487 <tt>ty2</tt> by applying either a zero extension or a truncation depending on
5488 the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
5489 size of a pointer then a truncation is done. If <tt>value</tt> is smaller
5490 than the size of a pointer then a zero extension is done. If they are the
5491 same size, nothing is done (<i>no-op cast</i>).</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5492
5493<h5>Example:</h5>
5494<pre>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5495 %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]5496 %Y = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
5497 %Z = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5498</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5499
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5500</div>
5501
5502<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5503<h4>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]5504 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5505</h4>
5506
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5507<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5508
5509<h5>Syntax:</h5>
5510<pre>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]5511 &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]5512</pre>
5513
5514<h5>Overview:</h5>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]5515<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5516 <tt>ty2</tt> without changing any bits.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5517
5518<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5519<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be a
5520 non-aggregate first class value, and a type to cast it to, which must also be
5521 a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes
5522 of <tt>value</tt> and the destination type, <tt>ty2</tt>, must be
5523 identical. If the source type is a pointer, the destination type must also be
5524 a pointer. This instruction supports bitwise conversion of vectors to
5525 integers and to vectors of other types (as long as they have the same
5526 size).</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5527
5528<h5>Semantics:</h5>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]5529<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5530 <tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
5531 this conversion. The conversion is done as if the <tt>value</tt> had been
5532 stored to memory and read back as type <tt>ty2</tt>. Pointer types may only
5533 be converted to other pointer types with this instruction. To convert
5534 pointers to other types, use the <a href="#i_inttoptr">inttoptr</a> or
5535 <a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5536
5537<h5>Example:</h5>
5538<pre>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5539 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5540 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5541 %Z = bitcast &lt;2 x int&gt; %V to i64; <i>; yields i64: %V</i>
Chris Lattner70de6632001-07-09 00:26:23 +0000[diff] [blame]5542</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5543
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5544</div>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5545
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5546</div>
5547
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5548<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5549<h3>
5550 <a name="otherops">Other Operations</a>
5551</h3>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5552
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5553<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5554
5555<p>The instructions in this category are the "miscellaneous" instructions, which
5556 defy better classification.</p>
5557
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5558<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5559<h4>
5560 <a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
5561</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5562
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5563<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5564
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5565<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5566<pre>
5567 &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]5568</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5569
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5570<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5571<p>The '<tt>icmp</tt>' instruction returns a boolean value or a vector of
5572 boolean values based on comparison of its two integer, integer vector, or
5573 pointer operands.</p>
5574
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5575<h5>Arguments:</h5>
5576<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5577 the condition code indicating the kind of comparison to perform. It is not a
5578 value, just a keyword. The possible condition code are:</p>
5579
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5580<ol>
5581 <li><tt>eq</tt>: equal</li>
5582 <li><tt>ne</tt>: not equal </li>
5583 <li><tt>ugt</tt>: unsigned greater than</li>
5584 <li><tt>uge</tt>: unsigned greater or equal</li>
5585 <li><tt>ult</tt>: unsigned less than</li>
5586 <li><tt>ule</tt>: unsigned less or equal</li>
5587 <li><tt>sgt</tt>: signed greater than</li>
5588 <li><tt>sge</tt>: signed greater or equal</li>
5589 <li><tt>slt</tt>: signed less than</li>
5590 <li><tt>sle</tt>: signed less or equal</li>
5591</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5592
Chris Lattnerc0f423a2007-01-15 01:54:13 +0000[diff] [blame]5593<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5594 <a href="#t_pointer">pointer</a> or integer <a href="#t_vector">vector</a>
5595 typed. They must also be identical types.</p>
5596
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5597<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5598<p>The '<tt>icmp</tt>' compares <tt>op1</tt> and <tt>op2</tt> according to the
5599 condition code given as <tt>cond</tt>. The comparison performed always yields
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]5600 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]5601 result, as follows:</p>
5602
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5603<ol>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5604 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5605 <tt>false</tt> otherwise. No sign interpretation is necessary or
5606 performed.</li>
5607
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5608 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5609 <tt>false</tt> otherwise. No sign interpretation is necessary or
5610 performed.</li>
5611
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5612 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5613 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5614
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5615 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5616 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5617 to <tt>op2</tt>.</li>
5618
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5619 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5620 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5621
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5622 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5623 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5624
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5625 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5626 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5627
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5628 <li><tt>sge</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5629 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5630 to <tt>op2</tt>.</li>
5631
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5632 <li><tt>slt</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5633 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5634
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5635 <li><tt>sle</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5636 <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]5637</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5638
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5639<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5640 values are compared as if they were integers.</p>
5641
5642<p>If the operands are integer vectors, then they are compared element by
5643 element. The result is an <tt>i1</tt> vector with the same number of elements
5644 as the values being compared. Otherwise, the result is an <tt>i1</tt>.</p>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5645
5646<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5647<pre>
5648 &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5649 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
5650 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
5651 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
5652 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
5653 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5654</pre>
Dan Gohmana5127ab2009-01-22 01:39:38 +0000[diff] [blame]5655
5656<p>Note that the code generator does not yet support vector types with
5657 the <tt>icmp</tt> instruction.</p>
5658
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5659</div>
5660
5661<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5662<h4>
5663 <a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
5664</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5665
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5666<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5667
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5668<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5669<pre>
5670 &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]5671</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5672
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5673<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5674<p>The '<tt>fcmp</tt>' instruction returns a boolean value or vector of boolean
5675 values based on comparison of its operands.</p>
5676
5677<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]5678(<a href="#t_integer"><tt>i1</tt></a>).</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5679
5680<p>If the operands are floating point vectors, then the result type is a vector
5681 of boolean with the same number of elements as the operands being
5682 compared.</p>
5683
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5684<h5>Arguments:</h5>
5685<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5686 the condition code indicating the kind of comparison to perform. It is not a
5687 value, just a keyword. The possible condition code are:</p>
5688
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5689<ol>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5690 <li><tt>false</tt>: no comparison, always returns false</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5691 <li><tt>oeq</tt>: ordered and equal</li>
5692 <li><tt>ogt</tt>: ordered and greater than </li>
5693 <li><tt>oge</tt>: ordered and greater than or equal</li>
5694 <li><tt>olt</tt>: ordered and less than </li>
5695 <li><tt>ole</tt>: ordered and less than or equal</li>
5696 <li><tt>one</tt>: ordered and not equal</li>
5697 <li><tt>ord</tt>: ordered (no nans)</li>
5698 <li><tt>ueq</tt>: unordered or equal</li>
5699 <li><tt>ugt</tt>: unordered or greater than </li>
5700 <li><tt>uge</tt>: unordered or greater than or equal</li>
5701 <li><tt>ult</tt>: unordered or less than </li>
5702 <li><tt>ule</tt>: unordered or less than or equal</li>
5703 <li><tt>une</tt>: unordered or not equal</li>
5704 <li><tt>uno</tt>: unordered (either nans)</li>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5705 <li><tt>true</tt>: no comparison, always returns true</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5706</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5707
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5708<p><i>Ordered</i> means that neither operand is a QNAN while
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5709 <i>unordered</i> means that either operand may be a QNAN.</p>
5710
5711<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be either
5712 a <a href="#t_floating">floating point</a> type or
5713 a <a href="#t_vector">vector</a> of floating point type. They must have
5714 identical types.</p>
5715
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5716<h5>Semantics:</h5>
Gabor Greif0f75ad02008-08-07 21:46:00 +0000[diff] [blame]5717<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5718 according to the condition code given as <tt>cond</tt>. If the operands are
5719 vectors, then the vectors are compared element by element. Each comparison
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]5720 performed always yields an <a href="#t_integer">i1</a> result, as
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5721 follows:</p>
5722
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5723<ol>
5724 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5725
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5726 <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]5727 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5728
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5729 <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]5730 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5731
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5732 <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]5733 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5734
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5735 <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]5736 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5737
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5738 <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]5739 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5740
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5741 <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]5742 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5743
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5744 <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]5745
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5746 <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]5747 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5748
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5749 <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]5750 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5751
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5752 <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]5753 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5754
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5755 <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]5756 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5757
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5758 <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]5759 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5760
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5761 <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]5762 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5763
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5764 <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]5765
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5766 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
5767</ol>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5768
5769<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5770<pre>
5771 &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]5772 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
5773 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
5774 &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]5775</pre>
Dan Gohmana5127ab2009-01-22 01:39:38 +0000[diff] [blame]5776
5777<p>Note that the code generator does not yet support vector types with
5778 the <tt>fcmp</tt> instruction.</p>
5779
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5780</div>
5781
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5782<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5783<h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5784 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5785</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5786
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5787<div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5788
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5789<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5790<pre>
5791 &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...
5792</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5793
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5794<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5795<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in the
5796 SSA graph representing the function.</p>
5797
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5798<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5799<p>The type of the incoming values is specified with the first type field. After
5800 this, the '<tt>phi</tt>' instruction takes a list of pairs as arguments, with
5801 one pair for each predecessor basic block of the current block. Only values
5802 of <a href="#t_firstclass">first class</a> type may be used as the value
5803 arguments to the PHI node. Only labels may be used as the label
5804 arguments.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5805
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5806<p>There must be no non-phi instructions between the start of a basic block and
5807 the PHI instructions: i.e. PHI instructions must be first in a basic
5808 block.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5809
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5810<p>For the purposes of the SSA form, the use of each incoming value is deemed to
5811 occur on the edge from the corresponding predecessor block to the current
5812 block (but after any definition of an '<tt>invoke</tt>' instruction's return
5813 value on the same edge).</p>
Jay Foad1a4eea52009-06-03 10:20:10 +0000[diff] [blame]5814
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5815<h5>Semantics:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5816<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5817 specified by the pair corresponding to the predecessor basic block that
5818 executed just prior to the current block.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5819
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5820<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5821<pre>
5822Loop: ; Infinite loop that counts from 0 on up...
5823 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
5824 %nextindvar = add i32 %indvar, 1
5825 br label %Loop
5826</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5827
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5828</div>
5829
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5830<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5831<h4>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5832 <a name="i_select">'<tt>select</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5833</h4>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5834
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5835<div>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5836
5837<h5>Syntax:</h5>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5838<pre>
Dan Gohmanc579d972008-09-09 01:02:47 +0000[diff] [blame]5839 &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>
5840
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]5841 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5842</pre>
5843
5844<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5845<p>The '<tt>select</tt>' instruction is used to choose one value based on a
5846 condition, without branching.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5847
5848
5849<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5850<p>The '<tt>select</tt>' instruction requires an 'i1' value or a vector of 'i1'
5851 values indicating the condition, and two values of the
5852 same <a href="#t_firstclass">first class</a> type. If the val1/val2 are
5853 vectors and the condition is a scalar, then entire vectors are selected, not
5854 individual elements.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5855
5856<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5857<p>If the condition is an i1 and it evaluates to 1, the instruction returns the
5858 first value argument; otherwise, it returns the second value argument.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5859
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5860<p>If the condition is a vector of i1, then the value arguments must be vectors
5861 of the same size, and the selection is done element by element.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5862
5863<h5>Example:</h5>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5864<pre>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]5865 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5866</pre>
Dan Gohmana5127ab2009-01-22 01:39:38 +0000[diff] [blame]5867
5868<p>Note that the code generator does not yet support conditions
5869 with vector type.</p>
5870
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5871</div>
5872
Robert Bocchinof72fdfe2006-01-15 20:48:27 +0000[diff] [blame]5873<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5874<h4>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5875 <a name="i_call">'<tt>call</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5876</h4>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5877
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5878<div>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5879
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5880<h5>Syntax:</h5>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5881<pre>
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]5882 &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]5883</pre>
5884
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5885<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5886<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5887
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5888<h5>Arguments:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5889<p>This instruction requires several arguments:</p>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5890
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]5891<ol>
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]5892 <li>The optional "tail" marker indicates that the callee function does not
5893 access any allocas or varargs in the caller. Note that calls may be
5894 marked "tail" even if they do not occur before
5895 a <a href="#i_ret"><tt>ret</tt></a> instruction. If the "tail" marker is
5896 present, the function call is eligible for tail call optimization,
5897 but <a href="CodeGenerator.html#tailcallopt">might not in fact be
Evan Cheng59676492010-03-08 21:05:02 +0000[diff] [blame]5898 optimized into a jump</a>. The code generator may optimize calls marked
5899 "tail" with either 1) automatic <a href="CodeGenerator.html#sibcallopt">
5900 sibling call optimization</a> when the caller and callee have
5901 matching signatures, or 2) forced tail call optimization when the
5902 following extra requirements are met:
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]5903 <ul>
5904 <li>Caller and callee both have the calling
5905 convention <tt>fastcc</tt>.</li>
5906 <li>The call is in tail position (ret immediately follows call and ret
5907 uses value of call or is void).</li>
5908 <li>Option <tt>-tailcallopt</tt> is enabled,
Dan Gohman6232f732010-03-02 01:08:11 +0000[diff] [blame]5909 or <code>llvm::GuaranteedTailCallOpt</code> is <code>true</code>.</li>
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]5910 <li><a href="CodeGenerator.html#tailcallopt">Platform specific
5911 constraints are met.</a></li>
5912 </ul>
5913 </li>
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]5914
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5915 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
5916 convention</a> the call should use. If none is specified, the call
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]5917 defaults to using C calling conventions. The calling convention of the
5918 call must match the calling convention of the target function, or else the
5919 behavior is undefined.</li>
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]5920
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5921 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
5922 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
5923 '<tt>inreg</tt>' attributes are valid here.</li>
5924
5925 <li>'<tt>ty</tt>': the type of the call instruction itself which is also the
5926 type of the return value. Functions that return no value are marked
5927 <tt><a href="#t_void">void</a></tt>.</li>
5928
5929 <li>'<tt>fnty</tt>': shall be the signature of the pointer to function value
5930 being invoked. The argument types must match the types implied by this
5931 signature. This type can be omitted if the function is not varargs and if
5932 the function type does not return a pointer to a function.</li>
5933
5934 <li>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
5935 be invoked. In most cases, this is a direct function invocation, but
5936 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
5937 to function value.</li>
5938
5939 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]5940 signature argument types and parameter attributes. All arguments must be
5941 of <a href="#t_firstclass">first class</a> type. If the function
5942 signature indicates the function accepts a variable number of arguments,
5943 the extra arguments can be specified.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5944
5945 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
5946 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
5947 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]5948</ol>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5949
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5950<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5951<p>The '<tt>call</tt>' instruction is used to cause control flow to transfer to
5952 a specified function, with its incoming arguments bound to the specified
5953 values. Upon a '<tt><a href="#i_ret">ret</a></tt>' instruction in the called
5954 function, control flow continues with the instruction after the function
5955 call, and the return value of the function is bound to the result
5956 argument.</p>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5957
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5958<h5>Example:</h5>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5959<pre>
Nick Lewyckya9b13d52007-09-08 13:57:50 +0000[diff] [blame]5960 %retval = call i32 @test(i32 %argc)
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]5961 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) <i>; yields i32</i>
Chris Lattnerfb7c88d2008-03-21 17:24:17 +0000[diff] [blame]5962 %X = tail call i32 @foo() <i>; yields i32</i>
5963 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
5964 call void %foo(i8 97 signext)
Devang Pateld6cff512008-03-10 20:49:15 +0000[diff] [blame]5965
5966 %struct.A = type { i32, i8 }
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]5967 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohmancc3132e2008-10-04 19:00:07 +0000[diff] [blame]5968 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
5969 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattner6cbe8e92008-10-08 06:26:11 +0000[diff] [blame]5970 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijmaneefa7df2008-10-07 10:03:45 +0000[diff] [blame]5971 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5972</pre>
5973
Dale Johannesen68f971b2009-09-24 18:38:21 +0000[diff] [blame]5974<p>llvm treats calls to some functions with names and arguments that match the
Dale Johannesen722212d2009-09-25 17:04:42 +0000[diff] [blame]5975standard C99 library as being the C99 library functions, and may perform
5976optimizations or generate code for them under that assumption. This is
5977something we'd like to change in the future to provide better support for
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]5978freestanding environments and non-C-based languages.</p>
Dale Johannesen68f971b2009-09-24 18:38:21 +0000[diff] [blame]5979
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5980</div>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5981
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5982<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5983<h4>
Chris Lattner33337472006-01-13 23:26:01 +0000[diff] [blame]5984 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5985</h4>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5986
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5987<div>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5988
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]5989<h5>Syntax:</h5>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5990<pre>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]5991 &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]5992</pre>
5993
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]5994<h5>Overview:</h5>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]5995<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5996 the "variable argument" area of a function call. It is used to implement the
5997 <tt>va_arg</tt> macro in C.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5998
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]5999<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6000<p>This instruction takes a <tt>va_list*</tt> value and the type of the
6001 argument. It returns a value of the specified argument type and increments
6002 the <tt>va_list</tt> to point to the next argument. The actual type
6003 of <tt>va_list</tt> is target specific.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]6004
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]6005<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6006<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified type
6007 from the specified <tt>va_list</tt> and causes the <tt>va_list</tt> to point
6008 to the next argument. For more information, see the variable argument
6009 handling <a href="#int_varargs">Intrinsic Functions</a>.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]6010
6011<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]6012 take a variable number of arguments, for example, the <tt>vfprintf</tt>
6013 function.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]6014
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6015<p><tt>va_arg</tt> is an LLVM instruction instead of
6016 an <a href="#intrinsics">intrinsic function</a> because it takes a type as an
6017 argument.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]6018
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]6019<h5>Example:</h5>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]6020<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
6021
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6022<p>Note that the code generator does not yet fully support va_arg on many
6023 targets. Also, it does not currently support va_arg with aggregate types on
6024 any target.</p>
Dan Gohman3065b612009-01-12 23:12:39 +0000[diff] [blame]6025
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]6026</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6027
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]6028<!-- _______________________________________________________________________ -->
6029<h4>
6030 <a name="i_landingpad">'<tt>landingpad</tt>' Instruction</a>
6031</h4>
6032
6033<div>
6034
6035<h5>Syntax:</h5>
6036<pre>
Bill Wendling49bfb122011-08-08 08:06:05 +0000[diff] [blame]6037 &lt;resultval&gt; = landingpad &lt;somety&gt; personality &lt;type&gt; &lt;pers_fn&gt; &lt;clause&gt;+
6038 &lt;resultval&gt; = landingpad &lt;somety&gt; personality &lt;type&gt; &lt;pers_fn&gt; cleanup &lt;clause&gt;*
6039
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]6040 &lt;clause&gt; := catch &lt;type&gt; &lt;value&gt;
Bill Wendlingfae14752011-08-12 20:24:12 +0000[diff] [blame]6041 &lt;clause&gt; := filter &lt;array constant type&gt; &lt;array constant&gt;
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]6042</pre>
6043
6044<h5>Overview:</h5>
6045<p>The '<tt>landingpad</tt>' instruction is used by
6046 <a href="ExceptionHandling.html#overview">LLVM's exception handling
6047 system</a> to specify that a basic block is a landing pad &mdash; one where
6048 the exception lands, and corresponds to the code found in the
6049 <i><tt>catch</tt></i> portion of a <i><tt>try/catch</tt></i> sequence. It
6050 defines values supplied by the personality function (<tt>pers_fn</tt>) upon
6051 re-entry to the function. The <tt>resultval</tt> has the
6052 type <tt>somety</tt>.</p>
6053
6054<h5>Arguments:</h5>
6055<p>This instruction takes a <tt>pers_fn</tt> value. This is the personality
6056 function associated with the unwinding mechanism. The optional
6057 <tt>cleanup</tt> flag indicates that the landing pad block is a cleanup.</p>
6058
6059<p>A <tt>clause</tt> begins with the clause type &mdash; <tt>catch</tt>
Bill Wendlingfae14752011-08-12 20:24:12 +0000[diff] [blame]6060 or <tt>filter</tt> &mdash; and contains the global variable representing the
6061 "type" that may be caught or filtered respectively. Unlike the
6062 <tt>catch</tt> clause, the <tt>filter</tt> clause takes an array constant as
6063 its argument. Use "<tt>[0 x i8**] undef</tt>" for a filter which cannot
6064 throw. The '<tt>landingpad</tt>' instruction must contain <em>at least</em>
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]6065 one <tt>clause</tt> or the <tt>cleanup</tt> flag.</p>
6066
6067<h5>Semantics:</h5>
6068<p>The '<tt>landingpad</tt>' instruction defines the values which are set by the
6069 personality function (<tt>pers_fn</tt>) upon re-entry to the function, and
6070 therefore the "result type" of the <tt>landingpad</tt> instruction. As with
6071 calling conventions, how the personality function results are represented in
6072 LLVM IR is target specific.</p>
6073
Bill Wendling0524b8d2011-08-03 17:17:06 +0000[diff] [blame]6074<p>The clauses are applied in order from top to bottom. If two
6075 <tt>landingpad</tt> instructions are merged together through inlining, the
Bill Wendlinga503fc02011-08-08 07:58:58 +0000[diff] [blame]6076 clauses from the calling function are appended to the list of clauses.</p>
Bill Wendling0524b8d2011-08-03 17:17:06 +0000[diff] [blame]6077
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]6078<p>The <tt>landingpad</tt> instruction has several restrictions:</p>
6079
6080<ul>
6081 <li>A landing pad block is a basic block which is the unwind destination of an
6082 '<tt>invoke</tt>' instruction.</li>
6083 <li>A landing pad block must have a '<tt>landingpad</tt>' instruction as its
6084 first non-PHI instruction.</li>
6085 <li>There can be only one '<tt>landingpad</tt>' instruction within the landing
6086 pad block.</li>
6087 <li>A basic block that is not a landing pad block may not include a
6088 '<tt>landingpad</tt>' instruction.</li>
6089 <li>All '<tt>landingpad</tt>' instructions in a function must have the same
6090 personality function.</li>
6091</ul>
6092
6093<h5>Example:</h5>
6094<pre>
6095 ;; A landing pad which can catch an integer.
6096 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6097 catch i8** @_ZTIi
6098 ;; A landing pad that is a cleanup.
6099 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
Bill Wendlingfae14752011-08-12 20:24:12 +0000[diff] [blame]6100 cleanup
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]6101 ;; A landing pad which can catch an integer and can only throw a double.
6102 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6103 catch i8** @_ZTIi
Bill Wendlingfae14752011-08-12 20:24:12 +0000[diff] [blame]6104 filter [1 x i8**] [@_ZTId]
Bill Wendlingbbcb7cd2011-08-02 21:52:38 +0000[diff] [blame]6105</pre>
6106
6107</div>
6108
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6109</div>
6110
6111</div>
6112
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6113<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6114<h2><a name="intrinsics">Intrinsic Functions</a></h2>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]6115<!-- *********************************************************************** -->
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6116
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6117<div>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6118
6119<p>LLVM supports the notion of an "intrinsic function". These functions have
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6120 well known names and semantics and are required to follow certain
6121 restrictions. Overall, these intrinsics represent an extension mechanism for
6122 the LLVM language that does not require changing all of the transformations
6123 in LLVM when adding to the language (or the bitcode reader/writer, the
6124 parser, etc...).</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6125
John Criswell88190562005-05-16 16:17:45 +0000[diff] [blame]6126<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6127 prefix is reserved in LLVM for intrinsic names; thus, function names may not
6128 begin with this prefix. Intrinsic functions must always be external
6129 functions: you cannot define the body of intrinsic functions. Intrinsic
6130 functions may only be used in call or invoke instructions: it is illegal to
6131 take the address of an intrinsic function. Additionally, because intrinsic
6132 functions are part of the LLVM language, it is required if any are added that
6133 they be documented here.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6134
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6135<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents a
6136 family of functions that perform the same operation but on different data
6137 types. Because LLVM can represent over 8 million different integer types,
6138 overloading is used commonly to allow an intrinsic function to operate on any
6139 integer type. One or more of the argument types or the result type can be
6140 overloaded to accept any integer type. Argument types may also be defined as
6141 exactly matching a previous argument's type or the result type. This allows
6142 an intrinsic function which accepts multiple arguments, but needs all of them
6143 to be of the same type, to only be overloaded with respect to a single
6144 argument or the result.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6145
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6146<p>Overloaded intrinsics will have the names of its overloaded argument types
6147 encoded into its function name, each preceded by a period. Only those types
6148 which are overloaded result in a name suffix. Arguments whose type is matched
6149 against another type do not. For example, the <tt>llvm.ctpop</tt> function
6150 can take an integer of any width and returns an integer of exactly the same
6151 integer width. This leads to a family of functions such as
6152 <tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29
6153 %val)</tt>. Only one type, the return type, is overloaded, and only one type
6154 suffix is required. Because the argument's type is matched against the return
6155 type, it does not require its own name suffix.</p>
Reid Spencer4eefaab2007-04-01 08:04:23 +0000[diff] [blame]6156
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]6157<p>To learn how to add an intrinsic function, please see the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6158 <a href="ExtendingLLVM.html">Extending LLVM Guide</a>.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6159
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6160<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6161<h3>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6162 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6163</h3>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6164
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6165<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6166
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6167<p>Variable argument support is defined in LLVM with
6168 the <a href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
6169 intrinsic functions. These functions are related to the similarly named
6170 macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6171
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6172<p>All of these functions operate on arguments that use a target-specific value
6173 type "<tt>va_list</tt>". The LLVM assembly language reference manual does
6174 not define what this type is, so all transformations should be prepared to
6175 handle these functions regardless of the type used.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6176
Chris Lattner30b868d2006-05-15 17:26:46 +0000[diff] [blame]6177<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]6178 instruction and the variable argument handling intrinsic functions are
6179 used.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6180
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]6181<pre class="doc_code">
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]6182define i32 @test(i32 %X, ...) {
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6183 ; Initialize variable argument processing
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]6184 %ap = alloca i8*
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]6185 %ap2 = bitcast i8** %ap to i8*
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]6186 call void @llvm.va_start(i8* %ap2)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6187
6188 ; Read a single integer argument
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]6189 %tmp = va_arg i8** %ap, i32
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6190
6191 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]6192 %aq = alloca i8*
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]6193 %aq2 = bitcast i8** %aq to i8*
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]6194 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]6195 call void @llvm.va_end(i8* %aq2)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6196
6197 ; Stop processing of arguments.
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]6198 call void @llvm.va_end(i8* %ap2)
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]6199 ret i32 %tmp
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6200}
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]6201
6202declare void @llvm.va_start(i8*)
6203declare void @llvm.va_copy(i8*, i8*)
6204declare void @llvm.va_end(i8*)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6205</pre>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6206
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6207<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6208<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6209 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6210</h4>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6211
6212
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6213<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6214
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6215<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6216<pre>
6217 declare void %llvm.va_start(i8* &lt;arglist&gt;)
6218</pre>
6219
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6220<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6221<p>The '<tt>llvm.va_start</tt>' intrinsic initializes <tt>*&lt;arglist&gt;</tt>
6222 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]6223
6224<h5>Arguments:</h5>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]6225<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]6226
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6227<h5>Semantics:</h5>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]6228<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]6229 macro available in C. In a target-dependent way, it initializes
6230 the <tt>va_list</tt> element to which the argument points, so that the next
6231 call to <tt>va_arg</tt> will produce the first variable argument passed to
6232 the function. Unlike the C <tt>va_start</tt> macro, this intrinsic does not
6233 need to know the last argument of the function as the compiler can figure
6234 that out.</p>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]6235
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]6236</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6237
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6238<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6239<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6240 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6241</h4>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6242
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6243<div>
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]6244
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6245<h5>Syntax:</h5>
6246<pre>
6247 declare void @llvm.va_end(i8* &lt;arglist&gt;)
6248</pre>
6249
6250<h5>Overview:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]6251<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6252 which has been initialized previously
6253 with <tt><a href="#int_va_start">llvm.va_start</a></tt>
6254 or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]6255
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6256<h5>Arguments:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]6257<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]6258
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6259<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]6260<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]6261 macro available in C. In a target-dependent way, it destroys
6262 the <tt>va_list</tt> element to which the argument points. Calls
6263 to <a href="#int_va_start"><tt>llvm.va_start</tt></a>
6264 and <a href="#int_va_copy"> <tt>llvm.va_copy</tt></a> must be matched exactly
6265 with calls to <tt>llvm.va_end</tt>.</p>
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]6266
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]6267</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6268
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6269<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6270<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6271 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6272</h4>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6273
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6274<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6275
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6276<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6277<pre>
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]6278 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6279</pre>
6280
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6281<h5>Overview:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]6282<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6283 from the source argument list to the destination argument list.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6284
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6285<h5>Arguments:</h5>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]6286<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]6287 The second argument is a pointer to a <tt>va_list</tt> element to copy
6288 from.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6289
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]6290<h5>Semantics:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]6291<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]6292 macro available in C. In a target-dependent way, it copies the
6293 source <tt>va_list</tt> element into the destination <tt>va_list</tt>
6294 element. This intrinsic is necessary because
6295 the <tt><a href="#int_va_start"> llvm.va_start</a></tt> intrinsic may be
6296 arbitrarily complex and require, for example, memory allocation.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6297
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]6298</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6299
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6300</div>
6301
Bill Wendling537603b2011-07-31 06:45:03 +0000[diff] [blame]6302</div>
6303
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6304<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6305<h3>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6306 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6307</h3>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6308
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6309<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6310
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6311<p>LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]6312Collection</a> (GC) requires the implementation and generation of these
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6313intrinsics. These intrinsics allow identification of <a href="#int_gcroot">GC
6314roots on the stack</a>, as well as garbage collector implementations that
6315require <a href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a>
6316barriers. Front-ends for type-safe garbage collected languages should generate
6317these intrinsics to make use of the LLVM garbage collectors. For more details,
6318see <a href="GarbageCollection.html">Accurate Garbage Collection with
6319LLVM</a>.</p>
Christopher Lamb55c6d4f2007-12-17 01:00:21 +0000[diff] [blame]6320
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6321<p>The garbage collection intrinsics only operate on objects in the generic
6322 address space (address space zero).</p>
Christopher Lamb55c6d4f2007-12-17 01:00:21 +0000[diff] [blame]6323
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6324<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6325<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6326 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6327</h4>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6328
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6329<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6330
6331<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6332<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]6333 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6334</pre>
6335
6336<h5>Overview:</h5>
John Criswelldfe6a862004-12-10 15:51:16 +0000[diff] [blame]6337<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]6338 the code generator, and allows some metadata to be associated with it.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6339
6340<h5>Arguments:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6341<p>The first argument specifies the address of a stack object that contains the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6342 root pointer. The second pointer (which must be either a constant or a
6343 global value address) contains the meta-data to be associated with the
6344 root.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6345
6346<h5>Semantics:</h5>
Chris Lattner851b7712008-04-24 05:59:56 +0000[diff] [blame]6347<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]6348 location. At compile-time, the code generator generates information to allow
6349 the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
6350 intrinsic may only be used in a function which <a href="#gc">specifies a GC
6351 algorithm</a>.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6352
6353</div>
6354
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6355<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6356<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6357 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6358</h4>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6359
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6360<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6361
6362<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6363<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]6364 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6365</pre>
6366
6367<h5>Overview:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6368<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6369 locations, allowing garbage collector implementations that require read
6370 barriers.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6371
6372<h5>Arguments:</h5>
Chris Lattnerf9228072006-03-14 20:02:51 +0000[diff] [blame]6373<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]6374 allocated from the garbage collector. The first object is a pointer to the
6375 start of the referenced object, if needed by the language runtime (otherwise
6376 null).</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6377
6378<h5>Semantics:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6379<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6380 instruction, but may be replaced with substantially more complex code by the
6381 garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
6382 may only be used in a function which <a href="#gc">specifies a GC
6383 algorithm</a>.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6384
6385</div>
6386
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6387<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6388<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6389 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6390</h4>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6391
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6392<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6393
6394<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6395<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]6396 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6397</pre>
6398
6399<h5>Overview:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6400<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6401 locations, allowing garbage collector implementations that require write
6402 barriers (such as generational or reference counting collectors).</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6403
6404<h5>Arguments:</h5>
Chris Lattnerf9228072006-03-14 20:02:51 +0000[diff] [blame]6405<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]6406 object to store it to, and the third is the address of the field of Obj to
6407 store to. If the runtime does not require a pointer to the object, Obj may
6408 be null.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6409
6410<h5>Semantics:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6411<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6412 instruction, but may be replaced with substantially more complex code by the
6413 garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
6414 may only be used in a function which <a href="#gc">specifies a GC
6415 algorithm</a>.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6416
6417</div>
6418
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6419</div>
6420
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]6421<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6422<h3>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6423 <a name="int_codegen">Code Generator Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6424</h3>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6425
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6426<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6427
6428<p>These intrinsics are provided by LLVM to expose special features that may
6429 only be implemented with code generator support.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6430
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6431<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6432<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6433 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6434</h4>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6435
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6436<div>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6437
6438<h5>Syntax:</h5>
6439<pre>
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]6440 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6441</pre>
6442
6443<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6444<p>The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
6445 target-specific value indicating the return address of the current function
6446 or one of its callers.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6447
6448<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6449<p>The argument to this intrinsic indicates which function to return the address
6450 for. Zero indicates the calling function, one indicates its caller, etc.
6451 The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6452
6453<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6454<p>The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer
6455 indicating the return address of the specified call frame, or zero if it
6456 cannot be identified. The value returned by this intrinsic is likely to be
6457 incorrect or 0 for arguments other than zero, so it should only be used for
6458 debugging purposes.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6459
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6460<p>Note that calling this intrinsic does not prevent function inlining or other
6461 aggressive transformations, so the value returned may not be that of the
6462 obvious source-language caller.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6463
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6464</div>
6465
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6466<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6467<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6468 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6469</h4>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6470
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6471<div>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6472
6473<h5>Syntax:</h5>
6474<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6475 declare i8* @llvm.frameaddress(i32 &lt;level&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6476</pre>
6477
6478<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6479<p>The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
6480 target-specific frame pointer value for the specified stack frame.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6481
6482<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6483<p>The argument to this intrinsic indicates which function to return the frame
6484 pointer for. Zero indicates the calling function, one indicates its caller,
6485 etc. The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6486
6487<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6488<p>The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer
6489 indicating the frame address of the specified call frame, or zero if it
6490 cannot be identified. The value returned by this intrinsic is likely to be
6491 incorrect or 0 for arguments other than zero, so it should only be used for
6492 debugging purposes.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6493
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6494<p>Note that calling this intrinsic does not prevent function inlining or other
6495 aggressive transformations, so the value returned may not be that of the
6496 obvious source-language caller.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6497
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6498</div>
6499
Chris Lattnerc8a2c222005-02-28 19:24:19 +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_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6503</h4>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6504
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6505<div>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6506
6507<h5>Syntax:</h5>
6508<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6509 declare i8* @llvm.stacksave()
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6510</pre>
6511
6512<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6513<p>The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state
6514 of the function stack, for use
6515 with <a href="#int_stackrestore"> <tt>llvm.stackrestore</tt></a>. This is
6516 useful for implementing language features like scoped automatic variable
6517 sized arrays in C99.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6518
6519<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6520<p>This intrinsic returns a opaque pointer value that can be passed
6521 to <a href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When
6522 an <tt>llvm.stackrestore</tt> intrinsic is executed with a value saved
6523 from <tt>llvm.stacksave</tt>, it effectively restores the state of the stack
6524 to the state it was in when the <tt>llvm.stacksave</tt> intrinsic executed.
6525 In practice, this pops any <a href="#i_alloca">alloca</a> blocks from the
6526 stack that were allocated after the <tt>llvm.stacksave</tt> was executed.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6527
6528</div>
6529
6530<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6531<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6532 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6533</h4>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6534
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6535<div>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6536
6537<h5>Syntax:</h5>
6538<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6539 declare void @llvm.stackrestore(i8* %ptr)
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6540</pre>
6541
6542<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6543<p>The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
6544 the function stack to the state it was in when the
6545 corresponding <a href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic
6546 executed. This is useful for implementing language features like scoped
6547 automatic variable sized arrays in C99.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6548
6549<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6550<p>See the description
6551 for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6552
6553</div>
6554
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6555<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6556<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6557 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6558</h4>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6559
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6560<div>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6561
6562<h5>Syntax:</h5>
6563<pre>
Bruno Cardoso Lopesdc9ff3a2011-06-14 04:58:37 +0000[diff] [blame]6564 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]6565</pre>
6566
6567<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6568<p>The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to
6569 insert a prefetch instruction if supported; otherwise, it is a noop.
6570 Prefetches have no effect on the behavior of the program but can change its
6571 performance characteristics.</p>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6572
6573<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6574<p><tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the
6575 specifier determining if the fetch should be for a read (0) or write (1),
6576 and <tt>locality</tt> is a temporal locality specifier ranging from (0) - no
Bruno Cardoso Lopesdc9ff3a2011-06-14 04:58:37 +0000[diff] [blame]6577 locality, to (3) - extremely local keep in cache. The <tt>cache type</tt>
6578 specifies whether the prefetch is performed on the data (1) or instruction (0)
6579 cache. The <tt>rw</tt>, <tt>locality</tt> and <tt>cache type</tt> arguments
6580 must be constant integers.</p>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6581
6582<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6583<p>This intrinsic does not modify the behavior of the program. In particular,
6584 prefetches cannot trap and do not produce a value. On targets that support
6585 this intrinsic, the prefetch can provide hints to the processor cache for
6586 better performance.</p>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6587
6588</div>
6589
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6590<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6591<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6592 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6593</h4>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6594
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6595<div>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6596
6597<h5>Syntax:</h5>
6598<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]6599 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6600</pre>
6601
6602<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6603<p>The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program
6604 Counter (PC) in a region of code to simulators and other tools. The method
6605 is target specific, but it is expected that the marker will use exported
6606 symbols to transmit the PC of the marker. The marker makes no guarantees
6607 that it will remain with any specific instruction after optimizations. It is
6608 possible that the presence of a marker will inhibit optimizations. The
6609 intended use is to be inserted after optimizations to allow correlations of
6610 simulation runs.</p>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6611
6612<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6613<p><tt>id</tt> is a numerical id identifying the marker.</p>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6614
6615<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6616<p>This intrinsic does not modify the behavior of the program. Backends that do
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]6617 not support this intrinsic may ignore it.</p>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6618
6619</div>
6620
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6621<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6622<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6623 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6624</h4>
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6625
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6626<div>
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6627
6628<h5>Syntax:</h5>
6629<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]6630 declare i64 @llvm.readcyclecounter()
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6631</pre>
6632
6633<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6634<p>The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
6635 counter register (or similar low latency, high accuracy clocks) on those
6636 targets that support it. On X86, it should map to RDTSC. On Alpha, it
6637 should map to RPCC. As the backing counters overflow quickly (on the order
6638 of 9 seconds on alpha), this should only be used for small timings.</p>
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6639
6640<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6641<p>When directly supported, reading the cycle counter should not modify any
6642 memory. Implementations are allowed to either return a application specific
6643 value or a system wide value. On backends without support, this is lowered
6644 to a constant 0.</p>
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6645
6646</div>
6647
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6648</div>
6649
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6650<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6651<h3>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6652 <a name="int_libc">Standard C Library Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6653</h3>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6654
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6655<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6656
6657<p>LLVM provides intrinsics for a few important standard C library functions.
6658 These intrinsics allow source-language front-ends to pass information about
6659 the alignment of the pointer arguments to the code generator, providing
6660 opportunity for more efficient code generation.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6661
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6662<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6663<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6664 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6665</h4>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6666
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6667<div>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6668
6669<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6670<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]6671 integer bit width and for different address spaces. Not all targets support
6672 all bit widths however.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6673
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6674<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6675 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]6676 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6677 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]6678 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6679</pre>
6680
6681<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6682<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6683 source location to the destination location.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6684
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6685<p>Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6686 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6687 and the pointers can be in specified address spaces.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6688
6689<h5>Arguments:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6690
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6691<p>The first argument is a pointer to the destination, the second is a pointer
6692 to the source. The third argument is an integer argument specifying the
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6693 number of bytes to copy, the fourth argument is the alignment of the
6694 source and destination locations, and the fifth is a boolean indicating a
6695 volatile access.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6696
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]6697<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]6698 then the caller guarantees that both the source and destination pointers are
6699 aligned to that boundary.</p>
Chris Lattner4c67c482004-02-12 21:18:15 +0000[diff] [blame]6700
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]6701<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6702 <tt>llvm.memcpy</tt> call is a <a href="#volatile">volatile operation</a>.
6703 The detailed access behavior is not very cleanly specified and it is unwise
6704 to depend on it.</p>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6705
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6706<h5>Semantics:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6707
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6708<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6709 source location to the destination location, which are not allowed to
6710 overlap. It copies "len" bytes of memory over. If the argument is known to
6711 be aligned to some boundary, this can be specified as the fourth argument,
6712 otherwise it should be set to 0 or 1.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6713
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6714</div>
6715
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6716<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6717<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6718 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6719</h4>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6720
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6721<div>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6722
6723<h5>Syntax:</h5>
Chris Lattnerdd708342008-11-21 16:42:48 +0000[diff] [blame]6724<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]6725 width and for different address space. Not all targets support all bit
6726 widths however.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6727
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6728<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6729 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]6730 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6731 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]6732 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6733</pre>
6734
6735<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6736<p>The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the
6737 source location to the destination location. It is similar to the
6738 '<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to
6739 overlap.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6740
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6741<p>Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6742 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6743 and the pointers can be in specified address spaces.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6744
6745<h5>Arguments:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6746
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6747<p>The first argument is a pointer to the destination, the second is a pointer
6748 to the source. The third argument is an integer argument specifying the
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6749 number of bytes to copy, the fourth argument is the alignment of the
6750 source and destination locations, and the fifth is a boolean indicating a
6751 volatile access.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6752
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]6753<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]6754 then the caller guarantees that the source and destination pointers are
6755 aligned to that boundary.</p>
Chris Lattner4c67c482004-02-12 21:18:15 +0000[diff] [blame]6756
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]6757<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6758 <tt>llvm.memmove</tt> call is a <a href="#volatile">volatile operation</a>.
6759 The detailed access behavior is not very cleanly specified and it is unwise
6760 to depend on it.</p>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6761
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6762<h5>Semantics:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6763
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6764<p>The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the
6765 source location to the destination location, which may overlap. It copies
6766 "len" bytes of memory over. If the argument is known to be aligned to some
6767 boundary, this can be specified as the fourth argument, otherwise it should
6768 be set to 0 or 1.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6769
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6770</div>
6771
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6772<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6773<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6774 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6775</h4>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6776
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6777<div>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6778
6779<h5>Syntax:</h5>
Chris Lattnerdd708342008-11-21 16:42:48 +0000[diff] [blame]6780<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]6781 width and for different address spaces. However, not all targets support all
6782 bit widths.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6783
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6784<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6785 declare void @llvm.memset.p0i8.i32(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattner685db9d2010-04-08 00:54:34 +0000[diff] [blame]6786 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6787 declare void @llvm.memset.p0i8.i64(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattner685db9d2010-04-08 00:54:34 +0000[diff] [blame]6788 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6789</pre>
6790
6791<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6792<p>The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a
6793 particular byte value.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6794
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6795<p>Note that, unlike the standard libc function, the <tt>llvm.memset</tt>
John Criswellad05ae42010-07-30 16:30:28 +0000[diff] [blame]6796 intrinsic does not return a value and takes extra alignment/volatile
6797 arguments. Also, the destination can be in an arbitrary address space.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6798
6799<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6800<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]6801 byte value with which to fill it, the third argument is an integer argument
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6802 specifying the number of bytes to fill, and the fourth argument is the known
John Criswellad05ae42010-07-30 16:30:28 +0000[diff] [blame]6803 alignment of the destination location.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6804
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]6805<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]6806 then the caller guarantees that the destination pointer is aligned to that
6807 boundary.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6808
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]6809<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6810 <tt>llvm.memset</tt> call is a <a href="#volatile">volatile operation</a>.
6811 The detailed access behavior is not very cleanly specified and it is unwise
6812 to depend on it.</p>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6813
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6814<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6815<p>The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting
6816 at the destination location. If the argument is known to be aligned to some
6817 boundary, this can be specified as the fourth argument, otherwise it should
6818 be set to 0 or 1.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6819
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6820</div>
6821
Chris Lattner3b4f4372004-06-11 02:28:03 +0000[diff] [blame]6822<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6823<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6824 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6825</h4>
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6826
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6827<div>
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6828
6829<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6830<p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
6831 floating point or vector of floating point type. Not all targets support all
6832 types however.</p>
6833
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6834<pre>
Dale Johannesendd89d272007-10-02 17:47:38 +0000[diff] [blame]6835 declare float @llvm.sqrt.f32(float %Val)
6836 declare double @llvm.sqrt.f64(double %Val)
6837 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
6838 declare fp128 @llvm.sqrt.f128(fp128 %Val)
6839 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6840</pre>
6841
6842<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6843<p>The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
6844 returning the same value as the libm '<tt>sqrt</tt>' functions would.
6845 Unlike <tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined
6846 behavior for negative numbers other than -0.0 (which allows for better
6847 optimization, because there is no need to worry about errno being
6848 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]6849
6850<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6851<p>The argument and return value are floating point numbers of the same
6852 type.</p>
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6853
6854<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6855<p>This function returns the sqrt of the specified operand if it is a
6856 nonnegative floating point number.</p>
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6857
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6858</div>
6859
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6860<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6861<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6862 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6863</h4>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6864
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6865<div>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6866
6867<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6868<p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
6869 floating point or vector of floating point type. Not all targets support all
6870 types however.</p>
6871
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6872<pre>
Dale Johannesendd89d272007-10-02 17:47:38 +0000[diff] [blame]6873 declare float @llvm.powi.f32(float %Val, i32 %power)
6874 declare double @llvm.powi.f64(double %Val, i32 %power)
6875 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
6876 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
6877 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6878</pre>
6879
6880<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6881<p>The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
6882 specified (positive or negative) power. The order of evaluation of
6883 multiplications is not defined. When a vector of floating point type is
6884 used, the second argument remains a scalar integer value.</p>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6885
6886<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6887<p>The second argument is an integer power, and the first is a value to raise to
6888 that power.</p>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6889
6890<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6891<p>This function returns the first value raised to the second power with an
6892 unspecified sequence of rounding operations.</p>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6893
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6894</div>
6895
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6896<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6897<h4>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6898 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6899</h4>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6900
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6901<div>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6902
6903<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6904<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
6905 floating point or vector of floating point type. Not all targets support all
6906 types however.</p>
6907
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6908<pre>
6909 declare float @llvm.sin.f32(float %Val)
6910 declare double @llvm.sin.f64(double %Val)
6911 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
6912 declare fp128 @llvm.sin.f128(fp128 %Val)
6913 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
6914</pre>
6915
6916<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6917<p>The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6918
6919<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6920<p>The argument and return value are floating point numbers of the same
6921 type.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6922
6923<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6924<p>This function returns the sine of the specified operand, returning the same
6925 values as the libm <tt>sin</tt> functions would, and handles error conditions
6926 in the same way.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6927
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6928</div>
6929
6930<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6931<h4>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6932 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6933</h4>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6934
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6935<div>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6936
6937<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6938<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
6939 floating point or vector of floating point type. Not all targets support all
6940 types however.</p>
6941
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6942<pre>
6943 declare float @llvm.cos.f32(float %Val)
6944 declare double @llvm.cos.f64(double %Val)
6945 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
6946 declare fp128 @llvm.cos.f128(fp128 %Val)
6947 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
6948</pre>
6949
6950<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6951<p>The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6952
6953<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6954<p>The argument and return value are floating point numbers of the same
6955 type.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6956
6957<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6958<p>This function returns the cosine of the specified operand, returning the same
6959 values as the libm <tt>cos</tt> functions would, and handles error conditions
6960 in the same way.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6961
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6962</div>
6963
6964<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6965<h4>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6966 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6967</h4>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6968
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6969<div>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6970
6971<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6972<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
6973 floating point or vector of floating point type. Not all targets support all
6974 types however.</p>
6975
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6976<pre>
6977 declare float @llvm.pow.f32(float %Val, float %Power)
6978 declare double @llvm.pow.f64(double %Val, double %Power)
6979 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
6980 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
6981 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
6982</pre>
6983
6984<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6985<p>The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
6986 specified (positive or negative) power.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6987
6988<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6989<p>The second argument is a floating point power, and the first is a value to
6990 raise to that power.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6991
6992<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6993<p>This function returns the first value raised to the second power, returning
6994 the same values as the libm <tt>pow</tt> functions would, and handles error
6995 conditions in the same way.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6996
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6997</div>
6998
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6999</div>
7000
Dan Gohman911fa902011-05-23 21:13:03 +0000[diff] [blame]7001<!-- _______________________________________________________________________ -->
7002<h4>
7003 <a name="int_exp">'<tt>llvm.exp.*</tt>' Intrinsic</a>
7004</h4>
7005
7006<div>
7007
7008<h5>Syntax:</h5>
7009<p>This is an overloaded intrinsic. You can use <tt>llvm.exp</tt> on any
7010 floating point or vector of floating point type. Not all targets support all
7011 types however.</p>
7012
7013<pre>
7014 declare float @llvm.exp.f32(float %Val)
7015 declare double @llvm.exp.f64(double %Val)
7016 declare x86_fp80 @llvm.exp.f80(x86_fp80 %Val)
7017 declare fp128 @llvm.exp.f128(fp128 %Val)
7018 declare ppc_fp128 @llvm.exp.ppcf128(ppc_fp128 %Val)
7019</pre>
7020
7021<h5>Overview:</h5>
7022<p>The '<tt>llvm.exp.*</tt>' intrinsics perform the exp function.</p>
7023
7024<h5>Arguments:</h5>
7025<p>The argument and return value are floating point numbers of the same
7026 type.</p>
7027
7028<h5>Semantics:</h5>
7029<p>This function returns the same values as the libm <tt>exp</tt> functions
7030 would, and handles error conditions in the same way.</p>
7031
7032</div>
7033
7034<!-- _______________________________________________________________________ -->
7035<h4>
7036 <a name="int_log">'<tt>llvm.log.*</tt>' Intrinsic</a>
7037</h4>
7038
7039<div>
7040
7041<h5>Syntax:</h5>
7042<p>This is an overloaded intrinsic. You can use <tt>llvm.log</tt> on any
7043 floating point or vector of floating point type. Not all targets support all
7044 types however.</p>
7045
7046<pre>
7047 declare float @llvm.log.f32(float %Val)
7048 declare double @llvm.log.f64(double %Val)
7049 declare x86_fp80 @llvm.log.f80(x86_fp80 %Val)
7050 declare fp128 @llvm.log.f128(fp128 %Val)
7051 declare ppc_fp128 @llvm.log.ppcf128(ppc_fp128 %Val)
7052</pre>
7053
7054<h5>Overview:</h5>
7055<p>The '<tt>llvm.log.*</tt>' intrinsics perform the log function.</p>
7056
7057<h5>Arguments:</h5>
7058<p>The argument and return value are floating point numbers of the same
7059 type.</p>
7060
7061<h5>Semantics:</h5>
7062<p>This function returns the same values as the libm <tt>log</tt> functions
7063 would, and handles error conditions in the same way.</p>
7064
Cameron Zwarichf03fa182011-07-08 21:39:21 +0000[diff] [blame]7065<h4>
7066 <a name="int_fma">'<tt>llvm.fma.*</tt>' Intrinsic</a>
7067</h4>
7068
7069<div>
7070
7071<h5>Syntax:</h5>
7072<p>This is an overloaded intrinsic. You can use <tt>llvm.fma</tt> on any
7073 floating point or vector of floating point type. Not all targets support all
7074 types however.</p>
7075
7076<pre>
7077 declare float @llvm.fma.f32(float %a, float %b, float %c)
7078 declare double @llvm.fma.f64(double %a, double %b, double %c)
7079 declare x86_fp80 @llvm.fma.f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c)
7080 declare fp128 @llvm.fma.f128(fp128 %a, fp128 %b, fp128 %c)
7081 declare ppc_fp128 @llvm.fma.ppcf128(ppc_fp128 %a, ppc_fp128 %b, ppc_fp128 %c)
7082</pre>
7083
7084<h5>Overview:</h5>
Cameron Zwaricha32fd212011-07-08 22:13:55 +0000[diff] [blame]7085<p>The '<tt>llvm.fma.*</tt>' intrinsics perform the fused multiply-add
Cameron Zwarichf03fa182011-07-08 21:39:21 +0000[diff] [blame]7086 operation.</p>
7087
7088<h5>Arguments:</h5>
7089<p>The argument and return value are floating point numbers of the same
7090 type.</p>
7091
7092<h5>Semantics:</h5>
7093<p>This function returns the same values as the libm <tt>fma</tt> functions
7094 would.</p>
7095
Dan Gohman911fa902011-05-23 21:13:03 +0000[diff] [blame]7096</div>
7097
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7098<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7099<h3>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]7100 <a name="int_manip">Bit Manipulation Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7101</h3>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7102
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7103<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7104
7105<p>LLVM provides intrinsics for a few important bit manipulation operations.
7106 These allow efficient code generation for some algorithms.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7107
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7108<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7109<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]7110 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7111</h4>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]7112
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7113<div>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]7114
7115<h5>Syntax:</h5>
Reid Spencer4eefaab2007-04-01 08:04:23 +0000[diff] [blame]7116<p>This is an overloaded intrinsic function. You can use bswap on any integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7117 type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
7118
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]7119<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]7120 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
7121 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
7122 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]7123</pre>
7124
7125<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7126<p>The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
7127 values with an even number of bytes (positive multiple of 16 bits). These
7128 are useful for performing operations on data that is not in the target's
7129 native byte order.</p>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]7130
7131<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7132<p>The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
7133 and low byte of the input i16 swapped. Similarly,
7134 the <tt>llvm.bswap.i32</tt> intrinsic returns an i32 value that has the four
7135 bytes of the input i32 swapped, so that if the input bytes are numbered 0, 1,
7136 2, 3 then the returned i32 will have its bytes in 3, 2, 1, 0 order.
7137 The <tt>llvm.bswap.i48</tt>, <tt>llvm.bswap.i64</tt> and other intrinsics
7138 extend this concept to additional even-byte lengths (6 bytes, 8 bytes and
7139 more, respectively).</p>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]7140
7141</div>
7142
7143<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7144<h4>
Reid Spencerb4f9a6f2006-01-16 21:12:35 +0000[diff] [blame]7145 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7146</h4>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7147
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7148<div>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7149
7150<h5>Syntax:</h5>
Reid Spencer4eefaab2007-04-01 08:04:23 +0000[diff] [blame]7151<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]7152 width, or on any vector with integer elements. Not all targets support all
7153 bit widths or vector types, however.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7154
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7155<pre>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7156 declare i8 @llvm.ctpop.i8(i8 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]7157 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]7158 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]7159 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
7160 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]7161 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]7162</pre>
7163
7164<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7165<p>The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set
7166 in a value.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7167
7168<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7169<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]7170 integer type, or a vector with integer elements.
7171 The return type must match the argument type.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7172
7173<h5>Semantics:</h5>
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]7174<p>The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable, or within each
7175 element of a vector.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7176
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7177</div>
7178
7179<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7180<h4>
Chris Lattnerb748c672006-01-16 22:34:14 +0000[diff] [blame]7181 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7182</h4>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7183
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7184<div>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7185
7186<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7187<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]7188 integer bit width, or any vector whose elements are integers. Not all
7189 targets support all bit widths or vector types, however.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7190
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7191<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]7192 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
7193 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]7194 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]7195 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
7196 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]7197 declare &lt;2 x i32&gt; @llvm.ctlz.v2i32(&lt;2 x i32&gt; &lt;src;gt)
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7198</pre>
7199
7200<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7201<p>The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
7202 leading zeros in a variable.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7203
7204<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7205<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]7206 integer type, or any vector type with integer element type.
7207 The return type must match the argument type.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7208
7209<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7210<p>The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant)
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]7211 zeros in a variable, or within each element of the vector if the operation
7212 is of vector type. If the src == 0 then the result is the size in bits of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7213 the type of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7214
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]7215</div>
Chris Lattner3b4f4372004-06-11 02:28:03 +0000[diff] [blame]7216
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]7217<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7218<h4>
Chris Lattnerb748c672006-01-16 22:34:14 +0000[diff] [blame]7219 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7220</h4>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]7221
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7222<div>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]7223
7224<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7225<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]7226 integer bit width, or any vector of integer elements. Not all targets
7227 support all bit widths or vector types, however.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7228
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]7229<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]7230 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
7231 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]7232 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]7233 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
7234 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]7235 declase &lt;2 x i32&gt; @llvm.cttz.v2i32(&lt;2 x i32&gt; &lt;src&gt;)
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]7236</pre>
7237
7238<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7239<p>The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
7240 trailing zeros.</p>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]7241
7242<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7243<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]7244 integer type, or a vectory with integer element type.. The return type
7245 must match the argument type.</p>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]7246
7247<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7248<p>The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant)
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]7249 zeros in a variable, or within each element of a vector.
7250 If the src == 0 then the result is the size in bits of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7251 the type of src. For example, <tt>llvm.cttz(2) = 1</tt>.</p>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]7252
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]7253</div>
7254
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7255</div>
7256
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]7257<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7258<h3>
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]7259 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7260</h3>
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]7261
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7262<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7263
7264<p>LLVM provides intrinsics for some arithmetic with overflow operations.</p>
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]7265
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7266<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7267<h4>
7268 <a name="int_sadd_overflow">
7269 '<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics
7270 </a>
7271</h4>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7272
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7273<div>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7274
7275<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7276<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]7277 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7278
7279<pre>
7280 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
7281 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
7282 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
7283</pre>
7284
7285<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7286<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7287 a signed addition of the two arguments, and indicate whether an overflow
7288 occurred during the signed summation.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7289
7290<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7291<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]7292 be of integer types of any bit width, but they must have the same bit
7293 width. The second element of the result structure must be of
7294 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7295 undergo signed addition.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7296
7297<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7298<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7299 a signed addition of the two variables. They return a structure &mdash; the
7300 first element of which is the signed summation, and the second element of
7301 which is a bit specifying if the signed summation resulted in an
7302 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7303
7304<h5>Examples:</h5>
7305<pre>
7306 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
7307 %sum = extractvalue {i32, i1} %res, 0
7308 %obit = extractvalue {i32, i1} %res, 1
7309 br i1 %obit, label %overflow, label %normal
7310</pre>
7311
7312</div>
7313
7314<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7315<h4>
7316 <a name="int_uadd_overflow">
7317 '<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics
7318 </a>
7319</h4>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7320
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7321<div>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7322
7323<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7324<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]7325 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7326
7327<pre>
7328 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
7329 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
7330 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
7331</pre>
7332
7333<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7334<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7335 an unsigned addition of the two arguments, and indicate whether a carry
7336 occurred during the unsigned summation.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7337
7338<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7339<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]7340 be of integer types of any bit width, but they must have the same bit
7341 width. The second element of the result structure must be of
7342 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7343 undergo unsigned addition.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7344
7345<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7346<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7347 an unsigned addition of the two arguments. They return a structure &mdash;
7348 the first element of which is the sum, and the second element of which is a
7349 bit specifying if the unsigned summation resulted in a carry.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7350
7351<h5>Examples:</h5>
7352<pre>
7353 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
7354 %sum = extractvalue {i32, i1} %res, 0
7355 %obit = extractvalue {i32, i1} %res, 1
7356 br i1 %obit, label %carry, label %normal
7357</pre>
7358
7359</div>
7360
7361<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7362<h4>
7363 <a name="int_ssub_overflow">
7364 '<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics
7365 </a>
7366</h4>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7367
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7368<div>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7369
7370<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7371<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]7372 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7373
7374<pre>
7375 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
7376 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
7377 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
7378</pre>
7379
7380<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7381<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7382 a signed subtraction of the two arguments, and indicate whether an overflow
7383 occurred during the signed subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7384
7385<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7386<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]7387 be of integer types of any bit width, but they must have the same bit
7388 width. The second element of the result structure must be of
7389 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7390 undergo signed subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7391
7392<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7393<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7394 a signed subtraction of the two arguments. They return a structure &mdash;
7395 the first element of which is the subtraction, and the second element of
7396 which is a bit specifying if the signed subtraction resulted in an
7397 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7398
7399<h5>Examples:</h5>
7400<pre>
7401 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
7402 %sum = extractvalue {i32, i1} %res, 0
7403 %obit = extractvalue {i32, i1} %res, 1
7404 br i1 %obit, label %overflow, label %normal
7405</pre>
7406
7407</div>
7408
7409<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7410<h4>
7411 <a name="int_usub_overflow">
7412 '<tt>llvm.usub.with.overflow.*</tt>' Intrinsics
7413 </a>
7414</h4>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7415
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7416<div>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7417
7418<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7419<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]7420 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7421
7422<pre>
7423 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
7424 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
7425 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
7426</pre>
7427
7428<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7429<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7430 an unsigned subtraction of the two arguments, and indicate whether an
7431 overflow occurred during the unsigned subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7432
7433<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7434<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]7435 be of integer types of any bit width, but they must have the same bit
7436 width. The second element of the result structure must be of
7437 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7438 undergo unsigned subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7439
7440<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7441<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7442 an unsigned subtraction of the two arguments. They return a structure &mdash;
7443 the first element of which is the subtraction, and the second element of
7444 which is a bit specifying if the unsigned subtraction resulted in an
7445 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7446
7447<h5>Examples:</h5>
7448<pre>
7449 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
7450 %sum = extractvalue {i32, i1} %res, 0
7451 %obit = extractvalue {i32, i1} %res, 1
7452 br i1 %obit, label %overflow, label %normal
7453</pre>
7454
7455</div>
7456
7457<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7458<h4>
7459 <a name="int_smul_overflow">
7460 '<tt>llvm.smul.with.overflow.*</tt>' Intrinsics
7461 </a>
7462</h4>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7463
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7464<div>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7465
7466<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7467<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]7468 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7469
7470<pre>
7471 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
7472 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
7473 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
7474</pre>
7475
7476<h5>Overview:</h5>
7477
7478<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7479 a signed multiplication of the two arguments, and indicate whether an
7480 overflow occurred during the signed multiplication.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7481
7482<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7483<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]7484 be of integer types of any bit width, but they must have the same bit
7485 width. The second element of the result structure must be of
7486 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7487 undergo signed multiplication.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7488
7489<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7490<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7491 a signed multiplication of the two arguments. They return a structure &mdash;
7492 the first element of which is the multiplication, and the second element of
7493 which is a bit specifying if the signed multiplication resulted in an
7494 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7495
7496<h5>Examples:</h5>
7497<pre>
7498 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
7499 %sum = extractvalue {i32, i1} %res, 0
7500 %obit = extractvalue {i32, i1} %res, 1
7501 br i1 %obit, label %overflow, label %normal
7502</pre>
7503
Reid Spencer5bf54c82007-04-11 23:23:49 +0000[diff] [blame]7504</div>
7505
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7506<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7507<h4>
7508 <a name="int_umul_overflow">
7509 '<tt>llvm.umul.with.overflow.*</tt>' Intrinsics
7510 </a>
7511</h4>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7512
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7513<div>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7514
7515<h5>Syntax:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7516<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]7517 on any integer bit width.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7518
7519<pre>
7520 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
7521 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7522 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
7523</pre>
7524
7525<h5>Overview:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7526<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7527 a unsigned multiplication of the two arguments, and indicate whether an
7528 overflow occurred during the unsigned multiplication.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7529
7530<h5>Arguments:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7531<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]7532 be of integer types of any bit width, but they must have the same bit
7533 width. The second element of the result structure must be of
7534 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7535 undergo unsigned multiplication.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7536
7537<h5>Semantics:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7538<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7539 an unsigned multiplication of the two arguments. They return a structure
7540 &mdash; the first element of which is the multiplication, and the second
7541 element of which is a bit specifying if the unsigned multiplication resulted
7542 in an overflow.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7543
7544<h5>Examples:</h5>
7545<pre>
7546 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7547 %sum = extractvalue {i32, i1} %res, 0
7548 %obit = extractvalue {i32, i1} %res, 1
7549 br i1 %obit, label %overflow, label %normal
7550</pre>
7551
7552</div>
7553
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7554</div>
7555
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]7556<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7557<h3>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7558 <a name="int_fp16">Half Precision Floating Point Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7559</h3>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7560
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7561<div>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7562
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]7563<p>Half precision floating point is a storage-only format. This means that it is
7564 a dense encoding (in memory) but does not support computation in the
7565 format.</p>
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]7566
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]7567<p>This means that code must first load the half-precision floating point
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]7568 value as an i16, then convert it to float with <a
7569 href="#int_convert_from_fp16"><tt>llvm.convert.from.fp16</tt></a>.
7570 Computation can then be performed on the float value (including extending to
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]7571 double etc). To store the value back to memory, it is first converted to
7572 float if needed, then converted to i16 with
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]7573 <a href="#int_convert_to_fp16"><tt>llvm.convert.to.fp16</tt></a>, then
7574 storing as an i16 value.</p>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7575
7576<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7577<h4>
7578 <a name="int_convert_to_fp16">
7579 '<tt>llvm.convert.to.fp16</tt>' Intrinsic
7580 </a>
7581</h4>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7582
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7583<div>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7584
7585<h5>Syntax:</h5>
7586<pre>
7587 declare i16 @llvm.convert.to.fp16(f32 %a)
7588</pre>
7589
7590<h5>Overview:</h5>
7591<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
7592 a conversion from single precision floating point format to half precision
7593 floating point format.</p>
7594
7595<h5>Arguments:</h5>
7596<p>The intrinsic function contains single argument - the value to be
7597 converted.</p>
7598
7599<h5>Semantics:</h5>
7600<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
7601 a conversion from single precision floating point format to half precision
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]7602 floating point format. The return value is an <tt>i16</tt> which
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]7603 contains the converted number.</p>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7604
7605<h5>Examples:</h5>
7606<pre>
7607 %res = call i16 @llvm.convert.to.fp16(f32 %a)
7608 store i16 %res, i16* @x, align 2
7609</pre>
7610
7611</div>
7612
7613<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7614<h4>
7615 <a name="int_convert_from_fp16">
7616 '<tt>llvm.convert.from.fp16</tt>' Intrinsic
7617 </a>
7618</h4>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7619
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7620<div>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7621
7622<h5>Syntax:</h5>
7623<pre>
7624 declare f32 @llvm.convert.from.fp16(i16 %a)
7625</pre>
7626
7627<h5>Overview:</h5>
7628<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs
7629 a conversion from half precision floating point format to single precision
7630 floating point format.</p>
7631
7632<h5>Arguments:</h5>
7633<p>The intrinsic function contains single argument - the value to be
7634 converted.</p>
7635
7636<h5>Semantics:</h5>
7637<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs a
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]7638 conversion from half single precision floating point format to single
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]7639 precision floating point format. The input half-float value is represented by
7640 an <tt>i16</tt> value.</p>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7641
7642<h5>Examples:</h5>
7643<pre>
7644 %a = load i16* @x, align 2
7645 %res = call f32 @llvm.convert.from.fp16(i16 %a)
7646</pre>
7647
7648</div>
7649
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7650</div>
7651
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7652<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7653<h3>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]7654 <a name="int_debugger">Debugger Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7655</h3>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]7656
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7657<div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]7658
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7659<p>The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt>
7660 prefix), are described in
7661 the <a href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source
7662 Level Debugging</a> document.</p>
7663
7664</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]7665
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]7666<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7667<h3>
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]7668 <a name="int_eh">Exception Handling Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7669</h3>
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]7670
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7671<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7672
7673<p>The LLVM exception handling intrinsics (which all start with
7674 <tt>llvm.eh.</tt> prefix), are described in
7675 the <a href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
7676 Handling</a> document.</p>
7677
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]7678</div>
7679
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7680<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7681<h3>
Duncan Sands86e01192007-09-11 14:10:23 +0000[diff] [blame]7682 <a name="int_trampoline">Trampoline Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7683</h3>
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7684
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7685<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7686
7687<p>This intrinsic makes it possible to excise one parameter, marked with
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]7688 the <a href="#nest"><tt>nest</tt></a> attribute, from a function.
7689 The result is a callable
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7690 function pointer lacking the nest parameter - the caller does not need to
7691 provide a value for it. Instead, the value to use is stored in advance in a
7692 "trampoline", a block of memory usually allocated on the stack, which also
7693 contains code to splice the nest value into the argument list. This is used
7694 to implement the GCC nested function address extension.</p>
7695
7696<p>For example, if the function is
7697 <tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
7698 pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as
7699 follows:</p>
7700
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]7701<pre class="doc_code">
Duncan Sands86e01192007-09-11 14:10:23 +0000[diff] [blame]7702 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
7703 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7704 %p = call i8* @llvm.init.trampoline(i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8*), i8* %nval)
Duncan Sands86e01192007-09-11 14:10:23 +0000[diff] [blame]7705 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7706</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7707
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7708<p>The call <tt>%val = call i32 %fp(i32 %x, i32 %y)</tt> is then equivalent
7709 to <tt>%val = call i32 %f(i8* %nval, i32 %x, i32 %y)</tt>.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7710
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7711<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7712<h4>
7713 <a name="int_it">
7714 '<tt>llvm.init.trampoline</tt>' Intrinsic
7715 </a>
7716</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7717
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7718<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7719
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7720<h5>Syntax:</h5>
7721<pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7722 declare i8* @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]7723</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7724
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7725<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7726<p>This fills the memory pointed to by <tt>tramp</tt> with code and returns a
7727 function pointer suitable for executing it.</p>
7728
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7729<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7730<p>The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
7731 pointers. The <tt>tramp</tt> argument must point to a sufficiently large and
7732 sufficiently aligned block of memory; this memory is written to by the
7733 intrinsic. Note that the size and the alignment are target-specific - LLVM
7734 currently provides no portable way of determining them, so a front-end that
7735 generates this intrinsic needs to have some target-specific knowledge.
7736 The <tt>func</tt> argument must hold a function bitcast to
7737 an <tt>i8*</tt>.</p>
7738
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7739<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7740<p>The block of memory pointed to by <tt>tramp</tt> is filled with target
7741 dependent code, turning it into a function. A pointer to this function is
7742 returned, but needs to be bitcast to an <a href="#int_trampoline">appropriate
7743 function pointer type</a> before being called. The new function's signature
7744 is the same as that of <tt>func</tt> with any arguments marked with
7745 the <tt>nest</tt> attribute removed. At most one such <tt>nest</tt> argument
7746 is allowed, and it must be of pointer type. Calling the new function is
7747 equivalent to calling <tt>func</tt> with the same argument list, but
7748 with <tt>nval</tt> used for the missing <tt>nest</tt> argument. If, after
7749 calling <tt>llvm.init.trampoline</tt>, the memory pointed to
7750 by <tt>tramp</tt> is modified, then the effect of any later call to the
7751 returned function pointer is undefined.</p>
7752
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7753</div>
7754
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7755</div>
7756
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7757<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7758<h3>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7759 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7760</h3>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7761
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7762<div>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7763
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7764<p>These intrinsic functions expand the "universal IR" of LLVM to represent
7765 hardware constructs for atomic operations and memory synchronization. This
7766 provides an interface to the hardware, not an interface to the programmer. It
7767 is aimed at a low enough level to allow any programming models or APIs
7768 (Application Programming Interfaces) which need atomic behaviors to map
7769 cleanly onto it. It is also modeled primarily on hardware behavior. Just as
7770 hardware provides a "universal IR" for source languages, it also provides a
7771 starting point for developing a "universal" atomic operation and
7772 synchronization IR.</p>
7773
7774<p>These do <em>not</em> form an API such as high-level threading libraries,
7775 software transaction memory systems, atomic primitives, and intrinsic
7776 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
7777 application libraries. The hardware interface provided by LLVM should allow
7778 a clean implementation of all of these APIs and parallel programming models.
7779 No one model or paradigm should be selected above others unless the hardware
7780 itself ubiquitously does so.</p>
7781
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7782<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7783<h4>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7784 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7785</h4>
7786
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7787<div>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7788<h5>Syntax:</h5>
7789<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7790 declare void @llvm.memory.barrier(i1 &lt;ll&gt;, i1 &lt;ls&gt;, i1 &lt;sl&gt;, i1 &lt;ss&gt;, i1 &lt;device&gt;)
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7791</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7792
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7793<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7794<p>The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
7795 specific pairs of memory access types.</p>
7796
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7797<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7798<p>The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
7799 The first four arguments enables a specific barrier as listed below. The
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]7800 fifth argument specifies that the barrier applies to io or device or uncached
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7801 memory.</p>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7802
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7803<ul>
7804 <li><tt>ll</tt>: load-load barrier</li>
7805 <li><tt>ls</tt>: load-store barrier</li>
7806 <li><tt>sl</tt>: store-load barrier</li>
7807 <li><tt>ss</tt>: store-store barrier</li>
7808 <li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
7809</ul>
7810
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7811<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7812<p>This intrinsic causes the system to enforce some ordering constraints upon
7813 the loads and stores of the program. This barrier does not
7814 indicate <em>when</em> any events will occur, it only enforces
7815 an <em>order</em> in which they occur. For any of the specified pairs of load
7816 and store operations (f.ex. load-load, or store-load), all of the first
7817 operations preceding the barrier will complete before any of the second
7818 operations succeeding the barrier begin. Specifically the semantics for each
7819 pairing is as follows:</p>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7820
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7821<ul>
7822 <li><tt>ll</tt>: All loads before the barrier must complete before any load
7823 after the barrier begins.</li>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7824 <li><tt>ls</tt>: All loads before the barrier must complete before any
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7825 store after the barrier begins.</li>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7826 <li><tt>ss</tt>: All stores before the barrier must complete before any
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7827 store after the barrier begins.</li>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7828 <li><tt>sl</tt>: All stores before the barrier must complete before any
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7829 load after the barrier begins.</li>
7830</ul>
7831
7832<p>These semantics are applied with a logical "and" behavior when more than one
7833 is enabled in a single memory barrier intrinsic.</p>
7834
7835<p>Backends may implement stronger barriers than those requested when they do
7836 not support as fine grained a barrier as requested. Some architectures do
7837 not need all types of barriers and on such architectures, these become
7838 noops.</p>
7839
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7840<h5>Example:</h5>
7841<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7842%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7843%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7844 store i32 4, %ptr
7845
7846%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
Evan Cheng0ac49c62011-06-29 17:14:00 +0000[diff] [blame]7847 call void @llvm.memory.barrier(i1 false, i1 true, i1 false, i1 false, i1 true)
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7848 <i>; guarantee the above finishes</i>
7849 store i32 8, %ptr <i>; before this begins</i>
7850</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7851
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7852</div>
7853
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7854<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7855<h4>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7856 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7857</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7858
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7859<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7860
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7861<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7862<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
7863 any integer bit width and for different address spaces. Not all targets
7864 support all bit widths however.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7865
7866<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7867 declare i8 @llvm.atomic.cmp.swap.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt;)
7868 declare i16 @llvm.atomic.cmp.swap.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt;)
7869 declare i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt;)
7870 declare i64 @llvm.atomic.cmp.swap.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;cmp&gt;, i64 &lt;val&gt;)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7871</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7872
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7873<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7874<p>This loads a value in memory and compares it to a given value. If they are
7875 equal, it stores a new value into the memory.</p>
7876
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7877<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7878<p>The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result
7879 as well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
7880 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
7881 this integer type. While any bit width integer may be used, targets may only
7882 lower representations they support in hardware.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7883
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7884<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7885<p>This entire intrinsic must be executed atomically. It first loads the value
7886 in memory pointed to by <tt>ptr</tt> and compares it with the
7887 value <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the
7888 memory. The loaded value is yielded in all cases. This provides the
7889 equivalent of an atomic compare-and-swap operation within the SSA
7890 framework.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7891
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7892<h5>Examples:</h5>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7893<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7894%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7895%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7896 store i32 4, %ptr
7897
7898%val1 = add i32 4, 4
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7899%result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %ptr, i32 4, %val1)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7900 <i>; yields {i32}:result1 = 4</i>
7901%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
7902%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
7903
7904%val2 = add i32 1, 1
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7905%result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %ptr, i32 5, %val2)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7906 <i>; yields {i32}:result2 = 8</i>
7907%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
7908
7909%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
7910</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7911
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7912</div>
7913
7914<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7915<h4>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7916 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7917</h4>
7918
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7919<div>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7920<h5>Syntax:</h5>
7921
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7922<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
7923 integer bit width. Not all targets support all bit widths however.</p>
7924
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7925<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7926 declare i8 @llvm.atomic.swap.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;val&gt;)
7927 declare i16 @llvm.atomic.swap.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;val&gt;)
7928 declare i32 @llvm.atomic.swap.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;val&gt;)
7929 declare i64 @llvm.atomic.swap.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;val&gt;)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7930</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7931
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7932<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7933<p>This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
7934 the value from memory. It then stores the value in <tt>val</tt> in the memory
7935 at <tt>ptr</tt>.</p>
7936
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7937<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7938<p>The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both
7939 the <tt>val</tt> argument and the result must be integers of the same bit
7940 width. The first argument, <tt>ptr</tt>, must be a pointer to a value of this
7941 integer type. The targets may only lower integer representations they
7942 support.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7943
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7944<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7945<p>This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
7946 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
7947 equivalent of an atomic swap operation within the SSA framework.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7948
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7949<h5>Examples:</h5>
7950<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7951%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7952%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7953 store i32 4, %ptr
7954
7955%val1 = add i32 4, 4
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7956%result1 = call i32 @llvm.atomic.swap.i32.p0i32(i32* %ptr, i32 %val1)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7957 <i>; yields {i32}:result1 = 4</i>
7958%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
7959%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
7960
7961%val2 = add i32 1, 1
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7962%result2 = call i32 @llvm.atomic.swap.i32.p0i32(i32* %ptr, i32 %val2)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7963 <i>; yields {i32}:result2 = 8</i>
7964
7965%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
7966%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
7967</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7968
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7969</div>
7970
7971<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7972<h4>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7973 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7974</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7975
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7976<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7977
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7978<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7979<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.add</tt> on
7980 any integer bit width. Not all targets support all bit widths however.</p>
7981
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7982<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7983 declare i8 @llvm.atomic.load.add.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7984 declare i16 @llvm.atomic.load.add.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7985 declare i32 @llvm.atomic.load.add.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7986 declare i64 @llvm.atomic.load.add.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7987</pre>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7988
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7989<h5>Overview:</h5>
7990<p>This intrinsic adds <tt>delta</tt> to the value stored in memory
7991 at <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
7992
7993<h5>Arguments:</h5>
7994<p>The intrinsic takes two arguments, the first a pointer to an integer value
7995 and the second an integer value. The result is also an integer value. These
7996 integer types can have any bit width, but they must all have the same bit
7997 width. The targets may only lower integer representations they support.</p>
7998
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7999<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8000<p>This intrinsic does a series of operations atomically. It first loads the
8001 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
8002 to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]8003
8004<h5>Examples:</h5>
8005<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]8006%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
8007%ptr = bitcast i8* %mallocP to i32*
8008 store i32 4, %ptr
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8009%result1 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 4)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]8010 <i>; yields {i32}:result1 = 4</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8011%result2 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 2)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]8012 <i>; yields {i32}:result2 = 8</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8013%result3 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 5)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]8014 <i>; yields {i32}:result3 = 10</i>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8015%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]8016</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8017
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]8018</div>
8019
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8020<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8021<h4>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8022 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8023</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8024
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8025<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8026
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8027<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8028<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.sub</tt> on
8029 any integer bit width and for different address spaces. Not all targets
8030 support all bit widths however.</p>
8031
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8032<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8033 declare i8 @llvm.atomic.load.sub.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
8034 declare i16 @llvm.atomic.load.sub.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
8035 declare i32 @llvm.atomic.load.sub.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
8036 declare i64 @llvm.atomic.load.sub.i64.p0i32(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8037</pre>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8038
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8039<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]8040<p>This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8041 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
8042
8043<h5>Arguments:</h5>
8044<p>The intrinsic takes two arguments, the first a pointer to an integer value
8045 and the second an integer value. The result is also an integer value. These
8046 integer types can have any bit width, but they must all have the same bit
8047 width. The targets may only lower integer representations they support.</p>
8048
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8049<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8050<p>This intrinsic does a series of operations atomically. It first loads the
8051 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
8052 result to <tt>ptr</tt>. It yields the original value stored
8053 at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8054
8055<h5>Examples:</h5>
8056<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]8057%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
8058%ptr = bitcast i8* %mallocP to i32*
8059 store i32 8, %ptr
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8060%result1 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 4)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8061 <i>; yields {i32}:result1 = 8</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8062%result2 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 2)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8063 <i>; yields {i32}:result2 = 4</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8064%result3 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 5)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8065 <i>; yields {i32}:result3 = 2</i>
8066%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
8067</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8068
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8069</div>
8070
8071<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8072<h4>
8073 <a name="int_atomic_load_and">
8074 '<tt>llvm.atomic.load.and.*</tt>' Intrinsic
8075 </a>
8076 <br>
8077 <a name="int_atomic_load_nand">
8078 '<tt>llvm.atomic.load.nand.*</tt>' Intrinsic
8079 </a>
8080 <br>
8081 <a name="int_atomic_load_or">
8082 '<tt>llvm.atomic.load.or.*</tt>' Intrinsic
8083 </a>
8084 <br>
8085 <a name="int_atomic_load_xor">
8086 '<tt>llvm.atomic.load.xor.*</tt>' Intrinsic
8087 </a>
8088</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8089
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8090<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8091
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8092<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8093<p>These are overloaded intrinsics. You can
8094 use <tt>llvm.atomic.load_and</tt>, <tt>llvm.atomic.load_nand</tt>,
8095 <tt>llvm.atomic.load_or</tt>, and <tt>llvm.atomic.load_xor</tt> on any integer
8096 bit width and for different address spaces. Not all targets support all bit
8097 widths however.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8098
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8099<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8100 declare i8 @llvm.atomic.load.and.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
8101 declare i16 @llvm.atomic.load.and.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
8102 declare i32 @llvm.atomic.load.and.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
8103 declare i64 @llvm.atomic.load.and.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8104</pre>
8105
8106<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8107 declare i8 @llvm.atomic.load.or.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
8108 declare i16 @llvm.atomic.load.or.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
8109 declare i32 @llvm.atomic.load.or.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
8110 declare i64 @llvm.atomic.load.or.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8111</pre>
8112
8113<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8114 declare i8 @llvm.atomic.load.nand.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
8115 declare i16 @llvm.atomic.load.nand.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
8116 declare i32 @llvm.atomic.load.nand.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
8117 declare i64 @llvm.atomic.load.nand.i64.p0i32(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8118</pre>
8119
8120<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8121 declare i8 @llvm.atomic.load.xor.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
8122 declare i16 @llvm.atomic.load.xor.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
8123 declare i32 @llvm.atomic.load.xor.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
8124 declare i64 @llvm.atomic.load.xor.i64.p0i32(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8125</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8126
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8127<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8128<p>These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
8129 the value stored in memory at <tt>ptr</tt>. It yields the original value
8130 at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8131
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8132<h5>Arguments:</h5>
8133<p>These intrinsics take two arguments, the first a pointer to an integer value
8134 and the second an integer value. The result is also an integer value. These
8135 integer types can have any bit width, but they must all have the same bit
8136 width. The targets may only lower integer representations they support.</p>
8137
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8138<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8139<p>These intrinsics does a series of operations atomically. They first load the
8140 value stored at <tt>ptr</tt>. They then do the bitwise
8141 operation <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the
8142 original value stored at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8143
8144<h5>Examples:</h5>
8145<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]8146%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
8147%ptr = bitcast i8* %mallocP to i32*
8148 store i32 0x0F0F, %ptr
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8149%result0 = call i32 @llvm.atomic.load.nand.i32.p0i32(i32* %ptr, i32 0xFF)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8150 <i>; yields {i32}:result0 = 0x0F0F</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8151%result1 = call i32 @llvm.atomic.load.and.i32.p0i32(i32* %ptr, i32 0xFF)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8152 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8153%result2 = call i32 @llvm.atomic.load.or.i32.p0i32(i32* %ptr, i32 0F)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8154 <i>; yields {i32}:result2 = 0xF0</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8155%result3 = call i32 @llvm.atomic.load.xor.i32.p0i32(i32* %ptr, i32 0F)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8156 <i>; yields {i32}:result3 = FF</i>
8157%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
8158</pre>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8159
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8160</div>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8161
8162<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8163<h4>
8164 <a name="int_atomic_load_max">
8165 '<tt>llvm.atomic.load.max.*</tt>' Intrinsic
8166 </a>
8167 <br>
8168 <a name="int_atomic_load_min">
8169 '<tt>llvm.atomic.load.min.*</tt>' Intrinsic
8170 </a>
8171 <br>
8172 <a name="int_atomic_load_umax">
8173 '<tt>llvm.atomic.load.umax.*</tt>' Intrinsic
8174 </a>
8175 <br>
8176 <a name="int_atomic_load_umin">
8177 '<tt>llvm.atomic.load.umin.*</tt>' Intrinsic
8178 </a>
8179</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8180
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8181<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8182
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8183<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8184<p>These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
8185 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
8186 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
8187 address spaces. Not all targets support all bit widths however.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8188
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8189<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8190 declare i8 @llvm.atomic.load.max.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
8191 declare i16 @llvm.atomic.load.max.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
8192 declare i32 @llvm.atomic.load.max.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
8193 declare i64 @llvm.atomic.load.max.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8194</pre>
8195
8196<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8197 declare i8 @llvm.atomic.load.min.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
8198 declare i16 @llvm.atomic.load.min.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
8199 declare i32 @llvm.atomic.load.min.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
8200 declare i64 @llvm.atomic.load.min.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8201</pre>
8202
8203<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8204 declare i8 @llvm.atomic.load.umax.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
8205 declare i16 @llvm.atomic.load.umax.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
8206 declare i32 @llvm.atomic.load.umax.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
8207 declare i64 @llvm.atomic.load.umax.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8208</pre>
8209
8210<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8211 declare i8 @llvm.atomic.load.umin.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
8212 declare i16 @llvm.atomic.load.umin.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
8213 declare i32 @llvm.atomic.load.umin.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
8214 declare i64 @llvm.atomic.load.umin.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8215</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8216
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8217<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]8218<p>These intrinsics takes the signed or unsigned minimum or maximum of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8219 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
8220 original value at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8221
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8222<h5>Arguments:</h5>
8223<p>These intrinsics take two arguments, the first a pointer to an integer value
8224 and the second an integer value. The result is also an integer value. These
8225 integer types can have any bit width, but they must all have the same bit
8226 width. The targets may only lower integer representations they support.</p>
8227
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8228<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8229<p>These intrinsics does a series of operations atomically. They first load the
8230 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or
8231 max <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They
8232 yield the original value stored at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8233
8234<h5>Examples:</h5>
8235<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]8236%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
8237%ptr = bitcast i8* %mallocP to i32*
8238 store i32 7, %ptr
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8239%result0 = call i32 @llvm.atomic.load.min.i32.p0i32(i32* %ptr, i32 -2)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8240 <i>; yields {i32}:result0 = 7</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8241%result1 = call i32 @llvm.atomic.load.max.i32.p0i32(i32* %ptr, i32 8)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8242 <i>; yields {i32}:result1 = -2</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8243%result2 = call i32 @llvm.atomic.load.umin.i32.p0i32(i32* %ptr, i32 10)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8244 <i>; yields {i32}:result2 = 8</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8245%result3 = call i32 @llvm.atomic.load.umax.i32.p0i32(i32* %ptr, i32 30)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8246 <i>; yields {i32}:result3 = 8</i>
8247%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
8248</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8249
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]8250</div>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]8251
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8252</div>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8253
8254<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8255<h3>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8256 <a name="int_memorymarkers">Memory Use Markers</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8257</h3>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8258
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8259<div>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8260
8261<p>This class of intrinsics exists to information about the lifetime of memory
8262 objects and ranges where variables are immutable.</p>
8263
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8264<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8265<h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8266 <a name="int_lifetime_start">'<tt>llvm.lifetime.start</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8267</h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8268
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8269<div>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8270
8271<h5>Syntax:</h5>
8272<pre>
8273 declare void @llvm.lifetime.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
8274</pre>
8275
8276<h5>Overview:</h5>
8277<p>The '<tt>llvm.lifetime.start</tt>' intrinsic specifies the start of a memory
8278 object's lifetime.</p>
8279
8280<h5>Arguments:</h5>
Nick Lewycky9bc89042009-10-13 07:57:33 +0000[diff] [blame]8281<p>The first argument is a constant integer representing the size of the
8282 object, or -1 if it is variable sized. The second argument is a pointer to
8283 the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8284
8285<h5>Semantics:</h5>
8286<p>This intrinsic indicates that before this point in the code, the value of the
8287 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]8288 never be used and has an undefined value. A load from the pointer that
8289 precedes this intrinsic can be replaced with
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8290 <tt>'<a href="#undefvalues">undef</a>'</tt>.</p>
8291
8292</div>
8293
8294<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8295<h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8296 <a name="int_lifetime_end">'<tt>llvm.lifetime.end</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8297</h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8298
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8299<div>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8300
8301<h5>Syntax:</h5>
8302<pre>
8303 declare void @llvm.lifetime.end(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
8304</pre>
8305
8306<h5>Overview:</h5>
8307<p>The '<tt>llvm.lifetime.end</tt>' intrinsic specifies the end of a memory
8308 object's lifetime.</p>
8309
8310<h5>Arguments:</h5>
Nick Lewycky9bc89042009-10-13 07:57:33 +0000[diff] [blame]8311<p>The first argument is a constant integer representing the size of the
8312 object, or -1 if it is variable sized. The second argument is a pointer to
8313 the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8314
8315<h5>Semantics:</h5>
8316<p>This intrinsic indicates that after this point in the code, the value of the
8317 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
8318 never be used and has an undefined value. Any stores into the memory object
8319 following this intrinsic may be removed as dead.
8320
8321</div>
8322
8323<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8324<h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8325 <a name="int_invariant_start">'<tt>llvm.invariant.start</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8326</h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8327
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8328<div>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8329
8330<h5>Syntax:</h5>
8331<pre>
Nick Lewycky2965d3e2010-11-30 04:13:41 +0000[diff] [blame]8332 declare {}* @llvm.invariant.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8333</pre>
8334
8335<h5>Overview:</h5>
8336<p>The '<tt>llvm.invariant.start</tt>' intrinsic specifies that the contents of
8337 a memory object will not change.</p>
8338
8339<h5>Arguments:</h5>
Nick Lewycky9bc89042009-10-13 07:57:33 +0000[diff] [blame]8340<p>The first argument is a constant integer representing the size of the
8341 object, or -1 if it is variable sized. The second argument is a pointer to
8342 the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8343
8344<h5>Semantics:</h5>
8345<p>This intrinsic indicates that until an <tt>llvm.invariant.end</tt> that uses
8346 the return value, the referenced memory location is constant and
8347 unchanging.</p>
8348
8349</div>
8350
8351<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8352<h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8353 <a name="int_invariant_end">'<tt>llvm.invariant.end</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8354</h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8355
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8356<div>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8357
8358<h5>Syntax:</h5>
8359<pre>
8360 declare void @llvm.invariant.end({}* &lt;start&gt;, i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
8361</pre>
8362
8363<h5>Overview:</h5>
8364<p>The '<tt>llvm.invariant.end</tt>' intrinsic specifies that the contents of
8365 a memory object are mutable.</p>
8366
8367<h5>Arguments:</h5>
8368<p>The first argument is the matching <tt>llvm.invariant.start</tt> intrinsic.
Nick Lewycky9bc89042009-10-13 07:57:33 +0000[diff] [blame]8369 The second argument is a constant integer representing the size of the
8370 object, or -1 if it is variable sized and the third argument is a pointer
8371 to the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]8372
8373<h5>Semantics:</h5>
8374<p>This intrinsic indicates that the memory is mutable again.</p>
8375
8376</div>
8377
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8378</div>
8379
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]8380<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8381<h3>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]8382 <a name="int_general">General Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8383</h3>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]8384
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8385<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8386
8387<p>This class of intrinsics is designed to be generic and has no specific
8388 purpose.</p>
8389
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]8390<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8391<h4>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]8392 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8393</h4>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]8394
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8395<div>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]8396
8397<h5>Syntax:</h5>
8398<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8399 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]8400</pre>
8401
8402<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8403<p>The '<tt>llvm.var.annotation</tt>' intrinsic.</p>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]8404
8405<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8406<p>The first argument is a pointer to a value, the second is a pointer to a
8407 global string, the third is a pointer to a global string which is the source
8408 file name, and the last argument is the line number.</p>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]8409
8410<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8411<p>This intrinsic allows annotation of local variables with arbitrary strings.
8412 This can be useful for special purpose optimizations that want to look for
John Criswellf0d536a2011-08-19 16:57:55 +0000[diff] [blame]8413 these annotations. These have no other defined use; they are ignored by code
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8414 generation and optimization.</p>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]8415
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]8416</div>
8417
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]8418<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8419<h4>
Tanya Lattner0186a652007-09-21 23:57:59 +0000[diff] [blame]8420 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8421</h4>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]8422
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8423<div>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]8424
8425<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8426<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
8427 any integer bit width.</p>
8428
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]8429<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8430 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8431 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8432 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8433 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8434 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]8435</pre>
8436
8437<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8438<p>The '<tt>llvm.annotation</tt>' intrinsic.</p>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]8439
8440<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8441<p>The first argument is an integer value (result of some expression), the
8442 second is a pointer to a global string, the third is a pointer to a global
8443 string which is the source file name, and the last argument is the line
8444 number. It returns the value of the first argument.</p>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]8445
8446<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8447<p>This intrinsic allows annotations to be put on arbitrary expressions with
8448 arbitrary strings. This can be useful for special purpose optimizations that
John Criswellf0d536a2011-08-19 16:57:55 +0000[diff] [blame]8449 want to look for these annotations. These have no other defined use; they
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8450 are ignored by code generation and optimization.</p>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]8451
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]8452</div>
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]8453
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]8454<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8455<h4>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]8456 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8457</h4>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]8458
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8459<div>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]8460
8461<h5>Syntax:</h5>
8462<pre>
8463 declare void @llvm.trap()
8464</pre>
8465
8466<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8467<p>The '<tt>llvm.trap</tt>' intrinsic.</p>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]8468
8469<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8470<p>None.</p>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]8471
8472<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8473<p>This intrinsics is lowered to the target dependent trap instruction. If the
8474 target does not have a trap instruction, this intrinsic will be lowered to
8475 the call of the <tt>abort()</tt> function.</p>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]8476
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]8477</div>
8478
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]8479<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8480<h4>
Misha Brukman50de2b22008-11-22 23:55:29 +0000[diff] [blame]8481 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8482</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8483
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8484<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8485
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]8486<h5>Syntax:</h5>
8487<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8488 declare void @llvm.stackprotector(i8* &lt;guard&gt;, i8** &lt;slot&gt;)
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]8489</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8490
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]8491<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8492<p>The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and
8493 stores it onto the stack at <tt>slot</tt>. The stack slot is adjusted to
8494 ensure that it is placed on the stack before local variables.</p>
8495
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]8496<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8497<p>The <tt>llvm.stackprotector</tt> intrinsic requires two pointer
8498 arguments. The first argument is the value loaded from the stack
8499 guard <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt>
8500 that has enough space to hold the value of the guard.</p>
8501
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]8502<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8503<p>This intrinsic causes the prologue/epilogue inserter to force the position of
8504 the <tt>AllocaInst</tt> stack slot to be before local variables on the
8505 stack. This is to ensure that if a local variable on the stack is
8506 overwritten, it will destroy the value of the guard. When the function exits,
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]8507 the guard on the stack is checked against the original guard. If they are
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8508 different, then the program aborts by calling the <tt>__stack_chk_fail()</tt>
8509 function.</p>
8510
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]8511</div>
8512
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8513<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8514<h4>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8515 <a name="int_objectsize">'<tt>llvm.objectsize</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8516</h4>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8517
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8518<div>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8519
8520<h5>Syntax:</h5>
8521<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8522 declare i32 @llvm.objectsize.i32(i8* &lt;object&gt;, i1 &lt;type&gt;)
8523 declare i64 @llvm.objectsize.i64(i8* &lt;object&gt;, i1 &lt;type&gt;)
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8524</pre>
8525
8526<h5>Overview:</h5>
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]8527<p>The <tt>llvm.objectsize</tt> intrinsic is designed to provide information to
8528 the optimizers to determine at compile time whether a) an operation (like
8529 memcpy) will overflow a buffer that corresponds to an object, or b) that a
8530 runtime check for overflow isn't necessary. An object in this context means
8531 an allocation of a specific class, structure, array, or other object.</p>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8532
8533<h5>Arguments:</h5>
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]8534<p>The <tt>llvm.objectsize</tt> intrinsic takes two arguments. The first
Eric Christopher31e39bd2009-12-23 00:29:49 +0000[diff] [blame]8535 argument is a pointer to or into the <tt>object</tt>. The second argument
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]8536 is a boolean 0 or 1. This argument determines whether you want the
8537 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]8538 1, variables are not allowed.</p>
8539
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8540<h5>Semantics:</h5>
8541<p>The <tt>llvm.objectsize</tt> intrinsic is lowered to either a constant
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]8542 representing the size of the object concerned, or <tt>i32/i64 -1 or 0</tt>,
8543 depending on the <tt>type</tt> argument, if the size cannot be determined at
8544 compile time.</p>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8545
8546</div>
8547
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8548</div>
8549
8550</div>
8551
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]8552<!-- *********************************************************************** -->
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]8553<hr>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]8554<address>
8555 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman86242e12008-12-11 17:34:48 +0000[diff] [blame]8556 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]8557 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman86242e12008-12-11 17:34:48 +0000[diff] [blame]8558 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]8559
8560 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
NAKAMURA Takumica46f5a2011-04-09 02:13:37 +0000[diff] [blame]8561 <a href="http://llvm.org/">The LLVM Compiler Infrastructure</a><br>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]8562 Last modified: $Date$
8563</address>
Chris Lattnerb8f816e2008-01-04 04:33:49 +0000[diff] [blame]8564
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]8565</body>
8566</html>