Gitiles
Code Review Sign In
gerrit-public.fairphone.software / toolchain / llvm-project / 3770af58520212f1fba3474a7e2e3f79d8c762aa / . / llvm / docs / LangRef.html
blob: 7d8f14bca6b3035308128c606740c50272f99573 [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
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]15<div class="doc_title"> LLVM Language Reference Manual </div>
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 Wendling8693ef82009-07-20 02:41:50 +0000[diff] [blame]28 <li><a href="#linkage_internal">'<tt>internal</tt>' Linkage</a></li>
29 <li><a href="#linkage_available_externally">'<tt>available_externally</tt>' Linkage</a></li>
30 <li><a href="#linkage_linkonce">'<tt>linkonce</tt>' Linkage</a></li>
31 <li><a href="#linkage_common">'<tt>common</tt>' Linkage</a></li>
32 <li><a href="#linkage_weak">'<tt>weak</tt>' Linkage</a></li>
33 <li><a href="#linkage_appending">'<tt>appending</tt>' Linkage</a></li>
34 <li><a href="#linkage_externweak">'<tt>extern_weak</tt>' Linkage</a></li>
Chris Lattner80d73c72009-10-10 18:26:06 +0000[diff] [blame]35 <li><a href="#linkage_linkonce_odr">'<tt>linkonce_odr</tt>' Linkage</a></li>
Bill Wendling8693ef82009-07-20 02:41:50 +0000[diff] [blame]36 <li><a href="#linkage_weak">'<tt>weak_odr</tt>' Linkage</a></li>
37 <li><a href="#linkage_external">'<tt>externally visible</tt>' Linkage</a></li>
38 <li><a href="#linkage_dllimport">'<tt>dllimport</tt>' Linkage</a></li>
39 <li><a href="#linkage_dllexport">'<tt>dllexport</tt>' Linkage</a></li>
Bill Wendlinga3c6f6b2009-07-20 01:03:30 +0000[diff] [blame]40 </ol>
41 </li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]42 <li><a href="#callingconv">Calling Conventions</a></li>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]43 <li><a href="#namedtypes">Named Types</a></li>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]44 <li><a href="#globalvars">Global Variables</a></li>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]45 <li><a href="#functionstructure">Functions</a></li>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]46 <li><a href="#aliasstructure">Aliases</a></li>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]47 <li><a href="#namedmetadatastructure">Named Metadata</a></li>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]48 <li><a href="#paramattrs">Parameter Attributes</a></li>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]49 <li><a href="#fnattrs">Function Attributes</a></li>
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]50 <li><a href="#gc">Garbage Collector Names</a></li>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]51 <li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]52 <li><a href="#datalayout">Data Layout</a></li>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]53 <li><a href="#pointeraliasing">Pointer Aliasing Rules</a></li>
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]54 <li><a href="#volatile">Volatile Memory Accesses</a></li>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]55 </ol>
56 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]57 <li><a href="#typesystem">Type System</a>
58 <ol>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]59 <li><a href="#t_classifications">Type Classifications</a></li>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]60 <li><a href="#t_primitive">Primitive Types</a>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]61 <ol>
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]62 <li><a href="#t_integer">Integer Type</a></li>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]63 <li><a href="#t_floating">Floating Point Types</a></li>
64 <li><a href="#t_void">Void Type</a></li>
65 <li><a href="#t_label">Label Type</a></li>
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]66 <li><a href="#t_metadata">Metadata Type</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]67 </ol>
68 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]69 <li><a href="#t_derived">Derived Types</a>
70 <ol>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]71 <li><a href="#t_aggregate">Aggregate Types</a>
72 <ol>
73 <li><a href="#t_array">Array Type</a></li>
74 <li><a href="#t_struct">Structure Type</a></li>
75 <li><a href="#t_pstruct">Packed Structure Type</a></li>
76 <li><a href="#t_union">Union Type</a></li>
77 <li><a href="#t_vector">Vector Type</a></li>
78 </ol>
79 </li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]80 <li><a href="#t_function">Function Type</a></li>
81 <li><a href="#t_pointer">Pointer Type</a></li>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]82 <li><a href="#t_opaque">Opaque Type</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]83 </ol>
84 </li>
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]85 <li><a href="#t_uprefs">Type Up-references</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]86 </ol>
87 </li>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]88 <li><a href="#constants">Constants</a>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]89 <ol>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]90 <li><a href="#simpleconstants">Simple Constants</a></li>
Chris Lattner361bfcd2009-02-28 18:32:25 +0000[diff] [blame]91 <li><a href="#complexconstants">Complex Constants</a></li>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]92 <li><a href="#globalconstants">Global Variable and Function Addresses</a></li>
93 <li><a href="#undefvalues">Undefined Values</a></li>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]94 <li><a href="#trapvalues">Trap Values</a></li>
Chris Lattner2bfd3202009-10-27 21:19:13 +0000[diff] [blame]95 <li><a href="#blockaddress">Addresses of Basic Blocks</a></li>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]96 <li><a href="#constantexprs">Constant Expressions</a></li>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]97 </ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]98 </li>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]99 <li><a href="#othervalues">Other Values</a>
100 <ol>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]101 <li><a href="#inlineasm">Inline Assembler Expressions</a></li>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]102 <li><a href="#metadata">Metadata Nodes and Metadata Strings</a></li>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]103 </ol>
104 </li>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]105 <li><a href="#intrinsic_globals">Intrinsic Global Variables</a>
106 <ol>
107 <li><a href="#intg_used">The '<tt>llvm.used</tt>' Global Variable</a></li>
Chris Lattner58f9bb22009-07-20 06:14:25 +0000[diff] [blame]108 <li><a href="#intg_compiler_used">The '<tt>llvm.compiler.used</tt>'
109 Global Variable</a></li>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]110 <li><a href="#intg_global_ctors">The '<tt>llvm.global_ctors</tt>'
111 Global Variable</a></li>
112 <li><a href="#intg_global_dtors">The '<tt>llvm.global_dtors</tt>'
113 Global Variable</a></li>
114 </ol>
115 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]116 <li><a href="#instref">Instruction Reference</a>
117 <ol>
118 <li><a href="#terminators">Terminator Instructions</a>
119 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]120 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
121 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]122 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]123 <li><a href="#i_indirectbr">'<tt>indirectbr</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]124 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]125 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]126 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]127 </ol>
128 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]129 <li><a href="#binaryops">Binary Operations</a>
130 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]131 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]132 <li><a href="#i_fadd">'<tt>fadd</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]133 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]134 <li><a href="#i_fsub">'<tt>fsub</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]135 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]136 <li><a href="#i_fmul">'<tt>fmul</tt>' Instruction</a></li>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]137 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
138 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
139 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]140 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
141 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
142 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]143 </ol>
144 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]145 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
146 <ol>
Reid Spencer2ab01932007-02-02 13:57:07 +0000[diff] [blame]147 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
148 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
149 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]150 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]151 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]152 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]153 </ol>
154 </li>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]155 <li><a href="#vectorops">Vector Operations</a>
156 <ol>
157 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
158 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
159 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]160 </ol>
161 </li>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]162 <li><a href="#aggregateops">Aggregate Operations</a>
163 <ol>
164 <li><a href="#i_extractvalue">'<tt>extractvalue</tt>' Instruction</a></li>
165 <li><a href="#i_insertvalue">'<tt>insertvalue</tt>' Instruction</a></li>
166 </ol>
167 </li>
Chris Lattner6ab66722006-08-15 00:45:58 +0000[diff] [blame]168 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]169 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]170 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
Robert Bocchino820bc75b2006-02-17 21:18:08 +0000[diff] [blame]171 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
172 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
173 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]174 </ol>
175 </li>
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]176 <li><a href="#convertops">Conversion Operations</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]177 <ol>
178 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
179 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
180 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
181 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
182 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]183 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
184 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
185 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
186 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]187 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
188 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]189 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]190 </ol>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]191 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]192 <li><a href="#otherops">Other Operations</a>
193 <ol>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]194 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
195 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]196 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]197 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]198 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Chris Lattner33337472006-01-13 23:26:01 +0000[diff] [blame]199 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]200 </ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]201 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]202 </ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]203 </li>
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]204 <li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]205 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]206 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
207 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]208 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
209 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
210 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]211 </ol>
212 </li>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]213 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
214 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]215 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
216 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
217 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]218 </ol>
219 </li>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]220 <li><a href="#int_codegen">Code Generator Intrinsics</a>
221 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]222 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
223 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
224 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
225 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
226 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
227 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
Dan Gohmane58f7b32010-05-26 21:56:15 +0000[diff] [blame]228 <li><a href="#int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswellaa1c3c12004-04-09 16:43:20 +0000[diff] [blame]229 </ol>
230 </li>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]231 <li><a href="#int_libc">Standard C Library Intrinsics</a>
232 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]233 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
234 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
235 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
236 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
237 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]238 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
239 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
240 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]241 </ol>
242 </li>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]243 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]244 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]245 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattnerb748c672006-01-16 22:34:14 +0000[diff] [blame]246 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
247 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
248 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]249 </ol>
250 </li>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]251 <li><a href="#int_overflow">Arithmetic with Overflow Intrinsics</a>
252 <ol>
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]253 <li><a href="#int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt> Intrinsics</a></li>
254 <li><a href="#int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt> Intrinsics</a></li>
255 <li><a href="#int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt> Intrinsics</a></li>
256 <li><a href="#int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt> Intrinsics</a></li>
257 <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]258 <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]259 </ol>
260 </li>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]261 <li><a href="#int_fp16">Half Precision Floating Point Intrinsics</a>
262 <ol>
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]263 <li><a href="#int_convert_to_fp16">'<tt>llvm.convert.to.fp16</tt>' Intrinsic</a></li>
264 <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]265 </ol>
266 </li>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]267 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]268 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sands86e01192007-09-11 14:10:23 +0000[diff] [blame]269 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]270 <ol>
271 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]272 </ol>
273 </li>
Bill Wendlingf85850f2008-11-18 22:10:53 +0000[diff] [blame]274 <li><a href="#int_atomics">Atomic intrinsics</a>
275 <ol>
276 <li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
277 <li><a href="#int_atomic_cmp_swap"><tt>llvm.atomic.cmp.swap</tt></a></li>
278 <li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
279 <li><a href="#int_atomic_load_add"><tt>llvm.atomic.load.add</tt></a></li>
280 <li><a href="#int_atomic_load_sub"><tt>llvm.atomic.load.sub</tt></a></li>
281 <li><a href="#int_atomic_load_and"><tt>llvm.atomic.load.and</tt></a></li>
282 <li><a href="#int_atomic_load_nand"><tt>llvm.atomic.load.nand</tt></a></li>
283 <li><a href="#int_atomic_load_or"><tt>llvm.atomic.load.or</tt></a></li>
284 <li><a href="#int_atomic_load_xor"><tt>llvm.atomic.load.xor</tt></a></li>
285 <li><a href="#int_atomic_load_max"><tt>llvm.atomic.load.max</tt></a></li>
286 <li><a href="#int_atomic_load_min"><tt>llvm.atomic.load.min</tt></a></li>
287 <li><a href="#int_atomic_load_umax"><tt>llvm.atomic.load.umax</tt></a></li>
288 <li><a href="#int_atomic_load_umin"><tt>llvm.atomic.load.umin</tt></a></li>
289 </ol>
290 </li>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]291 <li><a href="#int_memorymarkers">Memory Use Markers</a>
292 <ol>
293 <li><a href="#int_lifetime_start"><tt>llvm.lifetime.start</tt></a></li>
294 <li><a href="#int_lifetime_end"><tt>llvm.lifetime.end</tt></a></li>
295 <li><a href="#int_invariant_start"><tt>llvm.invariant.start</tt></a></li>
296 <li><a href="#int_invariant_end"><tt>llvm.invariant.end</tt></a></li>
297 </ol>
298 </li>
Reid Spencer5b2cb0f2007-07-20 19:59:11 +0000[diff] [blame]299 <li><a href="#int_general">General intrinsics</a>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]300 <ol>
Reid Spencer5b2cb0f2007-07-20 19:59:11 +0000[diff] [blame]301 <li><a href="#int_var_annotation">
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]302 '<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]303 <li><a href="#int_annotation">
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]304 '<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]305 <li><a href="#int_trap">
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]306 '<tt>llvm.trap</tt>' Intrinsic</a></li>
307 <li><a href="#int_stackprotector">
308 '<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]309 <li><a href="#int_objectsize">
310 '<tt>llvm.objectsize</tt>' Intrinsic</a></li>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]311 </ol>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]312 </li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]313 </ol>
314 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]315</ol>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]316
317<div class="doc_author">
318 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
319 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]320</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]321
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]322<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]323<div class="doc_section"> <a name="abstract">Abstract </a></div>
324<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]325
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]326<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]327
328<p>This document is a reference manual for the LLVM assembly language. LLVM is
329 a Static Single Assignment (SSA) based representation that provides type
330 safety, low-level operations, flexibility, and the capability of representing
331 'all' high-level languages cleanly. It is the common code representation
332 used throughout all phases of the LLVM compilation strategy.</p>
333
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]334</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]335
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]336<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]337<div class="doc_section"> <a name="introduction">Introduction</a> </div>
338<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]339
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]340<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]341
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]342<p>The LLVM code representation is designed to be used in three different forms:
343 as an in-memory compiler IR, as an on-disk bitcode representation (suitable
344 for fast loading by a Just-In-Time compiler), and as a human readable
345 assembly language representation. This allows LLVM to provide a powerful
346 intermediate representation for efficient compiler transformations and
347 analysis, while providing a natural means to debug and visualize the
348 transformations. The three different forms of LLVM are all equivalent. This
349 document describes the human readable representation and notation.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]350
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]351<p>The LLVM representation aims to be light-weight and low-level while being
352 expressive, typed, and extensible at the same time. It aims to be a
353 "universal IR" of sorts, by being at a low enough level that high-level ideas
354 may be cleanly mapped to it (similar to how microprocessors are "universal
355 IR's", allowing many source languages to be mapped to them). By providing
356 type information, LLVM can be used as the target of optimizations: for
357 example, through pointer analysis, it can be proven that a C automatic
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]358 variable is never accessed outside of the current function, allowing it to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]359 be promoted to a simple SSA value instead of a memory location.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]360
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]361</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]362
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]363<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]364<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]365
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]366<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]367
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]368<p>It is important to note that this document describes 'well formed' LLVM
369 assembly language. There is a difference between what the parser accepts and
370 what is considered 'well formed'. For example, the following instruction is
371 syntactically okay, but not well formed:</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]372
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]373<div class="doc_code">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]374<pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]375%x = <a href="#i_add">add</a> i32 1, %x
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]376</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]377</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]378
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]379<p>because the definition of <tt>%x</tt> does not dominate all of its uses. The
380 LLVM infrastructure provides a verification pass that may be used to verify
381 that an LLVM module is well formed. This pass is automatically run by the
382 parser after parsing input assembly and by the optimizer before it outputs
383 bitcode. The violations pointed out by the verifier pass indicate bugs in
384 transformation passes or input to the parser.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]385
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]386</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]387
Chris Lattner87a3dbe2007-10-03 17:34:29 +0000[diff] [blame]388<!-- Describe the typesetting conventions here. -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]389
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]390<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]391<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]392<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]393
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]394<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]395
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]396<p>LLVM identifiers come in two basic types: global and local. Global
397 identifiers (functions, global variables) begin with the <tt>'@'</tt>
398 character. Local identifiers (register names, types) begin with
399 the <tt>'%'</tt> character. Additionally, there are three different formats
400 for identifiers, for different purposes:</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]401
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]402<ol>
Reid Spencerb23b65f2007-08-07 14:34:28 +0000[diff] [blame]403 <li>Named values are represented as a string of characters with their prefix.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]404 For example, <tt>%foo</tt>, <tt>@DivisionByZero</tt>,
405 <tt>%a.really.long.identifier</tt>. The actual regular expression used is
406 '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'. Identifiers which require
407 other characters in their names can be surrounded with quotes. Special
408 characters may be escaped using <tt>"\xx"</tt> where <tt>xx</tt> is the
409 ASCII code for the character in hexadecimal. In this way, any character
410 can be used in a name value, even quotes themselves.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]411
Reid Spencerb23b65f2007-08-07 14:34:28 +0000[diff] [blame]412 <li>Unnamed values are represented as an unsigned numeric value with their
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]413 prefix. For example, <tt>%12</tt>, <tt>@2</tt>, <tt>%44</tt>.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]414
Reid Spencer8f08d802004-12-09 18:02:53 +0000[diff] [blame]415 <li>Constants, which are described in a <a href="#constants">section about
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]416 constants</a>, below.</li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]417</ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]418
Reid Spencerb23b65f2007-08-07 14:34:28 +0000[diff] [blame]419<p>LLVM requires that values start with a prefix for two reasons: Compilers
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]420 don't need to worry about name clashes with reserved words, and the set of
421 reserved words may be expanded in the future without penalty. Additionally,
422 unnamed identifiers allow a compiler to quickly come up with a temporary
423 variable without having to avoid symbol table conflicts.</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]424
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]425<p>Reserved words in LLVM are very similar to reserved words in other
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]426 languages. There are keywords for different opcodes
427 ('<tt><a href="#i_add">add</a></tt>',
428 '<tt><a href="#i_bitcast">bitcast</a></tt>',
429 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names
430 ('<tt><a href="#t_void">void</a></tt>',
431 '<tt><a href="#t_primitive">i32</a></tt>', etc...), and others. These
432 reserved words cannot conflict with variable names, because none of them
433 start with a prefix character (<tt>'%'</tt> or <tt>'@'</tt>).</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]434
435<p>Here is an example of LLVM code to multiply the integer variable
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]436 '<tt>%X</tt>' by 8:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]437
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]438<p>The easy way:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]439
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]440<div class="doc_code">
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]441<pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]442%result = <a href="#i_mul">mul</a> i32 %X, 8
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]443</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]444</div>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]445
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]446<p>After strength reduction:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]447
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]448<div class="doc_code">
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]449<pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]450%result = <a href="#i_shl">shl</a> i32 %X, i8 3
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]451</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]452</div>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]453
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]454<p>And the hard way:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]455
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]456<div class="doc_code">
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]457<pre>
Gabor Greifbd0328f2009-10-28 13:05:07 +0000[diff] [blame]458%0 = <a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
459%1 = <a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]460%result = <a href="#i_add">add</a> i32 %1, %1
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]461</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]462</div>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]463
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]464<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several important
465 lexical features of LLVM:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]466
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]467<ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]468 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]469 line.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]470
471 <li>Unnamed temporaries are created when the result of a computation is not
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]472 assigned to a named value.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]473
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]474 <li>Unnamed temporaries are numbered sequentially</li>
475</ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]476
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]477<p>It also shows a convention that we follow in this document. When
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]478 demonstrating instructions, we will follow an instruction with a comment that
479 defines the type and name of value produced. Comments are shown in italic
480 text.</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]481
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]482</div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]483
484<!-- *********************************************************************** -->
485<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
486<!-- *********************************************************************** -->
487
488<!-- ======================================================================= -->
489<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
490</div>
491
492<div class="doc_text">
493
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]494<p>LLVM programs are composed of "Module"s, each of which is a translation unit
495 of the input programs. Each module consists of functions, global variables,
496 and symbol table entries. Modules may be combined together with the LLVM
497 linker, which merges function (and global variable) definitions, resolves
498 forward declarations, and merges symbol table entries. Here is an example of
499 the "hello world" module:</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]500
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]501<div class="doc_code">
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]502<pre>
503<i>; Declare the string constant as a global constant.</i>
504<a href="#identifiers">@.LC0</a> = <a href="#linkage_internal">internal</a> <a href="#globalvars">constant</a> <a href="#t_array">[13 x i8]</a> c"hello world\0A\00" <i>; [13 x i8]*</i>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]505
506<i>; External declaration of the puts function</i>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]507<a href="#functionstructure">declare</a> i32 @puts(i8*) <i>; i32 (i8*)* </i>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]508
509<i>; Definition of main function</i>
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]510define i32 @main() { <i>; i32()* </i>
511 <i>; Convert [13 x i8]* to i8 *...</i>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]512 %cast210 = <a href="#i_getelementptr">getelementptr</a> [13 x i8]* @.LC0, i64 0, i64 0 <i>; i8*</i>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]513
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]514 <i>; Call puts function to write out the string to stdout.</i>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]515 <a href="#i_call">call</a> i32 @puts(i8* %cast210) <i>; i32</i>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]516 <a href="#i_ret">ret</a> i32 0<br>}
517
518<i>; Named metadata</i>
519!1 = metadata !{i32 41}
520!foo = !{!1, null}
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]521</pre>
522</div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]523
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]524<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]525 "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>" function,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]526 a <a href="#functionstructure">function definition</a> for
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]527 "<tt>main</tt>" and <a href="#namedmetadatastructure">named metadata</a>
528 "<tt>foo"</tt>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]529
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]530<p>In general, a module is made up of a list of global values, where both
531 functions and global variables are global values. Global values are
532 represented by a pointer to a memory location (in this case, a pointer to an
533 array of char, and a pointer to a function), and have one of the
534 following <a href="#linkage">linkage types</a>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]535
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]536</div>
537
538<!-- ======================================================================= -->
539<div class="doc_subsection">
540 <a name="linkage">Linkage Types</a>
541</div>
542
543<div class="doc_text">
544
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]545<p>All Global Variables and Functions have one of the following types of
546 linkage:</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]547
548<dl>
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]549 <dt><tt><b><a name="linkage_private">private</a></b></tt></dt>
Bill Wendling03bcd6e2010-07-01 21:55:59 +0000[diff] [blame]550 <dd>Global values with "<tt>private</tt>" linkage are only directly accessible
551 by objects in the current module. In particular, linking code into a
552 module with an private global value may cause the private to be renamed as
553 necessary to avoid collisions. Because the symbol is private to the
554 module, all references can be updated. This doesn't show up in any symbol
555 table in the object file.</dd>
Rafael Espindola6de96a12009-01-15 20:18:42 +0000[diff] [blame]556
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]557 <dt><tt><b><a name="linkage_linker_private">linker_private</a></b></tt></dt>
Bill Wendling03bcd6e2010-07-01 21:55:59 +0000[diff] [blame]558 <dd>Similar to <tt>private</tt>, but the symbol is passed through the
559 assembler and evaluated by the linker. Unlike normal strong symbols, they
560 are removed by the linker from the final linked image (executable or
561 dynamic library).</dd>
562
563 <dt><tt><b><a name="linkage_linker_private_weak">linker_private_weak</a></b></tt></dt>
564 <dd>Similar to "<tt>linker_private</tt>", but the symbol is weak. Note that
565 <tt>linker_private_weak</tt> symbols are subject to coalescing by the
566 linker. The symbols are removed by the linker from the final linked image
567 (executable or dynamic library).</dd>
Bill Wendlinga3c6f6b2009-07-20 01:03:30 +0000[diff] [blame]568
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]569 <dt><tt><b><a name="linkage_internal">internal</a></b></tt></dt>
Bill Wendling36321712010-06-29 22:34:52 +0000[diff] [blame]570 <dd>Similar to private, but the value shows as a local symbol
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]571 (<tt>STB_LOCAL</tt> in the case of ELF) in the object file. This
572 corresponds to the notion of the '<tt>static</tt>' keyword in C.</dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]573
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]574 <dt><tt><b><a name="linkage_available_externally">available_externally</a></b></tt></dt>
Chris Lattner184f1be2009-04-13 05:44:34 +0000[diff] [blame]575 <dd>Globals with "<tt>available_externally</tt>" linkage are never emitted
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]576 into the object file corresponding to the LLVM module. They exist to
577 allow inlining and other optimizations to take place given knowledge of
578 the definition of the global, which is known to be somewhere outside the
579 module. Globals with <tt>available_externally</tt> linkage are allowed to
580 be discarded at will, and are otherwise the same as <tt>linkonce_odr</tt>.
581 This linkage type is only allowed on definitions, not declarations.</dd>
Chris Lattner184f1be2009-04-13 05:44:34 +0000[diff] [blame]582
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]583 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt></dt>
Chris Lattnere20b4702007-01-14 06:51:48 +0000[diff] [blame]584 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
Chris Lattner0de4caa2010-01-09 19:15:14 +0000[diff] [blame]585 the same name when linkage occurs. This can be used to implement
586 some forms of inline functions, templates, or other code which must be
587 generated in each translation unit that uses it, but where the body may
588 be overridden with a more definitive definition later. Unreferenced
589 <tt>linkonce</tt> globals are allowed to be discarded. Note that
590 <tt>linkonce</tt> linkage does not actually allow the optimizer to
591 inline the body of this function into callers because it doesn't know if
592 this definition of the function is the definitive definition within the
593 program or whether it will be overridden by a stronger definition.
594 To enable inlining and other optimizations, use "<tt>linkonce_odr</tt>"
595 linkage.</dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]596
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]597 <dt><tt><b><a name="linkage_weak">weak</a></b></tt></dt>
Chris Lattnerd0554882009-08-05 05:21:07 +0000[diff] [blame]598 <dd>"<tt>weak</tt>" linkage has the same merging semantics as
599 <tt>linkonce</tt> linkage, except that unreferenced globals with
600 <tt>weak</tt> linkage may not be discarded. This is used for globals that
601 are declared "weak" in C source code.</dd>
602
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]603 <dt><tt><b><a name="linkage_common">common</a></b></tt></dt>
Chris Lattnerd0554882009-08-05 05:21:07 +0000[diff] [blame]604 <dd>"<tt>common</tt>" linkage is most similar to "<tt>weak</tt>" linkage, but
605 they are used for tentative definitions in C, such as "<tt>int X;</tt>" at
606 global scope.
607 Symbols with "<tt>common</tt>" linkage are merged in the same way as
608 <tt>weak symbols</tt>, and they may not be deleted if unreferenced.
Chris Lattner0aff0b22009-08-05 05:41:44 +0000[diff] [blame]609 <tt>common</tt> symbols may not have an explicit section,
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]610 must have a zero initializer, and may not be marked '<a
Chris Lattner0aff0b22009-08-05 05:41:44 +0000[diff] [blame]611 href="#globalvars"><tt>constant</tt></a>'. Functions and aliases may not
612 have common linkage.</dd>
Chris Lattnerd0554882009-08-05 05:21:07 +0000[diff] [blame]613
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]614
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]615 <dt><tt><b><a name="linkage_appending">appending</a></b></tt></dt>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]616 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]617 pointer to array type. When two global variables with appending linkage
618 are linked together, the two global arrays are appended together. This is
619 the LLVM, typesafe, equivalent of having the system linker append together
620 "sections" with identical names when .o files are linked.</dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]621
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]622 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]623 <dd>The semantics of this linkage follow the ELF object file model: the symbol
624 is weak until linked, if not linked, the symbol becomes null instead of
625 being an undefined reference.</dd>
Anton Korobeynikova0554d92007-01-12 19:20:47 +0000[diff] [blame]626
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]627 <dt><tt><b><a name="linkage_linkonce_odr">linkonce_odr</a></b></tt></dt>
628 <dt><tt><b><a name="linkage_weak_odr">weak_odr</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]629 <dd>Some languages allow differing globals to be merged, such as two functions
630 with different semantics. Other languages, such as <tt>C++</tt>, ensure
Bill Wendling03bcd6e2010-07-01 21:55:59 +0000[diff] [blame]631 that only equivalent globals are ever merged (the "one definition rule"
632 &mdash; "ODR"). Such languages can use the <tt>linkonce_odr</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]633 and <tt>weak_odr</tt> linkage types to indicate that the global will only
634 be merged with equivalent globals. These linkage types are otherwise the
635 same as their non-<tt>odr</tt> versions.</dd>
Duncan Sands12da8ce2009-03-07 15:45:40 +0000[diff] [blame]636
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]637 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]638 <dd>If none of the above identifiers are used, the global is externally
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]639 visible, meaning that it participates in linkage and can be used to
640 resolve external symbol references.</dd>
Reid Spencer7972c472007-04-11 23:49:50 +0000[diff] [blame]641</dl>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]642
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]643<p>The next two types of linkage are targeted for Microsoft Windows platform
644 only. They are designed to support importing (exporting) symbols from (to)
645 DLLs (Dynamic Link Libraries).</p>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]646
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]647<dl>
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]648 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt></dt>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]649 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]650 or variable via a global pointer to a pointer that is set up by the DLL
651 exporting the symbol. On Microsoft Windows targets, the pointer name is
652 formed by combining <code>__imp_</code> and the function or variable
653 name.</dd>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]654
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]655 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt></dt>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]656 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]657 pointer to a pointer in a DLL, so that it can be referenced with the
658 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
659 name is formed by combining <code>__imp_</code> and the function or
660 variable name.</dd>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]661</dl>
662
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]663<p>For example, since the "<tt>.LC0</tt>" variable is defined to be internal, if
664 another module defined a "<tt>.LC0</tt>" variable and was linked with this
665 one, one of the two would be renamed, preventing a collision. Since
666 "<tt>main</tt>" and "<tt>puts</tt>" are external (i.e., lacking any linkage
667 declarations), they are accessible outside of the current module.</p>
668
669<p>It is illegal for a function <i>declaration</i> to have any linkage type
670 other than "externally visible", <tt>dllimport</tt>
671 or <tt>extern_weak</tt>.</p>
672
Duncan Sands12da8ce2009-03-07 15:45:40 +0000[diff] [blame]673<p>Aliases can have only <tt>external</tt>, <tt>internal</tt>, <tt>weak</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]674 or <tt>weak_odr</tt> linkages.</p>
675
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]676</div>
677
678<!-- ======================================================================= -->
679<div class="doc_subsection">
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]680 <a name="callingconv">Calling Conventions</a>
681</div>
682
683<div class="doc_text">
684
685<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]686 and <a href="#i_invoke">invokes</a> can all have an optional calling
687 convention specified for the call. The calling convention of any pair of
688 dynamic caller/callee must match, or the behavior of the program is
689 undefined. The following calling conventions are supported by LLVM, and more
690 may be added in the future:</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]691
692<dl>
693 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]694 <dd>This calling convention (the default if no other calling convention is
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]695 specified) matches the target C calling conventions. This calling
696 convention supports varargs function calls and tolerates some mismatch in
697 the declared prototype and implemented declaration of the function (as
698 does normal C).</dd>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]699
700 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]701 <dd>This calling convention attempts to make calls as fast as possible
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]702 (e.g. by passing things in registers). This calling convention allows the
703 target to use whatever tricks it wants to produce fast code for the
704 target, without having to conform to an externally specified ABI
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]705 (Application Binary Interface).
706 <a href="CodeGenerator.html#tailcallopt">Tail calls can only be optimized
Chris Lattnera179e4d2010-03-11 00:22:57 +0000[diff] [blame]707 when this or the GHC convention is used.</a> This calling convention
708 does not support varargs and requires the prototype of all callees to
709 exactly match the prototype of the function definition.</dd>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]710
711 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]712 <dd>This calling convention attempts to make code in the caller as efficient
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]713 as possible under the assumption that the call is not commonly executed.
714 As such, these calls often preserve all registers so that the call does
715 not break any live ranges in the caller side. This calling convention
716 does not support varargs and requires the prototype of all callees to
717 exactly match the prototype of the function definition.</dd>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]718
Chris Lattnera179e4d2010-03-11 00:22:57 +0000[diff] [blame]719 <dt><b>"<tt>cc <em>10</em></tt>" - GHC convention</b>:</dt>
720 <dd>This calling convention has been implemented specifically for use by the
721 <a href="http://www.haskell.org/ghc">Glasgow Haskell Compiler (GHC)</a>.
722 It passes everything in registers, going to extremes to achieve this by
723 disabling callee save registers. This calling convention should not be
724 used lightly but only for specific situations such as an alternative to
725 the <em>register pinning</em> performance technique often used when
726 implementing functional programming languages.At the moment only X86
727 supports this convention and it has the following limitations:
728 <ul>
729 <li>On <em>X86-32</em> only supports up to 4 bit type parameters. No
730 floating point types are supported.</li>
731 <li>On <em>X86-64</em> only supports up to 10 bit type parameters and
732 6 floating point parameters.</li>
733 </ul>
734 This calling convention supports
735 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> but
736 requires both the caller and callee are using it.
737 </dd>
738
Chris Lattner573f64e2005-05-07 01:46:40 +0000[diff] [blame]739 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]740 <dd>Any calling convention may be specified by number, allowing
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]741 target-specific calling conventions to be used. Target specific calling
742 conventions start at 64.</dd>
Chris Lattner573f64e2005-05-07 01:46:40 +0000[diff] [blame]743</dl>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]744
745<p>More calling conventions can be added/defined on an as-needed basis, to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]746 support Pascal conventions or any other well-known target-independent
747 convention.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]748
749</div>
750
751<!-- ======================================================================= -->
752<div class="doc_subsection">
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]753 <a name="visibility">Visibility Styles</a>
754</div>
755
756<div class="doc_text">
757
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]758<p>All Global Variables and Functions have one of the following visibility
759 styles:</p>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]760
761<dl>
762 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]763 <dd>On targets that use the ELF object file format, default visibility means
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]764 that the declaration is visible to other modules and, in shared libraries,
765 means that the declared entity may be overridden. On Darwin, default
766 visibility means that the declaration is visible to other modules. Default
767 visibility corresponds to "external linkage" in the language.</dd>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]768
769 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]770 <dd>Two declarations of an object with hidden visibility refer to the same
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]771 object if they are in the same shared object. Usually, hidden visibility
772 indicates that the symbol will not be placed into the dynamic symbol
773 table, so no other module (executable or shared library) can reference it
774 directly.</dd>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]775
Anton Korobeynikov39f3cff2007-04-29 18:35:00 +0000[diff] [blame]776 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
Anton Korobeynikov39f3cff2007-04-29 18:35:00 +0000[diff] [blame]777 <dd>On ELF, protected visibility indicates that the symbol will be placed in
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]778 the dynamic symbol table, but that references within the defining module
779 will bind to the local symbol. That is, the symbol cannot be overridden by
780 another module.</dd>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]781</dl>
782
783</div>
784
785<!-- ======================================================================= -->
786<div class="doc_subsection">
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]787 <a name="namedtypes">Named Types</a>
788</div>
789
790<div class="doc_text">
791
792<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]793 it easier to read the IR and make the IR more condensed (particularly when
794 recursive types are involved). An example of a name specification is:</p>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]795
796<div class="doc_code">
797<pre>
798%mytype = type { %mytype*, i32 }
799</pre>
800</div>
801
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]802<p>You may give a name to any <a href="#typesystem">type</a> except
803 "<a href="t_void">void</a>". Type name aliases may be used anywhere a type
804 is expected with the syntax "%mytype".</p>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]805
806<p>Note that type names are aliases for the structural type that they indicate,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]807 and that you can therefore specify multiple names for the same type. This
808 often leads to confusing behavior when dumping out a .ll file. Since LLVM IR
809 uses structural typing, the name is not part of the type. When printing out
810 LLVM IR, the printer will pick <em>one name</em> to render all types of a
811 particular shape. This means that if you have code where two different
812 source types end up having the same LLVM type, that the dumper will sometimes
813 print the "wrong" or unexpected type. This is an important design point and
814 isn't going to change.</p>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]815
816</div>
817
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]818<!-- ======================================================================= -->
819<div class="doc_subsection">
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]820 <a name="globalvars">Global Variables</a>
821</div>
822
823<div class="doc_text">
824
Chris Lattner5d5aede2005-02-12 19:30:21 +0000[diff] [blame]825<p>Global variables define regions of memory allocated at compilation time
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]826 instead of run-time. Global variables may optionally be initialized, may
827 have an explicit section to be placed in, and may have an optional explicit
828 alignment specified. A variable may be defined as "thread_local", which
829 means that it will not be shared by threads (each thread will have a
830 separated copy of the variable). A variable may be defined as a global
831 "constant," which indicates that the contents of the variable
832 will <b>never</b> be modified (enabling better optimization, allowing the
833 global data to be placed in the read-only section of an executable, etc).
834 Note that variables that need runtime initialization cannot be marked
835 "constant" as there is a store to the variable.</p>
Chris Lattner5d5aede2005-02-12 19:30:21 +0000[diff] [blame]836
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]837<p>LLVM explicitly allows <em>declarations</em> of global variables to be marked
838 constant, even if the final definition of the global is not. This capability
839 can be used to enable slightly better optimization of the program, but
840 requires the language definition to guarantee that optimizations based on the
841 'constantness' are valid for the translation units that do not include the
842 definition.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]843
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]844<p>As SSA values, global variables define pointer values that are in scope
845 (i.e. they dominate) all basic blocks in the program. Global variables
846 always define a pointer to their "content" type because they describe a
847 region of memory, and all memory objects in LLVM are accessed through
848 pointers.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]849
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]850<p>A global variable may be declared to reside in a target-specific numbered
851 address space. For targets that support them, address spaces may affect how
852 optimizations are performed and/or what target instructions are used to
853 access the variable. The default address space is zero. The address space
854 qualifier must precede any other attributes.</p>
Christopher Lamb308121c2007-12-11 09:31:00 +0000[diff] [blame]855
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]856<p>LLVM allows an explicit section to be specified for globals. If the target
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]857 supports it, it will emit globals to the section specified.</p>
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]858
Chris Lattner78e00bc2010-04-28 00:13:42 +0000[diff] [blame]859<p>An explicit alignment may be specified for a global, which must be a power
860 of 2. If not present, or if the alignment is set to zero, the alignment of
861 the global is set by the target to whatever it feels convenient. If an
862 explicit alignment is specified, the global is forced to have exactly that
Chris Lattner4bd85e42010-04-28 00:31:12 +0000[diff] [blame]863 alignment. Targets and optimizers are not allowed to over-align the global
864 if the global has an assigned section. In this case, the extra alignment
865 could be observable: for example, code could assume that the globals are
866 densely packed in their section and try to iterate over them as an array,
867 alignment padding would break this iteration.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]868
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]869<p>For example, the following defines a global in a numbered address space with
870 an initializer, section, and alignment:</p>
Chris Lattner5760c502007-01-14 00:27:09 +0000[diff] [blame]871
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]872<div class="doc_code">
Chris Lattner5760c502007-01-14 00:27:09 +0000[diff] [blame]873<pre>
Dan Gohmanaaa679b2009-01-11 00:40:00 +0000[diff] [blame]874@G = addrspace(5) constant float 1.0, section "foo", align 4
Chris Lattner5760c502007-01-14 00:27:09 +0000[diff] [blame]875</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]876</div>
Chris Lattner5760c502007-01-14 00:27:09 +0000[diff] [blame]877
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]878</div>
879
880
881<!-- ======================================================================= -->
882<div class="doc_subsection">
883 <a name="functionstructure">Functions</a>
884</div>
885
886<div class="doc_text">
887
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]888<p>LLVM function definitions consist of the "<tt>define</tt>" keyword, an
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]889 optional <a href="#linkage">linkage type</a>, an optional
890 <a href="#visibility">visibility style</a>, an optional
891 <a href="#callingconv">calling convention</a>, a return type, an optional
892 <a href="#paramattrs">parameter attribute</a> for the return type, a function
893 name, a (possibly empty) argument list (each with optional
894 <a href="#paramattrs">parameter attributes</a>), optional
895 <a href="#fnattrs">function attributes</a>, an optional section, an optional
896 alignment, an optional <a href="#gc">garbage collector name</a>, an opening
897 curly brace, a list of basic blocks, and a closing curly brace.</p>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]898
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]899<p>LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
900 optional <a href="#linkage">linkage type</a>, an optional
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]901 <a href="#visibility">visibility style</a>, an optional
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]902 <a href="#callingconv">calling convention</a>, a return type, an optional
903 <a href="#paramattrs">parameter attribute</a> for the return type, a function
904 name, a possibly empty list of arguments, an optional alignment, and an
905 optional <a href="#gc">garbage collector name</a>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]906
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]907<p>A function definition contains a list of basic blocks, forming the CFG
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]908 (Control Flow Graph) for the function. Each basic block may optionally start
909 with a label (giving the basic block a symbol table entry), contains a list
910 of instructions, and ends with a <a href="#terminators">terminator</a>
911 instruction (such as a branch or function return).</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]912
Chris Lattnera59fb102007-06-08 16:52:14 +0000[diff] [blame]913<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]914 executed on entrance to the function, and it is not allowed to have
915 predecessor basic blocks (i.e. there can not be any branches to the entry
916 block of a function). Because the block can have no predecessors, it also
917 cannot have any <a href="#i_phi">PHI nodes</a>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]918
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]919<p>LLVM allows an explicit section to be specified for functions. If the target
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]920 supports it, it will emit functions to the section specified.</p>
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]921
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]922<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]923 the alignment is set to zero, the alignment of the function is set by the
924 target to whatever it feels convenient. If an explicit alignment is
925 specified, the function is forced to have at least that much alignment. All
926 alignments must be a power of 2.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]927
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]928<h5>Syntax:</h5>
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]929<div class="doc_code">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]930<pre>
Chris Lattner0ae02092008-10-13 16:55:18 +0000[diff] [blame]931define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]932 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
933 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
934 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
935 [<a href="#gc">gc</a>] { ... }
936</pre>
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]937</div>
938
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]939</div>
940
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]941<!-- ======================================================================= -->
942<div class="doc_subsection">
943 <a name="aliasstructure">Aliases</a>
944</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]945
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]946<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]947
948<p>Aliases act as "second name" for the aliasee value (which can be either
949 function, global variable, another alias or bitcast of global value). Aliases
950 may have an optional <a href="#linkage">linkage type</a>, and an
951 optional <a href="#visibility">visibility style</a>.</p>
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]952
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]953<h5>Syntax:</h5>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]954<div class="doc_code">
Bill Wendling2d8b9a82007-05-29 09:42:13 +0000[diff] [blame]955<pre>
Duncan Sands7e99a942008-09-12 20:48:21 +0000[diff] [blame]956@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Bill Wendling2d8b9a82007-05-29 09:42:13 +0000[diff] [blame]957</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]958</div>
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]959
960</div>
961
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]962<!-- ======================================================================= -->
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]963<div class="doc_subsection">
964 <a name="namedmetadatastructure">Named Metadata</a>
965</div>
966
967<div class="doc_text">
968
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]969<p>Named metadata is a collection of metadata. <a href="#metadata">Metadata
970 nodes</a> (but not metadata strings) and null are the only valid operands for
971 a named metadata.</p>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]972
973<h5>Syntax:</h5>
974<div class="doc_code">
975<pre>
976!1 = metadata !{metadata !"one"}
977!name = !{null, !1}
978</pre>
979</div>
980
981</div>
982
983<!-- ======================================================================= -->
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]984<div class="doc_subsection"><a name="paramattrs">Parameter Attributes</a></div>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]985
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]986<div class="doc_text">
987
988<p>The return type and each parameter of a function type may have a set of
989 <i>parameter attributes</i> associated with them. Parameter attributes are
990 used to communicate additional information about the result or parameters of
991 a function. Parameter attributes are considered to be part of the function,
992 not of the function type, so functions with different parameter attributes
993 can have the same function type.</p>
994
995<p>Parameter attributes are simple keywords that follow the type specified. If
996 multiple parameter attributes are needed, they are space separated. For
997 example:</p>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]998
999<div class="doc_code">
1000<pre>
Nick Lewyckydac78d82009-02-15 23:06:14 +0000[diff] [blame]1001declare i32 @printf(i8* noalias nocapture, ...)
Chris Lattnerd2597d72008-10-04 18:33:34 +0000[diff] [blame]1002declare i32 @atoi(i8 zeroext)
1003declare signext i8 @returns_signed_char()
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]1004</pre>
1005</div>
1006
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1007<p>Note that any attributes for the function result (<tt>nounwind</tt>,
1008 <tt>readonly</tt>) come immediately after the argument list.</p>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1009
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1010<p>Currently, only the following parameter attributes are defined:</p>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000[diff] [blame]1011
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1012<dl>
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1013 <dt><tt><b>zeroext</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1014 <dd>This indicates to the code generator that the parameter or return value
1015 should be zero-extended to a 32-bit value by the caller (for a parameter)
1016 or the callee (for a return value).</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000[diff] [blame]1017
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1018 <dt><tt><b>signext</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1019 <dd>This indicates to the code generator that the parameter or return value
1020 should be sign-extended to a 32-bit value by the caller (for a parameter)
1021 or the callee (for a return value).</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000[diff] [blame]1022
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1023 <dt><tt><b>inreg</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1024 <dd>This indicates that this parameter or return value should be treated in a
1025 special target-dependent fashion during while emitting code for a function
1026 call or return (usually, by putting it in a register as opposed to memory,
1027 though some targets use it to distinguish between two different kinds of
1028 registers). Use of this attribute is target-specific.</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000[diff] [blame]1029
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1030 <dt><tt><b><a name="byval">byval</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1031 <dd>This indicates that the pointer parameter should really be passed by value
1032 to the function. The attribute implies that a hidden copy of the pointee
1033 is made between the caller and the callee, so the callee is unable to
1034 modify the value in the callee. This attribute is only valid on LLVM
1035 pointer arguments. It is generally used to pass structs and arrays by
1036 value, but is also valid on pointers to scalars. The copy is considered
1037 to belong to the caller not the callee (for example,
1038 <tt><a href="#readonly">readonly</a></tt> functions should not write to
1039 <tt>byval</tt> parameters). This is not a valid attribute for return
1040 values. The byval attribute also supports specifying an alignment with
1041 the align attribute. This has a target-specific effect on the code
1042 generator that usually indicates a desired alignment for the synthesized
1043 stack slot.</dd>
1044
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame^]1045 <dt><tt><b><a name="sret">sret</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1046 <dd>This indicates that the pointer parameter specifies the address of a
1047 structure that is the return value of the function in the source program.
1048 This pointer must be guaranteed by the caller to be valid: loads and
1049 stores to the structure may be assumed by the callee to not to trap. This
1050 may only be applied to the first parameter. This is not a valid attribute
1051 for return values. </dd>
1052
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame^]1053 <dt><tt><b><a name="noalias">noalias</a></b></tt></dt>
Dan Gohmandf12d082010-07-02 18:41:32 +0000[diff] [blame]1054 <dd>This indicates that pointer values
1055 <a href="#pointeraliasing"><i>based</i></a> on the argument or return
1056 value do not alias pointer values which are not <i>based</i> on it.
1057 The caller shares the responsibility with the callee for ensuring that
1058 these requirements are met.
1059 For further details, please see the discussion of the NoAlias response in
1060 <a href="AliasAnalysis.html#MustMayNo">alias analysis</a>.</dd>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1061
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame^]1062 <dt><tt><b><a name="nocapture">nocapture</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1063 <dd>This indicates that the callee does not make any copies of the pointer
1064 that outlive the callee itself. This is not a valid attribute for return
1065 values.</dd>
1066
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame^]1067 <dt><tt><b><a name="nest">nest</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1068 <dd>This indicates that the pointer parameter can be excised using the
1069 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
1070 attribute for return values.</dd>
1071</dl>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1072
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1073</div>
1074
1075<!-- ======================================================================= -->
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1076<div class="doc_subsection">
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1077 <a name="gc">Garbage Collector Names</a>
1078</div>
1079
1080<div class="doc_text">
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1081
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1082<p>Each function may specify a garbage collector name, which is simply a
1083 string:</p>
1084
1085<div class="doc_code">
1086<pre>
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1087define void @f() gc "name" { ... }
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1088</pre>
1089</div>
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1090
1091<p>The compiler declares the supported values of <i>name</i>. Specifying a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1092 collector which will cause the compiler to alter its output in order to
1093 support the named garbage collection algorithm.</p>
1094
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1095</div>
1096
1097<!-- ======================================================================= -->
1098<div class="doc_subsection">
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1099 <a name="fnattrs">Function Attributes</a>
Devang Patelcaacdba2008-09-04 23:05:13 +0000[diff] [blame]1100</div>
1101
1102<div class="doc_text">
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1103
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1104<p>Function attributes are set to communicate additional information about a
1105 function. Function attributes are considered to be part of the function, not
1106 of the function type, so functions with different parameter attributes can
1107 have the same function type.</p>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1108
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1109<p>Function attributes are simple keywords that follow the type specified. If
1110 multiple attributes are needed, they are space separated. For example:</p>
Devang Patelcaacdba2008-09-04 23:05:13 +0000[diff] [blame]1111
1112<div class="doc_code">
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1113<pre>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1114define void @f() noinline { ... }
1115define void @f() alwaysinline { ... }
1116define void @f() alwaysinline optsize { ... }
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1117define void @f() optsize { ... }
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1118</pre>
Devang Patelcaacdba2008-09-04 23:05:13 +0000[diff] [blame]1119</div>
1120
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1121<dl>
Charles Davisbe5557e2010-02-12 00:31:15 +0000[diff] [blame]1122 <dt><tt><b>alignstack(&lt;<em>n</em>&gt;)</b></tt></dt>
1123 <dd>This attribute indicates that, when emitting the prologue and epilogue,
1124 the backend should forcibly align the stack pointer. Specify the
1125 desired alignment, which must be a power of two, in parentheses.
1126
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1127 <dt><tt><b>alwaysinline</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1128 <dd>This attribute indicates that the inliner should attempt to inline this
1129 function into callers whenever possible, ignoring any active inlining size
1130 threshold for this caller.</dd>
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1131
Jakob Stoklund Olesen74bb06c2010-02-06 01:16:28 +0000[diff] [blame]1132 <dt><tt><b>inlinehint</b></tt></dt>
1133 <dd>This attribute indicates that the source code contained a hint that inlining
1134 this function is desirable (such as the "inline" keyword in C/C++). It
1135 is just a hint; it imposes no requirements on the inliner.</dd>
1136
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1137 <dt><tt><b>noinline</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1138 <dd>This attribute indicates that the inliner should never inline this
1139 function in any situation. This attribute may not be used together with
1140 the <tt>alwaysinline</tt> attribute.</dd>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1141
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1142 <dt><tt><b>optsize</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1143 <dd>This attribute suggests that optimization passes and code generator passes
1144 make choices that keep the code size of this function low, and otherwise
1145 do optimizations specifically to reduce code size.</dd>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1146
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1147 <dt><tt><b>noreturn</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1148 <dd>This function attribute indicates that the function never returns
1149 normally. This produces undefined behavior at runtime if the function
1150 ever does dynamically return.</dd>
Bill Wendlinga8130172008-11-13 01:02:51 +0000[diff] [blame]1151
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1152 <dt><tt><b>nounwind</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1153 <dd>This function attribute indicates that the function never returns with an
1154 unwind or exceptional control flow. If the function does unwind, its
1155 runtime behavior is undefined.</dd>
Bill Wendling0f5541e2008-11-26 19:07:40 +0000[diff] [blame]1156
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1157 <dt><tt><b>readnone</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1158 <dd>This attribute indicates that the function computes its result (or decides
1159 to unwind an exception) based strictly on its arguments, without
1160 dereferencing any pointer arguments or otherwise accessing any mutable
1161 state (e.g. memory, control registers, etc) visible to caller functions.
1162 It does not write through any pointer arguments
1163 (including <tt><a href="#byval">byval</a></tt> arguments) and never
1164 changes any state visible to callers. This means that it cannot unwind
1165 exceptions by calling the <tt>C++</tt> exception throwing methods, but
1166 could use the <tt>unwind</tt> instruction.</dd>
Devang Patel310fd4a2009-06-12 19:45:19 +0000[diff] [blame]1167
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1168 <dt><tt><b><a name="readonly">readonly</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1169 <dd>This attribute indicates that the function does not write through any
1170 pointer arguments (including <tt><a href="#byval">byval</a></tt>
1171 arguments) or otherwise modify any state (e.g. memory, control registers,
1172 etc) visible to caller functions. It may dereference pointer arguments
1173 and read state that may be set in the caller. A readonly function always
1174 returns the same value (or unwinds an exception identically) when called
1175 with the same set of arguments and global state. It cannot unwind an
1176 exception by calling the <tt>C++</tt> exception throwing methods, but may
1177 use the <tt>unwind</tt> instruction.</dd>
Anton Korobeynikovc8ce7b082009-07-17 18:07:26 +0000[diff] [blame]1178
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1179 <dt><tt><b><a name="ssp">ssp</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1180 <dd>This attribute indicates that the function should emit a stack smashing
1181 protector. It is in the form of a "canary"&mdash;a random value placed on
1182 the stack before the local variables that's checked upon return from the
1183 function to see if it has been overwritten. A heuristic is used to
1184 determine if a function needs stack protectors or not.<br>
1185<br>
1186 If a function that has an <tt>ssp</tt> attribute is inlined into a
1187 function that doesn't have an <tt>ssp</tt> attribute, then the resulting
1188 function will have an <tt>ssp</tt> attribute.</dd>
1189
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1190 <dt><tt><b>sspreq</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1191 <dd>This attribute indicates that the function should <em>always</em> emit a
1192 stack smashing protector. This overrides
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1193 the <tt><a href="#ssp">ssp</a></tt> function attribute.<br>
1194<br>
1195 If a function that has an <tt>sspreq</tt> attribute is inlined into a
1196 function that doesn't have an <tt>sspreq</tt> attribute or which has
1197 an <tt>ssp</tt> attribute, then the resulting function will have
1198 an <tt>sspreq</tt> attribute.</dd>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1199
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1200 <dt><tt><b>noredzone</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1201 <dd>This attribute indicates that the code generator should not use a red
1202 zone, even if the target-specific ABI normally permits it.</dd>
1203
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1204 <dt><tt><b>noimplicitfloat</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1205 <dd>This attributes disables implicit floating point instructions.</dd>
1206
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1207 <dt><tt><b>naked</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1208 <dd>This attribute disables prologue / epilogue emission for the function.
1209 This can have very system-specific consequences.</dd>
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1210</dl>
1211
Devang Patelcaacdba2008-09-04 23:05:13 +0000[diff] [blame]1212</div>
1213
1214<!-- ======================================================================= -->
1215<div class="doc_subsection">
Chris Lattner93564892006-04-08 04:40:53 +0000[diff] [blame]1216 <a name="moduleasm">Module-Level Inline Assembly</a>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1217</div>
1218
1219<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1220
1221<p>Modules may contain "module-level inline asm" blocks, which corresponds to
1222 the GCC "file scope inline asm" blocks. These blocks are internally
1223 concatenated by LLVM and treated as a single unit, but may be separated in
1224 the <tt>.ll</tt> file if desired. The syntax is very simple:</p>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1225
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]1226<div class="doc_code">
1227<pre>
1228module asm "inline asm code goes here"
1229module asm "more can go here"
1230</pre>
1231</div>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1232
1233<p>The strings can contain any character by escaping non-printable characters.
1234 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]1235 for the number.</p>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1236
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1237<p>The inline asm code is simply printed to the machine code .s file when
1238 assembly code is generated.</p>
1239
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1240</div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1241
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1242<!-- ======================================================================= -->
1243<div class="doc_subsection">
1244 <a name="datalayout">Data Layout</a>
1245</div>
1246
1247<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1248
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1249<p>A module may specify a target specific data layout string that specifies how
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1250 data is to be laid out in memory. The syntax for the data layout is
1251 simply:</p>
1252
1253<div class="doc_code">
1254<pre>
1255target datalayout = "<i>layout specification</i>"
1256</pre>
1257</div>
1258
1259<p>The <i>layout specification</i> consists of a list of specifications
1260 separated by the minus sign character ('-'). Each specification starts with
1261 a letter and may include other information after the letter to define some
1262 aspect of the data layout. The specifications accepted are as follows:</p>
1263
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1264<dl>
1265 <dt><tt>E</tt></dt>
1266 <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]1267 bits with the most significance have the lowest address location.</dd>
1268
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1269 <dt><tt>e</tt></dt>
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]1270 <dd>Specifies that the target lays out data in little-endian form. That is,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1271 the bits with the least significance have the lowest address
1272 location.</dd>
1273
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1274 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1275 <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]1276 <i>preferred</i> alignments. All sizes are in bits. Specifying
1277 the <i>pref</i> alignment is optional. If omitted, the
1278 preceding <tt>:</tt> should be omitted too.</dd>
1279
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1280 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1281 <dd>This specifies the alignment for an integer type of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1282 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1283
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1284 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1285 <dd>This specifies the alignment for a vector type of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1286 <i>size</i>.</dd>
1287
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1288 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1289 <dd>This specifies the alignment for a floating point type of a given bit
Dale Johannesence522852010-05-28 18:54:47 +0000[diff] [blame]1290 <i>size</i>. Only values of <i>size</i> that are supported by the target
1291 will work. 32 (float) and 64 (double) are supported on all targets;
1292 80 or 128 (different flavors of long double) are also supported on some
1293 targets.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1294
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1295 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1296 <dd>This specifies the alignment for an aggregate type of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1297 <i>size</i>.</dd>
1298
Daniel Dunbar7921a592009-06-08 22:17:53 +0000[diff] [blame]1299 <dt><tt>s<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1300 <dd>This specifies the alignment for a stack object of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1301 <i>size</i>.</dd>
Chris Lattnera381eff2009-11-07 09:35:34 +0000[diff] [blame]1302
1303 <dt><tt>n<i>size1</i>:<i>size2</i>:<i>size3</i>...</tt></dt>
1304 <dd>This specifies a set of native integer widths for the target CPU
1305 in bits. For example, it might contain "n32" for 32-bit PowerPC,
1306 "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]1307 this set are considered to support most general arithmetic
Chris Lattnera381eff2009-11-07 09:35:34 +0000[diff] [blame]1308 operations efficiently.</dd>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1309</dl>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1310
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1311<p>When constructing the data layout for a given target, LLVM starts with a
Dan Gohman61110ae2010-04-28 00:36:01 +0000[diff] [blame]1312 default set of specifications which are then (possibly) overridden by the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1313 specifications in the <tt>datalayout</tt> keyword. The default specifications
1314 are given in this list:</p>
1315
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1316<ul>
1317 <li><tt>E</tt> - big endian</li>
Dan Gohman8ad777d2010-02-23 02:44:03 +0000[diff] [blame]1318 <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]1319 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1320 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1321 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1322 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]1323 <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]1324 alignment of 64-bits</li>
1325 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1326 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1327 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1328 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1329 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
Daniel Dunbar7921a592009-06-08 22:17:53 +0000[diff] [blame]1330 <li><tt>s0:64:64</tt> - stack objects are 64-bit aligned</li>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1331</ul>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1332
1333<p>When LLVM is determining the alignment for a given type, it uses the
1334 following rules:</p>
1335
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1336<ol>
1337 <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]1338 specification is used.</li>
1339
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1340 <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]1341 smallest integer type that is larger than the bitwidth of the sought type
1342 is used. If none of the specifications are larger than the bitwidth then
1343 the the largest integer type is used. For example, given the default
1344 specifications above, the i7 type will use the alignment of i8 (next
1345 largest) while both i65 and i256 will use the alignment of i64 (largest
1346 specified).</li>
1347
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1348 <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]1349 largest vector type that is smaller than the sought vector type will be
1350 used as a fall back. This happens because &lt;128 x double&gt; can be
1351 implemented in terms of 64 &lt;2 x double&gt;, for example.</li>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1352</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1353
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1354</div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1355
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1356<!-- ======================================================================= -->
1357<div class="doc_subsection">
1358 <a name="pointeraliasing">Pointer Aliasing Rules</a>
1359</div>
1360
1361<div class="doc_text">
1362
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1363<p>Any memory access must be done through a pointer value associated
Andreas Bolkae39f0332009-07-27 20:37:10 +0000[diff] [blame]1364with an address range of the memory access, otherwise the behavior
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1365is undefined. Pointer values are associated with address ranges
1366according to the following rules:</p>
1367
1368<ul>
Dan Gohmandf12d082010-07-02 18:41:32 +0000[diff] [blame]1369 <li>A pointer value is associated with the addresses associated with
1370 any value it is <i>based</i> on.
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1371 <li>An address of a global variable is associated with the address
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1372 range of the variable's storage.</li>
1373 <li>The result value of an allocation instruction is associated with
1374 the address range of the allocated storage.</li>
1375 <li>A null pointer in the default address-space is associated with
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1376 no address.</li>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1377 <li>An integer constant other than zero or a pointer value returned
1378 from a function not defined within LLVM may be associated with address
1379 ranges allocated through mechanisms other than those provided by
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1380 LLVM. Such ranges shall not overlap with any ranges of addresses
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1381 allocated by mechanisms provided by LLVM.</li>
Dan Gohmandf12d082010-07-02 18:41:32 +0000[diff] [blame]1382</ul>
1383
1384<p>A pointer value is <i>based</i> on another pointer value according
1385 to the following rules:</p>
1386
1387<ul>
1388 <li>A pointer value formed from a
1389 <tt><a href="#i_getelementptr">getelementptr</a></tt> operation
1390 is <i>based</i> on the first operand of the <tt>getelementptr</tt>.</li>
1391 <li>The result value of a
1392 <tt><a href="#i_bitcast">bitcast</a></tt> is <i>based</i> on the operand
1393 of the <tt>bitcast</tt>.</li>
1394 <li>A pointer value formed by an
1395 <tt><a href="#i_inttoptr">inttoptr</a></tt> is <i>based</i> on all
1396 pointer values that contribute (directly or indirectly) to the
1397 computation of the pointer's value.</li>
1398 <li>The "<i>based</i> on" relationship is transitive.</li>
1399</ul>
1400
1401<p>Note that this definition of <i>"based"</i> is intentionally
1402 similar to the definition of <i>"based"</i> in C99, though it is
1403 slightly weaker.</p>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1404
1405<p>LLVM IR does not associate types with memory. The result type of a
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1406<tt><a href="#i_load">load</a></tt> merely indicates the size and
1407alignment of the memory from which to load, as well as the
Dan Gohman4eb47192010-06-17 19:23:50 +0000[diff] [blame]1408interpretation of the value. The first operand type of a
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1409<tt><a href="#i_store">store</a></tt> similarly only indicates the size
1410and alignment of the store.</p>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1411
1412<p>Consequently, type-based alias analysis, aka TBAA, aka
1413<tt>-fstrict-aliasing</tt>, is not applicable to general unadorned
1414LLVM IR. <a href="#metadata">Metadata</a> may be used to encode
1415additional information which specialized optimization passes may use
1416to implement type-based alias analysis.</p>
1417
1418</div>
1419
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]1420<!-- ======================================================================= -->
1421<div class="doc_subsection">
1422 <a name="volatile">Volatile Memory Accesses</a>
1423</div>
1424
1425<div class="doc_text">
1426
1427<p>Certain memory accesses, such as <a href="#i_load"><tt>load</tt></a>s, <a
1428href="#i_store"><tt>store</tt></a>s, and <a
1429href="#int_memcpy"><tt>llvm.memcpy</tt></a>s may be marked <tt>volatile</tt>.
1430The optimizers must not change the number of volatile operations or change their
1431order of execution relative to other volatile operations. The optimizers
1432<i>may</i> change the order of volatile operations relative to non-volatile
1433operations. This is not Java's "volatile" and has no cross-thread
1434synchronization behavior.</p>
1435
1436</div>
1437
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1438<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1439<div class="doc_section"> <a name="typesystem">Type System</a> </div>
1440<!-- *********************************************************************** -->
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1441
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1442<div class="doc_text">
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1443
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1444<p>The LLVM type system is one of the most important features of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1445 intermediate representation. Being typed enables a number of optimizations
1446 to be performed on the intermediate representation directly, without having
1447 to do extra analyses on the side before the transformation. A strong type
1448 system makes it easier to read the generated code and enables novel analyses
1449 and transformations that are not feasible to perform on normal three address
1450 code representations.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1451
1452</div>
1453
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1454<!-- ======================================================================= -->
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1455<div class="doc_subsection"> <a name="t_classifications">Type
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1456Classifications</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1457
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1458<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1459
1460<p>The types fall into a few useful classifications:</p>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]1461
1462<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1463 <tbody>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1464 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1465 <tr>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1466 <td><a href="#t_integer">integer</a></td>
Reid Spencer138249b2007-05-16 18:44:01 +0000[diff] [blame]1467 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1468 </tr>
1469 <tr>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1470 <td><a href="#t_floating">floating point</a></td>
1471 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1472 </tr>
1473 <tr>
1474 <td><a name="t_firstclass">first class</a></td>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1475 <td><a href="#t_integer">integer</a>,
1476 <a href="#t_floating">floating point</a>,
1477 <a href="#t_pointer">pointer</a>,
Dan Gohman08783a882008-06-18 18:42:13 +0000[diff] [blame]1478 <a href="#t_vector">vector</a>,
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]1479 <a href="#t_struct">structure</a>,
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1480 <a href="#t_union">union</a>,
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]1481 <a href="#t_array">array</a>,
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1482 <a href="#t_label">label</a>,
1483 <a href="#t_metadata">metadata</a>.
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1484 </td>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1485 </tr>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1486 <tr>
1487 <td><a href="#t_primitive">primitive</a></td>
1488 <td><a href="#t_label">label</a>,
1489 <a href="#t_void">void</a>,
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1490 <a href="#t_floating">floating point</a>,
1491 <a href="#t_metadata">metadata</a>.</td>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1492 </tr>
1493 <tr>
1494 <td><a href="#t_derived">derived</a></td>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1495 <td><a href="#t_array">array</a>,
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1496 <a href="#t_function">function</a>,
1497 <a href="#t_pointer">pointer</a>,
1498 <a href="#t_struct">structure</a>,
1499 <a href="#t_pstruct">packed structure</a>,
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1500 <a href="#t_union">union</a>,
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1501 <a href="#t_vector">vector</a>,
1502 <a href="#t_opaque">opaque</a>.
Dan Gohman93bf60d2008-10-14 16:32:04 +0000[diff] [blame]1503 </td>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1504 </tr>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1505 </tbody>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1506</table>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]1507
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1508<p>The <a href="#t_firstclass">first class</a> types are perhaps the most
1509 important. Values of these types are the only ones which can be produced by
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1510 instructions.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1511
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1512</div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1513
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1514<!-- ======================================================================= -->
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1515<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Chris Lattner43542b32008-01-04 04:34:14 +0000[diff] [blame]1516
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1517<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1518
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1519<p>The primitive types are the fundamental building blocks of the LLVM
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1520 system.</p>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1521
Chris Lattner43542b32008-01-04 04:34:14 +0000[diff] [blame]1522</div>
1523
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1524<!-- _______________________________________________________________________ -->
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1525<div class="doc_subsubsection"> <a name="t_integer">Integer Type</a> </div>
1526
1527<div class="doc_text">
1528
1529<h5>Overview:</h5>
1530<p>The integer type is a very simple type that simply specifies an arbitrary
1531 bit width for the integer type desired. Any bit width from 1 bit to
1532 2<sup>23</sup>-1 (about 8 million) can be specified.</p>
1533
1534<h5>Syntax:</h5>
1535<pre>
1536 iN
1537</pre>
1538
1539<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1540 value.</p>
1541
1542<h5>Examples:</h5>
1543<table class="layout">
1544 <tr class="layout">
1545 <td class="left"><tt>i1</tt></td>
1546 <td class="left">a single-bit integer.</td>
1547 </tr>
1548 <tr class="layout">
1549 <td class="left"><tt>i32</tt></td>
1550 <td class="left">a 32-bit integer.</td>
1551 </tr>
1552 <tr class="layout">
1553 <td class="left"><tt>i1942652</tt></td>
1554 <td class="left">a really big integer of over 1 million bits.</td>
1555 </tr>
1556</table>
1557
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1558</div>
1559
1560<!-- _______________________________________________________________________ -->
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1561<div class="doc_subsubsection"> <a name="t_floating">Floating Point Types</a> </div>
1562
1563<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1564
1565<table>
1566 <tbody>
1567 <tr><th>Type</th><th>Description</th></tr>
1568 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1569 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1570 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1571 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1572 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1573 </tbody>
1574</table>
1575
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1576</div>
1577
1578<!-- _______________________________________________________________________ -->
1579<div class="doc_subsubsection"> <a name="t_void">Void Type</a> </div>
1580
1581<div class="doc_text">
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1582
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1583<h5>Overview:</h5>
1584<p>The void type does not represent any value and has no size.</p>
1585
1586<h5>Syntax:</h5>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1587<pre>
1588 void
1589</pre>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1590
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1591</div>
1592
1593<!-- _______________________________________________________________________ -->
1594<div class="doc_subsubsection"> <a name="t_label">Label Type</a> </div>
1595
1596<div class="doc_text">
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1597
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1598<h5>Overview:</h5>
1599<p>The label type represents code labels.</p>
1600
1601<h5>Syntax:</h5>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1602<pre>
1603 label
1604</pre>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1605
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1606</div>
1607
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1608<!-- _______________________________________________________________________ -->
1609<div class="doc_subsubsection"> <a name="t_metadata">Metadata Type</a> </div>
1610
1611<div class="doc_text">
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1612
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1613<h5>Overview:</h5>
Nick Lewycky93e06a52009-09-27 23:27:42 +0000[diff] [blame]1614<p>The metadata type represents embedded metadata. No derived types may be
1615 created from metadata except for <a href="#t_function">function</a>
1616 arguments.
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1617
1618<h5>Syntax:</h5>
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1619<pre>
1620 metadata
1621</pre>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1622
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1623</div>
1624
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1625
1626<!-- ======================================================================= -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1627<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1628
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1629<div class="doc_text">
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1630
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1631<p>The real power in LLVM comes from the derived types in the system. This is
1632 what allows a programmer to represent arrays, functions, pointers, and other
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1633 useful types. Each of these types contain one or more element types which
1634 may be a primitive type, or another derived type. For example, it is
1635 possible to have a two dimensional array, using an array as the element type
1636 of another array.</p>
Dan Gohman142ccc02009-01-24 15:58:40 +0000[diff] [blame]1637
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1638
1639</div>
1640
1641<!-- _______________________________________________________________________ -->
1642<div class="doc_subsubsection"> <a name="t_aggregate">Aggregate Types</a> </div>
1643
1644<div class="doc_text">
1645
1646<p>Aggregate Types are a subset of derived types that can contain multiple
1647 member types. <a href="#t_array">Arrays</a>,
1648 <a href="#t_struct">structs</a>, <a href="#t_vector">vectors</a> and
1649 <a href="#t_union">unions</a> are aggregate types.</p>
1650
1651</div>
1652
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]1653</div>
Reid Spencer138249b2007-05-16 18:44:01 +0000[diff] [blame]1654
1655<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1656<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1657
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1658<div class="doc_text">
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1659
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1660<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1661<p>The array type is a very simple derived type that arranges elements
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1662 sequentially in memory. The array type requires a size (number of elements)
1663 and an underlying data type.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1664
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]1665<h5>Syntax:</h5>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1666<pre>
1667 [&lt;# elements&gt; x &lt;elementtype&gt;]
1668</pre>
1669
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1670<p>The number of elements is a constant integer value; <tt>elementtype</tt> may
1671 be any type with a size.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1672
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]1673<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1674<table class="layout">
1675 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1676 <td class="left"><tt>[40 x i32]</tt></td>
1677 <td class="left">Array of 40 32-bit integer values.</td>
1678 </tr>
1679 <tr class="layout">
1680 <td class="left"><tt>[41 x i32]</tt></td>
1681 <td class="left">Array of 41 32-bit integer values.</td>
1682 </tr>
1683 <tr class="layout">
1684 <td class="left"><tt>[4 x i8]</tt></td>
1685 <td class="left">Array of 4 8-bit integer values.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1686 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1687</table>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1688<p>Here are some examples of multidimensional arrays:</p>
1689<table class="layout">
1690 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1691 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1692 <td class="left">3x4 array of 32-bit integer values.</td>
1693 </tr>
1694 <tr class="layout">
1695 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1696 <td class="left">12x10 array of single precision floating point values.</td>
1697 </tr>
1698 <tr class="layout">
1699 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1700 <td class="left">2x3x4 array of 16-bit integer values.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1701 </tr>
1702</table>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +0000[diff] [blame]1703
Dan Gohmanc74bc282009-11-09 19:01:53 +0000[diff] [blame]1704<p>There is no restriction on indexing beyond the end of the array implied by
1705 a static type (though there are restrictions on indexing beyond the bounds
1706 of an allocated object in some cases). This means that single-dimension
1707 'variable sized array' addressing can be implemented in LLVM with a zero
1708 length array type. An implementation of 'pascal style arrays' in LLVM could
1709 use the type "<tt>{ i32, [0 x float]}</tt>", for example.</p>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +0000[diff] [blame]1710
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1711</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1712
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1713<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1714<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1715
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1716<div class="doc_text">
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]1717
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1718<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1719<p>The function type can be thought of as a function signature. It consists of
1720 a return type and a list of formal parameter types. The return type of a
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1721 function type is a scalar type, a void type, a struct type, or a union
1722 type. If the return type is a struct type then all struct elements must be
1723 of first class types, and the struct must have at least one element.</p>
Devang Pateld6cff512008-03-10 20:49:15 +0000[diff] [blame]1724
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1725<h5>Syntax:</h5>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]1726<pre>
Nick Lewycky14d1ccc2009-09-27 07:55:32 +0000[diff] [blame]1727 &lt;returntype&gt; (&lt;parameter list&gt;)
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]1728</pre>
1729
John Criswell4c0cf7f2005-10-24 16:17:18 +0000[diff] [blame]1730<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]1731 specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1732 which indicates that the function takes a variable number of arguments.
1733 Variable argument functions can access their arguments with
1734 the <a href="#int_varargs">variable argument handling intrinsic</a>
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]1735 functions. '<tt>&lt;returntype&gt;</tt>' is any type except
Nick Lewycky93e06a52009-09-27 23:27:42 +0000[diff] [blame]1736 <a href="#t_label">label</a>.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]1737
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1738<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1739<table class="layout">
1740 <tr class="layout">
Reid Spencer58c08712006-12-31 07:18:34 +0000[diff] [blame]1741 <td class="left"><tt>i32 (i32)</tt></td>
1742 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1743 </td>
Reid Spencer58c08712006-12-31 07:18:34 +0000[diff] [blame]1744 </tr><tr class="layout">
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]1745 <td class="left"><tt>float&nbsp;(i16,&nbsp;i32&nbsp;*)&nbsp;*
Reid Spencer655dcc62006-12-31 07:20:23 +0000[diff] [blame]1746 </tt></td>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1747 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]1748 an <tt>i16</tt> and a <a href="#t_pointer">pointer</a> to <tt>i32</tt>,
1749 returning <tt>float</tt>.
Reid Spencer58c08712006-12-31 07:18:34 +0000[diff] [blame]1750 </td>
1751 </tr><tr class="layout">
1752 <td class="left"><tt>i32 (i8*, ...)</tt></td>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1753 <td class="left">A vararg function that takes at least one
1754 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1755 which returns an integer. This is the signature for <tt>printf</tt> in
Reid Spencer58c08712006-12-31 07:18:34 +0000[diff] [blame]1756 LLVM.
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1757 </td>
Devang Patele3dfc1c2008-03-24 05:35:41 +0000[diff] [blame]1758 </tr><tr class="layout">
1759 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Nick Lewycky14d1ccc2009-09-27 07:55:32 +0000[diff] [blame]1760 <td class="left">A function taking an <tt>i32</tt>, returning a
1761 <a href="#t_struct">structure</a> containing two <tt>i32</tt> values
Devang Patele3dfc1c2008-03-24 05:35:41 +0000[diff] [blame]1762 </td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1763 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1764</table>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]1765
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1766</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1767
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1768<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1769<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1770
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1771<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1772
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1773<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1774<p>The structure type is used to represent a collection of data members together
1775 in memory. The packing of the field types is defined to match the ABI of the
1776 underlying processor. The elements of a structure may be any type that has a
1777 size.</p>
1778
Jeffrey Yasskinf991bbb2010-01-11 19:19:26 +0000[diff] [blame]1779<p>Structures in memory are accessed using '<tt><a href="#i_load">load</a></tt>'
1780 and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field
1781 with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
1782 Structures in registers are accessed using the
1783 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' and
1784 '<tt><a href="#i_insertvalue">insertvalue</a></tt>' instructions.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1785<h5>Syntax:</h5>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1786<pre>
1787 { &lt;type list&gt; }
1788</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1789
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1790<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1791<table class="layout">
1792 <tr class="layout">
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]1793 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1794 <td class="left">A triple of three <tt>i32</tt> values</td>
1795 </tr><tr class="layout">
1796 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1797 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1798 second element is a <a href="#t_pointer">pointer</a> to a
1799 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1800 an <tt>i32</tt>.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1801 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1802</table>
Dan Gohman142ccc02009-01-24 15:58:40 +0000[diff] [blame]1803
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1804</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1805
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1806<!-- _______________________________________________________________________ -->
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]1807<div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
1808</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1809
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]1810<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1811
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]1812<h5>Overview:</h5>
1813<p>The packed structure type is used to represent a collection of data members
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1814 together in memory. There is no padding between fields. Further, the
1815 alignment of a packed structure is 1 byte. The elements of a packed
1816 structure may be any type that has a size.</p>
1817
1818<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt> and
1819 '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field with
1820 the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
1821
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]1822<h5>Syntax:</h5>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1823<pre>
1824 &lt; { &lt;type list&gt; } &gt;
1825</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1826
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]1827<h5>Examples:</h5>
1828<table class="layout">
1829 <tr class="layout">
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]1830 <td class="left"><tt>&lt; { i32, i32, i32 } &gt;</tt></td>
1831 <td class="left">A triple of three <tt>i32</tt> values</td>
1832 </tr><tr class="layout">
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1833 <td class="left">
1834<tt>&lt;&nbsp;{&nbsp;float,&nbsp;i32&nbsp;(i32)*&nbsp;}&nbsp;&gt;</tt></td>
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]1835 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1836 second element is a <a href="#t_pointer">pointer</a> to a
1837 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1838 an <tt>i32</tt>.</td>
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]1839 </tr>
1840</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1841
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]1842</div>
1843
1844<!-- _______________________________________________________________________ -->
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1845<div class="doc_subsubsection"> <a name="t_union">Union Type</a> </div>
1846
1847<div class="doc_text">
1848
1849<h5>Overview:</h5>
1850<p>A union type describes an object with size and alignment suitable for
1851 an object of any one of a given set of types (also known as an "untagged"
1852 union). It is similar in concept and usage to a
1853 <a href="#t_struct">struct</a>, except that all members of the union
1854 have an offset of zero. The elements of a union may be any type that has a
1855 size. Unions must have at least one member - empty unions are not allowed.
1856 </p>
1857
1858<p>The size of the union as a whole will be the size of its largest member,
1859 and the alignment requirements of the union as a whole will be the largest
1860 alignment requirement of any member.</p>
1861
Dan Gohman1ad14992010-02-25 16:51:31 +0000[diff] [blame]1862<p>Union members are accessed using '<tt><a href="#i_load">load</a></tt> and
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1863 '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field with
1864 the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
1865 Since all members are at offset zero, the getelementptr instruction does
1866 not affect the address, only the type of the resulting pointer.</p>
1867
1868<h5>Syntax:</h5>
1869<pre>
1870 union { &lt;type list&gt; }
1871</pre>
1872
1873<h5>Examples:</h5>
1874<table class="layout">
1875 <tr class="layout">
1876 <td class="left"><tt>union { i32, i32*, float }</tt></td>
1877 <td class="left">A union of three types: an <tt>i32</tt>, a pointer to
1878 an <tt>i32</tt>, and a <tt>float</tt>.</td>
1879 </tr><tr class="layout">
1880 <td class="left">
1881 <tt>union {&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1882 <td class="left">A union, where the first element is a <tt>float</tt> and the
1883 second element is a <a href="#t_pointer">pointer</a> to a
1884 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1885 an <tt>i32</tt>.</td>
1886 </tr>
1887</table>
1888
1889</div>
1890
1891<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1892<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
Chris Lattner4a67c912009-02-08 19:53:29 +0000[diff] [blame]1893
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1894<div class="doc_text">
1895
1896<h5>Overview:</h5>
Dan Gohman88481112010-02-25 16:50:07 +0000[diff] [blame]1897<p>The pointer type is used to specify memory locations.
1898 Pointers are commonly used to reference objects in memory.</p>
1899
1900<p>Pointer types may have an optional address space attribute defining the
1901 numbered address space where the pointed-to object resides. The default
1902 address space is number zero. The semantics of non-zero address
1903 spaces are target-specific.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1904
1905<p>Note that LLVM does not permit pointers to void (<tt>void*</tt>) nor does it
1906 permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</tt> instead.</p>
Chris Lattner4a67c912009-02-08 19:53:29 +0000[diff] [blame]1907
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]1908<h5>Syntax:</h5>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1909<pre>
1910 &lt;type&gt; *
1911</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1912
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]1913<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1914<table class="layout">
1915 <tr class="layout">
Dan Gohman623806e2009-01-04 23:44:43 +0000[diff] [blame]1916 <td class="left"><tt>[4 x i32]*</tt></td>
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1917 <td class="left">A <a href="#t_pointer">pointer</a> to <a
1918 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
1919 </tr>
1920 <tr class="layout">
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]1921 <td class="left"><tt>i32 (i32*) *</tt></td>
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1922 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1923 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1924 <tt>i32</tt>.</td>
1925 </tr>
1926 <tr class="layout">
1927 <td class="left"><tt>i32 addrspace(5)*</tt></td>
1928 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
1929 that resides in address space #5.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1930 </tr>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1931</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1932
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1933</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1934
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]1935<!-- _______________________________________________________________________ -->
Reid Spencer404a3252007-02-15 03:07:05 +0000[diff] [blame]1936<div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1937
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1938<div class="doc_text">
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1939
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]1940<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1941<p>A vector type is a simple derived type that represents a vector of elements.
1942 Vector types are used when multiple primitive data are operated in parallel
1943 using a single instruction (SIMD). A vector type requires a size (number of
Duncan Sands31c0e0e2009-11-27 13:38:03 +0000[diff] [blame]1944 elements) and an underlying primitive data type. Vector types are considered
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1945 <a href="#t_firstclass">first class</a>.</p>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1946
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]1947<h5>Syntax:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1948<pre>
1949 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1950</pre>
1951
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1952<p>The number of elements is a constant integer value; elementtype may be any
1953 integer or floating point type.</p>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1954
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]1955<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1956<table class="layout">
1957 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1958 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
1959 <td class="left">Vector of 4 32-bit integer values.</td>
1960 </tr>
1961 <tr class="layout">
1962 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
1963 <td class="left">Vector of 8 32-bit floating-point values.</td>
1964 </tr>
1965 <tr class="layout">
1966 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
1967 <td class="left">Vector of 2 64-bit integer values.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1968 </tr>
1969</table>
Dan Gohman142ccc02009-01-24 15:58:40 +0000[diff] [blame]1970
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1971</div>
1972
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1973<!-- _______________________________________________________________________ -->
1974<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1975<div class="doc_text">
1976
1977<h5>Overview:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1978<p>Opaque types are used to represent unknown types in the system. This
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1979 corresponds (for example) to the C notion of a forward declared structure
1980 type. In LLVM, opaque types can eventually be resolved to any type (not just
1981 a structure type).</p>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1982
1983<h5>Syntax:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1984<pre>
1985 opaque
1986</pre>
1987
1988<h5>Examples:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1989<table class="layout">
1990 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1991 <td class="left"><tt>opaque</tt></td>
1992 <td class="left">An opaque type.</td>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1993 </tr>
1994</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1995
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1996</div>
1997
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]1998<!-- ======================================================================= -->
1999<div class="doc_subsection">
2000 <a name="t_uprefs">Type Up-references</a>
2001</div>
2002
2003<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2004
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]2005<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2006<p>An "up reference" allows you to refer to a lexically enclosing type without
2007 requiring it to have a name. For instance, a structure declaration may
2008 contain a pointer to any of the types it is lexically a member of. Example
2009 of up references (with their equivalent as named type declarations)
2010 include:</p>
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]2011
2012<pre>
Chris Lattnerbf1d5452009-02-09 10:00:56 +0000[diff] [blame]2013 { \2 * } %x = type { %x* }
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]2014 { \2 }* %y = type { %y }*
2015 \1* %z = type %z*
2016</pre>
2017
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2018<p>An up reference is needed by the asmprinter for printing out cyclic types
2019 when there is no declared name for a type in the cycle. Because the
2020 asmprinter does not want to print out an infinite type string, it needs a
2021 syntax to handle recursive types that have no names (all names are optional
2022 in llvm IR).</p>
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]2023
2024<h5>Syntax:</h5>
2025<pre>
2026 \&lt;level&gt;
2027</pre>
2028
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2029<p>The level is the count of the lexical type that is being referred to.</p>
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]2030
2031<h5>Examples:</h5>
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]2032<table class="layout">
2033 <tr class="layout">
2034 <td class="left"><tt>\1*</tt></td>
2035 <td class="left">Self-referential pointer.</td>
2036 </tr>
2037 <tr class="layout">
2038 <td class="left"><tt>{ { \3*, i8 }, i32 }</tt></td>
2039 <td class="left">Recursive structure where the upref refers to the out-most
2040 structure.</td>
2041 </tr>
2042</table>
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]2043
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2044</div>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]2045
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2046<!-- *********************************************************************** -->
2047<div class="doc_section"> <a name="constants">Constants</a> </div>
2048<!-- *********************************************************************** -->
2049
2050<div class="doc_text">
2051
2052<p>LLVM has several different basic types of constants. This section describes
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2053 them all and their syntax.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2054
2055</div>
2056
2057<!-- ======================================================================= -->
Reid Spencer8f08d802004-12-09 18:02:53 +0000[diff] [blame]2058<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2059
2060<div class="doc_text">
2061
2062<dl>
2063 <dt><b>Boolean constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2064 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]2065 constants of the <tt><a href="#t_integer">i1</a></tt> type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2066
2067 <dt><b>Integer constants</b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2068 <dd>Standard integers (such as '4') are constants of
2069 the <a href="#t_integer">integer</a> type. Negative numbers may be used
2070 with integer types.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2071
2072 <dt><b>Floating point constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2073 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2074 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
2075 notation (see below). The assembler requires the exact decimal value of a
2076 floating-point constant. For example, the assembler accepts 1.25 but
2077 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
2078 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2079
2080 <dt><b>Null pointer constants</b></dt>
John Criswelldfe6a862004-12-10 15:51:16 +0000[diff] [blame]2081 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2082 and must be of <a href="#t_pointer">pointer type</a>.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2083</dl>
2084
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2085<p>The one non-intuitive notation for constants is the hexadecimal form of
2086 floating point constants. For example, the form '<tt>double
2087 0x432ff973cafa8000</tt>' is equivalent to (but harder to read than)
2088 '<tt>double 4.5e+15</tt>'. The only time hexadecimal floating point
2089 constants are required (and the only time that they are generated by the
2090 disassembler) is when a floating point constant must be emitted but it cannot
2091 be represented as a decimal floating point number in a reasonable number of
2092 digits. For example, NaN's, infinities, and other special values are
2093 represented in their IEEE hexadecimal format so that assembly and disassembly
2094 do not cause any bits to change in the constants.</p>
2095
Dale Johannesencd4a3012009-02-11 22:14:51 +0000[diff] [blame]2096<p>When using the hexadecimal form, constants of types float and double are
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2097 represented using the 16-digit form shown above (which matches the IEEE754
2098 representation for double); float values must, however, be exactly
2099 representable as IEE754 single precision. Hexadecimal format is always used
2100 for long double, and there are three forms of long double. The 80-bit format
2101 used by x86 is represented as <tt>0xK</tt> followed by 20 hexadecimal digits.
2102 The 128-bit format used by PowerPC (two adjacent doubles) is represented
2103 by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit format
2104 is represented by <tt>0xL</tt> followed by 32 hexadecimal digits; no
2105 currently supported target uses this format. Long doubles will only work if
2106 they match the long double format on your target. All hexadecimal formats
2107 are big-endian (sign bit at the left).</p>
2108
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2109</div>
2110
2111<!-- ======================================================================= -->
Chris Lattner361bfcd2009-02-28 18:32:25 +0000[diff] [blame]2112<div class="doc_subsection">
Bill Wendling972b7202009-07-20 02:32:41 +0000[diff] [blame]2113<a name="aggregateconstants"></a> <!-- old anchor -->
2114<a name="complexconstants">Complex Constants</a>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2115</div>
2116
2117<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2118
Chris Lattner361bfcd2009-02-28 18:32:25 +0000[diff] [blame]2119<p>Complex constants are a (potentially recursive) combination of simple
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2120 constants and smaller complex constants.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2121
2122<dl>
2123 <dt><b>Structure constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2124 <dd>Structure constants are represented with notation similar to structure
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2125 type definitions (a comma separated list of elements, surrounded by braces
2126 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
2127 where "<tt>@G</tt>" is declared as "<tt>@G = external global i32</tt>".
2128 Structure constants must have <a href="#t_struct">structure type</a>, and
2129 the number and types of elements must match those specified by the
2130 type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2131
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]2132 <dt><b>Union constants</b></dt>
2133 <dd>Union constants are represented with notation similar to a structure with
2134 a single element - that is, a single typed element surrounded
2135 by braces (<tt>{}</tt>)). For example: "<tt>{ i32 4 }</tt>". The
2136 <a href="#t_union">union type</a> can be initialized with a single-element
2137 struct as long as the type of the struct element matches the type of
2138 one of the union members.</dd>
2139
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2140 <dt><b>Array constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2141 <dd>Array constants are represented with notation similar to array type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2142 definitions (a comma separated list of elements, surrounded by square
2143 brackets (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74
2144 ]</tt>". Array constants must have <a href="#t_array">array type</a>, and
2145 the number and types of elements must match those specified by the
2146 type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2147
Reid Spencer404a3252007-02-15 03:07:05 +0000[diff] [blame]2148 <dt><b>Vector constants</b></dt>
Reid Spencer404a3252007-02-15 03:07:05 +0000[diff] [blame]2149 <dd>Vector constants are represented with notation similar to vector type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2150 definitions (a comma separated list of elements, surrounded by
2151 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32
2152 42, i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must
2153 have <a href="#t_vector">vector type</a>, and the number and types of
2154 elements must match those specified by the type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2155
2156 <dt><b>Zero initialization</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2157 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]2158 value to zero of <em>any</em> type, including scalar and
2159 <a href="#t_aggregate">aggregate</a> types.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2160 This is often used to avoid having to print large zero initializers
2161 (e.g. for large arrays) and is always exactly equivalent to using explicit
2162 zero initializers.</dd>
Nick Lewycky49f89192009-04-04 07:22:01 +0000[diff] [blame]2163
2164 <dt><b>Metadata node</b></dt>
Nick Lewycky8e2c4f42009-05-30 16:08:30 +0000[diff] [blame]2165 <dd>A metadata node is a structure-like constant with
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2166 <a href="#t_metadata">metadata type</a>. For example: "<tt>metadata !{
2167 i32 0, metadata !"test" }</tt>". Unlike other constants that are meant to
2168 be interpreted as part of the instruction stream, metadata is a place to
2169 attach additional information such as debug info.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2170</dl>
2171
2172</div>
2173
2174<!-- ======================================================================= -->
2175<div class="doc_subsection">
2176 <a name="globalconstants">Global Variable and Function Addresses</a>
2177</div>
2178
2179<div class="doc_text">
2180
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2181<p>The addresses of <a href="#globalvars">global variables</a>
2182 and <a href="#functionstructure">functions</a> are always implicitly valid
2183 (link-time) constants. These constants are explicitly referenced when
2184 the <a href="#identifiers">identifier for the global</a> is used and always
2185 have <a href="#t_pointer">pointer</a> type. For example, the following is a
2186 legal LLVM file:</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2187
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2188<div class="doc_code">
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2189<pre>
Chris Lattner00538a12007-06-06 18:28:13 +0000[diff] [blame]2190@X = global i32 17
2191@Y = global i32 42
2192@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2193</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2194</div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2195
2196</div>
2197
2198<!-- ======================================================================= -->
Reid Spencer641f5c92004-12-09 18:13:12 +0000[diff] [blame]2199<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2200<div class="doc_text">
2201
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2202<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]2203 indicates that the user of the value may receive an unspecified bit-pattern.
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2204 Undefined values may be of any type (other than label or void) and be used
2205 anywhere a constant is permitted.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2206
Chris Lattner92ada5d2009-09-11 01:49:31 +0000[diff] [blame]2207<p>Undefined values are useful because they indicate to the compiler that the
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2208 program is well defined no matter what value is used. This gives the
2209 compiler more freedom to optimize. Here are some examples of (potentially
2210 surprising) transformations that are valid (in pseudo IR):</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2211
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2212
2213<div class="doc_code">
2214<pre>
2215 %A = add %X, undef
2216 %B = sub %X, undef
2217 %C = xor %X, undef
2218Safe:
2219 %A = undef
2220 %B = undef
2221 %C = undef
2222</pre>
2223</div>
2224
2225<p>This is safe because all of the output bits are affected by the undef bits.
2226Any output bit can have a zero or one depending on the input bits.</p>
2227
2228<div class="doc_code">
2229<pre>
2230 %A = or %X, undef
2231 %B = and %X, undef
2232Safe:
2233 %A = -1
2234 %B = 0
2235Unsafe:
2236 %A = undef
2237 %B = undef
2238</pre>
2239</div>
2240
2241<p>These logical operations have bits that are not always affected by the input.
2242For example, if "%X" has a zero bit, then the output of the 'and' operation will
2243always be a zero, no matter what the corresponding bit from the undef is. As
Chris Lattner92ada5d2009-09-11 01:49:31 +0000[diff] [blame]2244such, it is unsafe to optimize or assume that the result of the and is undef.
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2245However, it is safe to assume that all bits of the undef could be 0, and
2246optimize the and to 0. Likewise, it is safe to assume that all the bits of
2247the undef operand to the or could be set, allowing the or to be folded to
Chris Lattner92ada5d2009-09-11 01:49:31 +0000[diff] [blame]2248-1.</p>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2249
2250<div class="doc_code">
2251<pre>
2252 %A = select undef, %X, %Y
2253 %B = select undef, 42, %Y
2254 %C = select %X, %Y, undef
2255Safe:
2256 %A = %X (or %Y)
2257 %B = 42 (or %Y)
2258 %C = %Y
2259Unsafe:
2260 %A = undef
2261 %B = undef
2262 %C = undef
2263</pre>
2264</div>
2265
2266<p>This set of examples show that undefined select (and conditional branch)
2267conditions can go "either way" but they have to come from one of the two
2268operands. In the %A example, if %X and %Y were both known to have a clear low
2269bit, then %A would have to have a cleared low bit. However, in the %C example,
2270the optimizer is allowed to assume that the undef operand could be the same as
2271%Y, allowing the whole select to be eliminated.</p>
2272
2273
2274<div class="doc_code">
2275<pre>
2276 %A = xor undef, undef
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2277
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2278 %B = undef
2279 %C = xor %B, %B
2280
2281 %D = undef
2282 %E = icmp lt %D, 4
2283 %F = icmp gte %D, 4
2284
2285Safe:
2286 %A = undef
2287 %B = undef
2288 %C = undef
2289 %D = undef
2290 %E = undef
2291 %F = undef
2292</pre>
2293</div>
2294
2295<p>This example points out that two undef operands are not necessarily the same.
2296This can be surprising to people (and also matches C semantics) where they
2297assume that "X^X" is always zero, even if X is undef. This isn't true for a
2298number of reasons, but the short answer is that an undef "variable" can
2299arbitrarily change its value over its "live range". This is true because the
2300"variable" doesn't actually <em>have a live range</em>. Instead, the value is
2301logically read from arbitrary registers that happen to be around when needed,
Benjamin Kramer0f420382009-10-12 14:46:08 +0000[diff] [blame]2302so the value is not necessarily consistent over time. In fact, %A and %C need
Chris Lattner6760e542009-09-08 15:13:16 +0000[diff] [blame]2303to have the same semantics or the core LLVM "replace all uses with" concept
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2304would not hold.</p>
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2305
2306<div class="doc_code">
2307<pre>
2308 %A = fdiv undef, %X
2309 %B = fdiv %X, undef
2310Safe:
2311 %A = undef
2312b: unreachable
2313</pre>
2314</div>
2315
2316<p>These examples show the crucial difference between an <em>undefined
2317value</em> and <em>undefined behavior</em>. An undefined value (like undef) is
2318allowed to have an arbitrary bit-pattern. This means that the %A operation
2319can be constant folded to undef because the undef could be an SNaN, and fdiv is
2320not (currently) defined on SNaN's. However, in the second example, we can make
2321a more aggressive assumption: because the undef is allowed to be an arbitrary
2322value, we are allowed to assume that it could be zero. Since a divide by zero
Chris Lattner10ff0c12009-09-08 19:45:34 +0000[diff] [blame]2323has <em>undefined behavior</em>, we are allowed to assume that the operation
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2324does not execute at all. This allows us to delete the divide and all code after
2325it: since the undefined operation "can't happen", the optimizer can assume that
2326it occurs in dead code.
2327</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2328
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2329<div class="doc_code">
2330<pre>
2331a: store undef -> %X
2332b: store %X -> undef
2333Safe:
2334a: &lt;deleted&gt;
2335b: unreachable
2336</pre>
2337</div>
2338
2339<p>These examples reiterate the fdiv example: a store "of" an undefined value
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2340can be assumed to not have any effect: we can assume that the value is
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2341overwritten with bits that happen to match what was already there. However, a
2342store "to" an undefined location could clobber arbitrary memory, therefore, it
2343has undefined behavior.</p>
2344
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2345</div>
2346
2347<!-- ======================================================================= -->
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2348<div class="doc_subsection"><a name="trapvalues">Trap Values</a></div>
2349<div class="doc_text">
2350
Dan Gohmanb2a709b2010-04-26 20:21:21 +0000[diff] [blame]2351<p>Trap values are similar to <a href="#undefvalues">undef values</a>, however
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2352 instead of representing an unspecified bit pattern, they represent the
2353 fact that an instruction or constant expression which cannot evoke side
2354 effects has nevertheless detected a condition which results in undefined
Dan Gohmanb2a709b2010-04-26 20:21:21 +0000[diff] [blame]2355 behavior.</p>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2356
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2357<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]2358 only exist when produced by operations such as
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2359 <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]2360
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2361<p>Trap value behavior is defined in terms of value <i>dependence</i>:</p>
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2362
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2363<p>
2364<ul>
2365<li>Values other than <a href="#i_phi"><tt>phi</tt></a> nodes depend on
2366 their operands.</li>
2367
2368<li><a href="#i_phi"><tt>Phi</tt></a> nodes depend on the operand corresponding
2369 to their dynamic predecessor basic block.</li>
2370
2371<li>Function arguments depend on the corresponding actual argument values in
2372 the dynamic callers of their functions.</li>
2373
2374<li><a href="#i_call"><tt>Call</tt></a> instructions depend on the
2375 <a href="#i_ret"><tt>ret</tt></a> instructions that dynamically transfer
2376 control back to them.</li>
2377
Dan Gohman7292a752010-05-03 14:55:22 +0000[diff] [blame]2378<li><a href="#i_invoke"><tt>Invoke</tt></a> instructions depend on the
2379 <a href="#i_ret"><tt>ret</tt></a>, <a href="#i_unwind"><tt>unwind</tt></a>,
2380 or exception-throwing call instructions that dynamically transfer control
2381 back to them.</li>
2382
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2383<li>Non-volatile loads and stores depend on the most recent stores to all of the
2384 referenced memory addresses, following the order in the IR
2385 (including loads and stores implied by intrinsics such as
2386 <a href="#int_memcpy"><tt>@llvm.memcpy</tt></a>.)</li>
2387
Dan Gohman3513ea52010-05-03 14:59:34 +0000[diff] [blame]2388<!-- TODO: In the case of multiple threads, this only applies if the store
2389 "happens-before" the load or store. -->
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2390
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2391<!-- TODO: floating-point exception state -->
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2392
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2393<li>An instruction with externally visible side effects depends on the most
2394 recent preceding instruction with externally visible side effects, following
2395 the order in the IR. (This includes volatile loads and stores.)</li>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2396
Dan Gohman7292a752010-05-03 14:55:22 +0000[diff] [blame]2397<li>An instruction <i>control-depends</i> on a
2398 <a href="#terminators">terminator instruction</a>
2399 if the terminator instruction has multiple successors and the instruction
2400 is always executed when control transfers to one of the successors, and
2401 may not be executed when control is transfered to another.</li>
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2402
2403<li>Dependence is transitive.</li>
2404
2405</ul>
2406</p>
2407
2408<p>Whenever a trap value is generated, all values which depend on it evaluate
2409 to trap. If they have side effects, the evoke their side effects as if each
2410 operand with a trap value were undef. If they have externally-visible side
2411 effects, the behavior is undefined.</p>
2412
2413<p>Here are some examples:</p>
Dan Gohman48a25882010-04-26 20:54:53 +0000[diff] [blame]2414
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2415<div class="doc_code">
2416<pre>
2417entry:
2418 %trap = sub nuw i32 0, 1 ; Results in a trap value.
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2419 %still_trap = and i32 %trap, 0 ; Whereas (and i32 undef, 0) would return 0.
2420 %trap_yet_again = getelementptr i32* @h, i32 %still_trap
2421 store i32 0, i32* %trap_yet_again ; undefined behavior
2422
2423 store i32 %trap, i32* @g ; Trap value conceptually stored to memory.
2424 %trap2 = load i32* @g ; Returns a trap value, not just undef.
2425
2426 volatile store i32 %trap, i32* @g ; External observation; undefined behavior.
2427
2428 %narrowaddr = bitcast i32* @g to i16*
2429 %wideaddr = bitcast i32* @g to i64*
2430 %trap3 = load 16* %narrowaddr ; Returns a trap value.
2431 %trap4 = load i64* %widaddr ; Returns a trap value.
2432
2433 %cmp = icmp i32 slt %trap, 0 ; Returns a trap value.
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2434 %br i1 %cmp, %true, %end ; Branch to either destination.
2435
2436true:
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2437 volatile store i32 0, i32* @g ; This is control-dependent on %cmp, so
2438 ; it has undefined behavior.
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2439 br label %end
2440
2441end:
2442 %p = phi i32 [ 0, %entry ], [ 1, %true ]
2443 ; Both edges into this PHI are
2444 ; control-dependent on %cmp, so this
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2445 ; always results in a trap value.
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2446
2447 volatile store i32 0, i32* @g ; %end is control-equivalent to %entry
2448 ; so this is defined (ignoring earlier
2449 ; undefined behavior in this example).
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2450</pre>
2451</div>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2452
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2453</div>
2454
2455<!-- ======================================================================= -->
Chris Lattner2bfd3202009-10-27 21:19:13 +0000[diff] [blame]2456<div class="doc_subsection"><a name="blockaddress">Addresses of Basic
2457 Blocks</a></div>
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2458<div class="doc_text">
2459
Chris Lattneraa99c942009-11-01 01:27:45 +0000[diff] [blame]2460<p><b><tt>blockaddress(@function, %block)</tt></b></p>
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2461
2462<p>The '<tt>blockaddress</tt>' constant computes the address of the specified
Chris Lattner5c5f0ac2009-10-27 21:49:40 +0000[diff] [blame]2463 basic block in the specified function, and always has an i8* type. Taking
Chris Lattneraa99c942009-11-01 01:27:45 +0000[diff] [blame]2464 the address of the entry block is illegal.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2465
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2466<p>This value only has defined behavior when used as an operand to the
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]2467 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>' instruction or for comparisons
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2468 against null. Pointer equality tests between labels addresses is undefined
2469 behavior - though, again, comparison against null is ok, and no label is
Chris Lattner2bfd3202009-10-27 21:19:13 +0000[diff] [blame]2470 equal to the null pointer. This may also be passed around as an opaque
2471 pointer sized value as long as the bits are not inspected. This allows
Chris Lattnerda37b302009-10-27 21:44:20 +0000[diff] [blame]2472 <tt>ptrtoint</tt> and arithmetic to be performed on these values so long as
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]2473 the original value is reconstituted before the <tt>indirectbr</tt>.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2474
Chris Lattner2bfd3202009-10-27 21:19:13 +0000[diff] [blame]2475<p>Finally, some targets may provide defined semantics when
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2476 using the value as the operand to an inline assembly, but that is target
2477 specific.
2478 </p>
2479
2480</div>
2481
2482
2483<!-- ======================================================================= -->
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2484<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
2485</div>
2486
2487<div class="doc_text">
2488
2489<p>Constant expressions are used to allow expressions involving other constants
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2490 to be used as constants. Constant expressions may be of
2491 any <a href="#t_firstclass">first class</a> type and may involve any LLVM
2492 operation that does not have side effects (e.g. load and call are not
2493 supported). The following is the syntax for constant expressions:</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2494
2495<dl>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2496 <dt><b><tt>trunc (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2497 <dd>Truncate a constant to another type. The bit size of CST must be larger
2498 than the bit size of TYPE. Both types must be integers.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2499
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2500 <dt><b><tt>zext (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2501 <dd>Zero extend a constant to another type. The bit size of CST must be
2502 smaller or equal to the bit size of TYPE. Both types must be
2503 integers.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2504
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2505 <dt><b><tt>sext (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2506 <dd>Sign extend a constant to another type. The bit size of CST must be
2507 smaller or equal to the bit size of TYPE. Both types must be
2508 integers.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2509
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2510 <dt><b><tt>fptrunc (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2511 <dd>Truncate a floating point constant to another floating point type. The
2512 size of CST must be larger than the size of TYPE. Both types must be
2513 floating point.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2514
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2515 <dt><b><tt>fpext (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2516 <dd>Floating point extend a constant to another type. The size of CST must be
2517 smaller or equal to the size of TYPE. Both types must be floating
2518 point.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2519
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2520 <dt><b><tt>fptoui (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2521 <dd>Convert a floating point constant to the corresponding unsigned integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2522 constant. TYPE must be a scalar or vector integer type. CST must be of
2523 scalar or vector floating point type. Both CST and TYPE must be scalars,
2524 or vectors of the same number of elements. If the value won't fit in the
2525 integer type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2526
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2527 <dt><b><tt>fptosi (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2528 <dd>Convert a floating point constant to the corresponding signed integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2529 constant. TYPE must be a scalar or vector integer type. CST must be of
2530 scalar or vector floating point type. Both CST and TYPE must be scalars,
2531 or vectors of the same number of elements. If the value won't fit in the
2532 integer type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2533
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2534 <dt><b><tt>uitofp (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2535 <dd>Convert an unsigned integer constant to the corresponding floating point
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2536 constant. TYPE must be a scalar or vector floating point type. CST must be
2537 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2538 vectors of the same number of elements. If the value won't fit in the
2539 floating point type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2540
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2541 <dt><b><tt>sitofp (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2542 <dd>Convert a signed integer constant to the corresponding floating point
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2543 constant. TYPE must be a scalar or vector floating point type. CST must be
2544 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2545 vectors of the same number of elements. If the value won't fit in the
2546 floating point type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2547
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2548 <dt><b><tt>ptrtoint (CST to TYPE)</tt></b></dt>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]2549 <dd>Convert a pointer typed constant to the corresponding integer constant
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2550 <tt>TYPE</tt> must be an integer type. <tt>CST</tt> must be of pointer
2551 type. The <tt>CST</tt> value is zero extended, truncated, or unchanged to
2552 make it fit in <tt>TYPE</tt>.</dd>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]2553
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2554 <dt><b><tt>inttoptr (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2555 <dd>Convert a integer constant to a pointer constant. TYPE must be a pointer
2556 type. CST must be of integer type. The CST value is zero extended,
2557 truncated, or unchanged to make it fit in a pointer size. This one is
2558 <i>really</i> dangerous!</dd>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]2559
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2560 <dt><b><tt>bitcast (CST to TYPE)</tt></b></dt>
Chris Lattner789dee32009-02-28 18:27:03 +0000[diff] [blame]2561 <dd>Convert a constant, CST, to another TYPE. The constraints of the operands
2562 are the same as those for the <a href="#i_bitcast">bitcast
2563 instruction</a>.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2564
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2565 <dt><b><tt>getelementptr (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
2566 <dt><b><tt>getelementptr inbounds (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2567 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2568 constants. As with the <a href="#i_getelementptr">getelementptr</a>
2569 instruction, the index list may have zero or more indexes, which are
2570 required to make sense for the type of "CSTPTR".</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2571
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2572 <dt><b><tt>select (COND, VAL1, VAL2)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2573 <dd>Perform the <a href="#i_select">select operation</a> on constants.</dd>
Reid Spencer9965ee72006-12-04 19:23:19 +0000[diff] [blame]2574
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2575 <dt><b><tt>icmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer9965ee72006-12-04 19:23:19 +0000[diff] [blame]2576 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
2577
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2578 <dt><b><tt>fcmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer9965ee72006-12-04 19:23:19 +0000[diff] [blame]2579 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
Robert Bocchino7e97a6d2006-01-10 19:31:34 +0000[diff] [blame]2580
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2581 <dt><b><tt>extractelement (VAL, IDX)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2582 <dd>Perform the <a href="#i_extractelement">extractelement operation</a> on
2583 constants.</dd>
Robert Bocchino7e97a6d2006-01-10 19:31:34 +0000[diff] [blame]2584
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2585 <dt><b><tt>insertelement (VAL, ELT, IDX)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2586 <dd>Perform the <a href="#i_insertelement">insertelement operation</a> on
2587 constants.</dd>
Chris Lattner016a0e52006-04-08 00:13:41 +0000[diff] [blame]2588
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2589 <dt><b><tt>shufflevector (VEC1, VEC2, IDXMASK)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2590 <dd>Perform the <a href="#i_shufflevector">shufflevector operation</a> on
2591 constants.</dd>
Chris Lattner016a0e52006-04-08 00:13:41 +0000[diff] [blame]2592
Nick Lewycky9ab9a7f2010-05-29 06:44:15 +0000[diff] [blame]2593 <dt><b><tt>extractvalue (VAL, IDX0, IDX1, ...)</tt></b></dt>
2594 <dd>Perform the <a href="#i_extractvalue">extractvalue operation</a> on
2595 constants. The index list is interpreted in a similar manner as indices in
2596 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2597 index value must be specified.</dd>
2598
2599 <dt><b><tt>insertvalue (VAL, ELT, IDX0, IDX1, ...)</tt></b></dt>
2600 <dd>Perform the <a href="#i_insertvalue">insertvalue operation</a> on
2601 constants. The index list is interpreted in a similar manner as indices in
2602 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2603 index value must be specified.</dd>
2604
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2605 <dt><b><tt>OPCODE (LHS, RHS)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2606 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
2607 be any of the <a href="#binaryops">binary</a>
2608 or <a href="#bitwiseops">bitwise binary</a> operations. The constraints
2609 on operands are the same as those for the corresponding instruction
2610 (e.g. no bitwise operations on floating point values are allowed).</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2611</dl>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2612
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2613</div>
Chris Lattnerb1652612004-03-08 16:49:10 +0000[diff] [blame]2614
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2615<!-- *********************************************************************** -->
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2616<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
2617<!-- *********************************************************************** -->
2618
2619<!-- ======================================================================= -->
2620<div class="doc_subsection">
2621<a name="inlineasm">Inline Assembler Expressions</a>
2622</div>
2623
2624<div class="doc_text">
2625
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2626<p>LLVM supports inline assembler expressions (as opposed
2627 to <a href="#moduleasm"> Module-Level Inline Assembly</a>) through the use of
2628 a special value. This value represents the inline assembler as a string
2629 (containing the instructions to emit), a list of operand constraints (stored
Dale Johannesen63c94fe2009-10-13 21:56:55 +0000[diff] [blame]2630 as a string), a flag that indicates whether or not the inline asm
Dale Johannesen1cfb9582009-10-21 23:28:00 +0000[diff] [blame]2631 expression has side effects, and a flag indicating whether the function
2632 containing the asm needs to align its stack conservatively. An example
2633 inline assembler expression is:</p>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2634
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2635<div class="doc_code">
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2636<pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2637i32 (i32) asm "bswap $0", "=r,r"
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2638</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2639</div>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2640
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2641<p>Inline assembler expressions may <b>only</b> be used as the callee operand of
2642 a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we
2643 have:</p>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2644
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2645<div class="doc_code">
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2646<pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2647%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]2648</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2649</div>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2650
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2651<p>Inline asms with side effects not visible in the constraint list must be
2652 marked as having side effects. This is done through the use of the
2653 '<tt>sideeffect</tt>' keyword, like so:</p>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2654
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2655<div class="doc_code">
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2656<pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2657call void asm sideeffect "eieio", ""()
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2658</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2659</div>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2660
Dale Johannesen1cfb9582009-10-21 23:28:00 +0000[diff] [blame]2661<p>In some cases inline asms will contain code that will not work unless the
2662 stack is aligned in some way, such as calls or SSE instructions on x86,
2663 yet will not contain code that does that alignment within the asm.
2664 The compiler should make conservative assumptions about what the asm might
2665 contain and should generate its usual stack alignment code in the prologue
2666 if the '<tt>alignstack</tt>' keyword is present:</p>
Dale Johannesen63c94fe2009-10-13 21:56:55 +0000[diff] [blame]2667
2668<div class="doc_code">
2669<pre>
Dale Johannesen1cfb9582009-10-21 23:28:00 +0000[diff] [blame]2670call void asm alignstack "eieio", ""()
Dale Johannesen63c94fe2009-10-13 21:56:55 +0000[diff] [blame]2671</pre>
2672</div>
2673
2674<p>If both keywords appear the '<tt>sideeffect</tt>' keyword must come
2675 first.</p>
2676
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2677<p>TODO: The format of the asm and constraints string still need to be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2678 documented here. Constraints on what can be done (e.g. duplication, moving,
2679 etc need to be documented). This is probably best done by reference to
2680 another document that covers inline asm from a holistic perspective.</p>
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2681</div>
2682
2683<div class="doc_subsubsection">
2684<a name="inlineasm_md">Inline Asm Metadata</a>
2685</div>
2686
2687<div class="doc_text">
2688
2689<p>The call instructions that wrap inline asm nodes may have a "!srcloc" MDNode
2690 attached to it that contains a constant integer. If present, the code
2691 generator will use the integer as the location cookie value when report
2692 errors through the LLVMContext error reporting mechanisms. This allows a
Dan Gohman61110ae2010-04-28 00:36:01 +0000[diff] [blame]2693 front-end to correlate backend errors that occur with inline asm back to the
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2694 source code that produced it. For example:</p>
2695
2696<div class="doc_code">
2697<pre>
2698call void asm sideeffect "something bad", ""()<b>, !srcloc !42</b>
2699...
2700!42 = !{ i32 1234567 }
2701</pre>
2702</div>
2703
2704<p>It is up to the front-end to make sense of the magic numbers it places in the
2705 IR.</p>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2706
2707</div>
2708
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]2709<!-- ======================================================================= -->
2710<div class="doc_subsection"><a name="metadata">Metadata Nodes and Metadata
2711 Strings</a>
2712</div>
2713
2714<div class="doc_text">
2715
2716<p>LLVM IR allows metadata to be attached to instructions in the program that
2717 can convey extra information about the code to the optimizers and code
2718 generator. One example application of metadata is source-level debug
2719 information. There are two metadata primitives: strings and nodes. All
2720 metadata has the <tt>metadata</tt> type and is identified in syntax by a
2721 preceding exclamation point ('<tt>!</tt>').</p>
2722
2723<p>A metadata string is a string surrounded by double quotes. It can contain
2724 any character by escaping non-printable characters with "\xx" where "xx" is
2725 the two digit hex code. For example: "<tt>!"test\00"</tt>".</p>
2726
2727<p>Metadata nodes are represented with notation similar to structure constants
2728 (a comma separated list of elements, surrounded by braces and preceded by an
2729 exclamation point). For example: "<tt>!{ metadata !"test\00", i32
2730 10}</tt>". Metadata nodes can have any values as their operand.</p>
2731
2732<p>A <a href="#namedmetadatastructure">named metadata</a> is a collection of
2733 metadata nodes, which can be looked up in the module symbol table. For
2734 example: "<tt>!foo = metadata !{!4, !3}</tt>".
2735
Devang Patel9984bd62010-03-04 23:44:48 +0000[diff] [blame]2736<p>Metadata can be used as function arguments. Here <tt>llvm.dbg.value</tt>
2737 function is using two metadata arguments.
2738
2739 <div class="doc_code">
2740 <pre>
2741 call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2742 </pre>
2743 </div></p>
2744
2745<p>Metadata can be attached with an instruction. Here metadata <tt>!21</tt> is
2746 attached with <tt>add</tt> instruction using <tt>!dbg</tt> identifier.
2747
2748 <div class="doc_code">
2749 <pre>
2750 %indvar.next = add i64 %indvar, 1, !dbg !21
2751 </pre>
2752 </div></p>
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]2753</div>
2754
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2755
2756<!-- *********************************************************************** -->
2757<div class="doc_section">
2758 <a name="intrinsic_globals">Intrinsic Global Variables</a>
2759</div>
2760<!-- *********************************************************************** -->
2761
2762<p>LLVM has a number of "magic" global variables that contain data that affect
2763code generation or other IR semantics. These are documented here. All globals
Chris Lattner58f9bb22009-07-20 06:14:25 +0000[diff] [blame]2764of this sort should have a section specified as "<tt>llvm.metadata</tt>". This
2765section and all globals that start with "<tt>llvm.</tt>" are reserved for use
2766by LLVM.</p>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2767
2768<!-- ======================================================================= -->
2769<div class="doc_subsection">
2770<a name="intg_used">The '<tt>llvm.used</tt>' Global Variable</a>
2771</div>
2772
2773<div class="doc_text">
2774
2775<p>The <tt>@llvm.used</tt> global is an array with i8* element type which has <a
2776href="#linkage_appending">appending linkage</a>. This array contains a list of
2777pointers to global variables and functions which may optionally have a pointer
2778cast formed of bitcast or getelementptr. For example, a legal use of it is:</p>
2779
2780<pre>
2781 @X = global i8 4
2782 @Y = global i32 123
2783
2784 @llvm.used = appending global [2 x i8*] [
2785 i8* @X,
2786 i8* bitcast (i32* @Y to i8*)
2787 ], section "llvm.metadata"
2788</pre>
2789
2790<p>If a global variable appears in the <tt>@llvm.used</tt> list, then the
2791compiler, assembler, and linker are required to treat the symbol as if there is
2792a reference to the global that it cannot see. For example, if a variable has
2793internal linkage and no references other than that from the <tt>@llvm.used</tt>
2794list, it cannot be deleted. This is commonly used to represent references from
2795inline asms and other things the compiler cannot "see", and corresponds to
2796"attribute((used))" in GNU C.</p>
2797
2798<p>On some targets, the code generator must emit a directive to the assembler or
2799object file to prevent the assembler and linker from molesting the symbol.</p>
2800
2801</div>
2802
2803<!-- ======================================================================= -->
2804<div class="doc_subsection">
Chris Lattner58f9bb22009-07-20 06:14:25 +0000[diff] [blame]2805<a name="intg_compiler_used">The '<tt>llvm.compiler.used</tt>' Global Variable</a>
2806</div>
2807
2808<div class="doc_text">
2809
2810<p>The <tt>@llvm.compiler.used</tt> directive is the same as the
2811<tt>@llvm.used</tt> directive, except that it only prevents the compiler from
2812touching the symbol. On targets that support it, this allows an intelligent
2813linker to optimize references to the symbol without being impeded as it would be
2814by <tt>@llvm.used</tt>.</p>
2815
2816<p>This is a rare construct that should only be used in rare circumstances, and
2817should not be exposed to source languages.</p>
2818
2819</div>
2820
2821<!-- ======================================================================= -->
2822<div class="doc_subsection">
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2823<a name="intg_global_ctors">The '<tt>llvm.global_ctors</tt>' Global Variable</a>
2824</div>
2825
2826<div class="doc_text">
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]2827<pre>
2828%0 = type { i32, void ()* }
David Chisnallb492b812010-04-30 19:27:35 +0000[diff] [blame]2829@llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor }]
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]2830</pre>
2831<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.
2832</p>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2833
2834</div>
2835
2836<!-- ======================================================================= -->
2837<div class="doc_subsection">
2838<a name="intg_global_dtors">The '<tt>llvm.global_dtors</tt>' Global Variable</a>
2839</div>
2840
2841<div class="doc_text">
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]2842<pre>
2843%0 = type { i32, void ()* }
David Chisnallb492b812010-04-30 19:27:35 +0000[diff] [blame]2844@llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor }]
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]2845</pre>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2846
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]2847<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.
2848</p>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2849
2850</div>
2851
2852
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2853<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2854<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
2855<!-- *********************************************************************** -->
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2856
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2857<div class="doc_text">
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2858
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2859<p>The LLVM instruction set consists of several different classifications of
2860 instructions: <a href="#terminators">terminator
2861 instructions</a>, <a href="#binaryops">binary instructions</a>,
2862 <a href="#bitwiseops">bitwise binary instructions</a>,
2863 <a href="#memoryops">memory instructions</a>, and
2864 <a href="#otherops">other instructions</a>.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2865
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2866</div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2867
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2868<!-- ======================================================================= -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2869<div class="doc_subsection"> <a name="terminators">Terminator
2870Instructions</a> </div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2871
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2872<div class="doc_text">
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2873
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2874<p>As mentioned <a href="#functionstructure">previously</a>, every basic block
2875 in a program ends with a "Terminator" instruction, which indicates which
2876 block should be executed after the current block is finished. These
2877 terminator instructions typically yield a '<tt>void</tt>' value: they produce
2878 control flow, not values (the one exception being the
2879 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
2880
Duncan Sands626b0242010-04-15 20:35:54 +0000[diff] [blame]2881<p>There are seven different terminator instructions: the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2882 '<a href="#i_ret"><tt>ret</tt></a>' instruction, the
2883 '<a href="#i_br"><tt>br</tt></a>' instruction, the
2884 '<a href="#i_switch"><tt>switch</tt></a>' instruction, the
Bill Wendling33fef7e2009-11-02 00:25:26 +0000[diff] [blame]2885 '<a href="#i_indirectbr">'<tt>indirectbr</tt></a>' Instruction, the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2886 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the
2887 '<a href="#i_unwind"><tt>unwind</tt></a>' instruction, and the
2888 '<a href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2889
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2890</div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2891
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2892<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2893<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
2894Instruction</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2895
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2896<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2897
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2898<h5>Syntax:</h5>
Dan Gohmancc3132e2008-10-04 19:00:07 +0000[diff] [blame]2899<pre>
2900 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]2901 ret void <i>; Return from void function</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2902</pre>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]2903
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2904<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2905<p>The '<tt>ret</tt>' instruction is used to return control flow (and optionally
2906 a value) from a function back to the caller.</p>
2907
2908<p>There are two forms of the '<tt>ret</tt>' instruction: one that returns a
2909 value and then causes control flow, and one that just causes control flow to
2910 occur.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]2911
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2912<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2913<p>The '<tt>ret</tt>' instruction optionally accepts a single argument, the
2914 return value. The type of the return value must be a
2915 '<a href="#t_firstclass">first class</a>' type.</p>
Dan Gohmancc3132e2008-10-04 19:00:07 +0000[diff] [blame]2916
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2917<p>A function is not <a href="#wellformed">well formed</a> if it it has a
2918 non-void return type and contains a '<tt>ret</tt>' instruction with no return
2919 value or a return value with a type that does not match its type, or if it
2920 has a void return type and contains a '<tt>ret</tt>' instruction with a
2921 return value.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]2922
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2923<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2924<p>When the '<tt>ret</tt>' instruction is executed, control flow returns back to
2925 the calling function's context. If the caller is a
2926 "<a href="#i_call"><tt>call</tt></a>" instruction, execution continues at the
2927 instruction after the call. If the caller was an
2928 "<a href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues at
2929 the beginning of the "normal" destination block. If the instruction returns
2930 a value, that value shall set the call or invoke instruction's return
2931 value.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]2932
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2933<h5>Example:</h5>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]2934<pre>
2935 ret i32 5 <i>; Return an integer value of 5</i>
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]2936 ret void <i>; Return from a void function</i>
Bill Wendling050ee8f2009-02-28 22:12:54 +0000[diff] [blame]2937 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]2938</pre>
Dan Gohman3065b612009-01-12 23:12:39 +0000[diff] [blame]2939
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2940</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2941<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2942<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2943
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2944<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2945
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2946<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2947<pre>
2948 br i1 &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;<br> br label &lt;dest&gt; <i>; Unconditional branch</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2949</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2950
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2951<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2952<p>The '<tt>br</tt>' instruction is used to cause control flow to transfer to a
2953 different basic block in the current function. There are two forms of this
2954 instruction, corresponding to a conditional branch and an unconditional
2955 branch.</p>
2956
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2957<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2958<p>The conditional branch form of the '<tt>br</tt>' instruction takes a single
2959 '<tt>i1</tt>' value and two '<tt>label</tt>' values. The unconditional form
2960 of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>' value as a
2961 target.</p>
2962
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2963<h5>Semantics:</h5>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]2964<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2965 argument is evaluated. If the value is <tt>true</tt>, control flows to the
2966 '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
2967 control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
2968
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2969<h5>Example:</h5>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]2970<pre>
2971Test:
2972 %cond = <a href="#i_icmp">icmp</a> eq i32 %a, %b
2973 br i1 %cond, label %IfEqual, label %IfUnequal
2974IfEqual:
2975 <a href="#i_ret">ret</a> i32 1
2976IfUnequal:
2977 <a href="#i_ret">ret</a> i32 0
2978</pre>
2979
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2980</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2981
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2982<!-- _______________________________________________________________________ -->
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]2983<div class="doc_subsubsection">
2984 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
2985</div>
2986
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2987<div class="doc_text">
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]2988
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2989<h5>Syntax:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]2990<pre>
2991 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
2992</pre>
2993
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2994<h5>Overview:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]2995<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]2996 several different places. It is a generalization of the '<tt>br</tt>'
2997 instruction, allowing a branch to occur to one of many possible
2998 destinations.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]2999
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3000<h5>Arguments:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3001<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3002 comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination,
3003 and an array of pairs of comparison value constants and '<tt>label</tt>'s.
3004 The table is not allowed to contain duplicate constant entries.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3005
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3006<h5>Semantics:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3007<p>The <tt>switch</tt> instruction specifies a table of values and
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3008 destinations. When the '<tt>switch</tt>' instruction is executed, this table
3009 is searched for the given value. If the value is found, control flow is
Benjamin Kramer0f420382009-10-12 14:46:08 +0000[diff] [blame]3010 transferred to the corresponding destination; otherwise, control flow is
3011 transferred to the default destination.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3012
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3013<h5>Implementation:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3014<p>Depending on properties of the target machine and the particular
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3015 <tt>switch</tt> instruction, this instruction may be code generated in
3016 different ways. For example, it could be generated as a series of chained
3017 conditional branches or with a lookup table.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3018
3019<h5>Example:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3020<pre>
3021 <i>; Emulate a conditional br instruction</i>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]3022 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Dan Gohman623806e2009-01-04 23:44:43 +0000[diff] [blame]3023 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3024
3025 <i>; Emulate an unconditional br instruction</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]3026 switch i32 0, label %dest [ ]
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3027
3028 <i>; Implement a jump table:</i>
Dan Gohman623806e2009-01-04 23:44:43 +0000[diff] [blame]3029 switch i32 %val, label %otherwise [ i32 0, label %onzero
3030 i32 1, label %onone
3031 i32 2, label %ontwo ]
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3032</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3033
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3034</div>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3035
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3036
3037<!-- _______________________________________________________________________ -->
3038<div class="doc_subsubsection">
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]3039 <a name="i_indirectbr">'<tt>indirectbr</tt>' Instruction</a>
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3040</div>
3041
3042<div class="doc_text">
3043
3044<h5>Syntax:</h5>
3045<pre>
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]3046 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]3047</pre>
3048
3049<h5>Overview:</h5>
3050
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]3051<p>The '<tt>indirectbr</tt>' instruction implements an indirect branch to a label
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3052 within the current function, whose address is specified by
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]3053 "<tt>address</tt>". Address must be derived from a <a
3054 href="#blockaddress">blockaddress</a> constant.</p>
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3055
3056<h5>Arguments:</h5>
3057
3058<p>The '<tt>address</tt>' argument is the address of the label to jump to. The
3059 rest of the arguments indicate the full set of possible destinations that the
3060 address may point to. Blocks are allowed to occur multiple times in the
3061 destination list, though this isn't particularly useful.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3062
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3063<p>This destination list is required so that dataflow analysis has an accurate
3064 understanding of the CFG.</p>
3065
3066<h5>Semantics:</h5>
3067
3068<p>Control transfers to the block specified in the address argument. All
3069 possible destination blocks must be listed in the label list, otherwise this
3070 instruction has undefined behavior. This implies that jumps to labels
3071 defined in other functions have undefined behavior as well.</p>
3072
3073<h5>Implementation:</h5>
3074
3075<p>This is typically implemented with a jump through a register.</p>
3076
3077<h5>Example:</h5>
3078<pre>
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]3079 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3080</pre>
3081
3082</div>
3083
3084
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3085<!-- _______________________________________________________________________ -->
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3086<div class="doc_subsubsection">
3087 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
3088</div>
3089
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3090<div class="doc_text">
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3091
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3092<h5>Syntax:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3093<pre>
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]3094 &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]3095 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3096</pre>
3097
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]3098<h5>Overview:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3099<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3100 function, with the possibility of control flow transfer to either the
3101 '<tt>normal</tt>' label or the '<tt>exception</tt>' label. If the callee
3102 function returns with the "<tt><a href="#i_ret">ret</a></tt>" instruction,
3103 control flow will return to the "normal" label. If the callee (or any
3104 indirect callees) returns with the "<a href="#i_unwind"><tt>unwind</tt></a>"
3105 instruction, control is interrupted and continued at the dynamically nearest
3106 "exception" label.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3107
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3108<h5>Arguments:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3109<p>This instruction requires several arguments:</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3110
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3111<ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3112 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
3113 convention</a> the call should use. If none is specified, the call
3114 defaults to using C calling conventions.</li>
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]3115
3116 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3117 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
3118 '<tt>inreg</tt>' attributes are valid here.</li>
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]3119
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3120 <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]3121 function value being invoked. In most cases, this is a direct function
3122 invocation, but indirect <tt>invoke</tt>s are just as possible, branching
3123 off an arbitrary pointer to function value.</li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3124
3125 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3126 function to be invoked. </li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3127
3128 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]3129 signature argument types and parameter attributes. All arguments must be
3130 of <a href="#t_firstclass">first class</a> type. If the function
3131 signature indicates the function accepts a variable number of arguments,
3132 the extra arguments can be specified.</li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3133
3134 <li>'<tt>normal label</tt>': the label reached when the called function
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3135 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3136
3137 <li>'<tt>exception label</tt>': the label reached when a callee returns with
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3138 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3139
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]3140 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3141 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
3142 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3143</ol>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3144
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3145<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3146<p>This instruction is designed to operate as a standard
3147 '<tt><a href="#i_call">call</a></tt>' instruction in most regards. The
3148 primary difference is that it establishes an association with a label, which
3149 is used by the runtime library to unwind the stack.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3150
3151<p>This instruction is used in languages with destructors to ensure that proper
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3152 cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
3153 exception. Additionally, this is important for implementation of
3154 '<tt>catch</tt>' clauses in high-level languages that support them.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3155
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3156<p>For the purposes of the SSA form, the definition of the value returned by the
3157 '<tt>invoke</tt>' instruction is deemed to occur on the edge from the current
3158 block to the "normal" label. If the callee unwinds then no return value is
3159 available.</p>
Dan Gohman9069d892009-05-22 21:47:08 +0000[diff] [blame]3160
Chris Lattner97257f82010-01-15 18:08:37 +0000[diff] [blame]3161<p>Note that the code generator does not yet completely support unwind, and
3162that the invoke/unwind semantics are likely to change in future versions.</p>
3163
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3164<h5>Example:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3165<pre>
Nick Lewycky084ab472008-03-16 07:18:12 +0000[diff] [blame]3166 %retval = invoke i32 @Test(i32 15) to label %Continue
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]3167 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewycky084ab472008-03-16 07:18:12 +0000[diff] [blame]3168 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]3169 unwind label %TestCleanup <i>; {i32}:retval set</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3170</pre>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3171
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3172</div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3173
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]3174<!-- _______________________________________________________________________ -->
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3175
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3176<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
3177Instruction</a> </div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3178
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3179<div class="doc_text">
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3180
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]3181<h5>Syntax:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3182<pre>
3183 unwind
3184</pre>
3185
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]3186<h5>Overview:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3187<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3188 at the first callee in the dynamic call stack which used
3189 an <a href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call.
3190 This is primarily used to implement exception handling.</p>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3191
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]3192<h5>Semantics:</h5>
Chris Lattnerfe8519c2008-04-19 21:01:16 +0000[diff] [blame]3193<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3194 immediately halt. The dynamic call stack is then searched for the
3195 first <a href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack.
3196 Once found, execution continues at the "exceptional" destination block
3197 specified by the <tt>invoke</tt> instruction. If there is no <tt>invoke</tt>
3198 instruction in the dynamic call chain, undefined behavior results.</p>
3199
Chris Lattner97257f82010-01-15 18:08:37 +0000[diff] [blame]3200<p>Note that the code generator does not yet completely support unwind, and
3201that the invoke/unwind semantics are likely to change in future versions.</p>
3202
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3203</div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3204
3205<!-- _______________________________________________________________________ -->
3206
3207<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
3208Instruction</a> </div>
3209
3210<div class="doc_text">
3211
3212<h5>Syntax:</h5>
3213<pre>
3214 unreachable
3215</pre>
3216
3217<h5>Overview:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3218<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3219 instruction is used to inform the optimizer that a particular portion of the
3220 code is not reachable. This can be used to indicate that the code after a
3221 no-return function cannot be reached, and other facts.</p>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3222
3223<h5>Semantics:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3224<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3225
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3226</div>
3227
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3228<!-- ======================================================================= -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3229<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3230
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3231<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3232
3233<p>Binary operators are used to do most of the computation in a program. They
3234 require two operands of the same type, execute an operation on them, and
3235 produce a single value. The operands might represent multiple data, as is
3236 the case with the <a href="#t_vector">vector</a> data type. The result value
3237 has the same type as its operands.</p>
3238
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3239<p>There are several different binary operators:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3240
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3241</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3242
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3243<!-- _______________________________________________________________________ -->
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3244<div class="doc_subsubsection">
3245 <a name="i_add">'<tt>add</tt>' Instruction</a>
3246</div>
3247
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3248<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3249
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3250<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3251<pre>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3252 &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]3253 &lt;result&gt; = add nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3254 &lt;result&gt; = add nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3255 &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]3256</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3257
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3258<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3259<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3260
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3261<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3262<p>The two arguments to the '<tt>add</tt>' instruction must
3263 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3264 integer values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3265
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3266<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3267<p>The value produced is the integer sum of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3268
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3269<p>If the sum has unsigned overflow, the result returned is the mathematical
3270 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]3271
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3272<p>Because LLVM integers use a two's complement representation, this instruction
3273 is appropriate for both signed and unsigned integers.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3274
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3275<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3276 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3277 <tt>nsw</tt> keywords are present, the result value of the <tt>add</tt>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]3278 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3279 respectively, occurs.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3280
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3281<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3282<pre>
3283 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3284</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3285
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3286</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3287
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3288<!-- _______________________________________________________________________ -->
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3289<div class="doc_subsubsection">
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3290 <a name="i_fadd">'<tt>fadd</tt>' Instruction</a>
3291</div>
3292
3293<div class="doc_text">
3294
3295<h5>Syntax:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3296<pre>
3297 &lt;result&gt; = fadd &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3298</pre>
3299
3300<h5>Overview:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3301<p>The '<tt>fadd</tt>' instruction returns the sum of its two operands.</p>
3302
3303<h5>Arguments:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3304<p>The two arguments to the '<tt>fadd</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3305 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3306 floating point values. Both arguments must have identical types.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3307
3308<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3309<p>The value produced is the floating point sum of the two operands.</p>
3310
3311<h5>Example:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3312<pre>
3313 &lt;result&gt; = fadd float 4.0, %var <i>; yields {float}:result = 4.0 + %var</i>
3314</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3315
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3316</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3317
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3318<!-- _______________________________________________________________________ -->
3319<div class="doc_subsubsection">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3320 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
3321</div>
3322
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3323<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3324
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3325<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3326<pre>
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3327 &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]3328 &lt;result&gt; = sub nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3329 &lt;result&gt; = sub nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3330 &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]3331</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3332
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3333<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3334<p>The '<tt>sub</tt>' instruction returns the difference of its two
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3335 operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3336
3337<p>Note that the '<tt>sub</tt>' instruction is used to represent the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3338 '<tt>neg</tt>' instruction present in most other intermediate
3339 representations.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3340
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3341<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3342<p>The two arguments to the '<tt>sub</tt>' instruction must
3343 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3344 integer values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3345
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3346<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3347<p>The value produced is the integer difference of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3348
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3349<p>If the difference has unsigned overflow, the result returned is the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3350 mathematical result modulo 2<sup>n</sup>, where n is the bit width of the
3351 result.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3352
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3353<p>Because LLVM integers use a two's complement representation, this instruction
3354 is appropriate for both signed and unsigned integers.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3355
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3356<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3357 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3358 <tt>nsw</tt> keywords are present, the result value of the <tt>sub</tt>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]3359 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3360 respectively, occurs.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3361
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3362<h5>Example:</h5>
Bill Wendling2d8b9a82007-05-29 09:42:13 +0000[diff] [blame]3363<pre>
3364 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]3365 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3366</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3367
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3368</div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3369
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3370<!-- _______________________________________________________________________ -->
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3371<div class="doc_subsubsection">
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3372 <a name="i_fsub">'<tt>fsub</tt>' Instruction</a>
3373</div>
3374
3375<div class="doc_text">
3376
3377<h5>Syntax:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3378<pre>
3379 &lt;result&gt; = fsub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3380</pre>
3381
3382<h5>Overview:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3383<p>The '<tt>fsub</tt>' instruction returns the difference of its two
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3384 operands.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3385
3386<p>Note that the '<tt>fsub</tt>' instruction is used to represent the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3387 '<tt>fneg</tt>' instruction present in most other intermediate
3388 representations.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3389
3390<h5>Arguments:</h5>
Bill Wendling972b7202009-07-20 02:32:41 +0000[diff] [blame]3391<p>The two arguments to the '<tt>fsub</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3392 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3393 floating point values. Both arguments must have identical types.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3394
3395<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3396<p>The value produced is the floating point difference of the two operands.</p>
3397
3398<h5>Example:</h5>
3399<pre>
3400 &lt;result&gt; = fsub float 4.0, %var <i>; yields {float}:result = 4.0 - %var</i>
3401 &lt;result&gt; = fsub float -0.0, %val <i>; yields {float}:result = -%var</i>
3402</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3403
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3404</div>
3405
3406<!-- _______________________________________________________________________ -->
3407<div class="doc_subsubsection">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3408 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
3409</div>
3410
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3411<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3412
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3413<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3414<pre>
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3415 &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]3416 &lt;result&gt; = mul nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3417 &lt;result&gt; = mul nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3418 &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]3419</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3420
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3421<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3422<p>The '<tt>mul</tt>' instruction returns the product of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3423
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3424<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3425<p>The two arguments to the '<tt>mul</tt>' instruction must
3426 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3427 integer values. Both arguments must have identical types.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3428
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3429<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3430<p>The value produced is the integer product of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3431
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3432<p>If the result of the multiplication has unsigned overflow, the result
3433 returned is the mathematical result modulo 2<sup>n</sup>, where n is the bit
3434 width of the result.</p>
3435
3436<p>Because LLVM integers use a two's complement representation, and the result
3437 is the same width as the operands, this instruction returns the correct
3438 result for both signed and unsigned integers. If a full product
3439 (e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands should
3440 be sign-extended or zero-extended as appropriate to the width of the full
3441 product.</p>
3442
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3443<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3444 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3445 <tt>nsw</tt> keywords are present, the result value of the <tt>mul</tt>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]3446 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3447 respectively, occurs.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3448
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3449<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3450<pre>
3451 &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3452</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3453
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3454</div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3455
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3456<!-- _______________________________________________________________________ -->
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3457<div class="doc_subsubsection">
3458 <a name="i_fmul">'<tt>fmul</tt>' Instruction</a>
3459</div>
3460
3461<div class="doc_text">
3462
3463<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3464<pre>
3465 &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]3466</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3467
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3468<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3469<p>The '<tt>fmul</tt>' instruction returns the product of its two operands.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3470
3471<h5>Arguments:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3472<p>The two arguments to the '<tt>fmul</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3473 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3474 floating point values. Both arguments must have identical types.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3475
3476<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3477<p>The value produced is the floating point product of the two operands.</p>
3478
3479<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3480<pre>
3481 &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]3482</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3483
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3484</div>
3485
3486<!-- _______________________________________________________________________ -->
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3487<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
3488</a></div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3489
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3490<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3491
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3492<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3493<pre>
3494 &lt;result&gt; = udiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3495</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3496
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3497<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3498<p>The '<tt>udiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3499
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3500<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3501<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3502 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3503 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3504
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3505<h5>Semantics:</h5>
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3506<p>The value produced is the unsigned integer quotient of the two operands.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3507
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3508<p>Note that unsigned integer division and signed integer division are distinct
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3509 operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
3510
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3511<p>Division by zero leads to undefined behavior.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3512
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3513<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3514<pre>
3515 &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3516</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3517
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3518</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3519
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3520<!-- _______________________________________________________________________ -->
3521<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
3522</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3523
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3524<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3525
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3526<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3527<pre>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3528 &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]3529 &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]3530</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3531
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3532<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3533<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3534
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3535<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3536<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3537 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3538 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3539
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3540<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3541<p>The value produced is the signed integer quotient of the two operands rounded
3542 towards zero.</p>
3543
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3544<p>Note that signed integer division and unsigned integer division are distinct
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3545 operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
3546
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3547<p>Division by zero leads to undefined behavior. Overflow also leads to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3548 undefined behavior; this is a rare case, but can occur, for example, by doing
3549 a 32-bit division of -2147483648 by -1.</p>
3550
Dan Gohman71dfd782009-07-22 00:04:19 +0000[diff] [blame]3551<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohman57255802010-04-23 15:23:32 +0000[diff] [blame]3552 <tt>sdiv</tt> is a <a href="#trapvalues">trap value</a> if the result would
3553 be rounded or if overflow would occur.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3554
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3555<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3556<pre>
3557 &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3558</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3559
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3560</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3561
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3562<!-- _______________________________________________________________________ -->
3563<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3564Instruction</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3565
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3566<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3567
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3568<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3569<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +0000[diff] [blame]3570 &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]3571</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3572
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3573<h5>Overview:</h5>
3574<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3575
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3576<h5>Arguments:</h5>
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]3577<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3578 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3579 floating point values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3580
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3581<h5>Semantics:</h5>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3582<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3583
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3584<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3585<pre>
3586 &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]3587</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3588
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3589</div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3590
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3591<!-- _______________________________________________________________________ -->
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3592<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
3593</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3594
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3595<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3596
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3597<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3598<pre>
3599 &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]3600</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3601
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3602<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3603<p>The '<tt>urem</tt>' instruction returns the remainder from the unsigned
3604 division of its two arguments.</p>
3605
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3606<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3607<p>The two arguments to the '<tt>urem</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3608 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3609 values. Both arguments must have identical types.</p>
3610
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3611<h5>Semantics:</h5>
3612<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3613 This instruction always performs an unsigned division to get the
3614 remainder.</p>
3615
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3616<p>Note that unsigned integer remainder and signed integer remainder are
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3617 distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
3618
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3619<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3620
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3621<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3622<pre>
3623 &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3624</pre>
3625
3626</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3627
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3628<!-- _______________________________________________________________________ -->
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3629<div class="doc_subsubsection">
3630 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
3631</div>
3632
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3633<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3634
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3635<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3636<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +0000[diff] [blame]3637 &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]3638</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3639
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3640<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3641<p>The '<tt>srem</tt>' instruction returns the remainder from the signed
3642 division of its two operands. This instruction can also take
3643 <a href="#t_vector">vector</a> versions of the values in which case the
3644 elements must be integers.</p>
Chris Lattnerb8f816e2008-01-04 04:33:49 +0000[diff] [blame]3645
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3646<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3647<p>The two arguments to the '<tt>srem</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3648 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3649 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3650
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3651<h5>Semantics:</h5>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3652<p>This instruction returns the <i>remainder</i> of a division (where the result
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3653 has the same sign as the dividend, <tt>op1</tt>), not the <i>modulo</i>
3654 operator (where the result has the same sign as the divisor, <tt>op2</tt>) of
3655 a value. For more information about the difference,
3656 see <a href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
3657 Math Forum</a>. For a table of how this is implemented in various languages,
3658 please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
3659 Wikipedia: modulo operation</a>.</p>
3660
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3661<p>Note that signed integer remainder and unsigned integer remainder are
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3662 distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
3663
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3664<p>Taking the remainder of a division by zero leads to undefined behavior.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3665 Overflow also leads to undefined behavior; this is a rare case, but can
3666 occur, for example, by taking the remainder of a 32-bit division of
3667 -2147483648 by -1. (The remainder doesn't actually overflow, but this rule
3668 lets srem be implemented using instructions that return both the result of
3669 the division and the remainder.)</p>
3670
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3671<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3672<pre>
3673 &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3674</pre>
3675
3676</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3677
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3678<!-- _______________________________________________________________________ -->
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3679<div class="doc_subsubsection">
3680 <a name="i_frem">'<tt>frem</tt>' Instruction</a> </div>
3681
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3682<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3683
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3684<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3685<pre>
3686 &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]3687</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3688
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3689<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3690<p>The '<tt>frem</tt>' instruction returns the remainder from the division of
3691 its two operands.</p>
3692
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3693<h5>Arguments:</h5>
3694<p>The two arguments to the '<tt>frem</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3695 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3696 floating point values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3697
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3698<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3699<p>This instruction returns the <i>remainder</i> of a division. The remainder
3700 has the same sign as the dividend.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3701
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3702<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3703<pre>
3704 &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]3705</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3706
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3707</div>
Robert Bocchino820bc75b2006-02-17 21:18:08 +0000[diff] [blame]3708
Reid Spencer2ab01932007-02-02 13:57:07 +0000[diff] [blame]3709<!-- ======================================================================= -->
3710<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
3711Operations</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3712
Reid Spencer2ab01932007-02-02 13:57:07 +0000[diff] [blame]3713<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3714
3715<p>Bitwise binary operators are used to do various forms of bit-twiddling in a
3716 program. They are generally very efficient instructions and can commonly be
3717 strength reduced from other instructions. They require two operands of the
3718 same type, execute an operation on them, and produce a single value. The
3719 resulting value is the same type as its operands.</p>
3720
Reid Spencer2ab01932007-02-02 13:57:07 +0000[diff] [blame]3721</div>
3722
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3723<!-- _______________________________________________________________________ -->
3724<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
3725Instruction</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3726
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3727<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3728
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3729<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3730<pre>
3731 &lt;result&gt; = shl &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3732</pre>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3733
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3734<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3735<p>The '<tt>shl</tt>' instruction returns the first operand shifted to the left
3736 a specified number of bits.</p>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3737
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3738<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3739<p>Both arguments to the '<tt>shl</tt>' instruction must be the
3740 same <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3741 integer type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3742
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3743<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3744<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod
3745 2<sup>n</sup>, where <tt>n</tt> is the width of the result. If <tt>op2</tt>
3746 is (statically or dynamically) negative or equal to or larger than the number
3747 of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3748 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3749 shift amount in <tt>op2</tt>.</p>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3750
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3751<h5>Example:</h5>
3752<pre>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3753 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
3754 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
3755 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3756 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Mon P Wang4dd832d2008-12-09 05:46:39 +0000[diff] [blame]3757 &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]3758</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3759
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3760</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3761
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3762<!-- _______________________________________________________________________ -->
3763<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
3764Instruction</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3765
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3766<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3767
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3768<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3769<pre>
3770 &lt;result&gt; = lshr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3771</pre>
3772
3773<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3774<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
3775 operand shifted to the right a specified number of bits with zero fill.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3776
3777<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3778<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3779 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3780 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3781
3782<h5>Semantics:</h5>
3783<p>This instruction always performs a logical shift right operation. The most
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3784 significant bits of the result will be filled with zero bits after the shift.
3785 If <tt>op2</tt> is (statically or dynamically) equal to or larger than the
3786 number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3787 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3788 shift amount in <tt>op2</tt>.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3789
3790<h5>Example:</h5>
3791<pre>
3792 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
3793 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
3794 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
3795 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3796 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Mon P Wang4dd832d2008-12-09 05:46:39 +0000[diff] [blame]3797 &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]3798</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3799
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3800</div>
3801
Reid Spencer2ab01932007-02-02 13:57:07 +0000[diff] [blame]3802<!-- _______________________________________________________________________ -->
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3803<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
3804Instruction</a> </div>
3805<div class="doc_text">
3806
3807<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3808<pre>
3809 &lt;result&gt; = ashr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3810</pre>
3811
3812<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3813<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
3814 operand shifted to the right a specified number of bits with sign
3815 extension.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3816
3817<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3818<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3819 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3820 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3821
3822<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3823<p>This instruction always performs an arithmetic shift right operation, The
3824 most significant bits of the result will be filled with the sign bit
3825 of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
3826 larger than the number of bits in <tt>op1</tt>, the result is undefined. If
3827 the arguments are vectors, each vector element of <tt>op1</tt> is shifted by
3828 the corresponding shift amount in <tt>op2</tt>.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3829
3830<h5>Example:</h5>
3831<pre>
3832 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
3833 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
3834 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
3835 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3836 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Mon P Wang4dd832d2008-12-09 05:46:39 +0000[diff] [blame]3837 &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]3838</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3839
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3840</div>
3841
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3842<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3843<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
3844Instruction</a> </div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3845
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3846<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3847
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3848<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3849<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +0000[diff] [blame]3850 &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]3851</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3852
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3853<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3854<p>The '<tt>and</tt>' instruction returns the bitwise logical and of its two
3855 operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3856
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3857<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3858<p>The two arguments to the '<tt>and</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3859 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3860 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3861
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3862<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3863<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3864
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3865<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3866 <tbody>
3867 <tr>
3868 <td>In0</td>
3869 <td>In1</td>
3870 <td>Out</td>
3871 </tr>
3872 <tr>
3873 <td>0</td>
3874 <td>0</td>
3875 <td>0</td>
3876 </tr>
3877 <tr>
3878 <td>0</td>
3879 <td>1</td>
3880 <td>0</td>
3881 </tr>
3882 <tr>
3883 <td>1</td>
3884 <td>0</td>
3885 <td>0</td>
3886 </tr>
3887 <tr>
3888 <td>1</td>
3889 <td>1</td>
3890 <td>1</td>
3891 </tr>
3892 </tbody>
3893</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3894
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3895<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3896<pre>
3897 &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]3898 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
3899 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3900</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3901</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3902<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3903<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3904
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3905<div class="doc_text">
3906
3907<h5>Syntax:</h5>
3908<pre>
3909 &lt;result&gt; = or &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3910</pre>
3911
3912<h5>Overview:</h5>
3913<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive or of its
3914 two operands.</p>
3915
3916<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3917<p>The two arguments to the '<tt>or</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3918 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3919 values. Both arguments must have identical types.</p>
3920
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3921<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3922<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3923
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3924<table border="1" cellspacing="0" cellpadding="4">
3925 <tbody>
3926 <tr>
3927 <td>In0</td>
3928 <td>In1</td>
3929 <td>Out</td>
3930 </tr>
3931 <tr>
3932 <td>0</td>
3933 <td>0</td>
3934 <td>0</td>
3935 </tr>
3936 <tr>
3937 <td>0</td>
3938 <td>1</td>
3939 <td>1</td>
3940 </tr>
3941 <tr>
3942 <td>1</td>
3943 <td>0</td>
3944 <td>1</td>
3945 </tr>
3946 <tr>
3947 <td>1</td>
3948 <td>1</td>
3949 <td>1</td>
3950 </tr>
3951 </tbody>
3952</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3953
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3954<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3955<pre>
3956 &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]3957 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
3958 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3959</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3960
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3961</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3962
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3963<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3964<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
3965Instruction</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3966
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3967<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3968
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3969<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3970<pre>
3971 &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]3972</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3973
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3974<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3975<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive or of
3976 its two operands. The <tt>xor</tt> is used to implement the "one's
3977 complement" operation, which is the "~" operator in C.</p>
3978
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3979<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3980<p>The two arguments to the '<tt>xor</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3981 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3982 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3983
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3984<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3985<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3986
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3987<table border="1" cellspacing="0" cellpadding="4">
3988 <tbody>
3989 <tr>
3990 <td>In0</td>
3991 <td>In1</td>
3992 <td>Out</td>
3993 </tr>
3994 <tr>
3995 <td>0</td>
3996 <td>0</td>
3997 <td>0</td>
3998 </tr>
3999 <tr>
4000 <td>0</td>
4001 <td>1</td>
4002 <td>1</td>
4003 </tr>
4004 <tr>
4005 <td>1</td>
4006 <td>0</td>
4007 <td>1</td>
4008 </tr>
4009 <tr>
4010 <td>1</td>
4011 <td>1</td>
4012 <td>0</td>
4013 </tr>
4014 </tbody>
4015</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4016
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4017<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4018<pre>
4019 &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4020 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
4021 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
4022 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4023</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4024
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4025</div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4026
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4027<!-- ======================================================================= -->
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4028<div class="doc_subsection">
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4029 <a name="vectorops">Vector Operations</a>
4030</div>
4031
4032<div class="doc_text">
4033
4034<p>LLVM supports several instructions to represent vector operations in a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4035 target-independent manner. These instructions cover the element-access and
4036 vector-specific operations needed to process vectors effectively. While LLVM
4037 does directly support these vector operations, many sophisticated algorithms
4038 will want to use target-specific intrinsics to take full advantage of a
4039 specific target.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4040
4041</div>
4042
4043<!-- _______________________________________________________________________ -->
4044<div class="doc_subsubsection">
4045 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
4046</div>
4047
4048<div class="doc_text">
4049
4050<h5>Syntax:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4051<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4052 &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]4053</pre>
4054
4055<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4056<p>The '<tt>extractelement</tt>' instruction extracts a single scalar element
4057 from a vector at a specified index.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4058
4059
4060<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4061<p>The first operand of an '<tt>extractelement</tt>' instruction is a value
4062 of <a href="#t_vector">vector</a> type. The second operand is an index
4063 indicating the position from which to extract the element. The index may be
4064 a variable.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4065
4066<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4067<p>The result is a scalar of the same type as the element type of
4068 <tt>val</tt>. Its value is the value at position <tt>idx</tt> of
4069 <tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4070 results are undefined.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4071
4072<h5>Example:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4073<pre>
Gabor Greif03ab4dc2009-10-28 13:14:50 +0000[diff] [blame]4074 &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]4075</pre>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4076
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4077</div>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4078
4079<!-- _______________________________________________________________________ -->
4080<div class="doc_subsubsection">
4081 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
4082</div>
4083
4084<div class="doc_text">
4085
4086<h5>Syntax:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4087<pre>
Dan Gohman43ba0672008-05-12 23:38:42 +0000[diff] [blame]4088 &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]4089</pre>
4090
4091<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4092<p>The '<tt>insertelement</tt>' instruction inserts a scalar element into a
4093 vector at a specified index.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4094
4095<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4096<p>The first operand of an '<tt>insertelement</tt>' instruction is a value
4097 of <a href="#t_vector">vector</a> type. The second operand is a scalar value
4098 whose type must equal the element type of the first operand. The third
4099 operand is an index indicating the position at which to insert the value.
4100 The index may be a variable.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4101
4102<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4103<p>The result is a vector of the same type as <tt>val</tt>. Its element values
4104 are those of <tt>val</tt> except at position <tt>idx</tt>, where it gets the
4105 value <tt>elt</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4106 results are undefined.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4107
4108<h5>Example:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4109<pre>
Gabor Greif03ab4dc2009-10-28 13:14:50 +0000[diff] [blame]4110 &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]4111</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4112
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4113</div>
4114
4115<!-- _______________________________________________________________________ -->
4116<div class="doc_subsubsection">
4117 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
4118</div>
4119
4120<div class="doc_text">
4121
4122<h5>Syntax:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4123<pre>
Mon P Wang25f01062008-11-10 04:46:22 +0000[diff] [blame]4124 &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]4125</pre>
4126
4127<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4128<p>The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
4129 from two input vectors, returning a vector with the same element type as the
4130 input and length that is the same as the shuffle mask.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4131
4132<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4133<p>The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
4134 with types that match each other. The third argument is a shuffle mask whose
4135 element type is always 'i32'. The result of the instruction is a vector
4136 whose length is the same as the shuffle mask and whose element type is the
4137 same as the element type of the first two operands.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4138
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4139<p>The shuffle mask operand is required to be a constant vector with either
4140 constant integer or undef values.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4141
4142<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4143<p>The elements of the two input vectors are numbered from left to right across
4144 both of the vectors. The shuffle mask operand specifies, for each element of
4145 the result vector, which element of the two input vectors the result element
4146 gets. The element selector may be undef (meaning "don't care") and the
4147 second operand may be undef if performing a shuffle from only one vector.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4148
4149<h5>Example:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4150<pre>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4151 &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]4152 &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]4153 &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]4154 &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]4155 &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]4156 &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]4157 &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]4158 &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]4159</pre>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4160
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4161</div>
Tanya Lattnerb138bbe2006-04-14 19:24:33 +0000[diff] [blame]4162
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4163<!-- ======================================================================= -->
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4164<div class="doc_subsection">
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4165 <a name="aggregateops">Aggregate Operations</a>
4166</div>
4167
4168<div class="doc_text">
4169
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4170<p>LLVM supports several instructions for working with
4171 <a href="#t_aggregate">aggregate</a> values.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4172
4173</div>
4174
4175<!-- _______________________________________________________________________ -->
4176<div class="doc_subsubsection">
4177 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
4178</div>
4179
4180<div class="doc_text">
4181
4182<h5>Syntax:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4183<pre>
4184 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
4185</pre>
4186
4187<h5>Overview:</h5>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4188<p>The '<tt>extractvalue</tt>' instruction extracts the value of a member field
4189 from an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4190
4191<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4192<p>The first operand of an '<tt>extractvalue</tt>' instruction is a value
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4193 of <a href="#t_struct">struct</a>, <a href="#t_union">union</a> or
4194 <a href="#t_array">array</a> type. The operands are constant indices to
4195 specify which value to extract in a similar manner as indices in a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4196 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4197
4198<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4199<p>The result is the value at the position in the aggregate specified by the
4200 index operands.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4201
4202<h5>Example:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4203<pre>
Gabor Greif03ab4dc2009-10-28 13:14:50 +0000[diff] [blame]4204 &lt;result&gt; = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4205</pre>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4206
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4207</div>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4208
4209<!-- _______________________________________________________________________ -->
4210<div class="doc_subsubsection">
4211 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
4212</div>
4213
4214<div class="doc_text">
4215
4216<h5>Syntax:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4217<pre>
Jeffrey Yasskinf991bbb2010-01-11 19:19:26 +0000[diff] [blame]4218 &lt;result&gt; = insertvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;ty&gt; &lt;elt&gt;, &lt;idx&gt; <i>; yields &lt;aggregate type&gt;</i>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4219</pre>
4220
4221<h5>Overview:</h5>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4222<p>The '<tt>insertvalue</tt>' instruction inserts a value into a member field
4223 in an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4224
4225<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4226<p>The first operand of an '<tt>insertvalue</tt>' instruction is a value
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4227 of <a href="#t_struct">struct</a>, <a href="#t_union">union</a> or
4228 <a href="#t_array">array</a> type. The second operand is a first-class
4229 value to insert. The following operands are constant indices indicating
4230 the position at which to insert the value in a similar manner as indices in a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4231 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction. The
4232 value to insert must have the same type as the value identified by the
4233 indices.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4234
4235<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4236<p>The result is an aggregate of the same type as <tt>val</tt>. Its value is
4237 that of <tt>val</tt> except that the value at the position specified by the
4238 indices is that of <tt>elt</tt>.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4239
4240<h5>Example:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4241<pre>
Jeffrey Yasskinf991bbb2010-01-11 19:19:26 +0000[diff] [blame]4242 %agg1 = insertvalue {i32, float} undef, i32 1, 0 <i>; yields {i32 1, float undef}</i>
4243 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 <i>; yields {i32 1, float %val}</i>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4244</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4245
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4246</div>
4247
4248
4249<!-- ======================================================================= -->
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4250<div class="doc_subsection">
Chris Lattner6ab66722006-08-15 00:45:58 +0000[diff] [blame]4251 <a name="memoryops">Memory Access and Addressing Operations</a>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4252</div>
4253
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4254<div class="doc_text">
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4255
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4256<p>A key design point of an SSA-based representation is how it represents
4257 memory. In LLVM, no memory locations are in SSA form, which makes things
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]4258 very simple. This section describes how to read, write, and allocate
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4259 memory in LLVM.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4260
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4261</div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4262
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4263<!-- _______________________________________________________________________ -->
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4264<div class="doc_subsubsection">
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4265 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
4266</div>
4267
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4268<div class="doc_text">
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4269
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4270<h5>Syntax:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4271<pre>
Dan Gohman2140a742010-05-28 01:14:11 +0000[diff] [blame]4272 &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]4273</pre>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4274
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4275<h5>Overview:</h5>
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]4276<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4277 currently executing function, to be automatically released when this function
4278 returns to its caller. The object is always allocated in the generic address
4279 space (address space zero).</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4280
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4281<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4282<p>The '<tt>alloca</tt>' instruction
4283 allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt> bytes of memory on the
4284 runtime stack, returning a pointer of the appropriate type to the program.
4285 If "NumElements" is specified, it is the number of elements allocated,
4286 otherwise "NumElements" is defaulted to be one. If a constant alignment is
4287 specified, the value result of the allocation is guaranteed to be aligned to
4288 at least that boundary. If not specified, or if zero, the target can choose
4289 to align the allocation on any convenient boundary compatible with the
4290 type.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4291
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4292<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4293
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4294<h5>Semantics:</h5>
Bill Wendling9ee6a312009-05-08 20:49:29 +0000[diff] [blame]4295<p>Memory is allocated; a pointer is returned. The operation is undefined if
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4296 there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
4297 memory is automatically released when the function returns. The
4298 '<tt>alloca</tt>' instruction is commonly used to represent automatic
4299 variables that must have an address available. When the function returns
4300 (either with the <tt><a href="#i_ret">ret</a></tt>
4301 or <tt><a href="#i_unwind">unwind</a></tt> instructions), the memory is
4302 reclaimed. Allocating zero bytes is legal, but the result is undefined.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4303
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4304<h5>Example:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4305<pre>
Dan Gohman7a5acb52009-01-04 23:49:44 +0000[diff] [blame]4306 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
4307 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
4308 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
4309 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4310</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4311
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4312</div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4313
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4314<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]4315<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
4316Instruction</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4317
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4318<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4319
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4320<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4321<pre>
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4322 &lt;result&gt; = load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;]
4323 &lt;result&gt; = volatile load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;]
4324 !&lt;index&gt; = !{ i32 1 }
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4325</pre>
4326
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4327<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4328<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4329
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4330<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4331<p>The argument to the '<tt>load</tt>' instruction specifies the memory address
4332 from which to load. The pointer must point to
4333 a <a href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
4334 marked as <tt>volatile</tt>, then the optimizer is not allowed to modify the
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]4335 number or order of execution of this <tt>load</tt> with other <a
4336 href="#volatile">volatile operations</a>.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4337
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4338<p>The optional constant <tt>align</tt> argument specifies the alignment of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4339 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]4340 omitted <tt>align</tt> argument means that the operation has the preferential
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4341 alignment for the target. It is the responsibility of the code emitter to
4342 ensure that the alignment information is correct. Overestimating the
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4343 alignment results in undefined behavior. Underestimating the alignment may
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4344 produce less efficient code. An alignment of 1 is always safe.</p>
4345
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4346<p>The optional <tt>!nontemporal</tt> metadata must reference a single
4347 metatadata name &lt;index&gt; corresponding to a metadata node with
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]4348 one <tt>i32</tt> entry of value 1. The existence of
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4349 the <tt>!nontemporal</tt> metatadata on the instruction tells the optimizer
4350 and code generator that this load is not expected to be reused in the cache.
4351 The code generator may select special instructions to save cache bandwidth,
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]4352 such as the <tt>MOVNT</tt> instruction on x86.</p>
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4353
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4354<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4355<p>The location of memory pointed to is loaded. If the value being loaded is of
4356 scalar type then the number of bytes read does not exceed the minimum number
4357 of bytes needed to hold all bits of the type. For example, loading an
4358 <tt>i24</tt> reads at most three bytes. When loading a value of a type like
4359 <tt>i20</tt> with a size that is not an integral number of bytes, the result
4360 is undefined if the value was not originally written using a store of the
4361 same type.</p>
4362
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4363<h5>Examples:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4364<pre>
4365 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
4366 <a href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4367 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4368</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4369
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4370</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4371
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4372<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]4373<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
4374Instruction</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4375
Reid Spencera89fb182006-11-09 21:18:01 +0000[diff] [blame]4376<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4377
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4378<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4379<pre>
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4380 store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !<index>] <i>; yields {void}</i>
4381 volatile store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !<index>] <i>; yields {void}</i>
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4382</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4383
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4384<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4385<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4386
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4387<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4388<p>There are two arguments to the '<tt>store</tt>' instruction: a value to store
4389 and an address at which to store it. The type of the
4390 '<tt>&lt;pointer&gt;</tt>' operand must be a pointer to
4391 the <a href="#t_firstclass">first class</a> type of the
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]4392 '<tt>&lt;value&gt;</tt>' operand. If the <tt>store</tt> is marked as
4393 <tt>volatile</tt>, then the optimizer is not allowed to modify the number or
4394 order of execution of this <tt>store</tt> with other <a
4395 href="#volatile">volatile operations</a>.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4396
4397<p>The optional constant "align" argument specifies the alignment of the
4398 operation (that is, the alignment of the memory address). A value of 0 or an
4399 omitted "align" argument means that the operation has the preferential
4400 alignment for the target. It is the responsibility of the code emitter to
4401 ensure that the alignment information is correct. Overestimating the
4402 alignment results in an undefined behavior. Underestimating the alignment may
4403 produce less efficient code. An alignment of 1 is always safe.</p>
4404
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4405<p>The optional !nontemporal metadata must reference a single metatadata
4406 name <index> corresponding to a metadata node with one i32 entry of
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]4407 value 1. The existence of the !nontemporal metatadata on the
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4408 instruction tells the optimizer and code generator that this load is
4409 not expected to be reused in the cache. The code generator may
4410 select special instructions to save cache bandwidth, such as the
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]4411 MOVNT instruction on x86.</p>
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4412
4413
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]4414<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4415<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>' at the
4416 location specified by the '<tt>&lt;pointer&gt;</tt>' operand. If
4417 '<tt>&lt;value&gt;</tt>' is of scalar type then the number of bytes written
4418 does not exceed the minimum number of bytes needed to hold all bits of the
4419 type. For example, storing an <tt>i24</tt> writes at most three bytes. When
4420 writing a value of a type like <tt>i20</tt> with a size that is not an
4421 integral number of bytes, it is unspecified what happens to the extra bits
4422 that do not belong to the type, but they will typically be overwritten.</p>
4423
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4424<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4425<pre>
4426 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling8830ffe2007-10-22 05:10:05 +0000[diff] [blame]4427 store i32 3, i32* %ptr <i>; yields {void}</i>
4428 %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]4429</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4430
Reid Spencer443460a2006-11-09 21:15:49 +0000[diff] [blame]4431</div>
4432
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4433<!-- _______________________________________________________________________ -->
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4434<div class="doc_subsubsection">
4435 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
4436</div>
4437
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4438<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4439
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]4440<h5>Syntax:</h5>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4441<pre>
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]4442 &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]4443 &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]4444</pre>
4445
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]4446<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4447<p>The '<tt>getelementptr</tt>' instruction is used to get the address of a
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4448 subelement of an <a href="#t_aggregate">aggregate</a> data structure.
4449 It performs address calculation only and does not access memory.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4450
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]4451<h5>Arguments:</h5>
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]4452<p>The first argument is always a pointer, and forms the basis of the
Chris Lattnera40b9122009-07-29 06:44:13 +0000[diff] [blame]4453 calculation. The remaining arguments are indices that indicate which of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4454 elements of the aggregate object are indexed. The interpretation of each
4455 index is dependent on the type being indexed into. The first index always
4456 indexes the pointer value given as the first argument, the second index
4457 indexes a value of the type pointed to (not necessarily the value directly
4458 pointed to, since the first index can be non-zero), etc. The first type
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4459 indexed into must be a pointer value, subsequent types can be arrays,
4460 vectors, structs and unions. Note that subsequent types being indexed into
4461 can never be pointers, since that would require loading the pointer before
4462 continuing calculation.</p>
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]4463
4464<p>The type of each index argument depends on the type it is indexing into.
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4465 When indexing into a (optionally packed) structure or union, only <tt>i32</tt>
4466 integer <b>constants</b> are allowed. When indexing into an array, pointer
4467 or vector, integers of any width are allowed, and they are not required to be
Chris Lattnera40b9122009-07-29 06:44:13 +0000[diff] [blame]4468 constant.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4469
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4470<p>For example, let's consider a C code fragment and how it gets compiled to
4471 LLVM:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4472
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4473<div class="doc_code">
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4474<pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4475struct RT {
4476 char A;
Chris Lattnera446f1b2007-05-29 15:43:56 +0000[diff] [blame]4477 int B[10][20];
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4478 char C;
4479};
4480struct ST {
Chris Lattnera446f1b2007-05-29 15:43:56 +0000[diff] [blame]4481 int X;
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4482 double Y;
4483 struct RT Z;
4484};
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4485
Chris Lattnera446f1b2007-05-29 15:43:56 +0000[diff] [blame]4486int *foo(struct ST *s) {
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4487 return &amp;s[1].Z.B[5][13];
4488}
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4489</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4490</div>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4491
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4492<p>The LLVM code generated by the GCC frontend is:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4493
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4494<div class="doc_code">
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4495<pre>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]4496%RT = <a href="#namedtypes">type</a> { i8 , [10 x [20 x i32]], i8 }
4497%ST = <a href="#namedtypes">type</a> { i32, double, %RT }
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4498
Dan Gohman6b867702009-07-25 02:23:48 +0000[diff] [blame]4499define i32* @foo(%ST* %s) {
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4500entry:
4501 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
4502 ret i32* %reg
4503}
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4504</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4505</div>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4506
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]4507<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4508<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]4509 type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
4510 }</tt>' type, a structure. The second index indexes into the third element
4511 of the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
4512 i8 }</tt>' type, another structure. The third index indexes into the second
4513 element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
4514 array. The two dimensions of the array are subscripted into, yielding an
4515 '<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a
4516 pointer to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4517
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4518<p>Note that it is perfectly legal to index partially through a structure,
4519 returning a pointer to an inner element. Because of this, the LLVM code for
4520 the given testcase is equivalent to:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4521
4522<pre>
Dan Gohman6b867702009-07-25 02:23:48 +0000[diff] [blame]4523 define i32* @foo(%ST* %s) {
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4524 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]4525 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
4526 %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]4527 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
4528 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
4529 ret i32* %t5
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4530 }
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]4531</pre>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +0000[diff] [blame]4532
Dan Gohman1639c392009-07-27 21:53:46 +0000[diff] [blame]4533<p>If the <tt>inbounds</tt> keyword is present, the result value of the
Dan Gohman57255802010-04-23 15:23:32 +0000[diff] [blame]4534 <tt>getelementptr</tt> is a <a href="#trapvalues">trap value</a> if the
4535 base pointer is not an <i>in bounds</i> address of an allocated object,
4536 or if any of the addresses that would be formed by successive addition of
4537 the offsets implied by the indices to the base address with infinitely
4538 precise arithmetic are not an <i>in bounds</i> address of that allocated
4539 object. The <i>in bounds</i> addresses for an allocated object are all
4540 the addresses that point into the object, plus the address one byte past
4541 the end.</p>
Dan Gohman1639c392009-07-27 21:53:46 +0000[diff] [blame]4542
4543<p>If the <tt>inbounds</tt> keyword is not present, the offsets are added to
4544 the base address with silently-wrapping two's complement arithmetic, and
4545 the result value of the <tt>getelementptr</tt> may be outside the object
4546 pointed to by the base pointer. The result value may not necessarily be
4547 used to access memory though, even if it happens to point into allocated
4548 storage. See the <a href="#pointeraliasing">Pointer Aliasing Rules</a>
4549 section for more information.</p>
4550
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4551<p>The getelementptr instruction is often confusing. For some more insight into
4552 how it works, see <a href="GetElementPtr.html">the getelementptr FAQ</a>.</p>
Chris Lattner6ab66722006-08-15 00:45:58 +0000[diff] [blame]4553
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]4554<h5>Example:</h5>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4555<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4556 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]4557 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
4558 <i>; yields i8*:vptr</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]4559 %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]4560 <i>; yields i8*:eptr</i>
4561 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Sanjiv Gupta0c155e62009-04-25 07:27:44 +0000[diff] [blame]4562 <i>; yields i32*:iptr</i>
Sanjiv Gupta77abea02009-04-24 16:38:13 +0000[diff] [blame]4563 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4564</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4565
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4566</div>
Reid Spencer443460a2006-11-09 21:15:49 +0000[diff] [blame]4567
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4568<!-- ======================================================================= -->
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]4569<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4570</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4571
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4572<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4573
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]4574<p>The instructions in this category are the conversion instructions (casting)
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4575 which all take a single operand and a type. They perform various bit
4576 conversions on the operand.</p>
4577
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4578</div>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]4579
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]4580<!-- _______________________________________________________________________ -->
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]4581<div class="doc_subsubsection">
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4582 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
4583</div>
4584<div class="doc_text">
4585
4586<h5>Syntax:</h5>
4587<pre>
4588 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4589</pre>
4590
4591<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4592<p>The '<tt>trunc</tt>' instruction truncates its operand to the
4593 type <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4594
4595<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4596<p>The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
4597 be an <a href="#t_integer">integer</a> type, and a type that specifies the
4598 size and type of the result, which must be
4599 an <a href="#t_integer">integer</a> type. The bit size of <tt>value</tt> must
4600 be larger than the bit size of <tt>ty2</tt>. Equal sized types are not
4601 allowed.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4602
4603<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4604<p>The '<tt>trunc</tt>' instruction truncates the high order bits
4605 in <tt>value</tt> and converts the remaining bits to <tt>ty2</tt>. Since the
4606 source size must be larger than the destination size, <tt>trunc</tt> cannot
4607 be a <i>no-op cast</i>. It will always truncate bits.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4608
4609<h5>Example:</h5>
4610<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4611 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]4612 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
Gabor Greiff50fd572009-10-28 09:21:30 +0000[diff] [blame]4613 %Z = trunc i32 122 to i1 <i>; yields i1:false</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4614</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4615
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4616</div>
4617
4618<!-- _______________________________________________________________________ -->
4619<div class="doc_subsubsection">
4620 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
4621</div>
4622<div class="doc_text">
4623
4624<h5>Syntax:</h5>
4625<pre>
4626 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4627</pre>
4628
4629<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4630<p>The '<tt>zext</tt>' instruction zero extends its operand to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4631 <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4632
4633
4634<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4635<p>The '<tt>zext</tt>' instruction takes a value to cast, which must be of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4636 <a href="#t_integer">integer</a> type, and a type to cast it to, which must
4637 also be of <a href="#t_integer">integer</a> type. The bit size of the
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4638 <tt>value</tt> must be smaller than the bit size of the destination type,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4639 <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4640
4641<h5>Semantics:</h5>
4642<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]4643 bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4644
Reid Spencer07c9c682007-01-12 15:46:11 +0000[diff] [blame]4645<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]4646
4647<h5>Example:</h5>
4648<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4649 %X = zext i32 257 to i64 <i>; yields i64:257</i>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]4650 %Y = zext i1 true to i32 <i>; yields i32:1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4651</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4652
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4653</div>
4654
4655<!-- _______________________________________________________________________ -->
4656<div class="doc_subsubsection">
4657 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
4658</div>
4659<div class="doc_text">
4660
4661<h5>Syntax:</h5>
4662<pre>
4663 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4664</pre>
4665
4666<h5>Overview:</h5>
4667<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
4668
4669<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4670<p>The '<tt>sext</tt>' instruction takes a value to cast, which must be of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4671 <a href="#t_integer">integer</a> type, and a type to cast it to, which must
4672 also be of <a href="#t_integer">integer</a> type. The bit size of the
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4673 <tt>value</tt> must be smaller than the bit size of the destination type,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4674 <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4675
4676<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4677<p>The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
4678 bit (highest order bit) of the <tt>value</tt> until it reaches the bit size
4679 of the type <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4680
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]4681<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]4682
4683<h5>Example:</h5>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4684<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4685 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]4686 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4687</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4688
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4689</div>
4690
4691<!-- _______________________________________________________________________ -->
4692<div class="doc_subsubsection">
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4693 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
4694</div>
4695
4696<div class="doc_text">
4697
4698<h5>Syntax:</h5>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4699<pre>
4700 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4701</pre>
4702
4703<h5>Overview:</h5>
4704<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4705 <tt>ty2</tt>.</p>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4706
4707<h5>Arguments:</h5>
4708<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4709 point</a> value to cast and a <a href="#t_floating">floating point</a> type
4710 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]4711 <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]4712 <i>no-op cast</i>.</p>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4713
4714<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4715<p>The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4716 <a href="#t_floating">floating point</a> type to a smaller
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4717 <a href="#t_floating">floating point</a> type. If the value cannot fit
4718 within the destination type, <tt>ty2</tt>, then the results are
4719 undefined.</p>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4720
4721<h5>Example:</h5>
4722<pre>
4723 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
4724 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
4725</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4726
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4727</div>
4728
4729<!-- _______________________________________________________________________ -->
4730<div class="doc_subsubsection">
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4731 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
4732</div>
4733<div class="doc_text">
4734
4735<h5>Syntax:</h5>
4736<pre>
4737 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4738</pre>
4739
4740<h5>Overview:</h5>
4741<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]4742 floating point value.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4743
4744<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4745<p>The '<tt>fpext</tt>' instruction takes a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4746 <a href="#t_floating">floating point</a> <tt>value</tt> to cast, and
4747 a <a href="#t_floating">floating point</a> type to cast it to. The source
4748 type must be smaller than the destination type.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4749
4750<h5>Semantics:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4751<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4752 <a href="#t_floating">floating point</a> type to a larger
4753 <a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
4754 used to make a <i>no-op cast</i> because it always changes bits. Use
4755 <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]4756
4757<h5>Example:</h5>
4758<pre>
4759 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
4760 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
4761</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4762
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4763</div>
4764
4765<!-- _______________________________________________________________________ -->
4766<div class="doc_subsubsection">
Reid Spencer2eadb532007-01-21 00:29:26 +0000[diff] [blame]4767 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4768</div>
4769<div class="doc_text">
4770
4771<h5>Syntax:</h5>
4772<pre>
Reid Spencer753163d2007-07-31 14:40:14 +0000[diff] [blame]4773 &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]4774</pre>
4775
4776<h5>Overview:</h5>
Reid Spencer753163d2007-07-31 14:40:14 +0000[diff] [blame]4777<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4778 unsigned integer equivalent of type <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4779
4780<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4781<p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
4782 scalar or vector <a href="#t_floating">floating point</a> value, and a type
4783 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4784 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4785 vector integer type with the same number of elements as <tt>ty</tt></p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4786
4787<h5>Semantics:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4788<p>The '<tt>fptoui</tt>' instruction converts its
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4789 <a href="#t_floating">floating point</a> operand into the nearest (rounding
4790 towards zero) unsigned integer value. If the value cannot fit
4791 in <tt>ty2</tt>, the results are undefined.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4792
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4793<h5>Example:</h5>
4794<pre>
Reid Spencer753163d2007-07-31 14:40:14 +0000[diff] [blame]4795 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner5b95a172007-09-22 03:17:52 +0000[diff] [blame]4796 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Gabor Greiff50fd572009-10-28 09:21:30 +0000[diff] [blame]4797 %Z = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4798</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4799
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4800</div>
4801
4802<!-- _______________________________________________________________________ -->
4803<div class="doc_subsubsection">
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4804 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4805</div>
4806<div class="doc_text">
4807
4808<h5>Syntax:</h5>
4809<pre>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4810 &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]4811</pre>
4812
4813<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4814<p>The '<tt>fptosi</tt>' instruction converts
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4815 <a href="#t_floating">floating point</a> <tt>value</tt> to
4816 type <tt>ty2</tt>.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]4817
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]4818<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4819<p>The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
4820 scalar or vector <a href="#t_floating">floating point</a> value, and a type
4821 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4822 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4823 vector integer type with the same number of elements as <tt>ty</tt></p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]4824
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]4825<h5>Semantics:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4826<p>The '<tt>fptosi</tt>' instruction converts its
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4827 <a href="#t_floating">floating point</a> operand into the nearest (rounding
4828 towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
4829 the results are undefined.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]4830
Chris Lattner70de6632001-07-09 00:26:23 +0000[diff] [blame]4831<h5>Example:</h5>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]4832<pre>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]4833 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner5b95a172007-09-22 03:17:52 +0000[diff] [blame]4834 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Gabor Greiff50fd572009-10-28 09:21:30 +0000[diff] [blame]4835 %Z = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4836</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4837
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4838</div>
4839
4840<!-- _______________________________________________________________________ -->
4841<div class="doc_subsubsection">
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4842 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4843</div>
4844<div class="doc_text">
4845
4846<h5>Syntax:</h5>
4847<pre>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4848 &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]4849</pre>
4850
4851<h5>Overview:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4852<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4853 integer and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4854
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4855<h5>Arguments:</h5>
Nate Begemand4d45c22007-11-17 03:58:34 +0000[diff] [blame]4856<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]4857 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
4858 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4859 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4860 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4861
4862<h5>Semantics:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4863<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4864 integer quantity and converts it to the corresponding floating point
4865 value. If the value cannot fit in the floating point value, the results are
4866 undefined.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4867
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4868<h5>Example:</h5>
4869<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4870 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]4871 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4872</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4873
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4874</div>
4875
4876<!-- _______________________________________________________________________ -->
4877<div class="doc_subsubsection">
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4878 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4879</div>
4880<div class="doc_text">
4881
4882<h5>Syntax:</h5>
4883<pre>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4884 &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]4885</pre>
4886
4887<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4888<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed integer
4889 and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4890
4891<h5>Arguments:</h5>
Nate Begemand4d45c22007-11-17 03:58:34 +0000[diff] [blame]4892<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]4893 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
4894 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4895 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4896 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4897
4898<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4899<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed integer
4900 quantity and converts it to the corresponding floating point value. If the
4901 value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4902
4903<h5>Example:</h5>
4904<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4905 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]4906 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4907</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4908
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4909</div>
4910
4911<!-- _______________________________________________________________________ -->
4912<div class="doc_subsubsection">
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]4913 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
4914</div>
4915<div class="doc_text">
4916
4917<h5>Syntax:</h5>
4918<pre>
4919 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4920</pre>
4921
4922<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4923<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
4924 the integer type <tt>ty2</tt>.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]4925
4926<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4927<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
4928 must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
4929 <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]4930
4931<h5>Semantics:</h5>
4932<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4933 <tt>ty2</tt> by interpreting the pointer value as an integer and either
4934 truncating or zero extending that value to the size of the integer type. If
4935 <tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
4936 <tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
4937 are the same size, then nothing is done (<i>no-op cast</i>) other than a type
4938 change.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]4939
4940<h5>Example:</h5>
4941<pre>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]4942 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
4943 %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]4944</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4945
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]4946</div>
4947
4948<!-- _______________________________________________________________________ -->
4949<div class="doc_subsubsection">
4950 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
4951</div>
4952<div class="doc_text">
4953
4954<h5>Syntax:</h5>
4955<pre>
4956 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4957</pre>
4958
4959<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4960<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to a
4961 pointer type, <tt>ty2</tt>.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]4962
4963<h5>Arguments:</h5>
Duncan Sands16f122e2007-03-30 12:22:09 +0000[diff] [blame]4964<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4965 value to cast, and a type to cast it to, which must be a
4966 <a href="#t_pointer">pointer</a> type.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]4967
4968<h5>Semantics:</h5>
4969<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4970 <tt>ty2</tt> by applying either a zero extension or a truncation depending on
4971 the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
4972 size of a pointer then a truncation is done. If <tt>value</tt> is smaller
4973 than the size of a pointer then a zero extension is done. If they are the
4974 same size, nothing is done (<i>no-op cast</i>).</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]4975
4976<h5>Example:</h5>
4977<pre>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]4978 %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]4979 %Y = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
4980 %Z = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]4981</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4982
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]4983</div>
4984
4985<!-- _______________________________________________________________________ -->
4986<div class="doc_subsubsection">
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]4987 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4988</div>
4989<div class="doc_text">
4990
4991<h5>Syntax:</h5>
4992<pre>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]4993 &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]4994</pre>
4995
4996<h5>Overview:</h5>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]4997<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4998 <tt>ty2</tt> without changing any bits.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4999
5000<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5001<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be a
5002 non-aggregate first class value, and a type to cast it to, which must also be
5003 a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes
5004 of <tt>value</tt> and the destination type, <tt>ty2</tt>, must be
5005 identical. If the source type is a pointer, the destination type must also be
5006 a pointer. This instruction supports bitwise conversion of vectors to
5007 integers and to vectors of other types (as long as they have the same
5008 size).</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5009
5010<h5>Semantics:</h5>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]5011<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5012 <tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
5013 this conversion. The conversion is done as if the <tt>value</tt> had been
5014 stored to memory and read back as type <tt>ty2</tt>. Pointer types may only
5015 be converted to other pointer types with this instruction. To convert
5016 pointers to other types, use the <a href="#i_inttoptr">inttoptr</a> or
5017 <a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5018
5019<h5>Example:</h5>
5020<pre>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5021 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5022 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5023 %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]5024</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5025
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5026</div>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5027
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5028<!-- ======================================================================= -->
5029<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5030
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5031<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5032
5033<p>The instructions in this category are the "miscellaneous" instructions, which
5034 defy better classification.</p>
5035
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5036</div>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5037
5038<!-- _______________________________________________________________________ -->
5039<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
5040</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5041
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5042<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5043
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5044<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5045<pre>
5046 &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]5047</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5048
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5049<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5050<p>The '<tt>icmp</tt>' instruction returns a boolean value or a vector of
5051 boolean values based on comparison of its two integer, integer vector, or
5052 pointer operands.</p>
5053
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5054<h5>Arguments:</h5>
5055<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5056 the condition code indicating the kind of comparison to perform. It is not a
5057 value, just a keyword. The possible condition code are:</p>
5058
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5059<ol>
5060 <li><tt>eq</tt>: equal</li>
5061 <li><tt>ne</tt>: not equal </li>
5062 <li><tt>ugt</tt>: unsigned greater than</li>
5063 <li><tt>uge</tt>: unsigned greater or equal</li>
5064 <li><tt>ult</tt>: unsigned less than</li>
5065 <li><tt>ule</tt>: unsigned less or equal</li>
5066 <li><tt>sgt</tt>: signed greater than</li>
5067 <li><tt>sge</tt>: signed greater or equal</li>
5068 <li><tt>slt</tt>: signed less than</li>
5069 <li><tt>sle</tt>: signed less or equal</li>
5070</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5071
Chris Lattnerc0f423a2007-01-15 01:54:13 +0000[diff] [blame]5072<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5073 <a href="#t_pointer">pointer</a> or integer <a href="#t_vector">vector</a>
5074 typed. They must also be identical types.</p>
5075
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5076<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5077<p>The '<tt>icmp</tt>' compares <tt>op1</tt> and <tt>op2</tt> according to the
5078 condition code given as <tt>cond</tt>. The comparison performed always yields
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]5079 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]5080 result, as follows:</p>
5081
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5082<ol>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5083 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5084 <tt>false</tt> otherwise. No sign interpretation is necessary or
5085 performed.</li>
5086
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5087 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5088 <tt>false</tt> otherwise. No sign interpretation is necessary or
5089 performed.</li>
5090
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5091 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5092 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5093
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5094 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5095 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5096 to <tt>op2</tt>.</li>
5097
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5098 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5099 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5100
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5101 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5102 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5103
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5104 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5105 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5106
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5107 <li><tt>sge</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5108 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5109 to <tt>op2</tt>.</li>
5110
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5111 <li><tt>slt</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5112 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5113
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5114 <li><tt>sle</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5115 <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]5116</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5117
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5118<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5119 values are compared as if they were integers.</p>
5120
5121<p>If the operands are integer vectors, then they are compared element by
5122 element. The result is an <tt>i1</tt> vector with the same number of elements
5123 as the values being compared. Otherwise, the result is an <tt>i1</tt>.</p>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5124
5125<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5126<pre>
5127 &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5128 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
5129 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
5130 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
5131 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
5132 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5133</pre>
Dan Gohmana5127ab2009-01-22 01:39:38 +0000[diff] [blame]5134
5135<p>Note that the code generator does not yet support vector types with
5136 the <tt>icmp</tt> instruction.</p>
5137
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5138</div>
5139
5140<!-- _______________________________________________________________________ -->
5141<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
5142</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5143
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5144<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5145
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5146<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5147<pre>
5148 &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]5149</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5150
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5151<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5152<p>The '<tt>fcmp</tt>' instruction returns a boolean value or vector of boolean
5153 values based on comparison of its operands.</p>
5154
5155<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]5156(<a href="#t_integer"><tt>i1</tt></a>).</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5157
5158<p>If the operands are floating point vectors, then the result type is a vector
5159 of boolean with the same number of elements as the operands being
5160 compared.</p>
5161
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5162<h5>Arguments:</h5>
5163<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5164 the condition code indicating the kind of comparison to perform. It is not a
5165 value, just a keyword. The possible condition code are:</p>
5166
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5167<ol>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5168 <li><tt>false</tt>: no comparison, always returns false</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5169 <li><tt>oeq</tt>: ordered and equal</li>
5170 <li><tt>ogt</tt>: ordered and greater than </li>
5171 <li><tt>oge</tt>: ordered and greater than or equal</li>
5172 <li><tt>olt</tt>: ordered and less than </li>
5173 <li><tt>ole</tt>: ordered and less than or equal</li>
5174 <li><tt>one</tt>: ordered and not equal</li>
5175 <li><tt>ord</tt>: ordered (no nans)</li>
5176 <li><tt>ueq</tt>: unordered or equal</li>
5177 <li><tt>ugt</tt>: unordered or greater than </li>
5178 <li><tt>uge</tt>: unordered or greater than or equal</li>
5179 <li><tt>ult</tt>: unordered or less than </li>
5180 <li><tt>ule</tt>: unordered or less than or equal</li>
5181 <li><tt>une</tt>: unordered or not equal</li>
5182 <li><tt>uno</tt>: unordered (either nans)</li>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5183 <li><tt>true</tt>: no comparison, always returns true</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5184</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5185
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5186<p><i>Ordered</i> means that neither operand is a QNAN while
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5187 <i>unordered</i> means that either operand may be a QNAN.</p>
5188
5189<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be either
5190 a <a href="#t_floating">floating point</a> type or
5191 a <a href="#t_vector">vector</a> of floating point type. They must have
5192 identical types.</p>
5193
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5194<h5>Semantics:</h5>
Gabor Greif0f75ad02008-08-07 21:46:00 +0000[diff] [blame]5195<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5196 according to the condition code given as <tt>cond</tt>. If the operands are
5197 vectors, then the vectors are compared element by element. Each comparison
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]5198 performed always yields an <a href="#t_integer">i1</a> result, as
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5199 follows:</p>
5200
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5201<ol>
5202 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5203
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5204 <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]5205 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5206
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5207 <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]5208 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5209
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5210 <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]5211 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5212
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5213 <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]5214 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5215
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5216 <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]5217 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5218
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5219 <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]5220 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5221
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5222 <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]5223
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5224 <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]5225 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5226
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5227 <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]5228 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5229
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5230 <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]5231 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5232
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5233 <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]5234 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5235
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5236 <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]5237 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5238
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5239 <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]5240 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5241
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5242 <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]5243
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5244 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
5245</ol>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5246
5247<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5248<pre>
5249 &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]5250 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
5251 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
5252 &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]5253</pre>
Dan Gohmana5127ab2009-01-22 01:39:38 +0000[diff] [blame]5254
5255<p>Note that the code generator does not yet support vector types with
5256 the <tt>fcmp</tt> instruction.</p>
5257
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5258</div>
5259
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5260<!-- _______________________________________________________________________ -->
Nate Begemand2195702008-05-12 19:01:56 +0000[diff] [blame]5261<div class="doc_subsubsection">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5262 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
5263</div>
5264
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5265<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5266
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5267<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5268<pre>
5269 &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...
5270</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5271
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5272<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5273<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in the
5274 SSA graph representing the function.</p>
5275
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5276<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5277<p>The type of the incoming values is specified with the first type field. After
5278 this, the '<tt>phi</tt>' instruction takes a list of pairs as arguments, with
5279 one pair for each predecessor basic block of the current block. Only values
5280 of <a href="#t_firstclass">first class</a> type may be used as the value
5281 arguments to the PHI node. Only labels may be used as the label
5282 arguments.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5283
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5284<p>There must be no non-phi instructions between the start of a basic block and
5285 the PHI instructions: i.e. PHI instructions must be first in a basic
5286 block.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5287
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5288<p>For the purposes of the SSA form, the use of each incoming value is deemed to
5289 occur on the edge from the corresponding predecessor block to the current
5290 block (but after any definition of an '<tt>invoke</tt>' instruction's return
5291 value on the same edge).</p>
Jay Foad1a4eea52009-06-03 10:20:10 +0000[diff] [blame]5292
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5293<h5>Semantics:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5294<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5295 specified by the pair corresponding to the predecessor basic block that
5296 executed just prior to the current block.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5297
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5298<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5299<pre>
5300Loop: ; Infinite loop that counts from 0 on up...
5301 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
5302 %nextindvar = add i32 %indvar, 1
5303 br label %Loop
5304</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5305
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5306</div>
5307
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5308<!-- _______________________________________________________________________ -->
5309<div class="doc_subsubsection">
5310 <a name="i_select">'<tt>select</tt>' Instruction</a>
5311</div>
5312
5313<div class="doc_text">
5314
5315<h5>Syntax:</h5>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5316<pre>
Dan Gohmanc579d972008-09-09 01:02:47 +0000[diff] [blame]5317 &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>
5318
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]5319 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5320</pre>
5321
5322<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5323<p>The '<tt>select</tt>' instruction is used to choose one value based on a
5324 condition, without branching.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5325
5326
5327<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5328<p>The '<tt>select</tt>' instruction requires an 'i1' value or a vector of 'i1'
5329 values indicating the condition, and two values of the
5330 same <a href="#t_firstclass">first class</a> type. If the val1/val2 are
5331 vectors and the condition is a scalar, then entire vectors are selected, not
5332 individual elements.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5333
5334<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5335<p>If the condition is an i1 and it evaluates to 1, the instruction returns the
5336 first value argument; otherwise, it returns the second value argument.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5337
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5338<p>If the condition is a vector of i1, then the value arguments must be vectors
5339 of the same size, and the selection is done element by element.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5340
5341<h5>Example:</h5>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5342<pre>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]5343 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5344</pre>
Dan Gohmana5127ab2009-01-22 01:39:38 +0000[diff] [blame]5345
5346<p>Note that the code generator does not yet support conditions
5347 with vector type.</p>
5348
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5349</div>
5350
Robert Bocchinof72fdfe2006-01-15 20:48:27 +0000[diff] [blame]5351<!-- _______________________________________________________________________ -->
5352<div class="doc_subsubsection">
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5353 <a name="i_call">'<tt>call</tt>' Instruction</a>
5354</div>
5355
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5356<div class="doc_text">
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5357
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5358<h5>Syntax:</h5>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5359<pre>
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]5360 &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]5361</pre>
5362
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5363<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5364<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5365
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5366<h5>Arguments:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5367<p>This instruction requires several arguments:</p>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5368
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]5369<ol>
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]5370 <li>The optional "tail" marker indicates that the callee function does not
5371 access any allocas or varargs in the caller. Note that calls may be
5372 marked "tail" even if they do not occur before
5373 a <a href="#i_ret"><tt>ret</tt></a> instruction. If the "tail" marker is
5374 present, the function call is eligible for tail call optimization,
5375 but <a href="CodeGenerator.html#tailcallopt">might not in fact be
Evan Cheng59676492010-03-08 21:05:02 +0000[diff] [blame]5376 optimized into a jump</a>. The code generator may optimize calls marked
5377 "tail" with either 1) automatic <a href="CodeGenerator.html#sibcallopt">
5378 sibling call optimization</a> when the caller and callee have
5379 matching signatures, or 2) forced tail call optimization when the
5380 following extra requirements are met:
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]5381 <ul>
5382 <li>Caller and callee both have the calling
5383 convention <tt>fastcc</tt>.</li>
5384 <li>The call is in tail position (ret immediately follows call and ret
5385 uses value of call or is void).</li>
5386 <li>Option <tt>-tailcallopt</tt> is enabled,
Dan Gohman6232f732010-03-02 01:08:11 +0000[diff] [blame]5387 or <code>llvm::GuaranteedTailCallOpt</code> is <code>true</code>.</li>
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]5388 <li><a href="CodeGenerator.html#tailcallopt">Platform specific
5389 constraints are met.</a></li>
5390 </ul>
5391 </li>
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]5392
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5393 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
5394 convention</a> the call should use. If none is specified, the call
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]5395 defaults to using C calling conventions. The calling convention of the
5396 call must match the calling convention of the target function, or else the
5397 behavior is undefined.</li>
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]5398
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5399 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
5400 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
5401 '<tt>inreg</tt>' attributes are valid here.</li>
5402
5403 <li>'<tt>ty</tt>': the type of the call instruction itself which is also the
5404 type of the return value. Functions that return no value are marked
5405 <tt><a href="#t_void">void</a></tt>.</li>
5406
5407 <li>'<tt>fnty</tt>': shall be the signature of the pointer to function value
5408 being invoked. The argument types must match the types implied by this
5409 signature. This type can be omitted if the function is not varargs and if
5410 the function type does not return a pointer to a function.</li>
5411
5412 <li>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
5413 be invoked. In most cases, this is a direct function invocation, but
5414 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
5415 to function value.</li>
5416
5417 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]5418 signature argument types and parameter attributes. All arguments must be
5419 of <a href="#t_firstclass">first class</a> type. If the function
5420 signature indicates the function accepts a variable number of arguments,
5421 the extra arguments can be specified.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5422
5423 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
5424 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
5425 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]5426</ol>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5427
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5428<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5429<p>The '<tt>call</tt>' instruction is used to cause control flow to transfer to
5430 a specified function, with its incoming arguments bound to the specified
5431 values. Upon a '<tt><a href="#i_ret">ret</a></tt>' instruction in the called
5432 function, control flow continues with the instruction after the function
5433 call, and the return value of the function is bound to the result
5434 argument.</p>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5435
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5436<h5>Example:</h5>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5437<pre>
Nick Lewyckya9b13d52007-09-08 13:57:50 +0000[diff] [blame]5438 %retval = call i32 @test(i32 %argc)
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]5439 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) <i>; yields i32</i>
Chris Lattnerfb7c88d2008-03-21 17:24:17 +0000[diff] [blame]5440 %X = tail call i32 @foo() <i>; yields i32</i>
5441 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
5442 call void %foo(i8 97 signext)
Devang Pateld6cff512008-03-10 20:49:15 +0000[diff] [blame]5443
5444 %struct.A = type { i32, i8 }
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]5445 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohmancc3132e2008-10-04 19:00:07 +0000[diff] [blame]5446 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
5447 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattner6cbe8e92008-10-08 06:26:11 +0000[diff] [blame]5448 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijmaneefa7df2008-10-07 10:03:45 +0000[diff] [blame]5449 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5450</pre>
5451
Dale Johannesen68f971b2009-09-24 18:38:21 +0000[diff] [blame]5452<p>llvm treats calls to some functions with names and arguments that match the
Dale Johannesen722212d2009-09-25 17:04:42 +0000[diff] [blame]5453standard C99 library as being the C99 library functions, and may perform
5454optimizations or generate code for them under that assumption. This is
5455something we'd like to change in the future to provide better support for
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]5456freestanding environments and non-C-based languages.</p>
Dale Johannesen68f971b2009-09-24 18:38:21 +0000[diff] [blame]5457
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5458</div>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5459
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5460<!-- _______________________________________________________________________ -->
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5461<div class="doc_subsubsection">
Chris Lattner33337472006-01-13 23:26:01 +0000[diff] [blame]5462 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5463</div>
5464
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5465<div class="doc_text">
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5466
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]5467<h5>Syntax:</h5>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5468<pre>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]5469 &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]5470</pre>
5471
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]5472<h5>Overview:</h5>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]5473<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5474 the "variable argument" area of a function call. It is used to implement the
5475 <tt>va_arg</tt> macro in C.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5476
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]5477<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5478<p>This instruction takes a <tt>va_list*</tt> value and the type of the
5479 argument. It returns a value of the specified argument type and increments
5480 the <tt>va_list</tt> to point to the next argument. The actual type
5481 of <tt>va_list</tt> is target specific.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5482
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]5483<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5484<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified type
5485 from the specified <tt>va_list</tt> and causes the <tt>va_list</tt> to point
5486 to the next argument. For more information, see the variable argument
5487 handling <a href="#int_varargs">Intrinsic Functions</a>.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5488
5489<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]5490 take a variable number of arguments, for example, the <tt>vfprintf</tt>
5491 function.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5492
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5493<p><tt>va_arg</tt> is an LLVM instruction instead of
5494 an <a href="#intrinsics">intrinsic function</a> because it takes a type as an
5495 argument.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5496
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]5497<h5>Example:</h5>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5498<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
5499
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5500<p>Note that the code generator does not yet fully support va_arg on many
5501 targets. Also, it does not currently support va_arg with aggregate types on
5502 any target.</p>
Dan Gohman3065b612009-01-12 23:12:39 +0000[diff] [blame]5503
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5504</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5505
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5506<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]5507<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
5508<!-- *********************************************************************** -->
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5509
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5510<div class="doc_text">
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5511
5512<p>LLVM supports the notion of an "intrinsic function". These functions have
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5513 well known names and semantics and are required to follow certain
5514 restrictions. Overall, these intrinsics represent an extension mechanism for
5515 the LLVM language that does not require changing all of the transformations
5516 in LLVM when adding to the language (or the bitcode reader/writer, the
5517 parser, etc...).</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5518
John Criswell88190562005-05-16 16:17:45 +0000[diff] [blame]5519<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5520 prefix is reserved in LLVM for intrinsic names; thus, function names may not
5521 begin with this prefix. Intrinsic functions must always be external
5522 functions: you cannot define the body of intrinsic functions. Intrinsic
5523 functions may only be used in call or invoke instructions: it is illegal to
5524 take the address of an intrinsic function. Additionally, because intrinsic
5525 functions are part of the LLVM language, it is required if any are added that
5526 they be documented here.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5527
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5528<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents a
5529 family of functions that perform the same operation but on different data
5530 types. Because LLVM can represent over 8 million different integer types,
5531 overloading is used commonly to allow an intrinsic function to operate on any
5532 integer type. One or more of the argument types or the result type can be
5533 overloaded to accept any integer type. Argument types may also be defined as
5534 exactly matching a previous argument's type or the result type. This allows
5535 an intrinsic function which accepts multiple arguments, but needs all of them
5536 to be of the same type, to only be overloaded with respect to a single
5537 argument or the result.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5538
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5539<p>Overloaded intrinsics will have the names of its overloaded argument types
5540 encoded into its function name, each preceded by a period. Only those types
5541 which are overloaded result in a name suffix. Arguments whose type is matched
5542 against another type do not. For example, the <tt>llvm.ctpop</tt> function
5543 can take an integer of any width and returns an integer of exactly the same
5544 integer width. This leads to a family of functions such as
5545 <tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29
5546 %val)</tt>. Only one type, the return type, is overloaded, and only one type
5547 suffix is required. Because the argument's type is matched against the return
5548 type, it does not require its own name suffix.</p>
Reid Spencer4eefaab2007-04-01 08:04:23 +0000[diff] [blame]5549
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5550<p>To learn how to add an intrinsic function, please see the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5551 <a href="ExtendingLLVM.html">Extending LLVM Guide</a>.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5552
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5553</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5554
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5555<!-- ======================================================================= -->
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5556<div class="doc_subsection">
5557 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
5558</div>
5559
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5560<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5561
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5562<p>Variable argument support is defined in LLVM with
5563 the <a href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
5564 intrinsic functions. These functions are related to the similarly named
5565 macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5566
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5567<p>All of these functions operate on arguments that use a target-specific value
5568 type "<tt>va_list</tt>". The LLVM assembly language reference manual does
5569 not define what this type is, so all transformations should be prepared to
5570 handle these functions regardless of the type used.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5571
Chris Lattner30b868d2006-05-15 17:26:46 +0000[diff] [blame]5572<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]5573 instruction and the variable argument handling intrinsic functions are
5574 used.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5575
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]5576<div class="doc_code">
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5577<pre>
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]5578define i32 @test(i32 %X, ...) {
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5579 ; Initialize variable argument processing
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5580 %ap = alloca i8*
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]5581 %ap2 = bitcast i8** %ap to i8*
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]5582 call void @llvm.va_start(i8* %ap2)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5583
5584 ; Read a single integer argument
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5585 %tmp = va_arg i8** %ap, i32
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5586
5587 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5588 %aq = alloca i8*
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]5589 %aq2 = bitcast i8** %aq to i8*
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5590 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]5591 call void @llvm.va_end(i8* %aq2)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5592
5593 ; Stop processing of arguments.
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]5594 call void @llvm.va_end(i8* %ap2)
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5595 ret i32 %tmp
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5596}
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]5597
5598declare void @llvm.va_start(i8*)
5599declare void @llvm.va_copy(i8*, i8*)
5600declare void @llvm.va_end(i8*)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5601</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5602</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5603
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]5604</div>
5605
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5606<!-- _______________________________________________________________________ -->
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5607<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5608 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5609</div>
5610
5611
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5612<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5613
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5614<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5615<pre>
5616 declare void %llvm.va_start(i8* &lt;arglist&gt;)
5617</pre>
5618
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5619<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5620<p>The '<tt>llvm.va_start</tt>' intrinsic initializes <tt>*&lt;arglist&gt;</tt>
5621 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]5622
5623<h5>Arguments:</h5>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]5624<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]5625
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5626<h5>Semantics:</h5>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]5627<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]5628 macro available in C. In a target-dependent way, it initializes
5629 the <tt>va_list</tt> element to which the argument points, so that the next
5630 call to <tt>va_arg</tt> will produce the first variable argument passed to
5631 the function. Unlike the C <tt>va_start</tt> macro, this intrinsic does not
5632 need to know the last argument of the function as the compiler can figure
5633 that out.</p>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]5634
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5635</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5636
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5637<!-- _______________________________________________________________________ -->
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5638<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5639 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5640</div>
5641
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5642<div class="doc_text">
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]5643
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5644<h5>Syntax:</h5>
5645<pre>
5646 declare void @llvm.va_end(i8* &lt;arglist&gt;)
5647</pre>
5648
5649<h5>Overview:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5650<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5651 which has been initialized previously
5652 with <tt><a href="#int_va_start">llvm.va_start</a></tt>
5653 or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]5654
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5655<h5>Arguments:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5656<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]5657
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5658<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5659<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]5660 macro available in C. In a target-dependent way, it destroys
5661 the <tt>va_list</tt> element to which the argument points. Calls
5662 to <a href="#int_va_start"><tt>llvm.va_start</tt></a>
5663 and <a href="#int_va_copy"> <tt>llvm.va_copy</tt></a> must be matched exactly
5664 with calls to <tt>llvm.va_end</tt>.</p>
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]5665
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5666</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5667
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5668<!-- _______________________________________________________________________ -->
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5669<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5670 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5671</div>
5672
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5673<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5674
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5675<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5676<pre>
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]5677 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5678</pre>
5679
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5680<h5>Overview:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5681<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5682 from the source argument list to the destination argument list.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5683
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5684<h5>Arguments:</h5>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]5685<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]5686 The second argument is a pointer to a <tt>va_list</tt> element to copy
5687 from.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5688
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5689<h5>Semantics:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5690<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]5691 macro available in C. In a target-dependent way, it copies the
5692 source <tt>va_list</tt> element into the destination <tt>va_list</tt>
5693 element. This intrinsic is necessary because
5694 the <tt><a href="#int_va_start"> llvm.va_start</a></tt> intrinsic may be
5695 arbitrarily complex and require, for example, memory allocation.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5696
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5697</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5698
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5699<!-- ======================================================================= -->
5700<div class="doc_subsection">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5701 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
5702</div>
5703
5704<div class="doc_text">
5705
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5706<p>LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]5707Collection</a> (GC) requires the implementation and generation of these
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5708intrinsics. These intrinsics allow identification of <a href="#int_gcroot">GC
5709roots on the stack</a>, as well as garbage collector implementations that
5710require <a href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a>
5711barriers. Front-ends for type-safe garbage collected languages should generate
5712these intrinsics to make use of the LLVM garbage collectors. For more details,
5713see <a href="GarbageCollection.html">Accurate Garbage Collection with
5714LLVM</a>.</p>
Christopher Lamb55c6d4f2007-12-17 01:00:21 +0000[diff] [blame]5715
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5716<p>The garbage collection intrinsics only operate on objects in the generic
5717 address space (address space zero).</p>
Christopher Lamb55c6d4f2007-12-17 01:00:21 +0000[diff] [blame]5718
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5719</div>
5720
5721<!-- _______________________________________________________________________ -->
5722<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5723 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5724</div>
5725
5726<div class="doc_text">
5727
5728<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5729<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]5730 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5731</pre>
5732
5733<h5>Overview:</h5>
John Criswelldfe6a862004-12-10 15:51:16 +0000[diff] [blame]5734<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]5735 the code generator, and allows some metadata to be associated with it.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5736
5737<h5>Arguments:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5738<p>The first argument specifies the address of a stack object that contains the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5739 root pointer. The second pointer (which must be either a constant or a
5740 global value address) contains the meta-data to be associated with the
5741 root.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5742
5743<h5>Semantics:</h5>
Chris Lattner851b7712008-04-24 05:59:56 +0000[diff] [blame]5744<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]5745 location. At compile-time, the code generator generates information to allow
5746 the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
5747 intrinsic may only be used in a function which <a href="#gc">specifies a GC
5748 algorithm</a>.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5749
5750</div>
5751
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5752<!-- _______________________________________________________________________ -->
5753<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5754 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5755</div>
5756
5757<div class="doc_text">
5758
5759<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5760<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]5761 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5762</pre>
5763
5764<h5>Overview:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5765<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5766 locations, allowing garbage collector implementations that require read
5767 barriers.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5768
5769<h5>Arguments:</h5>
Chris Lattnerf9228072006-03-14 20:02:51 +0000[diff] [blame]5770<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]5771 allocated from the garbage collector. The first object is a pointer to the
5772 start of the referenced object, if needed by the language runtime (otherwise
5773 null).</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5774
5775<h5>Semantics:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5776<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5777 instruction, but may be replaced with substantially more complex code by the
5778 garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
5779 may only be used in a function which <a href="#gc">specifies a GC
5780 algorithm</a>.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5781
5782</div>
5783
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5784<!-- _______________________________________________________________________ -->
5785<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5786 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5787</div>
5788
5789<div class="doc_text">
5790
5791<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5792<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]5793 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5794</pre>
5795
5796<h5>Overview:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5797<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5798 locations, allowing garbage collector implementations that require write
5799 barriers (such as generational or reference counting collectors).</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5800
5801<h5>Arguments:</h5>
Chris Lattnerf9228072006-03-14 20:02:51 +0000[diff] [blame]5802<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]5803 object to store it to, and the third is the address of the field of Obj to
5804 store to. If the runtime does not require a pointer to the object, Obj may
5805 be null.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5806
5807<h5>Semantics:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5808<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5809 instruction, but may be replaced with substantially more complex code by the
5810 garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
5811 may only be used in a function which <a href="#gc">specifies a GC
5812 algorithm</a>.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5813
5814</div>
5815
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5816<!-- ======================================================================= -->
5817<div class="doc_subsection">
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5818 <a name="int_codegen">Code Generator Intrinsics</a>
5819</div>
5820
5821<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5822
5823<p>These intrinsics are provided by LLVM to expose special features that may
5824 only be implemented with code generator support.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5825
5826</div>
5827
5828<!-- _______________________________________________________________________ -->
5829<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5830 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5831</div>
5832
5833<div class="doc_text">
5834
5835<h5>Syntax:</h5>
5836<pre>
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]5837 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5838</pre>
5839
5840<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5841<p>The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
5842 target-specific value indicating the return address of the current function
5843 or one of its callers.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5844
5845<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5846<p>The argument to this intrinsic indicates which function to return the address
5847 for. Zero indicates the calling function, one indicates its caller, etc.
5848 The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5849
5850<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5851<p>The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer
5852 indicating the return address of the specified call frame, or zero if it
5853 cannot be identified. The value returned by this intrinsic is likely to be
5854 incorrect or 0 for arguments other than zero, so it should only be used for
5855 debugging purposes.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5856
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5857<p>Note that calling this intrinsic does not prevent function inlining or other
5858 aggressive transformations, so the value returned may not be that of the
5859 obvious source-language caller.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5860
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5861</div>
5862
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5863<!-- _______________________________________________________________________ -->
5864<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5865 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5866</div>
5867
5868<div class="doc_text">
5869
5870<h5>Syntax:</h5>
5871<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]5872 declare i8* @llvm.frameaddress(i32 &lt;level&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5873</pre>
5874
5875<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5876<p>The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
5877 target-specific frame pointer value for the specified stack frame.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5878
5879<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5880<p>The argument to this intrinsic indicates which function to return the frame
5881 pointer for. Zero indicates the calling function, one indicates its caller,
5882 etc. The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5883
5884<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5885<p>The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer
5886 indicating the frame address of the specified call frame, or zero if it
5887 cannot be identified. The value returned by this intrinsic is likely to be
5888 incorrect or 0 for arguments other than zero, so it should only be used for
5889 debugging purposes.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5890
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5891<p>Note that calling this intrinsic does not prevent function inlining or other
5892 aggressive transformations, so the value returned may not be that of the
5893 obvious source-language caller.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5894
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5895</div>
5896
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]5897<!-- _______________________________________________________________________ -->
5898<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5899 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]5900</div>
5901
5902<div class="doc_text">
5903
5904<h5>Syntax:</h5>
5905<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]5906 declare i8* @llvm.stacksave()
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]5907</pre>
5908
5909<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5910<p>The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state
5911 of the function stack, for use
5912 with <a href="#int_stackrestore"> <tt>llvm.stackrestore</tt></a>. This is
5913 useful for implementing language features like scoped automatic variable
5914 sized arrays in C99.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]5915
5916<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5917<p>This intrinsic returns a opaque pointer value that can be passed
5918 to <a href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When
5919 an <tt>llvm.stackrestore</tt> intrinsic is executed with a value saved
5920 from <tt>llvm.stacksave</tt>, it effectively restores the state of the stack
5921 to the state it was in when the <tt>llvm.stacksave</tt> intrinsic executed.
5922 In practice, this pops any <a href="#i_alloca">alloca</a> blocks from the
5923 stack that were allocated after the <tt>llvm.stacksave</tt> was executed.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]5924
5925</div>
5926
5927<!-- _______________________________________________________________________ -->
5928<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5929 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]5930</div>
5931
5932<div class="doc_text">
5933
5934<h5>Syntax:</h5>
5935<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]5936 declare void @llvm.stackrestore(i8* %ptr)
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]5937</pre>
5938
5939<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5940<p>The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
5941 the function stack to the state it was in when the
5942 corresponding <a href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic
5943 executed. This is useful for implementing language features like scoped
5944 automatic variable sized arrays in C99.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]5945
5946<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5947<p>See the description
5948 for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]5949
5950</div>
5951
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]5952<!-- _______________________________________________________________________ -->
5953<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5954 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]5955</div>
5956
5957<div class="doc_text">
5958
5959<h5>Syntax:</h5>
5960<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]5961 declare void @llvm.prefetch(i8* &lt;address&gt;, i32 &lt;rw&gt;, i32 &lt;locality&gt;)
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]5962</pre>
5963
5964<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5965<p>The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to
5966 insert a prefetch instruction if supported; otherwise, it is a noop.
5967 Prefetches have no effect on the behavior of the program but can change its
5968 performance characteristics.</p>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]5969
5970<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5971<p><tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the
5972 specifier determining if the fetch should be for a read (0) or write (1),
5973 and <tt>locality</tt> is a temporal locality specifier ranging from (0) - no
5974 locality, to (3) - extremely local keep in cache. The <tt>rw</tt>
5975 and <tt>locality</tt> arguments must be constant integers.</p>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]5976
5977<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5978<p>This intrinsic does not modify the behavior of the program. In particular,
5979 prefetches cannot trap and do not produce a value. On targets that support
5980 this intrinsic, the prefetch can provide hints to the processor cache for
5981 better performance.</p>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]5982
5983</div>
5984
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]5985<!-- _______________________________________________________________________ -->
5986<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5987 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]5988</div>
5989
5990<div class="doc_text">
5991
5992<h5>Syntax:</h5>
5993<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]5994 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]5995</pre>
5996
5997<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5998<p>The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program
5999 Counter (PC) in a region of code to simulators and other tools. The method
6000 is target specific, but it is expected that the marker will use exported
6001 symbols to transmit the PC of the marker. The marker makes no guarantees
6002 that it will remain with any specific instruction after optimizations. It is
6003 possible that the presence of a marker will inhibit optimizations. The
6004 intended use is to be inserted after optimizations to allow correlations of
6005 simulation runs.</p>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6006
6007<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6008<p><tt>id</tt> is a numerical id identifying the marker.</p>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6009
6010<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6011<p>This intrinsic does not modify the behavior of the program. Backends that do
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]6012 not support this intrinsic may ignore it.</p>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6013
6014</div>
6015
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6016<!-- _______________________________________________________________________ -->
6017<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6018 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6019</div>
6020
6021<div class="doc_text">
6022
6023<h5>Syntax:</h5>
6024<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]6025 declare i64 @llvm.readcyclecounter()
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6026</pre>
6027
6028<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6029<p>The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
6030 counter register (or similar low latency, high accuracy clocks) on those
6031 targets that support it. On X86, it should map to RDTSC. On Alpha, it
6032 should map to RPCC. As the backing counters overflow quickly (on the order
6033 of 9 seconds on alpha), this should only be used for small timings.</p>
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6034
6035<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6036<p>When directly supported, reading the cycle counter should not modify any
6037 memory. Implementations are allowed to either return a application specific
6038 value or a system wide value. On backends without support, this is lowered
6039 to a constant 0.</p>
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6040
6041</div>
6042
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6043<!-- ======================================================================= -->
6044<div class="doc_subsection">
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6045 <a name="int_libc">Standard C Library Intrinsics</a>
6046</div>
6047
6048<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6049
6050<p>LLVM provides intrinsics for a few important standard C library functions.
6051 These intrinsics allow source-language front-ends to pass information about
6052 the alignment of the pointer arguments to the code generator, providing
6053 opportunity for more efficient code generation.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6054
6055</div>
6056
6057<!-- _______________________________________________________________________ -->
6058<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6059 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6060</div>
6061
6062<div class="doc_text">
6063
6064<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6065<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]6066 integer bit width and for different address spaces. Not all targets support
6067 all bit widths however.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6068
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6069<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6070 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]6071 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6072 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]6073 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6074</pre>
6075
6076<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6077<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6078 source location to the destination location.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6079
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6080<p>Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6081 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6082 and the pointers can be in specified address spaces.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6083
6084<h5>Arguments:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6085
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6086<p>The first argument is a pointer to the destination, the second is a pointer
6087 to the source. The third argument is an integer argument specifying the
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6088 number of bytes to copy, the fourth argument is the alignment of the
6089 source and destination locations, and the fifth is a boolean indicating a
6090 volatile access.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6091
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]6092<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]6093 then the caller guarantees that both the source and destination pointers are
6094 aligned to that boundary.</p>
Chris Lattner4c67c482004-02-12 21:18:15 +0000[diff] [blame]6095
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]6096<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6097 <tt>llvm.memcpy</tt> call is a <a href="#volatile">volatile operation</a>.
6098 The detailed access behavior is not very cleanly specified and it is unwise
6099 to depend on it.</p>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6100
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6101<h5>Semantics:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6102
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6103<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6104 source location to the destination location, which are not allowed to
6105 overlap. It copies "len" bytes of memory over. If the argument is known to
6106 be aligned to some boundary, this can be specified as the fourth argument,
6107 otherwise it should be set to 0 or 1.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6108
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6109</div>
6110
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6111<!-- _______________________________________________________________________ -->
6112<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6113 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6114</div>
6115
6116<div class="doc_text">
6117
6118<h5>Syntax:</h5>
Chris Lattnerdd708342008-11-21 16:42:48 +0000[diff] [blame]6119<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]6120 width and for different address space. Not all targets support all bit
6121 widths however.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6122
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6123<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6124 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]6125 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6126 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]6127 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6128</pre>
6129
6130<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6131<p>The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the
6132 source location to the destination location. It is similar to the
6133 '<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to
6134 overlap.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6135
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6136<p>Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6137 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6138 and the pointers can be in specified address spaces.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6139
6140<h5>Arguments:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6141
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6142<p>The first argument is a pointer to the destination, the second is a pointer
6143 to the source. The third argument is an integer argument specifying the
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6144 number of bytes to copy, the fourth argument is the alignment of the
6145 source and destination locations, and the fifth is a boolean indicating a
6146 volatile access.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6147
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]6148<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]6149 then the caller guarantees that the source and destination pointers are
6150 aligned to that boundary.</p>
Chris Lattner4c67c482004-02-12 21:18:15 +0000[diff] [blame]6151
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]6152<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6153 <tt>llvm.memmove</tt> call is a <a href="#volatile">volatile operation</a>.
6154 The detailed access behavior is not very cleanly specified and it is unwise
6155 to depend on it.</p>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6156
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6157<h5>Semantics:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6158
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6159<p>The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the
6160 source location to the destination location, which may overlap. It copies
6161 "len" bytes of memory over. If the argument is known to be aligned to some
6162 boundary, this can be specified as the fourth argument, otherwise it should
6163 be set to 0 or 1.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6164
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6165</div>
6166
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6167<!-- _______________________________________________________________________ -->
6168<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6169 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6170</div>
6171
6172<div class="doc_text">
6173
6174<h5>Syntax:</h5>
Chris Lattnerdd708342008-11-21 16:42:48 +0000[diff] [blame]6175<p>This is an overloaded intrinsic. You can use llvm.memset on any integer bit
Mon P Wang508127b2010-04-07 06:35:53 +0000[diff] [blame]6176 width and for different address spaces. Not all targets support all bit
6177 widths however.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6178
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6179<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6180 declare void @llvm.memset.p0i8.i32(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattner685db9d2010-04-08 00:54:34 +0000[diff] [blame]6181 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6182 declare void @llvm.memset.p0i8.i64(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattner685db9d2010-04-08 00:54:34 +0000[diff] [blame]6183 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6184</pre>
6185
6186<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6187<p>The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a
6188 particular byte value.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6189
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6190<p>Note that, unlike the standard libc function, the <tt>llvm.memset</tt>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6191 intrinsic does not return a value, takes extra alignment/volatile arguments,
6192 and the destination can be in an arbitrary address space.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6193
6194<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6195<p>The first argument is a pointer to the destination to fill, the second is the
6196 byte value to fill it with, the third argument is an integer argument
6197 specifying the number of bytes to fill, and the fourth argument is the known
6198 alignment of destination location.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6199
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]6200<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]6201 then the caller guarantees that the destination pointer is aligned to that
6202 boundary.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6203
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]6204<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6205 <tt>llvm.memset</tt> call is a <a href="#volatile">volatile operation</a>.
6206 The detailed access behavior is not very cleanly specified and it is unwise
6207 to depend on it.</p>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6208
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6209<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6210<p>The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting
6211 at the destination location. If the argument is known to be aligned to some
6212 boundary, this can be specified as the fourth argument, otherwise it should
6213 be set to 0 or 1.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6214
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6215</div>
6216
Chris Lattner3b4f4372004-06-11 02:28:03 +0000[diff] [blame]6217<!-- _______________________________________________________________________ -->
6218<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6219 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6220</div>
6221
6222<div class="doc_text">
6223
6224<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6225<p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
6226 floating point or vector of floating point type. Not all targets support all
6227 types however.</p>
6228
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6229<pre>
Dale Johannesendd89d272007-10-02 17:47:38 +0000[diff] [blame]6230 declare float @llvm.sqrt.f32(float %Val)
6231 declare double @llvm.sqrt.f64(double %Val)
6232 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
6233 declare fp128 @llvm.sqrt.f128(fp128 %Val)
6234 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6235</pre>
6236
6237<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6238<p>The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
6239 returning the same value as the libm '<tt>sqrt</tt>' functions would.
6240 Unlike <tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined
6241 behavior for negative numbers other than -0.0 (which allows for better
6242 optimization, because there is no need to worry about errno being
6243 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]6244
6245<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6246<p>The argument and return value are floating point numbers of the same
6247 type.</p>
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6248
6249<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6250<p>This function returns the sqrt of the specified operand if it is a
6251 nonnegative floating point number.</p>
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6252
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6253</div>
6254
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6255<!-- _______________________________________________________________________ -->
6256<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6257 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6258</div>
6259
6260<div class="doc_text">
6261
6262<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6263<p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
6264 floating point or vector of floating point type. Not all targets support all
6265 types however.</p>
6266
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6267<pre>
Dale Johannesendd89d272007-10-02 17:47:38 +0000[diff] [blame]6268 declare float @llvm.powi.f32(float %Val, i32 %power)
6269 declare double @llvm.powi.f64(double %Val, i32 %power)
6270 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
6271 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
6272 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6273</pre>
6274
6275<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6276<p>The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
6277 specified (positive or negative) power. The order of evaluation of
6278 multiplications is not defined. When a vector of floating point type is
6279 used, the second argument remains a scalar integer value.</p>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6280
6281<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6282<p>The second argument is an integer power, and the first is a value to raise to
6283 that power.</p>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6284
6285<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6286<p>This function returns the first value raised to the second power with an
6287 unspecified sequence of rounding operations.</p>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6288
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6289</div>
6290
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6291<!-- _______________________________________________________________________ -->
6292<div class="doc_subsubsection">
6293 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
6294</div>
6295
6296<div class="doc_text">
6297
6298<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6299<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
6300 floating point or vector of floating point type. Not all targets support all
6301 types however.</p>
6302
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6303<pre>
6304 declare float @llvm.sin.f32(float %Val)
6305 declare double @llvm.sin.f64(double %Val)
6306 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
6307 declare fp128 @llvm.sin.f128(fp128 %Val)
6308 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
6309</pre>
6310
6311<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6312<p>The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6313
6314<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6315<p>The argument and return value are floating point numbers of the same
6316 type.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6317
6318<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6319<p>This function returns the sine of the specified operand, returning the same
6320 values as the libm <tt>sin</tt> functions would, and handles error conditions
6321 in the same way.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6322
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6323</div>
6324
6325<!-- _______________________________________________________________________ -->
6326<div class="doc_subsubsection">
6327 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
6328</div>
6329
6330<div class="doc_text">
6331
6332<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6333<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
6334 floating point or vector of floating point type. Not all targets support all
6335 types however.</p>
6336
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6337<pre>
6338 declare float @llvm.cos.f32(float %Val)
6339 declare double @llvm.cos.f64(double %Val)
6340 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
6341 declare fp128 @llvm.cos.f128(fp128 %Val)
6342 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
6343</pre>
6344
6345<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6346<p>The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6347
6348<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6349<p>The argument and return value are floating point numbers of the same
6350 type.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6351
6352<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6353<p>This function returns the cosine of the specified operand, returning the same
6354 values as the libm <tt>cos</tt> functions would, and handles error conditions
6355 in the same way.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6356
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6357</div>
6358
6359<!-- _______________________________________________________________________ -->
6360<div class="doc_subsubsection">
6361 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
6362</div>
6363
6364<div class="doc_text">
6365
6366<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6367<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
6368 floating point or vector of floating point type. Not all targets support all
6369 types however.</p>
6370
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6371<pre>
6372 declare float @llvm.pow.f32(float %Val, float %Power)
6373 declare double @llvm.pow.f64(double %Val, double %Power)
6374 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
6375 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
6376 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
6377</pre>
6378
6379<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6380<p>The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
6381 specified (positive or negative) power.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6382
6383<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6384<p>The second argument is a floating point power, and the first is a value to
6385 raise to that power.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6386
6387<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6388<p>This function returns the first value raised to the second power, returning
6389 the same values as the libm <tt>pow</tt> functions would, and handles error
6390 conditions in the same way.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6391
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6392</div>
6393
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6394<!-- ======================================================================= -->
6395<div class="doc_subsection">
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6396 <a name="int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6397</div>
6398
6399<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6400
6401<p>LLVM provides intrinsics for a few important bit manipulation operations.
6402 These allow efficient code generation for some algorithms.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6403
6404</div>
6405
6406<!-- _______________________________________________________________________ -->
6407<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6408 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6409</div>
6410
6411<div class="doc_text">
6412
6413<h5>Syntax:</h5>
Reid Spencer4eefaab2007-04-01 08:04:23 +0000[diff] [blame]6414<p>This is an overloaded intrinsic function. You can use bswap on any integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6415 type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
6416
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6417<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6418 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
6419 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
6420 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6421</pre>
6422
6423<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6424<p>The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
6425 values with an even number of bytes (positive multiple of 16 bits). These
6426 are useful for performing operations on data that is not in the target's
6427 native byte order.</p>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6428
6429<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6430<p>The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
6431 and low byte of the input i16 swapped. Similarly,
6432 the <tt>llvm.bswap.i32</tt> intrinsic returns an i32 value that has the four
6433 bytes of the input i32 swapped, so that if the input bytes are numbered 0, 1,
6434 2, 3 then the returned i32 will have its bytes in 3, 2, 1, 0 order.
6435 The <tt>llvm.bswap.i48</tt>, <tt>llvm.bswap.i64</tt> and other intrinsics
6436 extend this concept to additional even-byte lengths (6 bytes, 8 bytes and
6437 more, respectively).</p>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6438
6439</div>
6440
6441<!-- _______________________________________________________________________ -->
6442<div class="doc_subsubsection">
Reid Spencerb4f9a6f2006-01-16 21:12:35 +0000[diff] [blame]6443 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6444</div>
6445
6446<div class="doc_text">
6447
6448<h5>Syntax:</h5>
Reid Spencer4eefaab2007-04-01 08:04:23 +0000[diff] [blame]6449<p>This is an overloaded intrinsic. You can use llvm.ctpop on any integer bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6450 width. Not all targets support all bit widths however.</p>
6451
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6452<pre>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6453 declare i8 @llvm.ctpop.i8(i8 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6454 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]6455 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6456 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
6457 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6458</pre>
6459
6460<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6461<p>The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set
6462 in a value.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6463
6464<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6465<p>The only argument is the value to be counted. The argument may be of any
6466 integer type. The return type must match the argument type.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6467
6468<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6469<p>The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6470
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6471</div>
6472
6473<!-- _______________________________________________________________________ -->
6474<div class="doc_subsubsection">
Chris Lattnerb748c672006-01-16 22:34:14 +0000[diff] [blame]6475 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6476</div>
6477
6478<div class="doc_text">
6479
6480<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6481<p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
6482 integer bit width. Not all targets support all bit widths however.</p>
6483
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6484<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6485 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
6486 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]6487 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6488 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
6489 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6490</pre>
6491
6492<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6493<p>The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
6494 leading zeros in a variable.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6495
6496<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6497<p>The only argument is the value to be counted. The argument may be of any
6498 integer type. The return type must match the argument type.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6499
6500<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6501<p>The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant)
6502 zeros in a variable. If the src == 0 then the result is the size in bits of
6503 the type of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6504
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6505</div>
Chris Lattner3b4f4372004-06-11 02:28:03 +0000[diff] [blame]6506
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6507<!-- _______________________________________________________________________ -->
6508<div class="doc_subsubsection">
Chris Lattnerb748c672006-01-16 22:34:14 +0000[diff] [blame]6509 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6510</div>
6511
6512<div class="doc_text">
6513
6514<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6515<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
6516 integer bit width. Not all targets support all bit widths however.</p>
6517
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6518<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6519 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
6520 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]6521 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6522 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
6523 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6524</pre>
6525
6526<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6527<p>The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
6528 trailing zeros.</p>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6529
6530<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6531<p>The only argument is the value to be counted. The argument may be of any
6532 integer type. The return type must match the argument type.</p>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6533
6534<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6535<p>The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant)
6536 zeros in a variable. If the src == 0 then the result is the size in bits of
6537 the type of src. For example, <tt>llvm.cttz(2) = 1</tt>.</p>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6538
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6539</div>
6540
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]6541<!-- ======================================================================= -->
6542<div class="doc_subsection">
6543 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
6544</div>
6545
6546<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6547
6548<p>LLVM provides intrinsics for some arithmetic with overflow operations.</p>
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]6549
6550</div>
6551
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6552<!-- _______________________________________________________________________ -->
6553<div class="doc_subsubsection">
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]6554 <a name="int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics</a>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6555</div>
6556
6557<div class="doc_text">
6558
6559<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6560<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]6561 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6562
6563<pre>
6564 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
6565 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6566 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
6567</pre>
6568
6569<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6570<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6571 a signed addition of the two arguments, and indicate whether an overflow
6572 occurred during the signed summation.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6573
6574<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6575<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]6576 be of integer types of any bit width, but they must have the same bit
6577 width. The second element of the result structure must be of
6578 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6579 undergo signed addition.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6580
6581<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6582<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6583 a signed addition of the two variables. They return a structure &mdash; the
6584 first element of which is the signed summation, and the second element of
6585 which is a bit specifying if the signed summation resulted in an
6586 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6587
6588<h5>Examples:</h5>
6589<pre>
6590 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6591 %sum = extractvalue {i32, i1} %res, 0
6592 %obit = extractvalue {i32, i1} %res, 1
6593 br i1 %obit, label %overflow, label %normal
6594</pre>
6595
6596</div>
6597
6598<!-- _______________________________________________________________________ -->
6599<div class="doc_subsubsection">
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]6600 <a name="int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics</a>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6601</div>
6602
6603<div class="doc_text">
6604
6605<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6606<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]6607 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6608
6609<pre>
6610 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
6611 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6612 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
6613</pre>
6614
6615<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6616<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6617 an unsigned addition of the two arguments, and indicate whether a carry
6618 occurred during the unsigned summation.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6619
6620<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6621<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]6622 be of integer types of any bit width, but they must have the same bit
6623 width. The second element of the result structure must be of
6624 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6625 undergo unsigned addition.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6626
6627<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6628<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6629 an unsigned addition of the two arguments. They return a structure &mdash;
6630 the first element of which is the sum, and the second element of which is a
6631 bit specifying if the unsigned summation resulted in a carry.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6632
6633<h5>Examples:</h5>
6634<pre>
6635 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6636 %sum = extractvalue {i32, i1} %res, 0
6637 %obit = extractvalue {i32, i1} %res, 1
6638 br i1 %obit, label %carry, label %normal
6639</pre>
6640
6641</div>
6642
6643<!-- _______________________________________________________________________ -->
6644<div class="doc_subsubsection">
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]6645 <a name="int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics</a>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6646</div>
6647
6648<div class="doc_text">
6649
6650<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6651<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]6652 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6653
6654<pre>
6655 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
6656 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6657 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
6658</pre>
6659
6660<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6661<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6662 a signed subtraction of the two arguments, and indicate whether an overflow
6663 occurred during the signed subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6664
6665<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6666<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]6667 be of integer types of any bit width, but they must have the same bit
6668 width. The second element of the result structure must be of
6669 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6670 undergo signed subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6671
6672<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6673<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6674 a signed subtraction of the two arguments. They return a structure &mdash;
6675 the first element of which is the subtraction, and the second element of
6676 which is a bit specifying if the signed subtraction resulted in an
6677 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6678
6679<h5>Examples:</h5>
6680<pre>
6681 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6682 %sum = extractvalue {i32, i1} %res, 0
6683 %obit = extractvalue {i32, i1} %res, 1
6684 br i1 %obit, label %overflow, label %normal
6685</pre>
6686
6687</div>
6688
6689<!-- _______________________________________________________________________ -->
6690<div class="doc_subsubsection">
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]6691 <a name="int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt>' Intrinsics</a>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6692</div>
6693
6694<div class="doc_text">
6695
6696<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6697<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]6698 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6699
6700<pre>
6701 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
6702 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6703 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
6704</pre>
6705
6706<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6707<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6708 an unsigned subtraction of the two arguments, and indicate whether an
6709 overflow occurred during the unsigned subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6710
6711<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6712<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]6713 be of integer types of any bit width, but they must have the same bit
6714 width. The second element of the result structure must be of
6715 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6716 undergo unsigned subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6717
6718<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6719<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6720 an unsigned subtraction of the two arguments. They return a structure &mdash;
6721 the first element of which is the subtraction, and the second element of
6722 which is a bit specifying if the unsigned subtraction resulted in an
6723 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6724
6725<h5>Examples:</h5>
6726<pre>
6727 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6728 %sum = extractvalue {i32, i1} %res, 0
6729 %obit = extractvalue {i32, i1} %res, 1
6730 br i1 %obit, label %overflow, label %normal
6731</pre>
6732
6733</div>
6734
6735<!-- _______________________________________________________________________ -->
6736<div class="doc_subsubsection">
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]6737 <a name="int_smul_overflow">'<tt>llvm.smul.with.overflow.*</tt>' Intrinsics</a>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6738</div>
6739
6740<div class="doc_text">
6741
6742<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6743<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]6744 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6745
6746<pre>
6747 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
6748 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6749 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
6750</pre>
6751
6752<h5>Overview:</h5>
6753
6754<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6755 a signed multiplication of the two arguments, and indicate whether an
6756 overflow occurred during the signed multiplication.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6757
6758<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6759<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]6760 be of integer types of any bit width, but they must have the same bit
6761 width. The second element of the result structure must be of
6762 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6763 undergo signed multiplication.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6764
6765<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6766<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6767 a signed multiplication of the two arguments. They return a structure &mdash;
6768 the first element of which is the multiplication, and the second element of
6769 which is a bit specifying if the signed multiplication resulted in an
6770 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6771
6772<h5>Examples:</h5>
6773<pre>
6774 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6775 %sum = extractvalue {i32, i1} %res, 0
6776 %obit = extractvalue {i32, i1} %res, 1
6777 br i1 %obit, label %overflow, label %normal
6778</pre>
6779
Reid Spencer5bf54c82007-04-11 23:23:49 +0000[diff] [blame]6780</div>
6781
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6782<!-- _______________________________________________________________________ -->
6783<div class="doc_subsubsection">
6784 <a name="int_umul_overflow">'<tt>llvm.umul.with.overflow.*</tt>' Intrinsics</a>
6785</div>
6786
6787<div class="doc_text">
6788
6789<h5>Syntax:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6790<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]6791 on any integer bit width.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6792
6793<pre>
6794 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
6795 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6796 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
6797</pre>
6798
6799<h5>Overview:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6800<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6801 a unsigned multiplication of the two arguments, and indicate whether an
6802 overflow occurred during the unsigned multiplication.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6803
6804<h5>Arguments:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6805<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]6806 be of integer types of any bit width, but they must have the same bit
6807 width. The second element of the result structure must be of
6808 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6809 undergo unsigned multiplication.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6810
6811<h5>Semantics:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6812<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6813 an unsigned multiplication of the two arguments. They return a structure
6814 &mdash; the first element of which is the multiplication, and the second
6815 element of which is a bit specifying if the unsigned multiplication resulted
6816 in an overflow.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6817
6818<h5>Examples:</h5>
6819<pre>
6820 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6821 %sum = extractvalue {i32, i1} %res, 0
6822 %obit = extractvalue {i32, i1} %res, 1
6823 br i1 %obit, label %overflow, label %normal
6824</pre>
6825
6826</div>
6827
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6828<!-- ======================================================================= -->
6829<div class="doc_subsection">
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]6830 <a name="int_fp16">Half Precision Floating Point Intrinsics</a>
6831</div>
6832
6833<div class="doc_text">
6834
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]6835<p>Half precision floating point is a storage-only format. This means that it is
6836 a dense encoding (in memory) but does not support computation in the
6837 format.</p>
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]6838
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]6839<p>This means that code must first load the half-precision floating point
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]6840 value as an i16, then convert it to float with <a
6841 href="#int_convert_from_fp16"><tt>llvm.convert.from.fp16</tt></a>.
6842 Computation can then be performed on the float value (including extending to
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]6843 double etc). To store the value back to memory, it is first converted to
6844 float if needed, then converted to i16 with
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]6845 <a href="#int_convert_to_fp16"><tt>llvm.convert.to.fp16</tt></a>, then
6846 storing as an i16 value.</p>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]6847</div>
6848
6849<!-- _______________________________________________________________________ -->
6850<div class="doc_subsubsection">
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]6851 <a name="int_convert_to_fp16">'<tt>llvm.convert.to.fp16</tt>' Intrinsic</a>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]6852</div>
6853
6854<div class="doc_text">
6855
6856<h5>Syntax:</h5>
6857<pre>
6858 declare i16 @llvm.convert.to.fp16(f32 %a)
6859</pre>
6860
6861<h5>Overview:</h5>
6862<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
6863 a conversion from single precision floating point format to half precision
6864 floating point format.</p>
6865
6866<h5>Arguments:</h5>
6867<p>The intrinsic function contains single argument - the value to be
6868 converted.</p>
6869
6870<h5>Semantics:</h5>
6871<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
6872 a conversion from single precision floating point format to half precision
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]6873 floating point format. The return value is an <tt>i16</tt> which
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]6874 contains the converted number.</p>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]6875
6876<h5>Examples:</h5>
6877<pre>
6878 %res = call i16 @llvm.convert.to.fp16(f32 %a)
6879 store i16 %res, i16* @x, align 2
6880</pre>
6881
6882</div>
6883
6884<!-- _______________________________________________________________________ -->
6885<div class="doc_subsubsection">
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]6886 <a name="int_convert_from_fp16">'<tt>llvm.convert.from.fp16</tt>' Intrinsic</a>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]6887</div>
6888
6889<div class="doc_text">
6890
6891<h5>Syntax:</h5>
6892<pre>
6893 declare f32 @llvm.convert.from.fp16(i16 %a)
6894</pre>
6895
6896<h5>Overview:</h5>
6897<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs
6898 a conversion from half precision floating point format to single precision
6899 floating point format.</p>
6900
6901<h5>Arguments:</h5>
6902<p>The intrinsic function contains single argument - the value to be
6903 converted.</p>
6904
6905<h5>Semantics:</h5>
6906<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs a
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]6907 conversion from half single precision floating point format to single
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]6908 precision floating point format. The input half-float value is represented by
6909 an <tt>i16</tt> value.</p>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]6910
6911<h5>Examples:</h5>
6912<pre>
6913 %a = load i16* @x, align 2
6914 %res = call f32 @llvm.convert.from.fp16(i16 %a)
6915</pre>
6916
6917</div>
6918
6919<!-- ======================================================================= -->
6920<div class="doc_subsection">
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6921 <a name="int_debugger">Debugger Intrinsics</a>
6922</div>
6923
6924<div class="doc_text">
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6925
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6926<p>The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt>
6927 prefix), are described in
6928 the <a href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source
6929 Level Debugging</a> document.</p>
6930
6931</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6932
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]6933<!-- ======================================================================= -->
6934<div class="doc_subsection">
6935 <a name="int_eh">Exception Handling Intrinsics</a>
6936</div>
6937
6938<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6939
6940<p>The LLVM exception handling intrinsics (which all start with
6941 <tt>llvm.eh.</tt> prefix), are described in
6942 the <a href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
6943 Handling</a> document.</p>
6944
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]6945</div>
6946
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]6947<!-- ======================================================================= -->
6948<div class="doc_subsection">
Duncan Sands86e01192007-09-11 14:10:23 +0000[diff] [blame]6949 <a name="int_trampoline">Trampoline Intrinsic</a>
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]6950</div>
6951
6952<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6953
6954<p>This intrinsic makes it possible to excise one parameter, marked with
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame^]6955 the <a href="#nest"><tt>nest</tt></a> attribute, from a function.
6956 The result is a callable
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6957 function pointer lacking the nest parameter - the caller does not need to
6958 provide a value for it. Instead, the value to use is stored in advance in a
6959 "trampoline", a block of memory usually allocated on the stack, which also
6960 contains code to splice the nest value into the argument list. This is used
6961 to implement the GCC nested function address extension.</p>
6962
6963<p>For example, if the function is
6964 <tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
6965 pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as
6966 follows:</p>
6967
6968<div class="doc_code">
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]6969<pre>
Duncan Sands86e01192007-09-11 14:10:23 +0000[diff] [blame]6970 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
6971 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]6972 %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]6973 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]6974</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6975</div>
6976
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]6977<p>The call <tt>%val = call i32 %fp(i32 %x, i32 %y)</tt> is then equivalent
6978 to <tt>%val = call i32 %f(i8* %nval, i32 %x, i32 %y)</tt>.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6979
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]6980</div>
6981
6982<!-- _______________________________________________________________________ -->
6983<div class="doc_subsubsection">
6984 <a name="int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a>
6985</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6986
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]6987<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6988
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]6989<h5>Syntax:</h5>
6990<pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6991 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]6992</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6993
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]6994<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6995<p>This fills the memory pointed to by <tt>tramp</tt> with code and returns a
6996 function pointer suitable for executing it.</p>
6997
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]6998<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6999<p>The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
7000 pointers. The <tt>tramp</tt> argument must point to a sufficiently large and
7001 sufficiently aligned block of memory; this memory is written to by the
7002 intrinsic. Note that the size and the alignment are target-specific - LLVM
7003 currently provides no portable way of determining them, so a front-end that
7004 generates this intrinsic needs to have some target-specific knowledge.
7005 The <tt>func</tt> argument must hold a function bitcast to
7006 an <tt>i8*</tt>.</p>
7007
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7008<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7009<p>The block of memory pointed to by <tt>tramp</tt> is filled with target
7010 dependent code, turning it into a function. A pointer to this function is
7011 returned, but needs to be bitcast to an <a href="#int_trampoline">appropriate
7012 function pointer type</a> before being called. The new function's signature
7013 is the same as that of <tt>func</tt> with any arguments marked with
7014 the <tt>nest</tt> attribute removed. At most one such <tt>nest</tt> argument
7015 is allowed, and it must be of pointer type. Calling the new function is
7016 equivalent to calling <tt>func</tt> with the same argument list, but
7017 with <tt>nval</tt> used for the missing <tt>nest</tt> argument. If, after
7018 calling <tt>llvm.init.trampoline</tt>, the memory pointed to
7019 by <tt>tramp</tt> is modified, then the effect of any later call to the
7020 returned function pointer is undefined.</p>
7021
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7022</div>
7023
7024<!-- ======================================================================= -->
7025<div class="doc_subsection">
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7026 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
7027</div>
7028
7029<div class="doc_text">
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7030
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7031<p>These intrinsic functions expand the "universal IR" of LLVM to represent
7032 hardware constructs for atomic operations and memory synchronization. This
7033 provides an interface to the hardware, not an interface to the programmer. It
7034 is aimed at a low enough level to allow any programming models or APIs
7035 (Application Programming Interfaces) which need atomic behaviors to map
7036 cleanly onto it. It is also modeled primarily on hardware behavior. Just as
7037 hardware provides a "universal IR" for source languages, it also provides a
7038 starting point for developing a "universal" atomic operation and
7039 synchronization IR.</p>
7040
7041<p>These do <em>not</em> form an API such as high-level threading libraries,
7042 software transaction memory systems, atomic primitives, and intrinsic
7043 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
7044 application libraries. The hardware interface provided by LLVM should allow
7045 a clean implementation of all of these APIs and parallel programming models.
7046 No one model or paradigm should be selected above others unless the hardware
7047 itself ubiquitously does so.</p>
7048
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7049</div>
7050
7051<!-- _______________________________________________________________________ -->
7052<div class="doc_subsubsection">
7053 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
7054</div>
7055<div class="doc_text">
7056<h5>Syntax:</h5>
7057<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7058 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]7059</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7060
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7061<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7062<p>The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
7063 specific pairs of memory access types.</p>
7064
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7065<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7066<p>The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
7067 The first four arguments enables a specific barrier as listed below. The
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]7068 fifth argument specifies that the barrier applies to io or device or uncached
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7069 memory.</p>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7070
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7071<ul>
7072 <li><tt>ll</tt>: load-load barrier</li>
7073 <li><tt>ls</tt>: load-store barrier</li>
7074 <li><tt>sl</tt>: store-load barrier</li>
7075 <li><tt>ss</tt>: store-store barrier</li>
7076 <li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
7077</ul>
7078
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7079<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7080<p>This intrinsic causes the system to enforce some ordering constraints upon
7081 the loads and stores of the program. This barrier does not
7082 indicate <em>when</em> any events will occur, it only enforces
7083 an <em>order</em> in which they occur. For any of the specified pairs of load
7084 and store operations (f.ex. load-load, or store-load), all of the first
7085 operations preceding the barrier will complete before any of the second
7086 operations succeeding the barrier begin. Specifically the semantics for each
7087 pairing is as follows:</p>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7088
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7089<ul>
7090 <li><tt>ll</tt>: All loads before the barrier must complete before any load
7091 after the barrier begins.</li>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7092 <li><tt>ls</tt>: All loads before the barrier must complete before any
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7093 store after the barrier begins.</li>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7094 <li><tt>ss</tt>: All stores before the barrier must complete before any
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7095 store after the barrier begins.</li>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7096 <li><tt>sl</tt>: All stores before the barrier must complete before any
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7097 load after the barrier begins.</li>
7098</ul>
7099
7100<p>These semantics are applied with a logical "and" behavior when more than one
7101 is enabled in a single memory barrier intrinsic.</p>
7102
7103<p>Backends may implement stronger barriers than those requested when they do
7104 not support as fine grained a barrier as requested. Some architectures do
7105 not need all types of barriers and on such architectures, these become
7106 noops.</p>
7107
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7108<h5>Example:</h5>
7109<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7110%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7111%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7112 store i32 4, %ptr
7113
7114%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7115 call void @llvm.memory.barrier(i1 false, i1 true, i1 false, i1 false)
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7116 <i>; guarantee the above finishes</i>
7117 store i32 8, %ptr <i>; before this begins</i>
7118</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7119
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7120</div>
7121
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7122<!-- _______________________________________________________________________ -->
7123<div class="doc_subsubsection">
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7124 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7125</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7126
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7127<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7128
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7129<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7130<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
7131 any integer bit width and for different address spaces. Not all targets
7132 support all bit widths however.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7133
7134<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7135 declare i8 @llvm.atomic.cmp.swap.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt;)
7136 declare i16 @llvm.atomic.cmp.swap.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt;)
7137 declare i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt;)
7138 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]7139</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7140
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7141<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7142<p>This loads a value in memory and compares it to a given value. If they are
7143 equal, it stores a new value into the memory.</p>
7144
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7145<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7146<p>The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result
7147 as well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
7148 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
7149 this integer type. While any bit width integer may be used, targets may only
7150 lower representations they support in hardware.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7151
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7152<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7153<p>This entire intrinsic must be executed atomically. It first loads the value
7154 in memory pointed to by <tt>ptr</tt> and compares it with the
7155 value <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the
7156 memory. The loaded value is yielded in all cases. This provides the
7157 equivalent of an atomic compare-and-swap operation within the SSA
7158 framework.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7159
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7160<h5>Examples:</h5>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7161<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7162%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7163%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7164 store i32 4, %ptr
7165
7166%val1 = add i32 4, 4
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7167%result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %ptr, i32 4, %val1)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7168 <i>; yields {i32}:result1 = 4</i>
7169%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
7170%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
7171
7172%val2 = add i32 1, 1
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7173%result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %ptr, i32 5, %val2)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7174 <i>; yields {i32}:result2 = 8</i>
7175%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
7176
7177%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
7178</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7179
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7180</div>
7181
7182<!-- _______________________________________________________________________ -->
7183<div class="doc_subsubsection">
7184 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
7185</div>
7186<div class="doc_text">
7187<h5>Syntax:</h5>
7188
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7189<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
7190 integer bit width. Not all targets support all bit widths however.</p>
7191
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7192<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7193 declare i8 @llvm.atomic.swap.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;val&gt;)
7194 declare i16 @llvm.atomic.swap.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;val&gt;)
7195 declare i32 @llvm.atomic.swap.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;val&gt;)
7196 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]7197</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7198
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7199<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7200<p>This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
7201 the value from memory. It then stores the value in <tt>val</tt> in the memory
7202 at <tt>ptr</tt>.</p>
7203
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7204<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7205<p>The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both
7206 the <tt>val</tt> argument and the result must be integers of the same bit
7207 width. The first argument, <tt>ptr</tt>, must be a pointer to a value of this
7208 integer type. The targets may only lower integer representations they
7209 support.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7210
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7211<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7212<p>This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
7213 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
7214 equivalent of an atomic swap operation within the SSA framework.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7215
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7216<h5>Examples:</h5>
7217<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7218%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7219%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7220 store i32 4, %ptr
7221
7222%val1 = add i32 4, 4
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7223%result1 = call i32 @llvm.atomic.swap.i32.p0i32(i32* %ptr, i32 %val1)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7224 <i>; yields {i32}:result1 = 4</i>
7225%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
7226%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
7227
7228%val2 = add i32 1, 1
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7229%result2 = call i32 @llvm.atomic.swap.i32.p0i32(i32* %ptr, i32 %val2)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7230 <i>; yields {i32}:result2 = 8</i>
7231
7232%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
7233%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
7234</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7235
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7236</div>
7237
7238<!-- _______________________________________________________________________ -->
7239<div class="doc_subsubsection">
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7240 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7241
7242</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7243
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7244<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7245
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7246<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7247<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.add</tt> on
7248 any integer bit width. Not all targets support all bit widths however.</p>
7249
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7250<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7251 declare i8 @llvm.atomic.load.add.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7252 declare i16 @llvm.atomic.load.add.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7253 declare i32 @llvm.atomic.load.add.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7254 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]7255</pre>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7256
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7257<h5>Overview:</h5>
7258<p>This intrinsic adds <tt>delta</tt> to the value stored in memory
7259 at <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
7260
7261<h5>Arguments:</h5>
7262<p>The intrinsic takes two arguments, the first a pointer to an integer value
7263 and the second an integer value. The result is also an integer value. These
7264 integer types can have any bit width, but they must all have the same bit
7265 width. The targets may only lower integer representations they support.</p>
7266
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7267<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7268<p>This intrinsic does a series of operations atomically. It first loads the
7269 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
7270 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]7271
7272<h5>Examples:</h5>
7273<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7274%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7275%ptr = bitcast i8* %mallocP to i32*
7276 store i32 4, %ptr
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7277%result1 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 4)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7278 <i>; yields {i32}:result1 = 4</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7279%result2 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 2)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7280 <i>; yields {i32}:result2 = 8</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7281%result3 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 5)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7282 <i>; yields {i32}:result3 = 10</i>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7283%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7284</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7285
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7286</div>
7287
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7288<!-- _______________________________________________________________________ -->
7289<div class="doc_subsubsection">
7290 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
7291
7292</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7293
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7294<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7295
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7296<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7297<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.sub</tt> on
7298 any integer bit width and for different address spaces. Not all targets
7299 support all bit widths however.</p>
7300
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7301<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7302 declare i8 @llvm.atomic.load.sub.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7303 declare i16 @llvm.atomic.load.sub.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7304 declare i32 @llvm.atomic.load.sub.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7305 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]7306</pre>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7307
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7308<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7309<p>This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7310 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
7311
7312<h5>Arguments:</h5>
7313<p>The intrinsic takes two arguments, the first a pointer to an integer value
7314 and the second an integer value. The result is also an integer value. These
7315 integer types can have any bit width, but they must all have the same bit
7316 width. The targets may only lower integer representations they support.</p>
7317
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7318<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7319<p>This intrinsic does a series of operations atomically. It first loads the
7320 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
7321 result to <tt>ptr</tt>. It yields the original value stored
7322 at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7323
7324<h5>Examples:</h5>
7325<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7326%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7327%ptr = bitcast i8* %mallocP to i32*
7328 store i32 8, %ptr
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7329%result1 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 4)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7330 <i>; yields {i32}:result1 = 8</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7331%result2 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 2)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7332 <i>; yields {i32}:result2 = 4</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7333%result3 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 5)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7334 <i>; yields {i32}:result3 = 2</i>
7335%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
7336</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7337
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7338</div>
7339
7340<!-- _______________________________________________________________________ -->
7341<div class="doc_subsubsection">
7342 <a name="int_atomic_load_and">'<tt>llvm.atomic.load.and.*</tt>' Intrinsic</a><br>
7343 <a name="int_atomic_load_nand">'<tt>llvm.atomic.load.nand.*</tt>' Intrinsic</a><br>
7344 <a name="int_atomic_load_or">'<tt>llvm.atomic.load.or.*</tt>' Intrinsic</a><br>
7345 <a name="int_atomic_load_xor">'<tt>llvm.atomic.load.xor.*</tt>' Intrinsic</a><br>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7346</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7347
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7348<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7349
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7350<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7351<p>These are overloaded intrinsics. You can
7352 use <tt>llvm.atomic.load_and</tt>, <tt>llvm.atomic.load_nand</tt>,
7353 <tt>llvm.atomic.load_or</tt>, and <tt>llvm.atomic.load_xor</tt> on any integer
7354 bit width and for different address spaces. Not all targets support all bit
7355 widths however.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7356
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7357<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7358 declare i8 @llvm.atomic.load.and.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7359 declare i16 @llvm.atomic.load.and.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7360 declare i32 @llvm.atomic.load.and.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7361 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]7362</pre>
7363
7364<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7365 declare i8 @llvm.atomic.load.or.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7366 declare i16 @llvm.atomic.load.or.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7367 declare i32 @llvm.atomic.load.or.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7368 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]7369</pre>
7370
7371<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7372 declare i8 @llvm.atomic.load.nand.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7373 declare i16 @llvm.atomic.load.nand.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7374 declare i32 @llvm.atomic.load.nand.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7375 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]7376</pre>
7377
7378<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7379 declare i8 @llvm.atomic.load.xor.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7380 declare i16 @llvm.atomic.load.xor.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7381 declare i32 @llvm.atomic.load.xor.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7382 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]7383</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7384
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7385<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7386<p>These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
7387 the value stored in memory at <tt>ptr</tt>. It yields the original value
7388 at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7389
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7390<h5>Arguments:</h5>
7391<p>These intrinsics take two arguments, the first a pointer to an integer value
7392 and the second an integer value. The result is also an integer value. These
7393 integer types can have any bit width, but they must all have the same bit
7394 width. The targets may only lower integer representations they support.</p>
7395
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7396<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7397<p>These intrinsics does a series of operations atomically. They first load the
7398 value stored at <tt>ptr</tt>. They then do the bitwise
7399 operation <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the
7400 original value stored at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7401
7402<h5>Examples:</h5>
7403<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7404%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7405%ptr = bitcast i8* %mallocP to i32*
7406 store i32 0x0F0F, %ptr
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7407%result0 = call i32 @llvm.atomic.load.nand.i32.p0i32(i32* %ptr, i32 0xFF)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7408 <i>; yields {i32}:result0 = 0x0F0F</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7409%result1 = call i32 @llvm.atomic.load.and.i32.p0i32(i32* %ptr, i32 0xFF)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7410 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7411%result2 = call i32 @llvm.atomic.load.or.i32.p0i32(i32* %ptr, i32 0F)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7412 <i>; yields {i32}:result2 = 0xF0</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7413%result3 = call i32 @llvm.atomic.load.xor.i32.p0i32(i32* %ptr, i32 0F)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7414 <i>; yields {i32}:result3 = FF</i>
7415%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
7416</pre>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7417
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7418</div>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7419
7420<!-- _______________________________________________________________________ -->
7421<div class="doc_subsubsection">
7422 <a name="int_atomic_load_max">'<tt>llvm.atomic.load.max.*</tt>' Intrinsic</a><br>
7423 <a name="int_atomic_load_min">'<tt>llvm.atomic.load.min.*</tt>' Intrinsic</a><br>
7424 <a name="int_atomic_load_umax">'<tt>llvm.atomic.load.umax.*</tt>' Intrinsic</a><br>
7425 <a name="int_atomic_load_umin">'<tt>llvm.atomic.load.umin.*</tt>' Intrinsic</a><br>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7426</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7427
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7428<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7429
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7430<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7431<p>These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
7432 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
7433 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
7434 address spaces. Not all targets support all bit widths however.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7435
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7436<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7437 declare i8 @llvm.atomic.load.max.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7438 declare i16 @llvm.atomic.load.max.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7439 declare i32 @llvm.atomic.load.max.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7440 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]7441</pre>
7442
7443<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7444 declare i8 @llvm.atomic.load.min.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7445 declare i16 @llvm.atomic.load.min.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7446 declare i32 @llvm.atomic.load.min.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7447 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]7448</pre>
7449
7450<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7451 declare i8 @llvm.atomic.load.umax.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7452 declare i16 @llvm.atomic.load.umax.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7453 declare i32 @llvm.atomic.load.umax.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7454 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]7455</pre>
7456
7457<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7458 declare i8 @llvm.atomic.load.umin.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7459 declare i16 @llvm.atomic.load.umin.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7460 declare i32 @llvm.atomic.load.umin.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7461 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]7462</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7463
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7464<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7465<p>These intrinsics takes the signed or unsigned minimum or maximum of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7466 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
7467 original value at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7468
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7469<h5>Arguments:</h5>
7470<p>These intrinsics take two arguments, the first a pointer to an integer value
7471 and the second an integer value. The result is also an integer value. These
7472 integer types can have any bit width, but they must all have the same bit
7473 width. The targets may only lower integer representations they support.</p>
7474
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7475<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7476<p>These intrinsics does a series of operations atomically. They first load the
7477 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or
7478 max <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They
7479 yield the original value stored at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7480
7481<h5>Examples:</h5>
7482<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7483%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7484%ptr = bitcast i8* %mallocP to i32*
7485 store i32 7, %ptr
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7486%result0 = call i32 @llvm.atomic.load.min.i32.p0i32(i32* %ptr, i32 -2)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7487 <i>; yields {i32}:result0 = 7</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7488%result1 = call i32 @llvm.atomic.load.max.i32.p0i32(i32* %ptr, i32 8)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7489 <i>; yields {i32}:result1 = -2</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7490%result2 = call i32 @llvm.atomic.load.umin.i32.p0i32(i32* %ptr, i32 10)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7491 <i>; yields {i32}:result2 = 8</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7492%result3 = call i32 @llvm.atomic.load.umax.i32.p0i32(i32* %ptr, i32 30)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7493 <i>; yields {i32}:result3 = 8</i>
7494%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
7495</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7496
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7497</div>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7498
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7499
7500<!-- ======================================================================= -->
7501<div class="doc_subsection">
7502 <a name="int_memorymarkers">Memory Use Markers</a>
7503</div>
7504
7505<div class="doc_text">
7506
7507<p>This class of intrinsics exists to information about the lifetime of memory
7508 objects and ranges where variables are immutable.</p>
7509
7510</div>
7511
7512<!-- _______________________________________________________________________ -->
7513<div class="doc_subsubsection">
7514 <a name="int_lifetime_start">'<tt>llvm.lifetime.start</tt>' Intrinsic</a>
7515</div>
7516
7517<div class="doc_text">
7518
7519<h5>Syntax:</h5>
7520<pre>
7521 declare void @llvm.lifetime.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7522</pre>
7523
7524<h5>Overview:</h5>
7525<p>The '<tt>llvm.lifetime.start</tt>' intrinsic specifies the start of a memory
7526 object's lifetime.</p>
7527
7528<h5>Arguments:</h5>
Nick Lewycky9bc89042009-10-13 07:57:33 +0000[diff] [blame]7529<p>The first argument is a constant integer representing the size of the
7530 object, or -1 if it is variable sized. The second argument is a pointer to
7531 the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7532
7533<h5>Semantics:</h5>
7534<p>This intrinsic indicates that before this point in the code, the value of the
7535 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]7536 never be used and has an undefined value. A load from the pointer that
7537 precedes this intrinsic can be replaced with
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7538 <tt>'<a href="#undefvalues">undef</a>'</tt>.</p>
7539
7540</div>
7541
7542<!-- _______________________________________________________________________ -->
7543<div class="doc_subsubsection">
7544 <a name="int_lifetime_end">'<tt>llvm.lifetime.end</tt>' Intrinsic</a>
7545</div>
7546
7547<div class="doc_text">
7548
7549<h5>Syntax:</h5>
7550<pre>
7551 declare void @llvm.lifetime.end(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7552</pre>
7553
7554<h5>Overview:</h5>
7555<p>The '<tt>llvm.lifetime.end</tt>' intrinsic specifies the end of a memory
7556 object's lifetime.</p>
7557
7558<h5>Arguments:</h5>
Nick Lewycky9bc89042009-10-13 07:57:33 +0000[diff] [blame]7559<p>The first argument is a constant integer representing the size of the
7560 object, or -1 if it is variable sized. The second argument is a pointer to
7561 the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7562
7563<h5>Semantics:</h5>
7564<p>This intrinsic indicates that after this point in the code, the value of the
7565 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
7566 never be used and has an undefined value. Any stores into the memory object
7567 following this intrinsic may be removed as dead.
7568
7569</div>
7570
7571<!-- _______________________________________________________________________ -->
7572<div class="doc_subsubsection">
7573 <a name="int_invariant_start">'<tt>llvm.invariant.start</tt>' Intrinsic</a>
7574</div>
7575
7576<div class="doc_text">
7577
7578<h5>Syntax:</h5>
7579<pre>
7580 declare {}* @llvm.invariant.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;) readonly
7581</pre>
7582
7583<h5>Overview:</h5>
7584<p>The '<tt>llvm.invariant.start</tt>' intrinsic specifies that the contents of
7585 a memory object will not change.</p>
7586
7587<h5>Arguments:</h5>
Nick Lewycky9bc89042009-10-13 07:57:33 +0000[diff] [blame]7588<p>The first argument is a constant integer representing the size of the
7589 object, or -1 if it is variable sized. The second argument is a pointer to
7590 the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7591
7592<h5>Semantics:</h5>
7593<p>This intrinsic indicates that until an <tt>llvm.invariant.end</tt> that uses
7594 the return value, the referenced memory location is constant and
7595 unchanging.</p>
7596
7597</div>
7598
7599<!-- _______________________________________________________________________ -->
7600<div class="doc_subsubsection">
7601 <a name="int_invariant_end">'<tt>llvm.invariant.end</tt>' Intrinsic</a>
7602</div>
7603
7604<div class="doc_text">
7605
7606<h5>Syntax:</h5>
7607<pre>
7608 declare void @llvm.invariant.end({}* &lt;start&gt;, i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7609</pre>
7610
7611<h5>Overview:</h5>
7612<p>The '<tt>llvm.invariant.end</tt>' intrinsic specifies that the contents of
7613 a memory object are mutable.</p>
7614
7615<h5>Arguments:</h5>
7616<p>The first argument is the matching <tt>llvm.invariant.start</tt> intrinsic.
Nick Lewycky9bc89042009-10-13 07:57:33 +0000[diff] [blame]7617 The second argument is a constant integer representing the size of the
7618 object, or -1 if it is variable sized and the third argument is a pointer
7619 to the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7620
7621<h5>Semantics:</h5>
7622<p>This intrinsic indicates that the memory is mutable again.</p>
7623
7624</div>
7625
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7626<!-- ======================================================================= -->
7627<div class="doc_subsection">
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7628 <a name="int_general">General Intrinsics</a>
7629</div>
7630
7631<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7632
7633<p>This class of intrinsics is designed to be generic and has no specific
7634 purpose.</p>
7635
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7636</div>
7637
7638<!-- _______________________________________________________________________ -->
7639<div class="doc_subsubsection">
7640 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
7641</div>
7642
7643<div class="doc_text">
7644
7645<h5>Syntax:</h5>
7646<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7647 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]7648</pre>
7649
7650<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7651<p>The '<tt>llvm.var.annotation</tt>' intrinsic.</p>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7652
7653<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7654<p>The first argument is a pointer to a value, the second is a pointer to a
7655 global string, the third is a pointer to a global string which is the source
7656 file name, and the last argument is the line number.</p>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7657
7658<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7659<p>This intrinsic allows annotation of local variables with arbitrary strings.
7660 This can be useful for special purpose optimizations that want to look for
7661 these annotations. These have no other defined use, they are ignored by code
7662 generation and optimization.</p>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7663
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7664</div>
7665
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7666<!-- _______________________________________________________________________ -->
7667<div class="doc_subsubsection">
Tanya Lattner0186a652007-09-21 23:57:59 +0000[diff] [blame]7668 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7669</div>
7670
7671<div class="doc_text">
7672
7673<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7674<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
7675 any integer bit width.</p>
7676
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7677<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7678 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7679 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7680 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7681 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7682 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]7683</pre>
7684
7685<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7686<p>The '<tt>llvm.annotation</tt>' intrinsic.</p>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7687
7688<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7689<p>The first argument is an integer value (result of some expression), the
7690 second is a pointer to a global string, the third is a pointer to a global
7691 string which is the source file name, and the last argument is the line
7692 number. It returns the value of the first argument.</p>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7693
7694<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7695<p>This intrinsic allows annotations to be put on arbitrary expressions with
7696 arbitrary strings. This can be useful for special purpose optimizations that
7697 want to look for these annotations. These have no other defined use, they
7698 are ignored by code generation and optimization.</p>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7699
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7700</div>
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]7701
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]7702<!-- _______________________________________________________________________ -->
7703<div class="doc_subsubsection">
7704 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
7705</div>
7706
7707<div class="doc_text">
7708
7709<h5>Syntax:</h5>
7710<pre>
7711 declare void @llvm.trap()
7712</pre>
7713
7714<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7715<p>The '<tt>llvm.trap</tt>' intrinsic.</p>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]7716
7717<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7718<p>None.</p>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]7719
7720<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7721<p>This intrinsics is lowered to the target dependent trap instruction. If the
7722 target does not have a trap instruction, this intrinsic will be lowered to
7723 the call of the <tt>abort()</tt> function.</p>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]7724
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]7725</div>
7726
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7727<!-- _______________________________________________________________________ -->
7728<div class="doc_subsubsection">
Misha Brukman50de2b22008-11-22 23:55:29 +0000[diff] [blame]7729 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7730</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7731
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7732<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7733
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7734<h5>Syntax:</h5>
7735<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7736 declare void @llvm.stackprotector(i8* &lt;guard&gt;, i8** &lt;slot&gt;)
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7737</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7738
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7739<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7740<p>The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and
7741 stores it onto the stack at <tt>slot</tt>. The stack slot is adjusted to
7742 ensure that it is placed on the stack before local variables.</p>
7743
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7744<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7745<p>The <tt>llvm.stackprotector</tt> intrinsic requires two pointer
7746 arguments. The first argument is the value loaded from the stack
7747 guard <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt>
7748 that has enough space to hold the value of the guard.</p>
7749
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7750<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7751<p>This intrinsic causes the prologue/epilogue inserter to force the position of
7752 the <tt>AllocaInst</tt> stack slot to be before local variables on the
7753 stack. This is to ensure that if a local variable on the stack is
7754 overwritten, it will destroy the value of the guard. When the function exits,
7755 the guard on the stack is checked against the original guard. If they're
7756 different, then the program aborts by calling the <tt>__stack_chk_fail()</tt>
7757 function.</p>
7758
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7759</div>
7760
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]7761<!-- _______________________________________________________________________ -->
7762<div class="doc_subsubsection">
7763 <a name="int_objectsize">'<tt>llvm.objectsize</tt>' Intrinsic</a>
7764</div>
7765
7766<div class="doc_text">
7767
7768<h5>Syntax:</h5>
7769<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7770 declare i32 @llvm.objectsize.i32(i8* &lt;object&gt;, i1 &lt;type&gt;)
7771 declare i64 @llvm.objectsize.i64(i8* &lt;object&gt;, i1 &lt;type&gt;)
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]7772</pre>
7773
7774<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7775<p>The <tt>llvm.objectsize</tt> intrinsic is designed to provide information
Eric Christopher3070e162010-01-08 21:42:39 +0000[diff] [blame]7776 to the optimizers to discover at compile time either a) when an
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7777 operation like memcpy will either overflow a buffer that corresponds to
7778 an object, or b) to determine that a runtime check for overflow isn't
7779 necessary. An object in this context means an allocation of a
Eric Christopher31e39bd2009-12-23 00:29:49 +0000[diff] [blame]7780 specific class, structure, array, or other object.</p>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]7781
7782<h5>Arguments:</h5>
7783<p>The <tt>llvm.objectsize</tt> intrinsic takes two arguments. The first
Eric Christopher31e39bd2009-12-23 00:29:49 +0000[diff] [blame]7784 argument is a pointer to or into the <tt>object</tt>. The second argument
7785 is a boolean 0 or 1. This argument determines whether you want the
7786 maximum (0) or minimum (1) bytes remaining. This needs to be a literal 0 or
7787 1, variables are not allowed.</p>
7788
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]7789<h5>Semantics:</h5>
7790<p>The <tt>llvm.objectsize</tt> intrinsic is lowered to either a constant
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7791 representing the size of the object concerned or <tt>i32/i64 -1 or 0</tt>
7792 (depending on the <tt>type</tt> argument if the size cannot be determined
7793 at compile time.</p>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]7794
7795</div>
7796
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]7797<!-- *********************************************************************** -->
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]7798<hr>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]7799<address>
7800 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman86242e12008-12-11 17:34:48 +0000[diff] [blame]7801 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]7802 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman86242e12008-12-11 17:34:48 +0000[diff] [blame]7803 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]7804
7805 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
Reid Spencerca058542006-03-14 05:39:39 +0000[diff] [blame]7806 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]7807 Last modified: $Date$
7808</address>
Chris Lattnerb8f816e2008-01-04 04:33:49 +0000[diff] [blame]7809
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]7810</body>
7811</html>