Gitiles
Code Review Sign In
gerrit-public.fairphone.software / toolchain / llvm-project / 15e2d90746f8ca9f5d4b29cafc0a4493daeb54cb / . / llvm / docs / LangRef.html
blob: ec751504dab6a7c3007836e846e088a77e64732b [file] [log] [blame]
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3<html>
4<head>
5 <title>LLVM Assembly Language Reference Manual</title>
Reid Spencercb84e432004-08-26 20:44:00 +0000[diff] [blame]6 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
7 <meta name="author" content="Chris Lattner">
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]8 <meta name="description"
Reid Spencercb84e432004-08-26 20:44:00 +0000[diff] [blame]9 content="LLVM Assembly Language Reference Manual.">
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]10 <link rel="stylesheet" href="llvm.css" type="text/css">
11</head>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]12
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]13<body>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]14
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]15<h1>LLVM Language Reference Manual</h1>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]16<ol>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]17 <li><a href="#abstract">Abstract</a></li>
18 <li><a href="#introduction">Introduction</a></li>
19 <li><a href="#identifiers">Identifiers</a></li>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]20 <li><a href="#highlevel">High Level Structure</a>
21 <ol>
22 <li><a href="#modulestructure">Module Structure</a></li>
Bill Wendlinga3c6f6b2009-07-20 01:03:30 +0000[diff] [blame]23 <li><a href="#linkage">Linkage Types</a>
24 <ol>
Bill Wendling8693ef82009-07-20 02:41:50 +0000[diff] [blame]25 <li><a href="#linkage_private">'<tt>private</tt>' Linkage</a></li>
26 <li><a href="#linkage_linker_private">'<tt>linker_private</tt>' Linkage</a></li>
Bill Wendling03bcd6e2010-07-01 21:55:59 +0000[diff] [blame]27 <li><a href="#linkage_linker_private_weak">'<tt>linker_private_weak</tt>' Linkage</a></li>
Bill Wendling578ee402010-08-20 22:05:50 +0000[diff] [blame]28 <li><a href="#linkage_linker_private_weak_def_auto">'<tt>linker_private_weak_def_auto</tt>' Linkage</a></li>
Bill Wendling8693ef82009-07-20 02:41:50 +0000[diff] [blame]29 <li><a href="#linkage_internal">'<tt>internal</tt>' Linkage</a></li>
30 <li><a href="#linkage_available_externally">'<tt>available_externally</tt>' Linkage</a></li>
31 <li><a href="#linkage_linkonce">'<tt>linkonce</tt>' Linkage</a></li>
32 <li><a href="#linkage_common">'<tt>common</tt>' Linkage</a></li>
33 <li><a href="#linkage_weak">'<tt>weak</tt>' Linkage</a></li>
34 <li><a href="#linkage_appending">'<tt>appending</tt>' Linkage</a></li>
35 <li><a href="#linkage_externweak">'<tt>extern_weak</tt>' Linkage</a></li>
Chris Lattner80d73c72009-10-10 18:26:06 +0000[diff] [blame]36 <li><a href="#linkage_linkonce_odr">'<tt>linkonce_odr</tt>' Linkage</a></li>
Bill Wendling8693ef82009-07-20 02:41:50 +0000[diff] [blame]37 <li><a href="#linkage_weak">'<tt>weak_odr</tt>' Linkage</a></li>
38 <li><a href="#linkage_external">'<tt>externally visible</tt>' Linkage</a></li>
39 <li><a href="#linkage_dllimport">'<tt>dllimport</tt>' Linkage</a></li>
40 <li><a href="#linkage_dllexport">'<tt>dllexport</tt>' Linkage</a></li>
Bill Wendlinga3c6f6b2009-07-20 01:03:30 +0000[diff] [blame]41 </ol>
42 </li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]43 <li><a href="#callingconv">Calling Conventions</a></li>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]44 <li><a href="#namedtypes">Named Types</a></li>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]45 <li><a href="#globalvars">Global Variables</a></li>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]46 <li><a href="#functionstructure">Functions</a></li>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]47 <li><a href="#aliasstructure">Aliases</a></li>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]48 <li><a href="#namedmetadatastructure">Named Metadata</a></li>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]49 <li><a href="#paramattrs">Parameter Attributes</a></li>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]50 <li><a href="#fnattrs">Function Attributes</a></li>
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]51 <li><a href="#gc">Garbage Collector Names</a></li>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]52 <li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]53 <li><a href="#datalayout">Data Layout</a></li>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]54 <li><a href="#pointeraliasing">Pointer Aliasing Rules</a></li>
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]55 <li><a href="#volatile">Volatile Memory Accesses</a></li>
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]56 <li><a href="#memmodel">Memory Model for Concurrent Operations</a></li>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]57 </ol>
58 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]59 <li><a href="#typesystem">Type System</a>
60 <ol>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]61 <li><a href="#t_classifications">Type Classifications</a></li>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]62 <li><a href="#t_primitive">Primitive Types</a>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]63 <ol>
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]64 <li><a href="#t_integer">Integer Type</a></li>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]65 <li><a href="#t_floating">Floating Point Types</a></li>
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]66 <li><a href="#t_x86mmx">X86mmx Type</a></li>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]67 <li><a href="#t_void">Void Type</a></li>
68 <li><a href="#t_label">Label Type</a></li>
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]69 <li><a href="#t_metadata">Metadata Type</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]70 </ol>
71 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]72 <li><a href="#t_derived">Derived Types</a>
73 <ol>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]74 <li><a href="#t_aggregate">Aggregate Types</a>
75 <ol>
76 <li><a href="#t_array">Array Type</a></li>
77 <li><a href="#t_struct">Structure Type</a></li>
Chris Lattner2a843822011-07-23 19:59:08 +0000[diff] [blame]78 <li><a href="#t_opaque">Opaque Structure Types</a></li>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]79 <li><a href="#t_vector">Vector Type</a></li>
80 </ol>
81 </li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]82 <li><a href="#t_function">Function Type</a></li>
83 <li><a href="#t_pointer">Pointer Type</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]84 </ol>
85 </li>
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>
Dan Gohmane635c522011-05-27 00:36:31 +0000[diff] [blame]241 <li><a href="#int_exp">'<tt>llvm.exp.*</tt>' Intrinsic</a></li>
242 <li><a href="#int_log">'<tt>llvm.log.*</tt>' Intrinsic</a></li>
Cameron Zwarichf03fa182011-07-08 21:39:21 +0000[diff] [blame]243 <li><a href="#int_fma">'<tt>llvm.fma.*</tt>' Intrinsic</a></li>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]244 </ol>
245 </li>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]246 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]247 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]248 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattnerb748c672006-01-16 22:34:14 +0000[diff] [blame]249 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
250 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
251 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]252 </ol>
253 </li>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]254 <li><a href="#int_overflow">Arithmetic with Overflow Intrinsics</a>
255 <ol>
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]256 <li><a href="#int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt> Intrinsics</a></li>
257 <li><a href="#int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt> Intrinsics</a></li>
258 <li><a href="#int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt> Intrinsics</a></li>
259 <li><a href="#int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt> Intrinsics</a></li>
260 <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]261 <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]262 </ol>
263 </li>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]264 <li><a href="#int_fp16">Half Precision Floating Point Intrinsics</a>
265 <ol>
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]266 <li><a href="#int_convert_to_fp16">'<tt>llvm.convert.to.fp16</tt>' Intrinsic</a></li>
267 <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]268 </ol>
269 </li>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]270 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]271 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sands86e01192007-09-11 14:10:23 +0000[diff] [blame]272 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]273 <ol>
274 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]275 </ol>
276 </li>
Bill Wendlingf85850f2008-11-18 22:10:53 +0000[diff] [blame]277 <li><a href="#int_atomics">Atomic intrinsics</a>
278 <ol>
279 <li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
280 <li><a href="#int_atomic_cmp_swap"><tt>llvm.atomic.cmp.swap</tt></a></li>
281 <li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
282 <li><a href="#int_atomic_load_add"><tt>llvm.atomic.load.add</tt></a></li>
283 <li><a href="#int_atomic_load_sub"><tt>llvm.atomic.load.sub</tt></a></li>
284 <li><a href="#int_atomic_load_and"><tt>llvm.atomic.load.and</tt></a></li>
285 <li><a href="#int_atomic_load_nand"><tt>llvm.atomic.load.nand</tt></a></li>
286 <li><a href="#int_atomic_load_or"><tt>llvm.atomic.load.or</tt></a></li>
287 <li><a href="#int_atomic_load_xor"><tt>llvm.atomic.load.xor</tt></a></li>
288 <li><a href="#int_atomic_load_max"><tt>llvm.atomic.load.max</tt></a></li>
289 <li><a href="#int_atomic_load_min"><tt>llvm.atomic.load.min</tt></a></li>
290 <li><a href="#int_atomic_load_umax"><tt>llvm.atomic.load.umax</tt></a></li>
291 <li><a href="#int_atomic_load_umin"><tt>llvm.atomic.load.umin</tt></a></li>
292 </ol>
293 </li>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]294 <li><a href="#int_memorymarkers">Memory Use Markers</a>
295 <ol>
296 <li><a href="#int_lifetime_start"><tt>llvm.lifetime.start</tt></a></li>
297 <li><a href="#int_lifetime_end"><tt>llvm.lifetime.end</tt></a></li>
298 <li><a href="#int_invariant_start"><tt>llvm.invariant.start</tt></a></li>
299 <li><a href="#int_invariant_end"><tt>llvm.invariant.end</tt></a></li>
300 </ol>
301 </li>
Reid Spencer5b2cb0f2007-07-20 19:59:11 +0000[diff] [blame]302 <li><a href="#int_general">General intrinsics</a>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]303 <ol>
Reid Spencer5b2cb0f2007-07-20 19:59:11 +0000[diff] [blame]304 <li><a href="#int_var_annotation">
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]305 '<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]306 <li><a href="#int_annotation">
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]307 '<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]308 <li><a href="#int_trap">
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]309 '<tt>llvm.trap</tt>' Intrinsic</a></li>
310 <li><a href="#int_stackprotector">
311 '<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]312 <li><a href="#int_objectsize">
313 '<tt>llvm.objectsize</tt>' Intrinsic</a></li>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]314 </ol>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]315 </li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]316 </ol>
317 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]318</ol>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]319
320<div class="doc_author">
321 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
322 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]323</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]324
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]325<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]326<h2><a name="abstract">Abstract</a></h2>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]327<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]328
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]329<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]330
331<p>This document is a reference manual for the LLVM assembly language. LLVM is
332 a Static Single Assignment (SSA) based representation that provides type
333 safety, low-level operations, flexibility, and the capability of representing
334 'all' high-level languages cleanly. It is the common code representation
335 used throughout all phases of the LLVM compilation strategy.</p>
336
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]337</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]338
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]339<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]340<h2><a name="introduction">Introduction</a></h2>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]341<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]342
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]343<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]344
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]345<p>The LLVM code representation is designed to be used in three different forms:
346 as an in-memory compiler IR, as an on-disk bitcode representation (suitable
347 for fast loading by a Just-In-Time compiler), and as a human readable
348 assembly language representation. This allows LLVM to provide a powerful
349 intermediate representation for efficient compiler transformations and
350 analysis, while providing a natural means to debug and visualize the
351 transformations. The three different forms of LLVM are all equivalent. This
352 document describes the human readable representation and notation.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]353
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]354<p>The LLVM representation aims to be light-weight and low-level while being
355 expressive, typed, and extensible at the same time. It aims to be a
356 "universal IR" of sorts, by being at a low enough level that high-level ideas
357 may be cleanly mapped to it (similar to how microprocessors are "universal
358 IR's", allowing many source languages to be mapped to them). By providing
359 type information, LLVM can be used as the target of optimizations: for
360 example, through pointer analysis, it can be proven that a C automatic
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]361 variable is never accessed outside of the current function, allowing it to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]362 be promoted to a simple SSA value instead of a memory location.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]363
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]364<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]365<h4>
366 <a name="wellformed">Well-Formedness</a>
367</h4>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]368
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]369<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]370
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]371<p>It is important to note that this document describes 'well formed' LLVM
372 assembly language. There is a difference between what the parser accepts and
373 what is considered 'well formed'. For example, the following instruction is
374 syntactically okay, but not well formed:</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]375
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]376<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]377%x = <a href="#i_add">add</a> i32 1, %x
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]378</pre>
379
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]380<p>because the definition of <tt>%x</tt> does not dominate all of its uses. The
381 LLVM infrastructure provides a verification pass that may be used to verify
382 that an LLVM module is well formed. This pass is automatically run by the
383 parser after parsing input assembly and by the optimizer before it outputs
384 bitcode. The violations pointed out by the verifier pass indicate bugs in
385 transformation passes or input to the parser.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]386
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]387</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]388
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]389</div>
390
Chris Lattner87a3dbe2007-10-03 17:34:29 +0000[diff] [blame]391<!-- Describe the typesetting conventions here. -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]392
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]393<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]394<h2><a name="identifiers">Identifiers</a></h2>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]395<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]396
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]397<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]398
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]399<p>LLVM identifiers come in two basic types: global and local. Global
400 identifiers (functions, global variables) begin with the <tt>'@'</tt>
401 character. Local identifiers (register names, types) begin with
402 the <tt>'%'</tt> character. Additionally, there are three different formats
403 for identifiers, for different purposes:</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]404
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]405<ol>
Reid Spencerb23b65f2007-08-07 14:34:28 +0000[diff] [blame]406 <li>Named values are represented as a string of characters with their prefix.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]407 For example, <tt>%foo</tt>, <tt>@DivisionByZero</tt>,
408 <tt>%a.really.long.identifier</tt>. The actual regular expression used is
409 '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'. Identifiers which require
410 other characters in their names can be surrounded with quotes. Special
411 characters may be escaped using <tt>"\xx"</tt> where <tt>xx</tt> is the
412 ASCII code for the character in hexadecimal. In this way, any character
413 can be used in a name value, even quotes themselves.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]414
Reid Spencerb23b65f2007-08-07 14:34:28 +0000[diff] [blame]415 <li>Unnamed values are represented as an unsigned numeric value with their
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]416 prefix. For example, <tt>%12</tt>, <tt>@2</tt>, <tt>%44</tt>.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]417
Reid Spencer8f08d802004-12-09 18:02:53 +0000[diff] [blame]418 <li>Constants, which are described in a <a href="#constants">section about
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]419 constants</a>, below.</li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]420</ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]421
Reid Spencerb23b65f2007-08-07 14:34:28 +0000[diff] [blame]422<p>LLVM requires that values start with a prefix for two reasons: Compilers
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]423 don't need to worry about name clashes with reserved words, and the set of
424 reserved words may be expanded in the future without penalty. Additionally,
425 unnamed identifiers allow a compiler to quickly come up with a temporary
426 variable without having to avoid symbol table conflicts.</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]427
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]428<p>Reserved words in LLVM are very similar to reserved words in other
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]429 languages. There are keywords for different opcodes
430 ('<tt><a href="#i_add">add</a></tt>',
431 '<tt><a href="#i_bitcast">bitcast</a></tt>',
432 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names
433 ('<tt><a href="#t_void">void</a></tt>',
434 '<tt><a href="#t_primitive">i32</a></tt>', etc...), and others. These
435 reserved words cannot conflict with variable names, because none of them
436 start with a prefix character (<tt>'%'</tt> or <tt>'@'</tt>).</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]437
438<p>Here is an example of LLVM code to multiply the integer variable
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]439 '<tt>%X</tt>' by 8:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]440
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]441<p>The easy way:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]442
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]443<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]444%result = <a href="#i_mul">mul</a> i32 %X, 8
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]445</pre>
446
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]447<p>After strength reduction:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]448
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]449<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]450%result = <a href="#i_shl">shl</a> i32 %X, i8 3
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]451</pre>
452
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]453<p>And the hard way:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]454
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]455<pre class="doc_code">
Gabor Greifbd0328f2009-10-28 13:05:07 +0000[diff] [blame]456%0 = <a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
457%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]458%result = <a href="#i_add">add</a> i32 %1, %1
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]459</pre>
460
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]461<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several important
462 lexical features of LLVM:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]463
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]464<ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]465 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]466 line.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]467
468 <li>Unnamed temporaries are created when the result of a computation is not
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]469 assigned to a named value.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]470
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]471 <li>Unnamed temporaries are numbered sequentially</li>
472</ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]473
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]474<p>It also shows a convention that we follow in this document. When
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]475 demonstrating instructions, we will follow an instruction with a comment that
476 defines the type and name of value produced. Comments are shown in italic
477 text.</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]478
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]479</div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]480
481<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]482<h2><a name="highlevel">High Level Structure</a></h2>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]483<!-- *********************************************************************** -->
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]484<div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]485<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]486<h3>
487 <a name="modulestructure">Module Structure</a>
488</h3>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]489
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]490<div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]491
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]492<p>LLVM programs are composed of "Module"s, each of which is a translation unit
493 of the input programs. Each module consists of functions, global variables,
494 and symbol table entries. Modules may be combined together with the LLVM
495 linker, which merges function (and global variable) definitions, resolves
496 forward declarations, and merges symbol table entries. Here is an example of
497 the "hello world" module:</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]498
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]499<pre class="doc_code">
Chris Lattner54a7be72010-08-17 17:13:42 +0000[diff] [blame]500<i>; Declare the string constant as a global constant.</i>&nbsp;
Nick Lewyckyfea7ddc2011-01-29 01:09:53 +0000[diff] [blame]501<a href="#identifiers">@.LC0</a> = <a href="#linkage_internal">internal</a>&nbsp;<a href="#globalvars">constant</a>&nbsp;<a href="#t_array">[13 x i8]</a> c"hello world\0A\00" <i>; [13 x i8]*</i>&nbsp;
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]502
Chris Lattner54a7be72010-08-17 17:13:42 +0000[diff] [blame]503<i>; External declaration of the puts function</i>&nbsp;
504<a href="#functionstructure">declare</a> i32 @puts(i8*) <i>; i32 (i8*)* </i>&nbsp;
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]505
506<i>; Definition of main function</i>
Chris Lattner54a7be72010-08-17 17:13:42 +0000[diff] [blame]507define i32 @main() { <i>; i32()* </i>&nbsp;
508 <i>; Convert [13 x i8]* to i8 *...</i>&nbsp;
509 %cast210 = <a href="#i_getelementptr">getelementptr</a> [13 x i8]* @.LC0, i64 0, i64 0 <i>; i8*</i>&nbsp;
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]510
Chris Lattner54a7be72010-08-17 17:13:42 +0000[diff] [blame]511 <i>; Call puts function to write out the string to stdout.</i>&nbsp;
512 <a href="#i_call">call</a> i32 @puts(i8* %cast210) <i>; i32</i>&nbsp;
513 <a href="#i_ret">ret</a> i32 0&nbsp;
514}
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]515
516<i>; Named metadata</i>
517!1 = metadata !{i32 41}
518!foo = !{!1, null}
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]519</pre>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]520
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]521<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]522 "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>" function,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]523 a <a href="#functionstructure">function definition</a> for
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]524 "<tt>main</tt>" and <a href="#namedmetadatastructure">named metadata</a>
525 "<tt>foo"</tt>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]526
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]527<p>In general, a module is made up of a list of global values, where both
528 functions and global variables are global values. Global values are
529 represented by a pointer to a memory location (in this case, a pointer to an
530 array of char, and a pointer to a function), and have one of the
531 following <a href="#linkage">linkage types</a>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]532
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]533</div>
534
535<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]536<h3>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]537 <a name="linkage">Linkage Types</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]538</h3>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]539
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]540<div>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]541
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]542<p>All Global Variables and Functions have one of the following types of
543 linkage:</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]544
545<dl>
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]546 <dt><tt><b><a name="linkage_private">private</a></b></tt></dt>
Bill Wendling03bcd6e2010-07-01 21:55:59 +0000[diff] [blame]547 <dd>Global values with "<tt>private</tt>" linkage are only directly accessible
548 by objects in the current module. In particular, linking code into a
549 module with an private global value may cause the private to be renamed as
550 necessary to avoid collisions. Because the symbol is private to the
551 module, all references can be updated. This doesn't show up in any symbol
552 table in the object file.</dd>
Rafael Espindola6de96a12009-01-15 20:18:42 +0000[diff] [blame]553
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]554 <dt><tt><b><a name="linkage_linker_private">linker_private</a></b></tt></dt>
Bill Wendling03bcd6e2010-07-01 21:55:59 +0000[diff] [blame]555 <dd>Similar to <tt>private</tt>, but the symbol is passed through the
556 assembler and evaluated by the linker. Unlike normal strong symbols, they
557 are removed by the linker from the final linked image (executable or
558 dynamic library).</dd>
559
560 <dt><tt><b><a name="linkage_linker_private_weak">linker_private_weak</a></b></tt></dt>
561 <dd>Similar to "<tt>linker_private</tt>", but the symbol is weak. Note that
562 <tt>linker_private_weak</tt> symbols are subject to coalescing by the
563 linker. The symbols are removed by the linker from the final linked image
564 (executable or dynamic library).</dd>
Bill Wendlinga3c6f6b2009-07-20 01:03:30 +0000[diff] [blame]565
Bill Wendling578ee402010-08-20 22:05:50 +0000[diff] [blame]566 <dt><tt><b><a name="linkage_linker_private_weak_def_auto">linker_private_weak_def_auto</a></b></tt></dt>
567 <dd>Similar to "<tt>linker_private_weak</tt>", but it's known that the address
568 of the object is not taken. For instance, functions that had an inline
569 definition, but the compiler decided not to inline it. Note,
570 unlike <tt>linker_private</tt> and <tt>linker_private_weak</tt>,
571 <tt>linker_private_weak_def_auto</tt> may have only <tt>default</tt>
572 visibility. The symbols are removed by the linker from the final linked
573 image (executable or dynamic library).</dd>
574
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]575 <dt><tt><b><a name="linkage_internal">internal</a></b></tt></dt>
Bill Wendling36321712010-06-29 22:34:52 +0000[diff] [blame]576 <dd>Similar to private, but the value shows as a local symbol
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]577 (<tt>STB_LOCAL</tt> in the case of ELF) in the object file. This
578 corresponds to the notion of the '<tt>static</tt>' keyword in C.</dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]579
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]580 <dt><tt><b><a name="linkage_available_externally">available_externally</a></b></tt></dt>
Chris Lattner184f1be2009-04-13 05:44:34 +0000[diff] [blame]581 <dd>Globals with "<tt>available_externally</tt>" linkage are never emitted
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]582 into the object file corresponding to the LLVM module. They exist to
583 allow inlining and other optimizations to take place given knowledge of
584 the definition of the global, which is known to be somewhere outside the
585 module. Globals with <tt>available_externally</tt> linkage are allowed to
586 be discarded at will, and are otherwise the same as <tt>linkonce_odr</tt>.
587 This linkage type is only allowed on definitions, not declarations.</dd>
Chris Lattner184f1be2009-04-13 05:44:34 +0000[diff] [blame]588
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]589 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt></dt>
Chris Lattnere20b4702007-01-14 06:51:48 +0000[diff] [blame]590 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
Chris Lattner0de4caa2010-01-09 19:15:14 +0000[diff] [blame]591 the same name when linkage occurs. This can be used to implement
592 some forms of inline functions, templates, or other code which must be
593 generated in each translation unit that uses it, but where the body may
594 be overridden with a more definitive definition later. Unreferenced
595 <tt>linkonce</tt> globals are allowed to be discarded. Note that
596 <tt>linkonce</tt> linkage does not actually allow the optimizer to
597 inline the body of this function into callers because it doesn't know if
598 this definition of the function is the definitive definition within the
599 program or whether it will be overridden by a stronger definition.
600 To enable inlining and other optimizations, use "<tt>linkonce_odr</tt>"
601 linkage.</dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]602
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]603 <dt><tt><b><a name="linkage_weak">weak</a></b></tt></dt>
Chris Lattnerd0554882009-08-05 05:21:07 +0000[diff] [blame]604 <dd>"<tt>weak</tt>" linkage has the same merging semantics as
605 <tt>linkonce</tt> linkage, except that unreferenced globals with
606 <tt>weak</tt> linkage may not be discarded. This is used for globals that
607 are declared "weak" in C source code.</dd>
608
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]609 <dt><tt><b><a name="linkage_common">common</a></b></tt></dt>
Chris Lattnerd0554882009-08-05 05:21:07 +0000[diff] [blame]610 <dd>"<tt>common</tt>" linkage is most similar to "<tt>weak</tt>" linkage, but
611 they are used for tentative definitions in C, such as "<tt>int X;</tt>" at
612 global scope.
613 Symbols with "<tt>common</tt>" linkage are merged in the same way as
614 <tt>weak symbols</tt>, and they may not be deleted if unreferenced.
Chris Lattner0aff0b22009-08-05 05:41:44 +0000[diff] [blame]615 <tt>common</tt> symbols may not have an explicit section,
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]616 must have a zero initializer, and may not be marked '<a
Chris Lattner0aff0b22009-08-05 05:41:44 +0000[diff] [blame]617 href="#globalvars"><tt>constant</tt></a>'. Functions and aliases may not
618 have common linkage.</dd>
Chris Lattnerd0554882009-08-05 05:21:07 +0000[diff] [blame]619
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]620
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]621 <dt><tt><b><a name="linkage_appending">appending</a></b></tt></dt>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]622 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]623 pointer to array type. When two global variables with appending linkage
624 are linked together, the two global arrays are appended together. This is
625 the LLVM, typesafe, equivalent of having the system linker append together
626 "sections" with identical names when .o files are linked.</dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]627
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]628 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]629 <dd>The semantics of this linkage follow the ELF object file model: the symbol
630 is weak until linked, if not linked, the symbol becomes null instead of
631 being an undefined reference.</dd>
Anton Korobeynikova0554d92007-01-12 19:20:47 +0000[diff] [blame]632
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]633 <dt><tt><b><a name="linkage_linkonce_odr">linkonce_odr</a></b></tt></dt>
634 <dt><tt><b><a name="linkage_weak_odr">weak_odr</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]635 <dd>Some languages allow differing globals to be merged, such as two functions
636 with different semantics. Other languages, such as <tt>C++</tt>, ensure
Bill Wendling03bcd6e2010-07-01 21:55:59 +0000[diff] [blame]637 that only equivalent globals are ever merged (the "one definition rule"
638 &mdash; "ODR"). Such languages can use the <tt>linkonce_odr</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]639 and <tt>weak_odr</tt> linkage types to indicate that the global will only
640 be merged with equivalent globals. These linkage types are otherwise the
641 same as their non-<tt>odr</tt> versions.</dd>
Duncan Sands12da8ce2009-03-07 15:45:40 +0000[diff] [blame]642
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]643 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]644 <dd>If none of the above identifiers are used, the global is externally
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]645 visible, meaning that it participates in linkage and can be used to
646 resolve external symbol references.</dd>
Reid Spencer7972c472007-04-11 23:49:50 +0000[diff] [blame]647</dl>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]648
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]649<p>The next two types of linkage are targeted for Microsoft Windows platform
650 only. They are designed to support importing (exporting) symbols from (to)
651 DLLs (Dynamic Link Libraries).</p>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]652
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]653<dl>
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]654 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt></dt>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]655 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]656 or variable via a global pointer to a pointer that is set up by the DLL
657 exporting the symbol. On Microsoft Windows targets, the pointer name is
658 formed by combining <code>__imp_</code> and the function or variable
659 name.</dd>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]660
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]661 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt></dt>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]662 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]663 pointer to a pointer in a DLL, so that it can be referenced with the
664 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
665 name is formed by combining <code>__imp_</code> and the function or
666 variable name.</dd>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]667</dl>
668
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]669<p>For example, since the "<tt>.LC0</tt>" variable is defined to be internal, if
670 another module defined a "<tt>.LC0</tt>" variable and was linked with this
671 one, one of the two would be renamed, preventing a collision. Since
672 "<tt>main</tt>" and "<tt>puts</tt>" are external (i.e., lacking any linkage
673 declarations), they are accessible outside of the current module.</p>
674
675<p>It is illegal for a function <i>declaration</i> to have any linkage type
676 other than "externally visible", <tt>dllimport</tt>
677 or <tt>extern_weak</tt>.</p>
678
Duncan Sands12da8ce2009-03-07 15:45:40 +0000[diff] [blame]679<p>Aliases can have only <tt>external</tt>, <tt>internal</tt>, <tt>weak</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]680 or <tt>weak_odr</tt> linkages.</p>
681
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]682</div>
683
684<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]685<h3>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]686 <a name="callingconv">Calling Conventions</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]687</h3>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]688
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]689<div>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]690
691<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]692 and <a href="#i_invoke">invokes</a> can all have an optional calling
693 convention specified for the call. The calling convention of any pair of
694 dynamic caller/callee must match, or the behavior of the program is
695 undefined. The following calling conventions are supported by LLVM, and more
696 may be added in the future:</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]697
698<dl>
699 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]700 <dd>This calling convention (the default if no other calling convention is
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]701 specified) matches the target C calling conventions. This calling
702 convention supports varargs function calls and tolerates some mismatch in
703 the declared prototype and implemented declaration of the function (as
704 does normal C).</dd>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]705
706 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]707 <dd>This calling convention attempts to make calls as fast as possible
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]708 (e.g. by passing things in registers). This calling convention allows the
709 target to use whatever tricks it wants to produce fast code for the
710 target, without having to conform to an externally specified ABI
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]711 (Application Binary Interface).
712 <a href="CodeGenerator.html#tailcallopt">Tail calls can only be optimized
Chris Lattnera179e4d2010-03-11 00:22:57 +0000[diff] [blame]713 when this or the GHC convention is used.</a> This calling convention
714 does not support varargs and requires the prototype of all callees to
715 exactly match the prototype of the function definition.</dd>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]716
717 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]718 <dd>This calling convention attempts to make code in the caller as efficient
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]719 as possible under the assumption that the call is not commonly executed.
720 As such, these calls often preserve all registers so that the call does
721 not break any live ranges in the caller side. This calling convention
722 does not support varargs and requires the prototype of all callees to
723 exactly match the prototype of the function definition.</dd>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]724
Chris Lattnera179e4d2010-03-11 00:22:57 +0000[diff] [blame]725 <dt><b>"<tt>cc <em>10</em></tt>" - GHC convention</b>:</dt>
726 <dd>This calling convention has been implemented specifically for use by the
727 <a href="http://www.haskell.org/ghc">Glasgow Haskell Compiler (GHC)</a>.
728 It passes everything in registers, going to extremes to achieve this by
729 disabling callee save registers. This calling convention should not be
730 used lightly but only for specific situations such as an alternative to
731 the <em>register pinning</em> performance technique often used when
732 implementing functional programming languages.At the moment only X86
733 supports this convention and it has the following limitations:
734 <ul>
735 <li>On <em>X86-32</em> only supports up to 4 bit type parameters. No
736 floating point types are supported.</li>
737 <li>On <em>X86-64</em> only supports up to 10 bit type parameters and
738 6 floating point parameters.</li>
739 </ul>
740 This calling convention supports
741 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> but
742 requires both the caller and callee are using it.
743 </dd>
744
Chris Lattner573f64e2005-05-07 01:46:40 +0000[diff] [blame]745 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]746 <dd>Any calling convention may be specified by number, allowing
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]747 target-specific calling conventions to be used. Target specific calling
748 conventions start at 64.</dd>
Chris Lattner573f64e2005-05-07 01:46:40 +0000[diff] [blame]749</dl>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]750
751<p>More calling conventions can be added/defined on an as-needed basis, to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]752 support Pascal conventions or any other well-known target-independent
753 convention.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]754
755</div>
756
757<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]758<h3>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]759 <a name="visibility">Visibility Styles</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]760</h3>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]761
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]762<div>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]763
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]764<p>All Global Variables and Functions have one of the following visibility
765 styles:</p>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]766
767<dl>
768 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]769 <dd>On targets that use the ELF object file format, default visibility means
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]770 that the declaration is visible to other modules and, in shared libraries,
771 means that the declared entity may be overridden. On Darwin, default
772 visibility means that the declaration is visible to other modules. Default
773 visibility corresponds to "external linkage" in the language.</dd>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]774
775 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]776 <dd>Two declarations of an object with hidden visibility refer to the same
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]777 object if they are in the same shared object. Usually, hidden visibility
778 indicates that the symbol will not be placed into the dynamic symbol
779 table, so no other module (executable or shared library) can reference it
780 directly.</dd>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]781
Anton Korobeynikov39f3cff2007-04-29 18:35:00 +0000[diff] [blame]782 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
Anton Korobeynikov39f3cff2007-04-29 18:35:00 +0000[diff] [blame]783 <dd>On ELF, protected visibility indicates that the symbol will be placed in
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]784 the dynamic symbol table, but that references within the defining module
785 will bind to the local symbol. That is, the symbol cannot be overridden by
786 another module.</dd>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]787</dl>
788
789</div>
790
791<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]792<h3>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]793 <a name="namedtypes">Named Types</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]794</h3>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]795
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]796<div>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]797
798<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]799 it easier to read the IR and make the IR more condensed (particularly when
800 recursive types are involved). An example of a name specification is:</p>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]801
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]802<pre class="doc_code">
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]803%mytype = type { %mytype*, i32 }
804</pre>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]805
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]806<p>You may give a name to any <a href="#typesystem">type</a> except
Chris Lattner249b9762010-08-17 23:26:04 +0000[diff] [blame]807 "<a href="#t_void">void</a>". Type name aliases may be used anywhere a type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]808 is expected with the syntax "%mytype".</p>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]809
810<p>Note that type names are aliases for the structural type that they indicate,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]811 and that you can therefore specify multiple names for the same type. This
812 often leads to confusing behavior when dumping out a .ll file. Since LLVM IR
813 uses structural typing, the name is not part of the type. When printing out
814 LLVM IR, the printer will pick <em>one name</em> to render all types of a
815 particular shape. This means that if you have code where two different
816 source types end up having the same LLVM type, that the dumper will sometimes
817 print the "wrong" or unexpected type. This is an important design point and
818 isn't going to change.</p>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]819
820</div>
821
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]822<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]823<h3>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]824 <a name="globalvars">Global Variables</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]825</h3>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]826
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]827<div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]828
Chris Lattner5d5aede2005-02-12 19:30:21 +0000[diff] [blame]829<p>Global variables define regions of memory allocated at compilation time
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]830 instead of run-time. Global variables may optionally be initialized, may
831 have an explicit section to be placed in, and may have an optional explicit
832 alignment specified. A variable may be defined as "thread_local", which
833 means that it will not be shared by threads (each thread will have a
834 separated copy of the variable). A variable may be defined as a global
835 "constant," which indicates that the contents of the variable
836 will <b>never</b> be modified (enabling better optimization, allowing the
837 global data to be placed in the read-only section of an executable, etc).
838 Note that variables that need runtime initialization cannot be marked
839 "constant" as there is a store to the variable.</p>
Chris Lattner5d5aede2005-02-12 19:30:21 +0000[diff] [blame]840
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]841<p>LLVM explicitly allows <em>declarations</em> of global variables to be marked
842 constant, even if the final definition of the global is not. This capability
843 can be used to enable slightly better optimization of the program, but
844 requires the language definition to guarantee that optimizations based on the
845 'constantness' are valid for the translation units that do not include the
846 definition.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]847
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]848<p>As SSA values, global variables define pointer values that are in scope
849 (i.e. they dominate) all basic blocks in the program. Global variables
850 always define a pointer to their "content" type because they describe a
851 region of memory, and all memory objects in LLVM are accessed through
852 pointers.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]853
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame]854<p>Global variables can be marked with <tt>unnamed_addr</tt> which indicates
855 that the address is not significant, only the content. Constants marked
Rafael Espindolaf1ed7812011-01-15 08:20:57 +0000[diff] [blame]856 like this can be merged with other constants if they have the same
857 initializer. Note that a constant with significant address <em>can</em>
858 be merged with a <tt>unnamed_addr</tt> constant, the result being a
859 constant whose address is significant.</p>
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame]860
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]861<p>A global variable may be declared to reside in a target-specific numbered
862 address space. For targets that support them, address spaces may affect how
863 optimizations are performed and/or what target instructions are used to
864 access the variable. The default address space is zero. The address space
865 qualifier must precede any other attributes.</p>
Christopher Lamb308121c2007-12-11 09:31:00 +0000[diff] [blame]866
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]867<p>LLVM allows an explicit section to be specified for globals. If the target
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]868 supports it, it will emit globals to the section specified.</p>
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]869
Chris Lattner78e00bc2010-04-28 00:13:42 +0000[diff] [blame]870<p>An explicit alignment may be specified for a global, which must be a power
871 of 2. If not present, or if the alignment is set to zero, the alignment of
872 the global is set by the target to whatever it feels convenient. If an
873 explicit alignment is specified, the global is forced to have exactly that
Chris Lattner4bd85e42010-04-28 00:31:12 +0000[diff] [blame]874 alignment. Targets and optimizers are not allowed to over-align the global
875 if the global has an assigned section. In this case, the extra alignment
876 could be observable: for example, code could assume that the globals are
877 densely packed in their section and try to iterate over them as an array,
878 alignment padding would break this iteration.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]879
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]880<p>For example, the following defines a global in a numbered address space with
881 an initializer, section, and alignment:</p>
Chris Lattner5760c502007-01-14 00:27:09 +0000[diff] [blame]882
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]883<pre class="doc_code">
Dan Gohmanaaa679b2009-01-11 00:40:00 +0000[diff] [blame]884@G = addrspace(5) constant float 1.0, section "foo", align 4
Chris Lattner5760c502007-01-14 00:27:09 +0000[diff] [blame]885</pre>
886
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]887</div>
888
889
890<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]891<h3>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]892 <a name="functionstructure">Functions</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]893</h3>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]894
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]895<div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]896
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]897<p>LLVM function definitions consist of the "<tt>define</tt>" keyword, an
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]898 optional <a href="#linkage">linkage type</a>, an optional
899 <a href="#visibility">visibility style</a>, an optional
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame]900 <a href="#callingconv">calling convention</a>,
901 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]902 <a href="#paramattrs">parameter attribute</a> for the return type, a function
903 name, a (possibly empty) argument list (each with optional
904 <a href="#paramattrs">parameter attributes</a>), optional
905 <a href="#fnattrs">function attributes</a>, an optional section, an optional
906 alignment, an optional <a href="#gc">garbage collector name</a>, an opening
907 curly brace, a list of basic blocks, and a closing curly brace.</p>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]908
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]909<p>LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
910 optional <a href="#linkage">linkage type</a>, an optional
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]911 <a href="#visibility">visibility style</a>, an optional
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame]912 <a href="#callingconv">calling convention</a>,
913 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]914 <a href="#paramattrs">parameter attribute</a> for the return type, a function
915 name, a possibly empty list of arguments, an optional alignment, and an
916 optional <a href="#gc">garbage collector name</a>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]917
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]918<p>A function definition contains a list of basic blocks, forming the CFG
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]919 (Control Flow Graph) for the function. Each basic block may optionally start
920 with a label (giving the basic block a symbol table entry), contains a list
921 of instructions, and ends with a <a href="#terminators">terminator</a>
922 instruction (such as a branch or function return).</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]923
Chris Lattnera59fb102007-06-08 16:52:14 +0000[diff] [blame]924<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]925 executed on entrance to the function, and it is not allowed to have
926 predecessor basic blocks (i.e. there can not be any branches to the entry
927 block of a function). Because the block can have no predecessors, it also
928 cannot have any <a href="#i_phi">PHI nodes</a>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]929
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]930<p>LLVM allows an explicit section to be specified for functions. If the target
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]931 supports it, it will emit functions to the section specified.</p>
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]932
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]933<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]934 the alignment is set to zero, the alignment of the function is set by the
935 target to whatever it feels convenient. If an explicit alignment is
936 specified, the function is forced to have at least that much alignment. All
937 alignments must be a power of 2.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]938
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame]939<p>If the <tt>unnamed_addr</tt> attribute is given, the address is know to not
940 be significant and two identical functions can be merged</p>.
941
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]942<h5>Syntax:</h5>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]943<pre class="doc_code">
Chris Lattner0ae02092008-10-13 16:55:18 +0000[diff] [blame]944define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]945 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
946 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
947 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
948 [<a href="#gc">gc</a>] { ... }
949</pre>
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]950
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]951</div>
952
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]953<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]954<h3>
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]955 <a name="aliasstructure">Aliases</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]956</h3>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]957
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]958<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]959
960<p>Aliases act as "second name" for the aliasee value (which can be either
961 function, global variable, another alias or bitcast of global value). Aliases
962 may have an optional <a href="#linkage">linkage type</a>, and an
963 optional <a href="#visibility">visibility style</a>.</p>
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]964
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]965<h5>Syntax:</h5>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]966<pre class="doc_code">
Duncan Sands7e99a942008-09-12 20:48:21 +0000[diff] [blame]967@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Bill Wendling2d8b9a82007-05-29 09:42:13 +0000[diff] [blame]968</pre>
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]969
970</div>
971
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]972<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]973<h3>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]974 <a name="namedmetadatastructure">Named Metadata</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]975</h3>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]976
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]977<div>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]978
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]979<p>Named metadata is a collection of metadata. <a href="#metadata">Metadata
Dan Gohman093cb792010-07-21 18:54:18 +0000[diff] [blame]980 nodes</a> (but not metadata strings) are the only valid operands for
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]981 a named metadata.</p>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]982
983<h5>Syntax:</h5>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]984<pre class="doc_code">
Dan Gohman093cb792010-07-21 18:54:18 +0000[diff] [blame]985; Some unnamed metadata nodes, which are referenced by the named metadata.
986!0 = metadata !{metadata !"zero"}
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]987!1 = metadata !{metadata !"one"}
Dan Gohman093cb792010-07-21 18:54:18 +0000[diff] [blame]988!2 = metadata !{metadata !"two"}
Dan Gohman58cd65f2010-07-13 19:48:13 +0000[diff] [blame]989; A named metadata.
Dan Gohman093cb792010-07-21 18:54:18 +0000[diff] [blame]990!name = !{!0, !1, !2}
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]991</pre>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]992
993</div>
994
995<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]996<h3>
997 <a name="paramattrs">Parameter Attributes</a>
998</h3>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]999
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1000<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1001
1002<p>The return type and each parameter of a function type may have a set of
1003 <i>parameter attributes</i> associated with them. Parameter attributes are
1004 used to communicate additional information about the result or parameters of
1005 a function. Parameter attributes are considered to be part of the function,
1006 not of the function type, so functions with different parameter attributes
1007 can have the same function type.</p>
1008
1009<p>Parameter attributes are simple keywords that follow the type specified. If
1010 multiple parameter attributes are needed, they are space separated. For
1011 example:</p>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]1012
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1013<pre class="doc_code">
Nick Lewyckydac78d82009-02-15 23:06:14 +0000[diff] [blame]1014declare i32 @printf(i8* noalias nocapture, ...)
Chris Lattnerd2597d72008-10-04 18:33:34 +0000[diff] [blame]1015declare i32 @atoi(i8 zeroext)
1016declare signext i8 @returns_signed_char()
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]1017</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]1018
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1019<p>Note that any attributes for the function result (<tt>nounwind</tt>,
1020 <tt>readonly</tt>) come immediately after the argument list.</p>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1021
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1022<p>Currently, only the following parameter attributes are defined:</p>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000[diff] [blame]1023
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1024<dl>
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1025 <dt><tt><b>zeroext</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1026 <dd>This indicates to the code generator that the parameter or return value
Cameron Zwarichac106272011-03-16 22:20:18 +0000[diff] [blame]1027 should be zero-extended to the extent required by the target's ABI (which
1028 is usually 32-bits, but is 8-bits for a i1 on x86-64) by the caller (for a
1029 parameter) or the callee (for a return value).</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000[diff] [blame]1030
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1031 <dt><tt><b>signext</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1032 <dd>This indicates to the code generator that the parameter or return value
Cameron Zwarich341c36d2011-03-17 14:21:58 +0000[diff] [blame]1033 should be sign-extended to the extent required by the target's ABI (which
1034 is usually 32-bits) by the caller (for a parameter) or the callee (for a
1035 return value).</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000[diff] [blame]1036
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1037 <dt><tt><b>inreg</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1038 <dd>This indicates that this parameter or return value should be treated in a
1039 special target-dependent fashion during while emitting code for a function
1040 call or return (usually, by putting it in a register as opposed to memory,
1041 though some targets use it to distinguish between two different kinds of
1042 registers). Use of this attribute is target-specific.</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000[diff] [blame]1043
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1044 <dt><tt><b><a name="byval">byval</a></b></tt></dt>
Chris Lattnerd78dbee2010-11-20 23:49:06 +0000[diff] [blame]1045 <dd><p>This indicates that the pointer parameter should really be passed by
1046 value to the function. The attribute implies that a hidden copy of the
1047 pointee
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1048 is made between the caller and the callee, so the callee is unable to
1049 modify the value in the callee. This attribute is only valid on LLVM
1050 pointer arguments. It is generally used to pass structs and arrays by
1051 value, but is also valid on pointers to scalars. The copy is considered
1052 to belong to the caller not the callee (for example,
1053 <tt><a href="#readonly">readonly</a></tt> functions should not write to
1054 <tt>byval</tt> parameters). This is not a valid attribute for return
Chris Lattnerd78dbee2010-11-20 23:49:06 +0000[diff] [blame]1055 values.</p>
1056
1057 <p>The byval attribute also supports specifying an alignment with
1058 the align attribute. It indicates the alignment of the stack slot to
1059 form and the known alignment of the pointer specified to the call site. If
1060 the alignment is not specified, then the code generator makes a
1061 target-specific assumption.</p></dd>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1062
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]1063 <dt><tt><b><a name="sret">sret</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1064 <dd>This indicates that the pointer parameter specifies the address of a
1065 structure that is the return value of the function in the source program.
1066 This pointer must be guaranteed by the caller to be valid: loads and
1067 stores to the structure may be assumed by the callee to not to trap. This
1068 may only be applied to the first parameter. This is not a valid attribute
1069 for return values. </dd>
1070
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]1071 <dt><tt><b><a name="noalias">noalias</a></b></tt></dt>
Dan Gohmandf12d082010-07-02 18:41:32 +0000[diff] [blame]1072 <dd>This indicates that pointer values
1073 <a href="#pointeraliasing"><i>based</i></a> on the argument or return
Dan Gohmande256292010-07-02 23:46:54 +0000[diff] [blame]1074 value do not alias pointer values which are not <i>based</i> on it,
1075 ignoring certain "irrelevant" dependencies.
1076 For a call to the parent function, dependencies between memory
1077 references from before or after the call and from those during the call
1078 are "irrelevant" to the <tt>noalias</tt> keyword for the arguments and
1079 return value used in that call.
Dan Gohmandf12d082010-07-02 18:41:32 +0000[diff] [blame]1080 The caller shares the responsibility with the callee for ensuring that
1081 these requirements are met.
1082 For further details, please see the discussion of the NoAlias response in
Dan Gohman6c858db2010-07-06 15:26:33 +0000[diff] [blame]1083 <a href="AliasAnalysis.html#MustMayNo">alias analysis</a>.<br>
1084<br>
John McCall72ed8902010-07-06 21:07:14 +0000[diff] [blame]1085 Note that this definition of <tt>noalias</tt> is intentionally
1086 similar to the definition of <tt>restrict</tt> in C99 for function
Chris Lattner5eff9ca2010-07-06 20:51:35 +0000[diff] [blame]1087 arguments, though it is slightly weaker.
Dan Gohman6c858db2010-07-06 15:26:33 +0000[diff] [blame]1088<br>
1089 For function return values, C99's <tt>restrict</tt> is not meaningful,
1090 while LLVM's <tt>noalias</tt> is.
1091 </dd>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1092
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]1093 <dt><tt><b><a name="nocapture">nocapture</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1094 <dd>This indicates that the callee does not make any copies of the pointer
1095 that outlive the callee itself. This is not a valid attribute for return
1096 values.</dd>
1097
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]1098 <dt><tt><b><a name="nest">nest</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1099 <dd>This indicates that the pointer parameter can be excised using the
1100 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
1101 attribute for return values.</dd>
1102</dl>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1103
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1104</div>
1105
1106<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1107<h3>
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1108 <a name="gc">Garbage Collector Names</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1109</h3>
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1110
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1111<div>
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1112
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1113<p>Each function may specify a garbage collector name, which is simply a
1114 string:</p>
1115
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1116<pre class="doc_code">
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1117define void @f() gc "name" { ... }
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1118</pre>
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1119
1120<p>The compiler declares the supported values of <i>name</i>. Specifying a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1121 collector which will cause the compiler to alter its output in order to
1122 support the named garbage collection algorithm.</p>
1123
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1124</div>
1125
1126<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1127<h3>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1128 <a name="fnattrs">Function Attributes</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1129</h3>
Devang Patelcaacdba2008-09-04 23:05:13 +0000[diff] [blame]1130
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1131<div>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1132
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1133<p>Function attributes are set to communicate additional information about a
1134 function. Function attributes are considered to be part of the function, not
1135 of the function type, so functions with different parameter attributes can
1136 have the same function type.</p>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1137
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1138<p>Function attributes are simple keywords that follow the type specified. If
1139 multiple attributes are needed, they are space separated. For example:</p>
Devang Patelcaacdba2008-09-04 23:05:13 +0000[diff] [blame]1140
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1141<pre class="doc_code">
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1142define void @f() noinline { ... }
1143define void @f() alwaysinline { ... }
1144define void @f() alwaysinline optsize { ... }
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1145define void @f() optsize { ... }
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1146</pre>
Devang Patelcaacdba2008-09-04 23:05:13 +0000[diff] [blame]1147
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1148<dl>
Charles Davisbe5557e2010-02-12 00:31:15 +0000[diff] [blame]1149 <dt><tt><b>alignstack(&lt;<em>n</em>&gt;)</b></tt></dt>
1150 <dd>This attribute indicates that, when emitting the prologue and epilogue,
1151 the backend should forcibly align the stack pointer. Specify the
1152 desired alignment, which must be a power of two, in parentheses.
1153
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1154 <dt><tt><b>alwaysinline</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1155 <dd>This attribute indicates that the inliner should attempt to inline this
1156 function into callers whenever possible, ignoring any active inlining size
1157 threshold for this caller.</dd>
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1158
Charles Davis22fe1862010-10-25 15:37:09 +0000[diff] [blame]1159 <dt><tt><b>hotpatch</b></tt></dt>
Charles Davis1b2d3722010-10-25 16:29:03 +0000[diff] [blame]1160 <dd>This attribute indicates that the function should be 'hotpatchable',
Charles Davis74205252010-10-25 19:07:39 +0000[diff] [blame]1161 meaning the function can be patched and/or hooked even while it is
1162 loaded into memory. On x86, the function prologue will be preceded
1163 by six bytes of padding and will begin with a two-byte instruction.
1164 Most of the functions in the Windows system DLLs in Windows XP SP2 or
1165 higher were compiled in this fashion.</dd>
Charles Davis22fe1862010-10-25 15:37:09 +0000[diff] [blame]1166
Dan Gohman8bd11f12011-06-16 16:03:13 +0000[diff] [blame]1167 <dt><tt><b>nonlazybind</b></tt></dt>
1168 <dd>This attribute suppresses lazy symbol binding for the function. This
1169 may make calls to the function faster, at the cost of extra program
1170 startup time if the function is not called during program startup.</dd>
1171
Jakob Stoklund Olesen74bb06c2010-02-06 01:16:28 +0000[diff] [blame]1172 <dt><tt><b>inlinehint</b></tt></dt>
1173 <dd>This attribute indicates that the source code contained a hint that inlining
1174 this function is desirable (such as the "inline" keyword in C/C++). It
1175 is just a hint; it imposes no requirements on the inliner.</dd>
1176
Nick Lewycky14b58da2010-07-06 18:24:09 +0000[diff] [blame]1177 <dt><tt><b>naked</b></tt></dt>
1178 <dd>This attribute disables prologue / epilogue emission for the function.
1179 This can have very system-specific consequences.</dd>
1180
1181 <dt><tt><b>noimplicitfloat</b></tt></dt>
1182 <dd>This attributes disables implicit floating point instructions.</dd>
1183
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1184 <dt><tt><b>noinline</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1185 <dd>This attribute indicates that the inliner should never inline this
1186 function in any situation. This attribute may not be used together with
1187 the <tt>alwaysinline</tt> attribute.</dd>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1188
Nick Lewycky14b58da2010-07-06 18:24:09 +0000[diff] [blame]1189 <dt><tt><b>noredzone</b></tt></dt>
1190 <dd>This attribute indicates that the code generator should not use a red
1191 zone, even if the target-specific ABI normally permits it.</dd>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1192
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1193 <dt><tt><b>noreturn</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1194 <dd>This function attribute indicates that the function never returns
1195 normally. This produces undefined behavior at runtime if the function
1196 ever does dynamically return.</dd>
Bill Wendlinga8130172008-11-13 01:02:51 +0000[diff] [blame]1197
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1198 <dt><tt><b>nounwind</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1199 <dd>This function attribute indicates that the function never returns with an
1200 unwind or exceptional control flow. If the function does unwind, its
1201 runtime behavior is undefined.</dd>
Bill Wendling0f5541e2008-11-26 19:07:40 +0000[diff] [blame]1202
Nick Lewycky14b58da2010-07-06 18:24:09 +0000[diff] [blame]1203 <dt><tt><b>optsize</b></tt></dt>
1204 <dd>This attribute suggests that optimization passes and code generator passes
1205 make choices that keep the code size of this function low, and otherwise
1206 do optimizations specifically to reduce code size.</dd>
1207
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1208 <dt><tt><b>readnone</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1209 <dd>This attribute indicates that the function computes its result (or decides
1210 to unwind an exception) based strictly on its arguments, without
1211 dereferencing any pointer arguments or otherwise accessing any mutable
1212 state (e.g. memory, control registers, etc) visible to caller functions.
1213 It does not write through any pointer arguments
1214 (including <tt><a href="#byval">byval</a></tt> arguments) and never
1215 changes any state visible to callers. This means that it cannot unwind
1216 exceptions by calling the <tt>C++</tt> exception throwing methods, but
1217 could use the <tt>unwind</tt> instruction.</dd>
Devang Patel310fd4a2009-06-12 19:45:19 +0000[diff] [blame]1218
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1219 <dt><tt><b><a name="readonly">readonly</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1220 <dd>This attribute indicates that the function does not write through any
1221 pointer arguments (including <tt><a href="#byval">byval</a></tt>
1222 arguments) or otherwise modify any state (e.g. memory, control registers,
1223 etc) visible to caller functions. It may dereference pointer arguments
1224 and read state that may be set in the caller. A readonly function always
1225 returns the same value (or unwinds an exception identically) when called
1226 with the same set of arguments and global state. It cannot unwind an
1227 exception by calling the <tt>C++</tt> exception throwing methods, but may
1228 use the <tt>unwind</tt> instruction.</dd>
Anton Korobeynikovc8ce7b082009-07-17 18:07:26 +0000[diff] [blame]1229
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1230 <dt><tt><b><a name="ssp">ssp</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1231 <dd>This attribute indicates that the function should emit a stack smashing
1232 protector. It is in the form of a "canary"&mdash;a random value placed on
1233 the stack before the local variables that's checked upon return from the
1234 function to see if it has been overwritten. A heuristic is used to
1235 determine if a function needs stack protectors or not.<br>
1236<br>
1237 If a function that has an <tt>ssp</tt> attribute is inlined into a
1238 function that doesn't have an <tt>ssp</tt> attribute, then the resulting
1239 function will have an <tt>ssp</tt> attribute.</dd>
1240
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1241 <dt><tt><b>sspreq</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1242 <dd>This attribute indicates that the function should <em>always</em> emit a
1243 stack smashing protector. This overrides
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1244 the <tt><a href="#ssp">ssp</a></tt> function attribute.<br>
1245<br>
1246 If a function that has an <tt>sspreq</tt> attribute is inlined into a
1247 function that doesn't have an <tt>sspreq</tt> attribute or which has
1248 an <tt>ssp</tt> attribute, then the resulting function will have
1249 an <tt>sspreq</tt> attribute.</dd>
Rafael Espindola163d6752011-07-25 15:27:59 +0000[diff] [blame]1250
1251 <dt><tt><b><a name="uwtable">uwtable</a></b></tt></dt>
1252 <dd>This attribute indicates that the ABI being targeted requires that
1253 an unwind table entry be produce for this function even if we can
1254 show that no exceptions passes by it. This is normally the case for
1255 the ELF x86-64 abi, but it can be disabled for some compilation
1256 units.</dd>
1257
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1258</dl>
1259
Devang Patelcaacdba2008-09-04 23:05:13 +0000[diff] [blame]1260</div>
1261
1262<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1263<h3>
Chris Lattner93564892006-04-08 04:40:53 +0000[diff] [blame]1264 <a name="moduleasm">Module-Level Inline Assembly</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1265</h3>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1266
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1267<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1268
1269<p>Modules may contain "module-level inline asm" blocks, which corresponds to
1270 the GCC "file scope inline asm" blocks. These blocks are internally
1271 concatenated by LLVM and treated as a single unit, but may be separated in
1272 the <tt>.ll</tt> file if desired. The syntax is very simple:</p>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1273
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1274<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]1275module asm "inline asm code goes here"
1276module asm "more can go here"
1277</pre>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1278
1279<p>The strings can contain any character by escaping non-printable characters.
1280 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]1281 for the number.</p>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1282
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1283<p>The inline asm code is simply printed to the machine code .s file when
1284 assembly code is generated.</p>
1285
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1286</div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1287
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1288<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1289<h3>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1290 <a name="datalayout">Data Layout</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1291</h3>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1292
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1293<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1294
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1295<p>A module may specify a target specific data layout string that specifies how
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1296 data is to be laid out in memory. The syntax for the data layout is
1297 simply:</p>
1298
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1299<pre class="doc_code">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1300target datalayout = "<i>layout specification</i>"
1301</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1302
1303<p>The <i>layout specification</i> consists of a list of specifications
1304 separated by the minus sign character ('-'). Each specification starts with
1305 a letter and may include other information after the letter to define some
1306 aspect of the data layout. The specifications accepted are as follows:</p>
1307
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1308<dl>
1309 <dt><tt>E</tt></dt>
1310 <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]1311 bits with the most significance have the lowest address location.</dd>
1312
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1313 <dt><tt>e</tt></dt>
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]1314 <dd>Specifies that the target lays out data in little-endian form. That is,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1315 the bits with the least significance have the lowest address
1316 location.</dd>
1317
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1318 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1319 <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]1320 <i>preferred</i> alignments. All sizes are in bits. Specifying
1321 the <i>pref</i> alignment is optional. If omitted, the
1322 preceding <tt>:</tt> should be omitted too.</dd>
1323
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1324 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1325 <dd>This specifies the alignment for an integer type of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1326 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1327
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1328 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1329 <dd>This specifies the alignment for a vector type of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1330 <i>size</i>.</dd>
1331
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1332 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1333 <dd>This specifies the alignment for a floating point type of a given bit
Dale Johannesence522852010-05-28 18:54:47 +0000[diff] [blame]1334 <i>size</i>. Only values of <i>size</i> that are supported by the target
1335 will work. 32 (float) and 64 (double) are supported on all targets;
1336 80 or 128 (different flavors of long double) are also supported on some
1337 targets.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1338
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1339 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1340 <dd>This specifies the alignment for an aggregate type of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1341 <i>size</i>.</dd>
1342
Daniel Dunbar7921a592009-06-08 22:17:53 +0000[diff] [blame]1343 <dt><tt>s<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1344 <dd>This specifies the alignment for a stack object of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1345 <i>size</i>.</dd>
Chris Lattnera381eff2009-11-07 09:35:34 +0000[diff] [blame]1346
1347 <dt><tt>n<i>size1</i>:<i>size2</i>:<i>size3</i>...</tt></dt>
1348 <dd>This specifies a set of native integer widths for the target CPU
1349 in bits. For example, it might contain "n32" for 32-bit PowerPC,
1350 "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]1351 this set are considered to support most general arithmetic
Chris Lattnera381eff2009-11-07 09:35:34 +0000[diff] [blame]1352 operations efficiently.</dd>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1353</dl>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1354
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1355<p>When constructing the data layout for a given target, LLVM starts with a
Dan Gohman61110ae2010-04-28 00:36:01 +0000[diff] [blame]1356 default set of specifications which are then (possibly) overridden by the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1357 specifications in the <tt>datalayout</tt> keyword. The default specifications
1358 are given in this list:</p>
1359
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1360<ul>
1361 <li><tt>E</tt> - big endian</li>
Dan Gohman8ad777d2010-02-23 02:44:03 +0000[diff] [blame]1362 <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]1363 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1364 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1365 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1366 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]1367 <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]1368 alignment of 64-bits</li>
1369 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1370 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1371 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1372 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1373 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
Daniel Dunbar7921a592009-06-08 22:17:53 +0000[diff] [blame]1374 <li><tt>s0:64:64</tt> - stack objects are 64-bit aligned</li>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1375</ul>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1376
1377<p>When LLVM is determining the alignment for a given type, it uses the
1378 following rules:</p>
1379
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1380<ol>
1381 <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]1382 specification is used.</li>
1383
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1384 <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]1385 smallest integer type that is larger than the bitwidth of the sought type
1386 is used. If none of the specifications are larger than the bitwidth then
1387 the the largest integer type is used. For example, given the default
1388 specifications above, the i7 type will use the alignment of i8 (next
1389 largest) while both i65 and i256 will use the alignment of i64 (largest
1390 specified).</li>
1391
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1392 <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]1393 largest vector type that is smaller than the sought vector type will be
1394 used as a fall back. This happens because &lt;128 x double&gt; can be
1395 implemented in terms of 64 &lt;2 x double&gt;, for example.</li>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1396</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1397
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1398</div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1399
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1400<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1401<h3>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1402 <a name="pointeraliasing">Pointer Aliasing Rules</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1403</h3>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1404
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1405<div>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1406
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1407<p>Any memory access must be done through a pointer value associated
Andreas Bolkae39f0332009-07-27 20:37:10 +0000[diff] [blame]1408with an address range of the memory access, otherwise the behavior
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1409is undefined. Pointer values are associated with address ranges
1410according to the following rules:</p>
1411
1412<ul>
Dan Gohmandf12d082010-07-02 18:41:32 +0000[diff] [blame]1413 <li>A pointer value is associated with the addresses associated with
1414 any value it is <i>based</i> on.
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1415 <li>An address of a global variable is associated with the address
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1416 range of the variable's storage.</li>
1417 <li>The result value of an allocation instruction is associated with
1418 the address range of the allocated storage.</li>
1419 <li>A null pointer in the default address-space is associated with
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1420 no address.</li>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1421 <li>An integer constant other than zero or a pointer value returned
1422 from a function not defined within LLVM may be associated with address
1423 ranges allocated through mechanisms other than those provided by
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1424 LLVM. Such ranges shall not overlap with any ranges of addresses
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1425 allocated by mechanisms provided by LLVM.</li>
Dan Gohmandf12d082010-07-02 18:41:32 +0000[diff] [blame]1426</ul>
1427
1428<p>A pointer value is <i>based</i> on another pointer value according
1429 to the following rules:</p>
1430
1431<ul>
1432 <li>A pointer value formed from a
1433 <tt><a href="#i_getelementptr">getelementptr</a></tt> operation
1434 is <i>based</i> on the first operand of the <tt>getelementptr</tt>.</li>
1435 <li>The result value of a
1436 <tt><a href="#i_bitcast">bitcast</a></tt> is <i>based</i> on the operand
1437 of the <tt>bitcast</tt>.</li>
1438 <li>A pointer value formed by an
1439 <tt><a href="#i_inttoptr">inttoptr</a></tt> is <i>based</i> on all
1440 pointer values that contribute (directly or indirectly) to the
1441 computation of the pointer's value.</li>
1442 <li>The "<i>based</i> on" relationship is transitive.</li>
1443</ul>
1444
1445<p>Note that this definition of <i>"based"</i> is intentionally
1446 similar to the definition of <i>"based"</i> in C99, though it is
1447 slightly weaker.</p>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1448
1449<p>LLVM IR does not associate types with memory. The result type of a
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1450<tt><a href="#i_load">load</a></tt> merely indicates the size and
1451alignment of the memory from which to load, as well as the
Dan Gohman4eb47192010-06-17 19:23:50 +0000[diff] [blame]1452interpretation of the value. The first operand type of a
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1453<tt><a href="#i_store">store</a></tt> similarly only indicates the size
1454and alignment of the store.</p>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1455
1456<p>Consequently, type-based alias analysis, aka TBAA, aka
1457<tt>-fstrict-aliasing</tt>, is not applicable to general unadorned
1458LLVM IR. <a href="#metadata">Metadata</a> may be used to encode
1459additional information which specialized optimization passes may use
1460to implement type-based alias analysis.</p>
1461
1462</div>
1463
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]1464<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1465<h3>
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]1466 <a name="volatile">Volatile Memory Accesses</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1467</h3>
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]1468
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1469<div>
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]1470
1471<p>Certain memory accesses, such as <a href="#i_load"><tt>load</tt></a>s, <a
1472href="#i_store"><tt>store</tt></a>s, and <a
1473href="#int_memcpy"><tt>llvm.memcpy</tt></a>s may be marked <tt>volatile</tt>.
1474The optimizers must not change the number of volatile operations or change their
1475order of execution relative to other volatile operations. The optimizers
1476<i>may</i> change the order of volatile operations relative to non-volatile
1477operations. This is not Java's "volatile" and has no cross-thread
1478synchronization behavior.</p>
1479
1480</div>
1481
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1482<!-- ======================================================================= -->
1483<h3>
1484 <a name="memmodel">Memory Model for Concurrent Operations</a>
1485</h3>
1486
1487<div>
1488
1489<p>The LLVM IR does not define any way to start parallel threads of execution
1490or to register signal handlers. Nonetheless, there are platform-specific
1491ways to create them, and we define LLVM IR's behavior in their presence. This
1492model is inspired by the C++0x memory model.</p>
1493
1494<p>We define a <i>happens-before</i> partial order as the least partial order
1495that</p>
1496<ul>
1497 <li>Is a superset of single-thread program order, and</li>
1498 <li>When a <i>synchronizes-with</i> <tt>b</tt>, includes an edge from
1499 <tt>a</tt> to <tt>b</tt>. <i>Synchronizes-with</i> pairs are introduced
1500 by platform-specific techniques, like pthread locks, thread
1501 creation, thread joining, etc., and by the atomic operations described
1502 in the <a href="#int_atomics">Atomic intrinsics</a> section.</li>
1503</ul>
1504
1505<p>Note that program order does not introduce <i>happens-before</i> edges
1506between a thread and signals executing inside that thread.</p>
1507
1508<p>Every (defined) read operation (load instructions, memcpy, atomic
1509loads/read-modify-writes, etc.) <var>R</var> reads a series of bytes written by
1510(defined) write operations (store instructions, atomic
Eli Friedmanf12e4e92011-07-22 03:04:45 +0000[diff] [blame]1511stores/read-modify-writes, memcpy, etc.). For the purposes of this section,
1512initialized globals are considered to have a write of the initializer which is
1513atomic and happens before any other read or write of the memory in question.
1514For each byte of a read <var>R</var>, <var>R<sub>byte</sub></var> may see
1515any write to the same byte, except:</p>
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1516
1517<ul>
1518 <li>If <var>write<sub>1</sub></var> happens before
1519 <var>write<sub>2</sub></var>, and <var>write<sub>2</sub></var> happens
1520 before <var>R<sub>byte</sub></var>, then <var>R<sub>byte</sub></var>
Eli Friedmanf12e4e92011-07-22 03:04:45 +0000[diff] [blame]1521 does not see <var>write<sub>1</sub></var>.
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1522 <li>If <var>R<sub>byte</sub></var> happens before <var>write<sub>3</var>,
Eli Friedmanf12e4e92011-07-22 03:04:45 +0000[diff] [blame]1523 then <var>R<sub>byte</sub></var> does not see
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1524 <var>write<sub>3</sub></var>.
1525</ul>
1526
1527<p>Given that definition, <var>R<sub>byte</sub></var> is defined as follows:
1528<ul>
1529 <li>If there is no write to the same byte that happens before
1530 <var>R<sub>byte</sub></var>, <var>R<sub>byte</sub></var> returns
1531 <tt>undef</tt> for that byte.
Eli Friedmanf12e4e92011-07-22 03:04:45 +0000[diff] [blame]1532 <li>Otherwise, if <var>R<sub>byte</sub></var> may see exactly one write,
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1533 <var>R<sub>byte</sub></var> returns the value written by that
1534 write.</li>
Eli Friedmanf12e4e92011-07-22 03:04:45 +0000[diff] [blame]1535 <li>Otherwise, if <var>R</var> is atomic, and all the writes
1536 <var>R<sub>byte</sub></var> may see are atomic, it chooses one of the
1537 values written. See the <a href="#int_atomics">Atomic intrinsics</a>
1538 section for additional guarantees on how the choice is made.
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1539 <li>Otherwise <var>R<sub>byte</sub></var> returns <tt>undef</tt>.</li>
1540</ul>
1541
1542<p><var>R</var> returns the value composed of the series of bytes it read.
1543This implies that some bytes within the value may be <tt>undef</tt>
1544<b>without</b> the entire value being <tt>undef</tt>. Note that this only
1545defines the semantics of the operation; it doesn't mean that targets will
1546emit more than one instruction to read the series of bytes.</p>
1547
1548<p>Note that in cases where none of the atomic intrinsics are used, this model
1549places only one restriction on IR transformations on top of what is required
1550for single-threaded execution: introducing a store to a byte which might not
Eli Friedmanf12e4e92011-07-22 03:04:45 +0000[diff] [blame]1551otherwise be stored to can introduce undefined behavior. (Specifically, in
1552the case where another thread might write to and read from an address,
1553introducing a store can change a load that may see exactly one write into
1554a load that may see multiple writes.)</p>
Eli Friedman35b54aa2011-07-20 21:35:53 +0000[diff] [blame]1555
1556<!-- FIXME: This model assumes all targets where concurrency is relevant have
1557a byte-size store which doesn't affect adjacent bytes. As far as I can tell,
1558none of the backends currently in the tree fall into this category; however,
1559there might be targets which care. If there are, we want a paragraph
1560like the following:
1561
1562Targets may specify that stores narrower than a certain width are not
1563available; on such a target, for the purposes of this model, treat any
1564non-atomic write with an alignment or width less than the minimum width
1565as if it writes to the relevant surrounding bytes.
1566-->
1567
1568</div>
1569
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1570</div>
1571
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1572<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1573<h2><a name="typesystem">Type System</a></h2>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1574<!-- *********************************************************************** -->
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1575
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1576<div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1577
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1578<p>The LLVM type system is one of the most important features of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1579 intermediate representation. Being typed enables a number of optimizations
1580 to be performed on the intermediate representation directly, without having
1581 to do extra analyses on the side before the transformation. A strong type
1582 system makes it easier to read the generated code and enables novel analyses
1583 and transformations that are not feasible to perform on normal three address
1584 code representations.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1585
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1586<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1587<h3>
1588 <a name="t_classifications">Type Classifications</a>
1589</h3>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1590
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1591<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1592
1593<p>The types fall into a few useful classifications:</p>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]1594
1595<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1596 <tbody>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1597 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1598 <tr>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1599 <td><a href="#t_integer">integer</a></td>
Reid Spencer138249b2007-05-16 18:44:01 +0000[diff] [blame]1600 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1601 </tr>
1602 <tr>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1603 <td><a href="#t_floating">floating point</a></td>
1604 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1605 </tr>
1606 <tr>
1607 <td><a name="t_firstclass">first class</a></td>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1608 <td><a href="#t_integer">integer</a>,
1609 <a href="#t_floating">floating point</a>,
1610 <a href="#t_pointer">pointer</a>,
Dan Gohman08783a882008-06-18 18:42:13 +0000[diff] [blame]1611 <a href="#t_vector">vector</a>,
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]1612 <a href="#t_struct">structure</a>,
1613 <a href="#t_array">array</a>,
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1614 <a href="#t_label">label</a>,
1615 <a href="#t_metadata">metadata</a>.
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1616 </td>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1617 </tr>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1618 <tr>
1619 <td><a href="#t_primitive">primitive</a></td>
1620 <td><a href="#t_label">label</a>,
1621 <a href="#t_void">void</a>,
Tobias Grosser4c8c95b2010-12-28 20:29:31 +0000[diff] [blame]1622 <a href="#t_integer">integer</a>,
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1623 <a href="#t_floating">floating point</a>,
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]1624 <a href="#t_x86mmx">x86mmx</a>,
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1625 <a href="#t_metadata">metadata</a>.</td>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1626 </tr>
1627 <tr>
1628 <td><a href="#t_derived">derived</a></td>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1629 <td><a href="#t_array">array</a>,
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1630 <a href="#t_function">function</a>,
1631 <a href="#t_pointer">pointer</a>,
1632 <a href="#t_struct">structure</a>,
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1633 <a href="#t_vector">vector</a>,
1634 <a href="#t_opaque">opaque</a>.
Dan Gohman93bf60d2008-10-14 16:32:04 +0000[diff] [blame]1635 </td>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1636 </tr>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1637 </tbody>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1638</table>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]1639
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1640<p>The <a href="#t_firstclass">first class</a> types are perhaps the most
1641 important. Values of these types are the only ones which can be produced by
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1642 instructions.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1643
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1644</div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1645
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1646<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1647<h3>
1648 <a name="t_primitive">Primitive Types</a>
1649</h3>
Chris Lattner43542b32008-01-04 04:34:14 +0000[diff] [blame]1650
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1651<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1652
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1653<p>The primitive types are the fundamental building blocks of the LLVM
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1654 system.</p>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1655
1656<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1657<h4>
1658 <a name="t_integer">Integer Type</a>
1659</h4>
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1660
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1661<div>
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1662
1663<h5>Overview:</h5>
1664<p>The integer type is a very simple type that simply specifies an arbitrary
1665 bit width for the integer type desired. Any bit width from 1 bit to
1666 2<sup>23</sup>-1 (about 8 million) can be specified.</p>
1667
1668<h5>Syntax:</h5>
1669<pre>
1670 iN
1671</pre>
1672
1673<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1674 value.</p>
1675
1676<h5>Examples:</h5>
1677<table class="layout">
1678 <tr class="layout">
1679 <td class="left"><tt>i1</tt></td>
1680 <td class="left">a single-bit integer.</td>
1681 </tr>
1682 <tr class="layout">
1683 <td class="left"><tt>i32</tt></td>
1684 <td class="left">a 32-bit integer.</td>
1685 </tr>
1686 <tr class="layout">
1687 <td class="left"><tt>i1942652</tt></td>
1688 <td class="left">a really big integer of over 1 million bits.</td>
1689 </tr>
1690</table>
1691
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1692</div>
1693
1694<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1695<h4>
1696 <a name="t_floating">Floating Point Types</a>
1697</h4>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1698
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1699<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1700
1701<table>
1702 <tbody>
1703 <tr><th>Type</th><th>Description</th></tr>
1704 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1705 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1706 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1707 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1708 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1709 </tbody>
1710</table>
1711
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1712</div>
1713
1714<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1715<h4>
1716 <a name="t_x86mmx">X86mmx Type</a>
1717</h4>
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]1718
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1719<div>
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]1720
1721<h5>Overview:</h5>
1722<p>The x86mmx type represents a value held in an MMX register on an x86 machine. The operations allowed on it are quite limited: parameters and return values, load and store, and bitcast. User-specified MMX instructions are represented as intrinsic or asm calls with arguments and/or results of this type. There are no arrays, vectors or constants of this type.</p>
1723
1724<h5>Syntax:</h5>
1725<pre>
Dale Johannesenb1f0ff12010-10-01 01:07:02 +0000[diff] [blame]1726 x86mmx
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]1727</pre>
1728
1729</div>
1730
1731<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1732<h4>
1733 <a name="t_void">Void Type</a>
1734</h4>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1735
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1736<div>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1737
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1738<h5>Overview:</h5>
1739<p>The void type does not represent any value and has no size.</p>
1740
1741<h5>Syntax:</h5>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1742<pre>
1743 void
1744</pre>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1745
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1746</div>
1747
1748<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1749<h4>
1750 <a name="t_label">Label Type</a>
1751</h4>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1752
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1753<div>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1754
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1755<h5>Overview:</h5>
1756<p>The label type represents code labels.</p>
1757
1758<h5>Syntax:</h5>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1759<pre>
1760 label
1761</pre>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1762
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1763</div>
1764
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1765<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1766<h4>
1767 <a name="t_metadata">Metadata Type</a>
1768</h4>
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1769
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1770<div>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1771
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1772<h5>Overview:</h5>
Nick Lewycky93e06a52009-09-27 23:27:42 +0000[diff] [blame]1773<p>The metadata type represents embedded metadata. No derived types may be
1774 created from metadata except for <a href="#t_function">function</a>
1775 arguments.
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1776
1777<h5>Syntax:</h5>
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1778<pre>
1779 metadata
1780</pre>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1781
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1782</div>
1783
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1784</div>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1785
1786<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1787<h3>
1788 <a name="t_derived">Derived Types</a>
1789</h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1790
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1791<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1792
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1793<p>The real power in LLVM comes from the derived types in the system. This is
1794 what allows a programmer to represent arrays, functions, pointers, and other
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1795 useful types. Each of these types contain one or more element types which
1796 may be a primitive type, or another derived type. For example, it is
1797 possible to have a two dimensional array, using an array as the element type
1798 of another array.</p>
Dan Gohman142ccc02009-01-24 15:58:40 +0000[diff] [blame]1799
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]1800</div>
1801
1802
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1803<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1804<h4>
1805 <a name="t_aggregate">Aggregate Types</a>
1806</h4>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1807
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1808<div>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1809
1810<p>Aggregate Types are a subset of derived types that can contain multiple
1811 member types. <a href="#t_array">Arrays</a>,
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]1812 <a href="#t_struct">structs</a>, and <a href="#t_vector">vectors</a> are
1813 aggregate types.</p>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1814
1815</div>
1816
Reid Spencer138249b2007-05-16 18:44:01 +0000[diff] [blame]1817<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1818<h4>
1819 <a name="t_array">Array Type</a>
1820</h4>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1821
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1822<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1823
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1824<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1825<p>The array type is a very simple derived type that arranges elements
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1826 sequentially in memory. The array type requires a size (number of elements)
1827 and an underlying data type.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1828
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]1829<h5>Syntax:</h5>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1830<pre>
1831 [&lt;# elements&gt; x &lt;elementtype&gt;]
1832</pre>
1833
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1834<p>The number of elements is a constant integer value; <tt>elementtype</tt> may
1835 be any type with a size.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1836
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]1837<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1838<table class="layout">
1839 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1840 <td class="left"><tt>[40 x i32]</tt></td>
1841 <td class="left">Array of 40 32-bit integer values.</td>
1842 </tr>
1843 <tr class="layout">
1844 <td class="left"><tt>[41 x i32]</tt></td>
1845 <td class="left">Array of 41 32-bit integer values.</td>
1846 </tr>
1847 <tr class="layout">
1848 <td class="left"><tt>[4 x i8]</tt></td>
1849 <td class="left">Array of 4 8-bit integer values.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1850 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1851</table>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1852<p>Here are some examples of multidimensional arrays:</p>
1853<table class="layout">
1854 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1855 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1856 <td class="left">3x4 array of 32-bit integer values.</td>
1857 </tr>
1858 <tr class="layout">
1859 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1860 <td class="left">12x10 array of single precision floating point values.</td>
1861 </tr>
1862 <tr class="layout">
1863 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1864 <td class="left">2x3x4 array of 16-bit integer values.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1865 </tr>
1866</table>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +0000[diff] [blame]1867
Dan Gohmanc74bc282009-11-09 19:01:53 +0000[diff] [blame]1868<p>There is no restriction on indexing beyond the end of the array implied by
1869 a static type (though there are restrictions on indexing beyond the bounds
1870 of an allocated object in some cases). This means that single-dimension
1871 'variable sized array' addressing can be implemented in LLVM with a zero
1872 length array type. An implementation of 'pascal style arrays' in LLVM could
1873 use the type "<tt>{ i32, [0 x float]}</tt>", for example.</p>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +0000[diff] [blame]1874
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1875</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1876
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1877<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1878<h4>
1879 <a name="t_function">Function Type</a>
1880</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1881
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1882<div>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]1883
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1884<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1885<p>The function type can be thought of as a function signature. It consists of
1886 a return type and a list of formal parameter types. The return type of a
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]1887 function type is a first class type or a void type.</p>
Devang Pateld6cff512008-03-10 20:49:15 +0000[diff] [blame]1888
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1889<h5>Syntax:</h5>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]1890<pre>
Nick Lewycky14d1ccc2009-09-27 07:55:32 +0000[diff] [blame]1891 &lt;returntype&gt; (&lt;parameter list&gt;)
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]1892</pre>
1893
John Criswell4c0cf7f2005-10-24 16:17:18 +0000[diff] [blame]1894<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]1895 specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1896 which indicates that the function takes a variable number of arguments.
1897 Variable argument functions can access their arguments with
1898 the <a href="#int_varargs">variable argument handling intrinsic</a>
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]1899 functions. '<tt>&lt;returntype&gt;</tt>' is any type except
Nick Lewycky93e06a52009-09-27 23:27:42 +0000[diff] [blame]1900 <a href="#t_label">label</a>.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]1901
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1902<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1903<table class="layout">
1904 <tr class="layout">
Reid Spencer58c08712006-12-31 07:18:34 +0000[diff] [blame]1905 <td class="left"><tt>i32 (i32)</tt></td>
1906 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1907 </td>
Reid Spencer58c08712006-12-31 07:18:34 +0000[diff] [blame]1908 </tr><tr class="layout">
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]1909 <td class="left"><tt>float&nbsp;(i16,&nbsp;i32&nbsp;*)&nbsp;*
Reid Spencer655dcc62006-12-31 07:20:23 +0000[diff] [blame]1910 </tt></td>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1911 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]1912 an <tt>i16</tt> and a <a href="#t_pointer">pointer</a> to <tt>i32</tt>,
1913 returning <tt>float</tt>.
Reid Spencer58c08712006-12-31 07:18:34 +0000[diff] [blame]1914 </td>
1915 </tr><tr class="layout">
1916 <td class="left"><tt>i32 (i8*, ...)</tt></td>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1917 <td class="left">A vararg function that takes at least one
1918 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1919 which returns an integer. This is the signature for <tt>printf</tt> in
Reid Spencer58c08712006-12-31 07:18:34 +0000[diff] [blame]1920 LLVM.
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1921 </td>
Devang Patele3dfc1c2008-03-24 05:35:41 +0000[diff] [blame]1922 </tr><tr class="layout">
1923 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Nick Lewycky14d1ccc2009-09-27 07:55:32 +0000[diff] [blame]1924 <td class="left">A function taking an <tt>i32</tt>, returning a
1925 <a href="#t_struct">structure</a> containing two <tt>i32</tt> values
Devang Patele3dfc1c2008-03-24 05:35:41 +0000[diff] [blame]1926 </td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1927 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1928</table>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]1929
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1930</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1931
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1932<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1933<h4>
1934 <a name="t_struct">Structure Type</a>
1935</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1936
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1937<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1938
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1939<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1940<p>The structure type is used to represent a collection of data members together
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]1941 in memory. The elements of a structure may be any type that has a size.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1942
Jeffrey Yasskinf991bbb2010-01-11 19:19:26 +0000[diff] [blame]1943<p>Structures in memory are accessed using '<tt><a href="#i_load">load</a></tt>'
1944 and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field
1945 with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
1946 Structures in registers are accessed using the
1947 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' and
1948 '<tt><a href="#i_insertvalue">insertvalue</a></tt>' instructions.</p>
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]1949
1950<p>Structures may optionally be "packed" structures, which indicate that the
1951 alignment of the struct is one byte, and that there is no padding between
1952 the elements. In non-packed structs, padding between field types is defined
1953 by the target data string to match the underlying processor.</p>
1954
1955<p>Structures can either be "anonymous" or "named". An anonymous structure is
1956 defined inline with other types (e.g. <tt>{i32, i32}*</tt>) and a named types
1957 are always defined at the top level with a name. Anonmyous types are uniqued
1958 by their contents and can never be recursive since there is no way to write
1959 one. Named types can be recursive.
1960</p>
1961
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1962<h5>Syntax:</h5>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1963<pre>
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]1964 %T1 = type { &lt;type list&gt; } <i>; Named normal struct type</i>
1965 %T2 = type &lt;{ &lt;type list&gt; }&gt; <i>; Named packed struct type</i>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1966</pre>
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]1967
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1968<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1969<table class="layout">
1970 <tr class="layout">
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]1971 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1972 <td class="left">A triple of three <tt>i32</tt> values</td>
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]1973 </tr>
1974 <tr class="layout">
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]1975 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1976 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1977 second element is a <a href="#t_pointer">pointer</a> to a
1978 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1979 an <tt>i32</tt>.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1980 </tr>
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]1981 <tr class="layout">
1982 <td class="left"><tt>&lt;{ i8, i32 }&gt;</tt></td>
1983 <td class="left">A packed struct known to be 5 bytes in size.</td>
1984 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1985</table>
Dan Gohman142ccc02009-01-24 15:58:40 +0000[diff] [blame]1986
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1987</div>
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]1988
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1989<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1990<h4>
Chris Lattner2a843822011-07-23 19:59:08 +0000[diff] [blame]1991 <a name="t_opaque">Opaque Structure Types</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1992</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1993
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]1994<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1995
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]1996<h5>Overview:</h5>
Chris Lattner2a843822011-07-23 19:59:08 +0000[diff] [blame]1997<p>Opaque structure types are used to represent named structure types that do
1998 not have a body specified. This corresponds (for example) to the C notion of
1999 a forward declared structure.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2000
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]2001<h5>Syntax:</h5>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]2002<pre>
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2003 %X = type opaque
2004 %52 = type opaque
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]2005</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2006
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]2007<h5>Examples:</h5>
2008<table class="layout">
2009 <tr class="layout">
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2010 <td class="left"><tt>opaque</tt></td>
2011 <td class="left">An opaque type.</td>
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]2012 </tr>
2013</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2014
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]2015</div>
2016
Chris Lattnerb1ed91f2011-07-09 17:41:24 +0000[diff] [blame]2017
2018
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]2019<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2020<h4>
2021 <a name="t_pointer">Pointer Type</a>
2022</h4>
Chris Lattner4a67c912009-02-08 19:53:29 +0000[diff] [blame]2023
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2024<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2025
2026<h5>Overview:</h5>
Dan Gohman88481112010-02-25 16:50:07 +0000[diff] [blame]2027<p>The pointer type is used to specify memory locations.
2028 Pointers are commonly used to reference objects in memory.</p>
2029
2030<p>Pointer types may have an optional address space attribute defining the
2031 numbered address space where the pointed-to object resides. The default
2032 address space is number zero. The semantics of non-zero address
2033 spaces are target-specific.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2034
2035<p>Note that LLVM does not permit pointers to void (<tt>void*</tt>) nor does it
2036 permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</tt> instead.</p>
Chris Lattner4a67c912009-02-08 19:53:29 +0000[diff] [blame]2037
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]2038<h5>Syntax:</h5>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]2039<pre>
2040 &lt;type&gt; *
2041</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2042
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]2043<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2044<table class="layout">
2045 <tr class="layout">
Dan Gohman623806e2009-01-04 23:44:43 +0000[diff] [blame]2046 <td class="left"><tt>[4 x i32]*</tt></td>
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]2047 <td class="left">A <a href="#t_pointer">pointer</a> to <a
2048 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
2049 </tr>
2050 <tr class="layout">
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]2051 <td class="left"><tt>i32 (i32*) *</tt></td>
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]2052 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]2053 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]2054 <tt>i32</tt>.</td>
2055 </tr>
2056 <tr class="layout">
2057 <td class="left"><tt>i32 addrspace(5)*</tt></td>
2058 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
2059 that resides in address space #5.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2060 </tr>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2061</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2062
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2063</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2064
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]2065<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2066<h4>
2067 <a name="t_vector">Vector Type</a>
2068</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2069
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2070<div>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]2071
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]2072<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2073<p>A vector type is a simple derived type that represents a vector of elements.
2074 Vector types are used when multiple primitive data are operated in parallel
2075 using a single instruction (SIMD). A vector type requires a size (number of
Duncan Sands31c0e0e2009-11-27 13:38:03 +0000[diff] [blame]2076 elements) and an underlying primitive data type. Vector types are considered
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2077 <a href="#t_firstclass">first class</a>.</p>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]2078
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]2079<h5>Syntax:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]2080<pre>
2081 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
2082</pre>
2083
Chris Lattnerf11031a2010-10-10 18:20:35 +0000[diff] [blame]2084<p>The number of elements is a constant integer value larger than 0; elementtype
2085 may be any integer or floating point type. Vectors of size zero are not
2086 allowed, and pointers are not allowed as the element type.</p>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]2087
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]2088<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2089<table class="layout">
2090 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]2091 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
2092 <td class="left">Vector of 4 32-bit integer values.</td>
2093 </tr>
2094 <tr class="layout">
2095 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
2096 <td class="left">Vector of 8 32-bit floating-point values.</td>
2097 </tr>
2098 <tr class="layout">
2099 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
2100 <td class="left">Vector of 2 64-bit integer values.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]2101 </tr>
2102</table>
Dan Gohman142ccc02009-01-24 15:58:40 +0000[diff] [blame]2103
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2104</div>
2105
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2106<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2107<h2><a name="constants">Constants</a></h2>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2108<!-- *********************************************************************** -->
2109
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2110<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2111
2112<p>LLVM has several different basic types of constants. This section describes
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2113 them all and their syntax.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2114
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2115<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2116<h3>
2117 <a name="simpleconstants">Simple Constants</a>
2118</h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2119
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2120<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2121
2122<dl>
2123 <dt><b>Boolean constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2124 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]2125 constants of the <tt><a href="#t_integer">i1</a></tt> type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2126
2127 <dt><b>Integer constants</b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2128 <dd>Standard integers (such as '4') are constants of
2129 the <a href="#t_integer">integer</a> type. Negative numbers may be used
2130 with integer types.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2131
2132 <dt><b>Floating point constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2133 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2134 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
2135 notation (see below). The assembler requires the exact decimal value of a
2136 floating-point constant. For example, the assembler accepts 1.25 but
2137 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
2138 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2139
2140 <dt><b>Null pointer constants</b></dt>
John Criswelldfe6a862004-12-10 15:51:16 +0000[diff] [blame]2141 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2142 and must be of <a href="#t_pointer">pointer type</a>.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2143</dl>
2144
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2145<p>The one non-intuitive notation for constants is the hexadecimal form of
2146 floating point constants. For example, the form '<tt>double
2147 0x432ff973cafa8000</tt>' is equivalent to (but harder to read than)
2148 '<tt>double 4.5e+15</tt>'. The only time hexadecimal floating point
2149 constants are required (and the only time that they are generated by the
2150 disassembler) is when a floating point constant must be emitted but it cannot
2151 be represented as a decimal floating point number in a reasonable number of
2152 digits. For example, NaN's, infinities, and other special values are
2153 represented in their IEEE hexadecimal format so that assembly and disassembly
2154 do not cause any bits to change in the constants.</p>
2155
Dale Johannesencd4a3012009-02-11 22:14:51 +0000[diff] [blame]2156<p>When using the hexadecimal form, constants of types float and double are
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2157 represented using the 16-digit form shown above (which matches the IEEE754
2158 representation for double); float values must, however, be exactly
2159 representable as IEE754 single precision. Hexadecimal format is always used
2160 for long double, and there are three forms of long double. The 80-bit format
2161 used by x86 is represented as <tt>0xK</tt> followed by 20 hexadecimal digits.
2162 The 128-bit format used by PowerPC (two adjacent doubles) is represented
2163 by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit format
2164 is represented by <tt>0xL</tt> followed by 32 hexadecimal digits; no
2165 currently supported target uses this format. Long doubles will only work if
2166 they match the long double format on your target. All hexadecimal formats
2167 are big-endian (sign bit at the left).</p>
2168
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]2169<p>There are no constants of type x86mmx.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2170</div>
2171
2172<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2173<h3>
Bill Wendling972b7202009-07-20 02:32:41 +0000[diff] [blame]2174<a name="aggregateconstants"></a> <!-- old anchor -->
2175<a name="complexconstants">Complex Constants</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2176</h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2177
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2178<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2179
Chris Lattner361bfcd2009-02-28 18:32:25 +0000[diff] [blame]2180<p>Complex constants are a (potentially recursive) combination of simple
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2181 constants and smaller complex constants.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2182
2183<dl>
2184 <dt><b>Structure constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2185 <dd>Structure constants are represented with notation similar to structure
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2186 type definitions (a comma separated list of elements, surrounded by braces
2187 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
2188 where "<tt>@G</tt>" is declared as "<tt>@G = external global i32</tt>".
2189 Structure constants must have <a href="#t_struct">structure type</a>, and
2190 the number and types of elements must match those specified by the
2191 type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2192
2193 <dt><b>Array constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2194 <dd>Array constants are represented with notation similar to array type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2195 definitions (a comma separated list of elements, surrounded by square
2196 brackets (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74
2197 ]</tt>". Array constants must have <a href="#t_array">array type</a>, and
2198 the number and types of elements must match those specified by the
2199 type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2200
Reid Spencer404a3252007-02-15 03:07:05 +0000[diff] [blame]2201 <dt><b>Vector constants</b></dt>
Reid Spencer404a3252007-02-15 03:07:05 +0000[diff] [blame]2202 <dd>Vector constants are represented with notation similar to vector type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2203 definitions (a comma separated list of elements, surrounded by
2204 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32
2205 42, i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must
2206 have <a href="#t_vector">vector type</a>, and the number and types of
2207 elements must match those specified by the type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2208
2209 <dt><b>Zero initialization</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2210 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]2211 value to zero of <em>any</em> type, including scalar and
2212 <a href="#t_aggregate">aggregate</a> types.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2213 This is often used to avoid having to print large zero initializers
2214 (e.g. for large arrays) and is always exactly equivalent to using explicit
2215 zero initializers.</dd>
Nick Lewycky49f89192009-04-04 07:22:01 +0000[diff] [blame]2216
2217 <dt><b>Metadata node</b></dt>
Nick Lewycky8e2c4f42009-05-30 16:08:30 +0000[diff] [blame]2218 <dd>A metadata node is a structure-like constant with
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2219 <a href="#t_metadata">metadata type</a>. For example: "<tt>metadata !{
2220 i32 0, metadata !"test" }</tt>". Unlike other constants that are meant to
2221 be interpreted as part of the instruction stream, metadata is a place to
2222 attach additional information such as debug info.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2223</dl>
2224
2225</div>
2226
2227<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2228<h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2229 <a name="globalconstants">Global Variable and Function Addresses</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2230</h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2231
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2232<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2233
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2234<p>The addresses of <a href="#globalvars">global variables</a>
2235 and <a href="#functionstructure">functions</a> are always implicitly valid
2236 (link-time) constants. These constants are explicitly referenced when
2237 the <a href="#identifiers">identifier for the global</a> is used and always
2238 have <a href="#t_pointer">pointer</a> type. For example, the following is a
2239 legal LLVM file:</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2240
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2241<pre class="doc_code">
Chris Lattner00538a12007-06-06 18:28:13 +0000[diff] [blame]2242@X = global i32 17
2243@Y = global i32 42
2244@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2245</pre>
2246
2247</div>
2248
2249<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2250<h3>
2251 <a name="undefvalues">Undefined Values</a>
2252</h3>
2253
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2254<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2255
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2256<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]2257 indicates that the user of the value may receive an unspecified bit-pattern.
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2258 Undefined values may be of any type (other than '<tt>label</tt>'
2259 or '<tt>void</tt>') and be used anywhere a constant is permitted.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2260
Chris Lattner92ada5d2009-09-11 01:49:31 +0000[diff] [blame]2261<p>Undefined values are useful because they indicate to the compiler that the
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2262 program is well defined no matter what value is used. This gives the
2263 compiler more freedom to optimize. Here are some examples of (potentially
2264 surprising) transformations that are valid (in pseudo IR):</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2265
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2266
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2267<pre class="doc_code">
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2268 %A = add %X, undef
2269 %B = sub %X, undef
2270 %C = xor %X, undef
2271Safe:
2272 %A = undef
2273 %B = undef
2274 %C = undef
2275</pre>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2276
2277<p>This is safe because all of the output bits are affected by the undef bits.
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2278 Any output bit can have a zero or one depending on the input bits.</p>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2279
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2280<pre class="doc_code">
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2281 %A = or %X, undef
2282 %B = and %X, undef
2283Safe:
2284 %A = -1
2285 %B = 0
2286Unsafe:
2287 %A = undef
2288 %B = undef
2289</pre>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2290
2291<p>These logical operations have bits that are not always affected by the input.
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2292 For example, if <tt>%X</tt> has a zero bit, then the output of the
2293 '<tt>and</tt>' operation will always be a zero for that bit, no matter what
2294 the corresponding bit from the '<tt>undef</tt>' is. As such, it is unsafe to
2295 optimize or assume that the result of the '<tt>and</tt>' is '<tt>undef</tt>'.
2296 However, it is safe to assume that all bits of the '<tt>undef</tt>' could be
2297 0, and optimize the '<tt>and</tt>' to 0. Likewise, it is safe to assume that
2298 all the bits of the '<tt>undef</tt>' operand to the '<tt>or</tt>' could be
2299 set, allowing the '<tt>or</tt>' to be folded to -1.</p>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2300
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2301<pre class="doc_code">
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2302 %A = select undef, %X, %Y
2303 %B = select undef, 42, %Y
2304 %C = select %X, %Y, undef
2305Safe:
2306 %A = %X (or %Y)
2307 %B = 42 (or %Y)
2308 %C = %Y
2309Unsafe:
2310 %A = undef
2311 %B = undef
2312 %C = undef
2313</pre>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2314
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2315<p>This set of examples shows that undefined '<tt>select</tt>' (and conditional
2316 branch) conditions can go <em>either way</em>, but they have to come from one
2317 of the two operands. In the <tt>%A</tt> example, if <tt>%X</tt> and
2318 <tt>%Y</tt> were both known to have a clear low bit, then <tt>%A</tt> would
2319 have to have a cleared low bit. However, in the <tt>%C</tt> example, the
2320 optimizer is allowed to assume that the '<tt>undef</tt>' operand could be the
2321 same as <tt>%Y</tt>, allowing the whole '<tt>select</tt>' to be
2322 eliminated.</p>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2323
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2324<pre class="doc_code">
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2325 %A = xor undef, undef
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2326
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2327 %B = undef
2328 %C = xor %B, %B
2329
2330 %D = undef
2331 %E = icmp lt %D, 4
2332 %F = icmp gte %D, 4
2333
2334Safe:
2335 %A = undef
2336 %B = undef
2337 %C = undef
2338 %D = undef
2339 %E = undef
2340 %F = undef
2341</pre>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2342
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2343<p>This example points out that two '<tt>undef</tt>' operands are not
2344 necessarily the same. This can be surprising to people (and also matches C
2345 semantics) where they assume that "<tt>X^X</tt>" is always zero, even
2346 if <tt>X</tt> is undefined. This isn't true for a number of reasons, but the
2347 short answer is that an '<tt>undef</tt>' "variable" can arbitrarily change
2348 its value over its "live range". This is true because the variable doesn't
2349 actually <em>have a live range</em>. Instead, the value is logically read
2350 from arbitrary registers that happen to be around when needed, so the value
2351 is not necessarily consistent over time. In fact, <tt>%A</tt> and <tt>%C</tt>
2352 need to have the same semantics or the core LLVM "replace all uses with"
2353 concept would not hold.</p>
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2354
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2355<pre class="doc_code">
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2356 %A = fdiv undef, %X
2357 %B = fdiv %X, undef
2358Safe:
2359 %A = undef
2360b: unreachable
2361</pre>
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2362
2363<p>These examples show the crucial difference between an <em>undefined
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2364 value</em> and <em>undefined behavior</em>. An undefined value (like
2365 '<tt>undef</tt>') is allowed to have an arbitrary bit-pattern. This means that
2366 the <tt>%A</tt> operation can be constant folded to '<tt>undef</tt>', because
2367 the '<tt>undef</tt>' could be an SNaN, and <tt>fdiv</tt> is not (currently)
2368 defined on SNaN's. However, in the second example, we can make a more
2369 aggressive assumption: because the <tt>undef</tt> is allowed to be an
2370 arbitrary value, we are allowed to assume that it could be zero. Since a
2371 divide by zero has <em>undefined behavior</em>, we are allowed to assume that
2372 the operation does not execute at all. This allows us to delete the divide and
2373 all code after it. Because the undefined operation "can't happen", the
2374 optimizer can assume that it occurs in dead code.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2375
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2376<pre class="doc_code">
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2377a: store undef -> %X
2378b: store %X -> undef
2379Safe:
2380a: &lt;deleted&gt;
2381b: unreachable
2382</pre>
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2383
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2384<p>These examples reiterate the <tt>fdiv</tt> example: a store <em>of</em> an
2385 undefined value can be assumed to not have any effect; we can assume that the
2386 value is overwritten with bits that happen to match what was already there.
2387 However, a store <em>to</em> an undefined location could clobber arbitrary
2388 memory, therefore, it has undefined behavior.</p>
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2389
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2390</div>
2391
2392<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2393<h3>
2394 <a name="trapvalues">Trap Values</a>
2395</h3>
2396
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2397<div>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2398
Dan Gohmanb2a709b2010-04-26 20:21:21 +0000[diff] [blame]2399<p>Trap values are similar to <a href="#undefvalues">undef values</a>, however
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2400 instead of representing an unspecified bit pattern, they represent the
2401 fact that an instruction or constant expression which cannot evoke side
2402 effects has nevertheless detected a condition which results in undefined
Dan Gohmanb2a709b2010-04-26 20:21:21 +0000[diff] [blame]2403 behavior.</p>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2404
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2405<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]2406 only exist when produced by operations such as
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2407 <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]2408
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2409<p>Trap value behavior is defined in terms of value <i>dependence</i>:</p>
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2410
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2411<ul>
2412<li>Values other than <a href="#i_phi"><tt>phi</tt></a> nodes depend on
2413 their operands.</li>
2414
2415<li><a href="#i_phi"><tt>Phi</tt></a> nodes depend on the operand corresponding
2416 to their dynamic predecessor basic block.</li>
2417
2418<li>Function arguments depend on the corresponding actual argument values in
2419 the dynamic callers of their functions.</li>
2420
2421<li><a href="#i_call"><tt>Call</tt></a> instructions depend on the
2422 <a href="#i_ret"><tt>ret</tt></a> instructions that dynamically transfer
2423 control back to them.</li>
2424
Dan Gohman7292a752010-05-03 14:55:22 +0000[diff] [blame]2425<li><a href="#i_invoke"><tt>Invoke</tt></a> instructions depend on the
2426 <a href="#i_ret"><tt>ret</tt></a>, <a href="#i_unwind"><tt>unwind</tt></a>,
2427 or exception-throwing call instructions that dynamically transfer control
2428 back to them.</li>
2429
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2430<li>Non-volatile loads and stores depend on the most recent stores to all of the
2431 referenced memory addresses, following the order in the IR
2432 (including loads and stores implied by intrinsics such as
2433 <a href="#int_memcpy"><tt>@llvm.memcpy</tt></a>.)</li>
2434
Dan Gohman3513ea52010-05-03 14:59:34 +0000[diff] [blame]2435<!-- TODO: In the case of multiple threads, this only applies if the store
2436 "happens-before" the load or store. -->
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2437
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2438<!-- TODO: floating-point exception state -->
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2439
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2440<li>An instruction with externally visible side effects depends on the most
2441 recent preceding instruction with externally visible side effects, following
Dan Gohman6c858db2010-07-06 15:26:33 +0000[diff] [blame]2442 the order in the IR. (This includes
2443 <a href="#volatile">volatile operations</a>.)</li>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2444
Dan Gohman7292a752010-05-03 14:55:22 +0000[diff] [blame]2445<li>An instruction <i>control-depends</i> on a
2446 <a href="#terminators">terminator instruction</a>
2447 if the terminator instruction has multiple successors and the instruction
2448 is always executed when control transfers to one of the successors, and
Chris Lattner0ab5e2c2011-04-15 05:18:47 +0000[diff] [blame]2449 may not be executed when control is transferred to another.</li>
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2450
Dan Gohmanc8454ee2011-04-12 23:05:59 +0000[diff] [blame]2451<li>Additionally, an instruction also <i>control-depends</i> on a terminator
2452 instruction if the set of instructions it otherwise depends on would be
Chris Lattner0ab5e2c2011-04-15 05:18:47 +0000[diff] [blame]2453 different if the terminator had transferred control to a different
Dan Gohmanc8454ee2011-04-12 23:05:59 +0000[diff] [blame]2454 successor.</li>
2455
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2456<li>Dependence is transitive.</li>
2457
2458</ul>
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2459
2460<p>Whenever a trap value is generated, all values which depend on it evaluate
2461 to trap. If they have side effects, the evoke their side effects as if each
2462 operand with a trap value were undef. If they have externally-visible side
2463 effects, the behavior is undefined.</p>
2464
2465<p>Here are some examples:</p>
Dan Gohman48a25882010-04-26 20:54:53 +0000[diff] [blame]2466
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2467<pre class="doc_code">
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2468entry:
2469 %trap = sub nuw i32 0, 1 ; Results in a trap value.
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2470 %still_trap = and i32 %trap, 0 ; Whereas (and i32 undef, 0) would return 0.
2471 %trap_yet_again = getelementptr i32* @h, i32 %still_trap
2472 store i32 0, i32* %trap_yet_again ; undefined behavior
2473
2474 store i32 %trap, i32* @g ; Trap value conceptually stored to memory.
2475 %trap2 = load i32* @g ; Returns a trap value, not just undef.
2476
2477 volatile store i32 %trap, i32* @g ; External observation; undefined behavior.
2478
2479 %narrowaddr = bitcast i32* @g to i16*
2480 %wideaddr = bitcast i32* @g to i64*
Nick Lewycky9c876bf2011-05-16 19:29:30 +0000[diff] [blame]2481 %trap3 = load i16* %narrowaddr ; Returns a trap value.
2482 %trap4 = load i64* %wideaddr ; Returns a trap value.
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2483
Nick Lewycky9c876bf2011-05-16 19:29:30 +0000[diff] [blame]2484 %cmp = icmp slt i32 %trap, 0 ; Returns a trap value.
2485 br i1 %cmp, label %true, label %end ; Branch to either destination.
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2486
2487true:
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2488 volatile store i32 0, i32* @g ; This is control-dependent on %cmp, so
2489 ; it has undefined behavior.
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2490 br label %end
2491
2492end:
2493 %p = phi i32 [ 0, %entry ], [ 1, %true ]
2494 ; Both edges into this PHI are
2495 ; control-dependent on %cmp, so this
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2496 ; always results in a trap value.
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2497
Dan Gohmanc8454ee2011-04-12 23:05:59 +0000[diff] [blame]2498 volatile store i32 0, i32* @g ; This would depend on the store in %true
2499 ; if %cmp is true, or the store in %entry
2500 ; otherwise, so this is undefined behavior.
2501
Nick Lewycky9c876bf2011-05-16 19:29:30 +0000[diff] [blame]2502 br i1 %cmp, label %second_true, label %second_end
Dan Gohmanc8454ee2011-04-12 23:05:59 +0000[diff] [blame]2503 ; The same branch again, but this time the
2504 ; true block doesn't have side effects.
2505
2506second_true:
2507 ; No side effects!
Nick Lewycky9c876bf2011-05-16 19:29:30 +0000[diff] [blame]2508 ret void
Dan Gohmanc8454ee2011-04-12 23:05:59 +0000[diff] [blame]2509
2510second_end:
2511 volatile store i32 0, i32* @g ; This time, the instruction always depends
2512 ; on the store in %end. Also, it is
2513 ; control-equivalent to %end, so this is
Nick Lewycky9c876bf2011-05-16 19:29:30 +0000[diff] [blame]2514 ; well-defined (again, ignoring earlier
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2515 ; undefined behavior in this example).
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2516</pre>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2517
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2518</div>
2519
2520<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2521<h3>
2522 <a name="blockaddress">Addresses of Basic Blocks</a>
2523</h3>
2524
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2525<div>
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2526
Chris Lattneraa99c942009-11-01 01:27:45 +0000[diff] [blame]2527<p><b><tt>blockaddress(@function, %block)</tt></b></p>
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2528
2529<p>The '<tt>blockaddress</tt>' constant computes the address of the specified
Chris Lattner5c5f0ac2009-10-27 21:49:40 +0000[diff] [blame]2530 basic block in the specified function, and always has an i8* type. Taking
Chris Lattneraa99c942009-11-01 01:27:45 +0000[diff] [blame]2531 the address of the entry block is illegal.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2532
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2533<p>This value only has defined behavior when used as an operand to the
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2534 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>' instruction, or for
2535 comparisons against null. Pointer equality tests between labels addresses
2536 results in undefined behavior &mdash; though, again, comparison against null
2537 is ok, and no label is equal to the null pointer. This may be passed around
2538 as an opaque pointer sized value as long as the bits are not inspected. This
2539 allows <tt>ptrtoint</tt> and arithmetic to be performed on these values so
2540 long as the original value is reconstituted before the <tt>indirectbr</tt>
2541 instruction.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2542
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2543<p>Finally, some targets may provide defined semantics when using the value as
2544 the operand to an inline assembly, but that is target specific.</p>
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2545
2546</div>
2547
2548
2549<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2550<h3>
2551 <a name="constantexprs">Constant Expressions</a>
2552</h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2553
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2554<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2555
2556<p>Constant expressions are used to allow expressions involving other constants
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2557 to be used as constants. Constant expressions may be of
2558 any <a href="#t_firstclass">first class</a> type and may involve any LLVM
2559 operation that does not have side effects (e.g. load and call are not
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2560 supported). The following is the syntax for constant expressions:</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2561
2562<dl>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2563 <dt><b><tt>trunc (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2564 <dd>Truncate a constant to another type. The bit size of CST must be larger
2565 than the bit size of TYPE. Both types must be integers.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2566
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2567 <dt><b><tt>zext (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2568 <dd>Zero extend a constant to another type. The bit size of CST must be
Duncan Sandsa522e562010-07-13 12:06:14 +0000[diff] [blame]2569 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2570
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2571 <dt><b><tt>sext (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2572 <dd>Sign extend a constant to another type. The bit size of CST must be
Duncan Sandsa522e562010-07-13 12:06:14 +0000[diff] [blame]2573 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2574
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2575 <dt><b><tt>fptrunc (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2576 <dd>Truncate a floating point constant to another floating point type. The
2577 size of CST must be larger than the size of TYPE. Both types must be
2578 floating point.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2579
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2580 <dt><b><tt>fpext (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2581 <dd>Floating point extend a constant to another type. The size of CST must be
2582 smaller or equal to the size of TYPE. Both types must be floating
2583 point.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2584
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2585 <dt><b><tt>fptoui (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2586 <dd>Convert a floating point constant to the corresponding unsigned integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2587 constant. TYPE must be a scalar or vector integer type. CST must be of
2588 scalar or vector floating point type. Both CST and TYPE must be scalars,
2589 or vectors of the same number of elements. If the value won't fit in the
2590 integer type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2591
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2592 <dt><b><tt>fptosi (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2593 <dd>Convert a floating point constant to the corresponding signed integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2594 constant. TYPE must be a scalar or vector integer type. CST must be of
2595 scalar or vector floating point type. Both CST and TYPE must be scalars,
2596 or vectors of the same number of elements. If the value won't fit in the
2597 integer type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2598
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2599 <dt><b><tt>uitofp (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2600 <dd>Convert an unsigned integer constant to the corresponding floating point
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2601 constant. TYPE must be a scalar or vector floating point type. CST must be
2602 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2603 vectors of the same number of elements. If the value won't fit in the
2604 floating point type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2605
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2606 <dt><b><tt>sitofp (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2607 <dd>Convert a signed integer constant to the corresponding floating point
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2608 constant. TYPE must be a scalar or vector floating point type. CST must be
2609 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2610 vectors of the same number of elements. If the value won't fit in the
2611 floating point type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2612
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2613 <dt><b><tt>ptrtoint (CST to TYPE)</tt></b></dt>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]2614 <dd>Convert a pointer typed constant to the corresponding integer constant
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2615 <tt>TYPE</tt> must be an integer type. <tt>CST</tt> must be of pointer
2616 type. The <tt>CST</tt> value is zero extended, truncated, or unchanged to
2617 make it fit in <tt>TYPE</tt>.</dd>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]2618
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2619 <dt><b><tt>inttoptr (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2620 <dd>Convert a integer constant to a pointer constant. TYPE must be a pointer
2621 type. CST must be of integer type. The CST value is zero extended,
2622 truncated, or unchanged to make it fit in a pointer size. This one is
2623 <i>really</i> dangerous!</dd>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]2624
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2625 <dt><b><tt>bitcast (CST to TYPE)</tt></b></dt>
Chris Lattner789dee32009-02-28 18:27:03 +0000[diff] [blame]2626 <dd>Convert a constant, CST, to another TYPE. The constraints of the operands
2627 are the same as those for the <a href="#i_bitcast">bitcast
2628 instruction</a>.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2629
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2630 <dt><b><tt>getelementptr (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
2631 <dt><b><tt>getelementptr inbounds (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2632 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2633 constants. As with the <a href="#i_getelementptr">getelementptr</a>
2634 instruction, the index list may have zero or more indexes, which are
2635 required to make sense for the type of "CSTPTR".</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2636
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2637 <dt><b><tt>select (COND, VAL1, VAL2)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2638 <dd>Perform the <a href="#i_select">select operation</a> on constants.</dd>
Reid Spencer9965ee72006-12-04 19:23:19 +0000[diff] [blame]2639
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2640 <dt><b><tt>icmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer9965ee72006-12-04 19:23:19 +0000[diff] [blame]2641 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
2642
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2643 <dt><b><tt>fcmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer9965ee72006-12-04 19:23:19 +0000[diff] [blame]2644 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
Robert Bocchino7e97a6d2006-01-10 19:31:34 +0000[diff] [blame]2645
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2646 <dt><b><tt>extractelement (VAL, IDX)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2647 <dd>Perform the <a href="#i_extractelement">extractelement operation</a> on
2648 constants.</dd>
Robert Bocchino7e97a6d2006-01-10 19:31:34 +0000[diff] [blame]2649
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2650 <dt><b><tt>insertelement (VAL, ELT, IDX)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2651 <dd>Perform the <a href="#i_insertelement">insertelement operation</a> on
2652 constants.</dd>
Chris Lattner016a0e52006-04-08 00:13:41 +0000[diff] [blame]2653
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2654 <dt><b><tt>shufflevector (VEC1, VEC2, IDXMASK)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2655 <dd>Perform the <a href="#i_shufflevector">shufflevector operation</a> on
2656 constants.</dd>
Chris Lattner016a0e52006-04-08 00:13:41 +0000[diff] [blame]2657
Nick Lewycky9ab9a7f2010-05-29 06:44:15 +0000[diff] [blame]2658 <dt><b><tt>extractvalue (VAL, IDX0, IDX1, ...)</tt></b></dt>
2659 <dd>Perform the <a href="#i_extractvalue">extractvalue operation</a> on
2660 constants. The index list is interpreted in a similar manner as indices in
2661 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2662 index value must be specified.</dd>
2663
2664 <dt><b><tt>insertvalue (VAL, ELT, IDX0, IDX1, ...)</tt></b></dt>
2665 <dd>Perform the <a href="#i_insertvalue">insertvalue operation</a> on
2666 constants. The index list is interpreted in a similar manner as indices in
2667 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2668 index value must be specified.</dd>
2669
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2670 <dt><b><tt>OPCODE (LHS, RHS)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2671 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
2672 be any of the <a href="#binaryops">binary</a>
2673 or <a href="#bitwiseops">bitwise binary</a> operations. The constraints
2674 on operands are the same as those for the corresponding instruction
2675 (e.g. no bitwise operations on floating point values are allowed).</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2676</dl>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2677
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2678</div>
Chris Lattnerb1652612004-03-08 16:49:10 +0000[diff] [blame]2679
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2680</div>
2681
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2682<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2683<h2><a name="othervalues">Other Values</a></h2>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2684<!-- *********************************************************************** -->
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2685<div>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2686<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2687<h3>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2688<a name="inlineasm">Inline Assembler Expressions</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2689</h3>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2690
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2691<div>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2692
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2693<p>LLVM supports inline assembler expressions (as opposed
2694 to <a href="#moduleasm"> Module-Level Inline Assembly</a>) through the use of
2695 a special value. This value represents the inline assembler as a string
2696 (containing the instructions to emit), a list of operand constraints (stored
Dale Johannesen63c94fe2009-10-13 21:56:55 +0000[diff] [blame]2697 as a string), a flag that indicates whether or not the inline asm
Dale Johannesen1cfb9582009-10-21 23:28:00 +0000[diff] [blame]2698 expression has side effects, and a flag indicating whether the function
2699 containing the asm needs to align its stack conservatively. An example
2700 inline assembler expression is:</p>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2701
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2702<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2703i32 (i32) asm "bswap $0", "=r,r"
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2704</pre>
2705
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2706<p>Inline assembler expressions may <b>only</b> be used as the callee operand of
2707 a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we
2708 have:</p>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2709
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2710<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2711%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]2712</pre>
2713
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2714<p>Inline asms with side effects not visible in the constraint list must be
2715 marked as having side effects. This is done through the use of the
2716 '<tt>sideeffect</tt>' keyword, like so:</p>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2717
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2718<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2719call void asm sideeffect "eieio", ""()
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2720</pre>
2721
Dale Johannesen1cfb9582009-10-21 23:28:00 +0000[diff] [blame]2722<p>In some cases inline asms will contain code that will not work unless the
2723 stack is aligned in some way, such as calls or SSE instructions on x86,
2724 yet will not contain code that does that alignment within the asm.
2725 The compiler should make conservative assumptions about what the asm might
2726 contain and should generate its usual stack alignment code in the prologue
2727 if the '<tt>alignstack</tt>' keyword is present:</p>
Dale Johannesen63c94fe2009-10-13 21:56:55 +0000[diff] [blame]2728
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2729<pre class="doc_code">
Dale Johannesen1cfb9582009-10-21 23:28:00 +0000[diff] [blame]2730call void asm alignstack "eieio", ""()
Dale Johannesen63c94fe2009-10-13 21:56:55 +0000[diff] [blame]2731</pre>
Dale Johannesen63c94fe2009-10-13 21:56:55 +0000[diff] [blame]2732
2733<p>If both keywords appear the '<tt>sideeffect</tt>' keyword must come
2734 first.</p>
2735
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2736<p>TODO: The format of the asm and constraints string still need to be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2737 documented here. Constraints on what can be done (e.g. duplication, moving,
2738 etc need to be documented). This is probably best done by reference to
2739 another document that covers inline asm from a holistic perspective.</p>
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2740
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2741<h4>
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2742<a name="inlineasm_md">Inline Asm Metadata</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2743</h4>
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2744
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2745<div>
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2746
2747<p>The call instructions that wrap inline asm nodes may have a "!srcloc" MDNode
Chris Lattner79ffdc72010-11-17 08:20:42 +0000[diff] [blame]2748 attached to it that contains a list of constant integers. If present, the
2749 code generator will use the integer as the location cookie value when report
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2750 errors through the LLVMContext error reporting mechanisms. This allows a
Dan Gohman61110ae2010-04-28 00:36:01 +0000[diff] [blame]2751 front-end to correlate backend errors that occur with inline asm back to the
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2752 source code that produced it. For example:</p>
2753
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2754<pre class="doc_code">
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2755call void asm sideeffect "something bad", ""()<b>, !srcloc !42</b>
2756...
2757!42 = !{ i32 1234567 }
2758</pre>
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2759
2760<p>It is up to the front-end to make sense of the magic numbers it places in the
Chris Lattner79ffdc72010-11-17 08:20:42 +0000[diff] [blame]2761 IR. If the MDNode contains multiple constants, the code generator will use
2762 the one that corresponds to the line of the asm that the error occurs on.</p>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2763
2764</div>
2765
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2766</div>
2767
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]2768<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2769<h3>
2770 <a name="metadata">Metadata Nodes and Metadata Strings</a>
2771</h3>
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]2772
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2773<div>
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]2774
2775<p>LLVM IR allows metadata to be attached to instructions in the program that
2776 can convey extra information about the code to the optimizers and code
2777 generator. One example application of metadata is source-level debug
2778 information. There are two metadata primitives: strings and nodes. All
2779 metadata has the <tt>metadata</tt> type and is identified in syntax by a
2780 preceding exclamation point ('<tt>!</tt>').</p>
2781
2782<p>A metadata string is a string surrounded by double quotes. It can contain
2783 any character by escaping non-printable characters with "\xx" where "xx" is
2784 the two digit hex code. For example: "<tt>!"test\00"</tt>".</p>
2785
2786<p>Metadata nodes are represented with notation similar to structure constants
2787 (a comma separated list of elements, surrounded by braces and preceded by an
2788 exclamation point). For example: "<tt>!{ metadata !"test\00", i32
2789 10}</tt>". Metadata nodes can have any values as their operand.</p>
2790
2791<p>A <a href="#namedmetadatastructure">named metadata</a> is a collection of
2792 metadata nodes, which can be looked up in the module symbol table. For
2793 example: "<tt>!foo = metadata !{!4, !3}</tt>".
2794
Devang Patel9984bd62010-03-04 23:44:48 +0000[diff] [blame]2795<p>Metadata can be used as function arguments. Here <tt>llvm.dbg.value</tt>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2796 function is using two metadata arguments.</p>
Devang Patel9984bd62010-03-04 23:44:48 +0000[diff] [blame]2797
Bill Wendlingc0e10672011-03-02 02:17:11 +0000[diff] [blame]2798<div class="doc_code">
2799<pre>
2800call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2801</pre>
2802</div>
Devang Patel9984bd62010-03-04 23:44:48 +0000[diff] [blame]2803
2804<p>Metadata can be attached with an instruction. Here metadata <tt>!21</tt> is
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2805 attached with <tt>add</tt> instruction using <tt>!dbg</tt> identifier.</p>
Devang Patel9984bd62010-03-04 23:44:48 +0000[diff] [blame]2806
Bill Wendlingc0e10672011-03-02 02:17:11 +0000[diff] [blame]2807<div class="doc_code">
2808<pre>
2809%indvar.next = add i64 %indvar, 1, !dbg !21
2810</pre>
2811</div>
2812
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]2813</div>
2814
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2815</div>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2816
2817<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2818<h2>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2819 <a name="intrinsic_globals">Intrinsic Global Variables</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2820</h2>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2821<!-- *********************************************************************** -->
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2822<div>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2823<p>LLVM has a number of "magic" global variables that contain data that affect
2824code generation or other IR semantics. These are documented here. All globals
Chris Lattner58f9bb22009-07-20 06:14:25 +0000[diff] [blame]2825of this sort should have a section specified as "<tt>llvm.metadata</tt>". This
2826section and all globals that start with "<tt>llvm.</tt>" are reserved for use
2827by LLVM.</p>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2828
2829<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2830<h3>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2831<a name="intg_used">The '<tt>llvm.used</tt>' Global Variable</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2832</h3>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2833
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2834<div>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2835
2836<p>The <tt>@llvm.used</tt> global is an array with i8* element type which has <a
2837href="#linkage_appending">appending linkage</a>. This array contains a list of
2838pointers to global variables and functions which may optionally have a pointer
2839cast formed of bitcast or getelementptr. For example, a legal use of it is:</p>
2840
2841<pre>
2842 @X = global i8 4
2843 @Y = global i32 123
2844
2845 @llvm.used = appending global [2 x i8*] [
2846 i8* @X,
2847 i8* bitcast (i32* @Y to i8*)
2848 ], section "llvm.metadata"
2849</pre>
2850
2851<p>If a global variable appears in the <tt>@llvm.used</tt> list, then the
2852compiler, assembler, and linker are required to treat the symbol as if there is
2853a reference to the global that it cannot see. For example, if a variable has
2854internal linkage and no references other than that from the <tt>@llvm.used</tt>
2855list, it cannot be deleted. This is commonly used to represent references from
2856inline asms and other things the compiler cannot "see", and corresponds to
2857"attribute((used))" in GNU C.</p>
2858
2859<p>On some targets, the code generator must emit a directive to the assembler or
2860object file to prevent the assembler and linker from molesting the symbol.</p>
2861
2862</div>
2863
2864<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2865<h3>
2866 <a name="intg_compiler_used">
2867 The '<tt>llvm.compiler.used</tt>' Global Variable
2868 </a>
2869</h3>
Chris Lattner58f9bb22009-07-20 06:14:25 +0000[diff] [blame]2870
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2871<div>
Chris Lattner58f9bb22009-07-20 06:14:25 +0000[diff] [blame]2872
2873<p>The <tt>@llvm.compiler.used</tt> directive is the same as the
2874<tt>@llvm.used</tt> directive, except that it only prevents the compiler from
2875touching the symbol. On targets that support it, this allows an intelligent
2876linker to optimize references to the symbol without being impeded as it would be
2877by <tt>@llvm.used</tt>.</p>
2878
2879<p>This is a rare construct that should only be used in rare circumstances, and
2880should not be exposed to source languages.</p>
2881
2882</div>
2883
2884<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2885<h3>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2886<a name="intg_global_ctors">The '<tt>llvm.global_ctors</tt>' Global Variable</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2887</h3>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2888
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2889<div>
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]2890<pre>
2891%0 = type { i32, void ()* }
David Chisnallb492b812010-04-30 19:27:35 +0000[diff] [blame]2892@llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor }]
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]2893</pre>
2894<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.
2895</p>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2896
2897</div>
2898
2899<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2900<h3>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2901<a name="intg_global_dtors">The '<tt>llvm.global_dtors</tt>' Global Variable</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2902</h3>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2903
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2904<div>
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]2905<pre>
2906%0 = type { i32, void ()* }
David Chisnallb492b812010-04-30 19:27:35 +0000[diff] [blame]2907@llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor }]
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]2908</pre>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2909
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]2910<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.
2911</p>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2912
2913</div>
2914
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2915</div>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2916
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2917<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2918<h2><a name="instref">Instruction Reference</a></h2>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2919<!-- *********************************************************************** -->
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2920
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2921<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2922
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2923<p>The LLVM instruction set consists of several different classifications of
2924 instructions: <a href="#terminators">terminator
2925 instructions</a>, <a href="#binaryops">binary instructions</a>,
2926 <a href="#bitwiseops">bitwise binary instructions</a>,
2927 <a href="#memoryops">memory instructions</a>, and
2928 <a href="#otherops">other instructions</a>.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2929
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2930<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2931<h3>
2932 <a name="terminators">Terminator Instructions</a>
2933</h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2934
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2935<div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2936
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2937<p>As mentioned <a href="#functionstructure">previously</a>, every basic block
2938 in a program ends with a "Terminator" instruction, which indicates which
2939 block should be executed after the current block is finished. These
2940 terminator instructions typically yield a '<tt>void</tt>' value: they produce
2941 control flow, not values (the one exception being the
2942 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
2943
Duncan Sands626b0242010-04-15 20:35:54 +0000[diff] [blame]2944<p>There are seven different terminator instructions: the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2945 '<a href="#i_ret"><tt>ret</tt></a>' instruction, the
2946 '<a href="#i_br"><tt>br</tt></a>' instruction, the
2947 '<a href="#i_switch"><tt>switch</tt></a>' instruction, the
Bill Wendling33fef7e2009-11-02 00:25:26 +0000[diff] [blame]2948 '<a href="#i_indirectbr">'<tt>indirectbr</tt></a>' Instruction, the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2949 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the
2950 '<a href="#i_unwind"><tt>unwind</tt></a>' instruction, and the
2951 '<a href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2952
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2953<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2954<h4>
2955 <a name="i_ret">'<tt>ret</tt>' Instruction</a>
2956</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2957
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]2958<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2959
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2960<h5>Syntax:</h5>
Dan Gohmancc3132e2008-10-04 19:00:07 +0000[diff] [blame]2961<pre>
2962 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]2963 ret void <i>; Return from void function</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2964</pre>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]2965
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2966<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2967<p>The '<tt>ret</tt>' instruction is used to return control flow (and optionally
2968 a value) from a function back to the caller.</p>
2969
2970<p>There are two forms of the '<tt>ret</tt>' instruction: one that returns a
2971 value and then causes control flow, and one that just causes control flow to
2972 occur.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]2973
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2974<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2975<p>The '<tt>ret</tt>' instruction optionally accepts a single argument, the
2976 return value. The type of the return value must be a
2977 '<a href="#t_firstclass">first class</a>' type.</p>
Dan Gohmancc3132e2008-10-04 19:00:07 +0000[diff] [blame]2978
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2979<p>A function is not <a href="#wellformed">well formed</a> if it it has a
2980 non-void return type and contains a '<tt>ret</tt>' instruction with no return
2981 value or a return value with a type that does not match its type, or if it
2982 has a void return type and contains a '<tt>ret</tt>' instruction with a
2983 return value.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]2984
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2985<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2986<p>When the '<tt>ret</tt>' instruction is executed, control flow returns back to
2987 the calling function's context. If the caller is a
2988 "<a href="#i_call"><tt>call</tt></a>" instruction, execution continues at the
2989 instruction after the call. If the caller was an
2990 "<a href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues at
2991 the beginning of the "normal" destination block. If the instruction returns
2992 a value, that value shall set the call or invoke instruction's return
2993 value.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]2994
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2995<h5>Example:</h5>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]2996<pre>
2997 ret i32 5 <i>; Return an integer value of 5</i>
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]2998 ret void <i>; Return from a void function</i>
Bill Wendling050ee8f2009-02-28 22:12:54 +0000[diff] [blame]2999 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]3000</pre>
Dan Gohman3065b612009-01-12 23:12:39 +0000[diff] [blame]3001
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3002</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3003<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3004<h4>
3005 <a name="i_br">'<tt>br</tt>' Instruction</a>
3006</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3007
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3008<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3009
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3010<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3011<pre>
3012 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]3013</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3014
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3015<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3016<p>The '<tt>br</tt>' instruction is used to cause control flow to transfer to a
3017 different basic block in the current function. There are two forms of this
3018 instruction, corresponding to a conditional branch and an unconditional
3019 branch.</p>
3020
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3021<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3022<p>The conditional branch form of the '<tt>br</tt>' instruction takes a single
3023 '<tt>i1</tt>' value and two '<tt>label</tt>' values. The unconditional form
3024 of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>' value as a
3025 target.</p>
3026
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3027<h5>Semantics:</h5>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]3028<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3029 argument is evaluated. If the value is <tt>true</tt>, control flows to the
3030 '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
3031 control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
3032
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3033<h5>Example:</h5>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]3034<pre>
3035Test:
3036 %cond = <a href="#i_icmp">icmp</a> eq i32 %a, %b
3037 br i1 %cond, label %IfEqual, label %IfUnequal
3038IfEqual:
3039 <a href="#i_ret">ret</a> i32 1
3040IfUnequal:
3041 <a href="#i_ret">ret</a> i32 0
3042</pre>
3043
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3044</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3045
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3046<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3047<h4>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3048 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3049</h4>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3050
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3051<div>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3052
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3053<h5>Syntax:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3054<pre>
3055 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
3056</pre>
3057
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3058<h5>Overview:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3059<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]3060 several different places. It is a generalization of the '<tt>br</tt>'
3061 instruction, allowing a branch to occur to one of many possible
3062 destinations.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3063
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3064<h5>Arguments:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3065<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3066 comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination,
3067 and an array of pairs of comparison value constants and '<tt>label</tt>'s.
3068 The table is not allowed to contain duplicate constant entries.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3069
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3070<h5>Semantics:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3071<p>The <tt>switch</tt> instruction specifies a table of values and
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3072 destinations. When the '<tt>switch</tt>' instruction is executed, this table
3073 is searched for the given value. If the value is found, control flow is
Benjamin Kramer0f420382009-10-12 14:46:08 +0000[diff] [blame]3074 transferred to the corresponding destination; otherwise, control flow is
3075 transferred to the default destination.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3076
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3077<h5>Implementation:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3078<p>Depending on properties of the target machine and the particular
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3079 <tt>switch</tt> instruction, this instruction may be code generated in
3080 different ways. For example, it could be generated as a series of chained
3081 conditional branches or with a lookup table.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3082
3083<h5>Example:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3084<pre>
3085 <i>; Emulate a conditional br instruction</i>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]3086 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Dan Gohman623806e2009-01-04 23:44:43 +0000[diff] [blame]3087 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3088
3089 <i>; Emulate an unconditional br instruction</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]3090 switch i32 0, label %dest [ ]
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3091
3092 <i>; Implement a jump table:</i>
Dan Gohman623806e2009-01-04 23:44:43 +0000[diff] [blame]3093 switch i32 %val, label %otherwise [ i32 0, label %onzero
3094 i32 1, label %onone
3095 i32 2, label %ontwo ]
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3096</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3097
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3098</div>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3099
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3100
3101<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3102<h4>
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]3103 <a name="i_indirectbr">'<tt>indirectbr</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3104</h4>
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3105
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3106<div>
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3107
3108<h5>Syntax:</h5>
3109<pre>
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]3110 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]3111</pre>
3112
3113<h5>Overview:</h5>
3114
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]3115<p>The '<tt>indirectbr</tt>' instruction implements an indirect branch to a label
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3116 within the current function, whose address is specified by
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]3117 "<tt>address</tt>". Address must be derived from a <a
3118 href="#blockaddress">blockaddress</a> constant.</p>
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3119
3120<h5>Arguments:</h5>
3121
3122<p>The '<tt>address</tt>' argument is the address of the label to jump to. The
3123 rest of the arguments indicate the full set of possible destinations that the
3124 address may point to. Blocks are allowed to occur multiple times in the
3125 destination list, though this isn't particularly useful.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3126
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3127<p>This destination list is required so that dataflow analysis has an accurate
3128 understanding of the CFG.</p>
3129
3130<h5>Semantics:</h5>
3131
3132<p>Control transfers to the block specified in the address argument. All
3133 possible destination blocks must be listed in the label list, otherwise this
3134 instruction has undefined behavior. This implies that jumps to labels
3135 defined in other functions have undefined behavior as well.</p>
3136
3137<h5>Implementation:</h5>
3138
3139<p>This is typically implemented with a jump through a register.</p>
3140
3141<h5>Example:</h5>
3142<pre>
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]3143 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3144</pre>
3145
3146</div>
3147
3148
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3149<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3150<h4>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3151 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3152</h4>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3153
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3154<div>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3155
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3156<h5>Syntax:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3157<pre>
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]3158 &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]3159 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3160</pre>
3161
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]3162<h5>Overview:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3163<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3164 function, with the possibility of control flow transfer to either the
3165 '<tt>normal</tt>' label or the '<tt>exception</tt>' label. If the callee
3166 function returns with the "<tt><a href="#i_ret">ret</a></tt>" instruction,
3167 control flow will return to the "normal" label. If the callee (or any
3168 indirect callees) returns with the "<a href="#i_unwind"><tt>unwind</tt></a>"
3169 instruction, control is interrupted and continued at the dynamically nearest
3170 "exception" label.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3171
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3172<h5>Arguments:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3173<p>This instruction requires several arguments:</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3174
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3175<ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3176 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
3177 convention</a> the call should use. If none is specified, the call
3178 defaults to using C calling conventions.</li>
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]3179
3180 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3181 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
3182 '<tt>inreg</tt>' attributes are valid here.</li>
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]3183
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3184 <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]3185 function value being invoked. In most cases, this is a direct function
3186 invocation, but indirect <tt>invoke</tt>s are just as possible, branching
3187 off an arbitrary pointer to function value.</li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3188
3189 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3190 function to be invoked. </li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3191
3192 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]3193 signature argument types and parameter attributes. All arguments must be
3194 of <a href="#t_firstclass">first class</a> type. If the function
3195 signature indicates the function accepts a variable number of arguments,
3196 the extra arguments can be specified.</li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3197
3198 <li>'<tt>normal label</tt>': the label reached when the called function
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3199 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3200
3201 <li>'<tt>exception label</tt>': the label reached when a callee returns with
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3202 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3203
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]3204 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3205 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
3206 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3207</ol>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3208
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3209<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3210<p>This instruction is designed to operate as a standard
3211 '<tt><a href="#i_call">call</a></tt>' instruction in most regards. The
3212 primary difference is that it establishes an association with a label, which
3213 is used by the runtime library to unwind the stack.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3214
3215<p>This instruction is used in languages with destructors to ensure that proper
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3216 cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
3217 exception. Additionally, this is important for implementation of
3218 '<tt>catch</tt>' clauses in high-level languages that support them.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3219
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3220<p>For the purposes of the SSA form, the definition of the value returned by the
3221 '<tt>invoke</tt>' instruction is deemed to occur on the edge from the current
3222 block to the "normal" label. If the callee unwinds then no return value is
3223 available.</p>
Dan Gohman9069d892009-05-22 21:47:08 +0000[diff] [blame]3224
Chris Lattner97257f82010-01-15 18:08:37 +0000[diff] [blame]3225<p>Note that the code generator does not yet completely support unwind, and
3226that the invoke/unwind semantics are likely to change in future versions.</p>
3227
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3228<h5>Example:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3229<pre>
Nick Lewycky084ab472008-03-16 07:18:12 +0000[diff] [blame]3230 %retval = invoke i32 @Test(i32 15) to label %Continue
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]3231 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewycky084ab472008-03-16 07:18:12 +0000[diff] [blame]3232 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]3233 unwind label %TestCleanup <i>; {i32}:retval set</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3234</pre>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3235
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3236</div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3237
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]3238<!-- _______________________________________________________________________ -->
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3239
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3240<h4>
3241 <a name="i_unwind">'<tt>unwind</tt>' Instruction</a>
3242</h4>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3243
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3244<div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3245
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]3246<h5>Syntax:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3247<pre>
3248 unwind
3249</pre>
3250
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]3251<h5>Overview:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3252<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3253 at the first callee in the dynamic call stack which used
3254 an <a href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call.
3255 This is primarily used to implement exception handling.</p>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3256
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]3257<h5>Semantics:</h5>
Chris Lattnerfe8519c2008-04-19 21:01:16 +0000[diff] [blame]3258<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3259 immediately halt. The dynamic call stack is then searched for the
3260 first <a href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack.
3261 Once found, execution continues at the "exceptional" destination block
3262 specified by the <tt>invoke</tt> instruction. If there is no <tt>invoke</tt>
3263 instruction in the dynamic call chain, undefined behavior results.</p>
3264
Chris Lattner97257f82010-01-15 18:08:37 +0000[diff] [blame]3265<p>Note that the code generator does not yet completely support unwind, and
3266that the invoke/unwind semantics are likely to change in future versions.</p>
3267
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3268</div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3269
3270<!-- _______________________________________________________________________ -->
3271
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3272<h4>
3273 <a name="i_unreachable">'<tt>unreachable</tt>' Instruction</a>
3274</h4>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3275
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3276<div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3277
3278<h5>Syntax:</h5>
3279<pre>
3280 unreachable
3281</pre>
3282
3283<h5>Overview:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3284<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3285 instruction is used to inform the optimizer that a particular portion of the
3286 code is not reachable. This can be used to indicate that the code after a
3287 no-return function cannot be reached, and other facts.</p>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3288
3289<h5>Semantics:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3290<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3291
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3292</div>
3293
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3294</div>
3295
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3296<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3297<h3>
3298 <a name="binaryops">Binary Operations</a>
3299</h3>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3300
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3301<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3302
3303<p>Binary operators are used to do most of the computation in a program. They
3304 require two operands of the same type, execute an operation on them, and
3305 produce a single value. The operands might represent multiple data, as is
3306 the case with the <a href="#t_vector">vector</a> data type. The result value
3307 has the same type as its operands.</p>
3308
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3309<p>There are several different binary operators:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3310
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3311<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3312<h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3313 <a name="i_add">'<tt>add</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3314</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3315
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3316<div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3317
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3318<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3319<pre>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3320 &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]3321 &lt;result&gt; = add nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3322 &lt;result&gt; = add nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3323 &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]3324</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3325
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3326<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3327<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3328
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3329<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3330<p>The two arguments to the '<tt>add</tt>' instruction must
3331 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3332 integer values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3333
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3334<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3335<p>The value produced is the integer sum of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3336
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3337<p>If the sum has unsigned overflow, the result returned is the mathematical
3338 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]3339
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3340<p>Because LLVM integers use a two's complement representation, this instruction
3341 is appropriate for both signed and unsigned integers.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3342
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3343<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3344 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3345 <tt>nsw</tt> keywords are present, the result value of the <tt>add</tt>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]3346 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3347 respectively, occurs.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3348
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3349<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3350<pre>
3351 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3352</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3353
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3354</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3355
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3356<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3357<h4>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3358 <a name="i_fadd">'<tt>fadd</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3359</h4>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3360
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3361<div>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3362
3363<h5>Syntax:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3364<pre>
3365 &lt;result&gt; = fadd &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3366</pre>
3367
3368<h5>Overview:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3369<p>The '<tt>fadd</tt>' instruction returns the sum of its two operands.</p>
3370
3371<h5>Arguments:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3372<p>The two arguments to the '<tt>fadd</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3373 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3374 floating point values. Both arguments must have identical types.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3375
3376<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3377<p>The value produced is the floating point sum of the two operands.</p>
3378
3379<h5>Example:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3380<pre>
3381 &lt;result&gt; = fadd float 4.0, %var <i>; yields {float}:result = 4.0 + %var</i>
3382</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3383
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3384</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3385
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3386<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3387<h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3388 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3389</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3390
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3391<div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3392
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3393<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3394<pre>
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3395 &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]3396 &lt;result&gt; = sub nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3397 &lt;result&gt; = sub nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3398 &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]3399</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3400
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3401<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3402<p>The '<tt>sub</tt>' instruction returns the difference of its two
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3403 operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3404
3405<p>Note that the '<tt>sub</tt>' instruction is used to represent the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3406 '<tt>neg</tt>' instruction present in most other intermediate
3407 representations.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3408
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3409<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3410<p>The two arguments to the '<tt>sub</tt>' instruction must
3411 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3412 integer values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3413
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3414<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3415<p>The value produced is the integer difference of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3416
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3417<p>If the difference has unsigned overflow, the result returned is the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3418 mathematical result modulo 2<sup>n</sup>, where n is the bit width of the
3419 result.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3420
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3421<p>Because LLVM integers use a two's complement representation, this instruction
3422 is appropriate for both signed and unsigned integers.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3423
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3424<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3425 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3426 <tt>nsw</tt> keywords are present, the result value of the <tt>sub</tt>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]3427 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3428 respectively, occurs.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3429
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3430<h5>Example:</h5>
Bill Wendling2d8b9a82007-05-29 09:42:13 +0000[diff] [blame]3431<pre>
3432 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]3433 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3434</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3435
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3436</div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3437
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3438<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3439<h4>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3440 <a name="i_fsub">'<tt>fsub</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3441</h4>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3442
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3443<div>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3444
3445<h5>Syntax:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3446<pre>
3447 &lt;result&gt; = fsub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3448</pre>
3449
3450<h5>Overview:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3451<p>The '<tt>fsub</tt>' instruction returns the difference of its two
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3452 operands.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3453
3454<p>Note that the '<tt>fsub</tt>' instruction is used to represent the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3455 '<tt>fneg</tt>' instruction present in most other intermediate
3456 representations.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3457
3458<h5>Arguments:</h5>
Bill Wendling972b7202009-07-20 02:32:41 +0000[diff] [blame]3459<p>The two arguments to the '<tt>fsub</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3460 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3461 floating point values. Both arguments must have identical types.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3462
3463<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3464<p>The value produced is the floating point difference of the two operands.</p>
3465
3466<h5>Example:</h5>
3467<pre>
3468 &lt;result&gt; = fsub float 4.0, %var <i>; yields {float}:result = 4.0 - %var</i>
3469 &lt;result&gt; = fsub float -0.0, %val <i>; yields {float}:result = -%var</i>
3470</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3471
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3472</div>
3473
3474<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3475<h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3476 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3477</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3478
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3479<div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3480
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3481<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3482<pre>
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3483 &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]3484 &lt;result&gt; = mul nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3485 &lt;result&gt; = mul nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3486 &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]3487</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3488
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3489<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3490<p>The '<tt>mul</tt>' instruction returns the product of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3491
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3492<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3493<p>The two arguments to the '<tt>mul</tt>' instruction must
3494 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3495 integer values. Both arguments must have identical types.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3496
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3497<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3498<p>The value produced is the integer product of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3499
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3500<p>If the result of the multiplication has unsigned overflow, the result
3501 returned is the mathematical result modulo 2<sup>n</sup>, where n is the bit
3502 width of the result.</p>
3503
3504<p>Because LLVM integers use a two's complement representation, and the result
3505 is the same width as the operands, this instruction returns the correct
3506 result for both signed and unsigned integers. If a full product
3507 (e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands should
3508 be sign-extended or zero-extended as appropriate to the width of the full
3509 product.</p>
3510
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3511<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3512 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3513 <tt>nsw</tt> keywords are present, the result value of the <tt>mul</tt>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]3514 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3515 respectively, occurs.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3516
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3517<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3518<pre>
3519 &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3520</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3521
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3522</div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3523
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3524<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3525<h4>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3526 <a name="i_fmul">'<tt>fmul</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3527</h4>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3528
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3529<div>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3530
3531<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3532<pre>
3533 &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]3534</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3535
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3536<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3537<p>The '<tt>fmul</tt>' instruction returns the product of its two operands.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3538
3539<h5>Arguments:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3540<p>The two arguments to the '<tt>fmul</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3541 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3542 floating point values. Both arguments must have identical types.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3543
3544<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3545<p>The value produced is the floating point product of the two operands.</p>
3546
3547<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3548<pre>
3549 &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]3550</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3551
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3552</div>
3553
3554<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3555<h4>
3556 <a name="i_udiv">'<tt>udiv</tt>' Instruction</a>
3557</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3558
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3559<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3560
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3561<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3562<pre>
Chris Lattner35315d02011-02-06 21:44:57 +0000[diff] [blame]3563 &lt;result&gt; = udiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3564 &lt;result&gt; = udiv exact &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3565</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3566
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3567<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3568<p>The '<tt>udiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3569
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3570<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3571<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3572 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3573 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3574
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3575<h5>Semantics:</h5>
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3576<p>The value produced is the unsigned integer quotient of the two operands.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3577
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3578<p>Note that unsigned integer division and signed integer division are distinct
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3579 operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
3580
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3581<p>Division by zero leads to undefined behavior.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3582
Chris Lattner35315d02011-02-06 21:44:57 +0000[diff] [blame]3583<p>If the <tt>exact</tt> keyword is present, the result value of the
3584 <tt>udiv</tt> is a <a href="#trapvalues">trap value</a> if %op1 is not a
3585 multiple of %op2 (as such, "((a udiv exact b) mul b) == a").</p>
3586
3587
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3588<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3589<pre>
3590 &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3591</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3592
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3593</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3594
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3595<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3596<h4>
3597 <a name="i_sdiv">'<tt>sdiv</tt>' Instruction</a>
3598</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3599
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3600<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3601
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3602<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3603<pre>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3604 &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]3605 &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]3606</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3607
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3608<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3609<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3610
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3611<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3612<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3613 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3614 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3615
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3616<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3617<p>The value produced is the signed integer quotient of the two operands rounded
3618 towards zero.</p>
3619
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3620<p>Note that signed integer division and unsigned integer division are distinct
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3621 operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
3622
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3623<p>Division by zero leads to undefined behavior. Overflow also leads to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3624 undefined behavior; this is a rare case, but can occur, for example, by doing
3625 a 32-bit division of -2147483648 by -1.</p>
3626
Dan Gohman71dfd782009-07-22 00:04:19 +0000[diff] [blame]3627<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohman57255802010-04-23 15:23:32 +0000[diff] [blame]3628 <tt>sdiv</tt> is a <a href="#trapvalues">trap value</a> if the result would
Dan Gohmane501ff72010-07-11 00:08:34 +0000[diff] [blame]3629 be rounded.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3630
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3631<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3632<pre>
3633 &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3634</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3635
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3636</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3637
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3638<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3639<h4>
3640 <a name="i_fdiv">'<tt>fdiv</tt>' Instruction</a>
3641</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3642
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3643<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3644
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3645<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3646<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +0000[diff] [blame]3647 &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]3648</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3649
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3650<h5>Overview:</h5>
3651<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3652
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3653<h5>Arguments:</h5>
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]3654<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3655 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3656 floating point values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3657
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3658<h5>Semantics:</h5>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3659<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3660
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3661<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3662<pre>
3663 &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]3664</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3665
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3666</div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3667
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3668<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3669<h4>
3670 <a name="i_urem">'<tt>urem</tt>' Instruction</a>
3671</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3672
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3673<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3674
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3675<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3676<pre>
3677 &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]3678</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3679
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3680<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3681<p>The '<tt>urem</tt>' instruction returns the remainder from the unsigned
3682 division of its two arguments.</p>
3683
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3684<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3685<p>The two arguments to the '<tt>urem</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3686 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3687 values. Both arguments must have identical types.</p>
3688
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3689<h5>Semantics:</h5>
3690<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3691 This instruction always performs an unsigned division to get the
3692 remainder.</p>
3693
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3694<p>Note that unsigned integer remainder and signed integer remainder are
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3695 distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
3696
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3697<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3698
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3699<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3700<pre>
3701 &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3702</pre>
3703
3704</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3705
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3706<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3707<h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3708 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3709</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3710
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3711<div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3712
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3713<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3714<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +0000[diff] [blame]3715 &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]3716</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3717
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3718<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3719<p>The '<tt>srem</tt>' instruction returns the remainder from the signed
3720 division of its two operands. This instruction can also take
3721 <a href="#t_vector">vector</a> versions of the values in which case the
3722 elements must be integers.</p>
Chris Lattnerb8f816e2008-01-04 04:33:49 +0000[diff] [blame]3723
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3724<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3725<p>The two arguments to the '<tt>srem</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3726 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3727 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3728
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3729<h5>Semantics:</h5>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3730<p>This instruction returns the <i>remainder</i> of a division (where the result
Duncan Sands2769c6e2011-03-07 09:12:24 +0000[diff] [blame]3731 is either zero or has the same sign as the dividend, <tt>op1</tt>), not the
3732 <i>modulo</i> operator (where the result is either zero or has the same sign
3733 as the divisor, <tt>op2</tt>) of a value.
3734 For more information about the difference,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3735 see <a href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
3736 Math Forum</a>. For a table of how this is implemented in various languages,
3737 please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
3738 Wikipedia: modulo operation</a>.</p>
3739
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3740<p>Note that signed integer remainder and unsigned integer remainder are
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3741 distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
3742
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3743<p>Taking the remainder of a division by zero leads to undefined behavior.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3744 Overflow also leads to undefined behavior; this is a rare case, but can
3745 occur, for example, by taking the remainder of a 32-bit division of
3746 -2147483648 by -1. (The remainder doesn't actually overflow, but this rule
3747 lets srem be implemented using instructions that return both the result of
3748 the division and the remainder.)</p>
3749
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3750<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3751<pre>
3752 &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3753</pre>
3754
3755</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3756
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3757<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3758<h4>
3759 <a name="i_frem">'<tt>frem</tt>' Instruction</a>
3760</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3761
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3762<div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3763
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3764<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3765<pre>
3766 &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]3767</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3768
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3769<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3770<p>The '<tt>frem</tt>' instruction returns the remainder from the division of
3771 its two operands.</p>
3772
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3773<h5>Arguments:</h5>
3774<p>The two arguments to the '<tt>frem</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3775 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3776 floating point values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3777
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3778<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3779<p>This instruction returns the <i>remainder</i> of a division. The remainder
3780 has the same sign as the dividend.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3781
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3782<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3783<pre>
3784 &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]3785</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3786
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3787</div>
Robert Bocchino820bc75b2006-02-17 21:18:08 +0000[diff] [blame]3788
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3789</div>
3790
Reid Spencer2ab01932007-02-02 13:57:07 +0000[diff] [blame]3791<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3792<h3>
3793 <a name="bitwiseops">Bitwise Binary Operations</a>
3794</h3>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3795
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3796<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3797
3798<p>Bitwise binary operators are used to do various forms of bit-twiddling in a
3799 program. They are generally very efficient instructions and can commonly be
3800 strength reduced from other instructions. They require two operands of the
3801 same type, execute an operation on them, and produce a single value. The
3802 resulting value is the same type as its operands.</p>
3803
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3804<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3805<h4>
3806 <a name="i_shl">'<tt>shl</tt>' Instruction</a>
3807</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3808
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3809<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3810
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3811<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3812<pre>
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]3813 &lt;result&gt; = shl &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3814 &lt;result&gt; = shl nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3815 &lt;result&gt; = shl nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3816 &lt;result&gt; = shl nuw nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3817</pre>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3818
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3819<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3820<p>The '<tt>shl</tt>' instruction returns the first operand shifted to the left
3821 a specified number of bits.</p>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3822
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3823<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3824<p>Both arguments to the '<tt>shl</tt>' instruction must be the
3825 same <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3826 integer type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3827
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3828<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3829<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod
3830 2<sup>n</sup>, where <tt>n</tt> is the width of the result. If <tt>op2</tt>
3831 is (statically or dynamically) negative or equal to or larger than the number
3832 of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3833 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3834 shift amount in <tt>op2</tt>.</p>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3835
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]3836<p>If the <tt>nuw</tt> keyword is present, then the shift produces a
3837 <a href="#trapvalues">trap value</a> if it shifts out any non-zero bits. If
Chris Lattnerf10dfdc2011-02-09 16:44:44 +0000[diff] [blame]3838 the <tt>nsw</tt> keyword is present, then the shift produces a
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]3839 <a href="#trapvalues">trap value</a> if it shifts out any bits that disagree
3840 with the resultant sign bit. As such, NUW/NSW have the same semantics as
3841 they would if the shift were expressed as a mul instruction with the same
3842 nsw/nuw bits in (mul %op1, (shl 1, %op2)).</p>
3843
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3844<h5>Example:</h5>
3845<pre>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3846 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
3847 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
3848 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3849 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Mon P Wang4dd832d2008-12-09 05:46:39 +0000[diff] [blame]3850 &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]3851</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3852
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3853</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3854
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3855<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3856<h4>
3857 <a name="i_lshr">'<tt>lshr</tt>' Instruction</a>
3858</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3859
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3860<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3861
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3862<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3863<pre>
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]3864 &lt;result&gt; = lshr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3865 &lt;result&gt; = lshr exact &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3866</pre>
3867
3868<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3869<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
3870 operand shifted to the right a specified number of bits with zero fill.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3871
3872<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3873<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3874 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3875 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3876
3877<h5>Semantics:</h5>
3878<p>This instruction always performs a logical shift right operation. The most
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3879 significant bits of the result will be filled with zero bits after the shift.
3880 If <tt>op2</tt> is (statically or dynamically) equal to or larger than the
3881 number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3882 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3883 shift amount in <tt>op2</tt>.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3884
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]3885<p>If the <tt>exact</tt> keyword is present, the result value of the
3886 <tt>lshr</tt> is a <a href="#trapvalues">trap value</a> if any of the bits
3887 shifted out are non-zero.</p>
3888
3889
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3890<h5>Example:</h5>
3891<pre>
3892 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
3893 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
3894 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
3895 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3896 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Mon P Wang4dd832d2008-12-09 05:46:39 +0000[diff] [blame]3897 &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]3898</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3899
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3900</div>
3901
Reid Spencer2ab01932007-02-02 13:57:07 +0000[diff] [blame]3902<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3903<h4>
3904 <a name="i_ashr">'<tt>ashr</tt>' Instruction</a>
3905</h4>
3906
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3907<div>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3908
3909<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3910<pre>
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]3911 &lt;result&gt; = ashr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3912 &lt;result&gt; = ashr exact &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3913</pre>
3914
3915<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3916<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
3917 operand shifted to the right a specified number of bits with sign
3918 extension.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3919
3920<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3921<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3922 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3923 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3924
3925<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3926<p>This instruction always performs an arithmetic shift right operation, The
3927 most significant bits of the result will be filled with the sign bit
3928 of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
3929 larger than the number of bits in <tt>op1</tt>, the result is undefined. If
3930 the arguments are vectors, each vector element of <tt>op1</tt> is shifted by
3931 the corresponding shift amount in <tt>op2</tt>.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3932
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]3933<p>If the <tt>exact</tt> keyword is present, the result value of the
3934 <tt>ashr</tt> is a <a href="#trapvalues">trap value</a> if any of the bits
3935 shifted out are non-zero.</p>
3936
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3937<h5>Example:</h5>
3938<pre>
3939 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
3940 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
3941 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
3942 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3943 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Mon P Wang4dd832d2008-12-09 05:46:39 +0000[diff] [blame]3944 &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]3945</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3946
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3947</div>
3948
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3949<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3950<h4>
3951 <a name="i_and">'<tt>and</tt>' Instruction</a>
3952</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3953
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]3954<div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3955
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3956<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3957<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +0000[diff] [blame]3958 &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]3959</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3960
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3961<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3962<p>The '<tt>and</tt>' instruction returns the bitwise logical and of its two
3963 operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3964
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3965<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3966<p>The two arguments to the '<tt>and</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3967 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3968 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3969
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3970<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3971<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3972
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3973<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3974 <tbody>
3975 <tr>
3976 <td>In0</td>
3977 <td>In1</td>
3978 <td>Out</td>
3979 </tr>
3980 <tr>
3981 <td>0</td>
3982 <td>0</td>
3983 <td>0</td>
3984 </tr>
3985 <tr>
3986 <td>0</td>
3987 <td>1</td>
3988 <td>0</td>
3989 </tr>
3990 <tr>
3991 <td>1</td>
3992 <td>0</td>
3993 <td>0</td>
3994 </tr>
3995 <tr>
3996 <td>1</td>
3997 <td>1</td>
3998 <td>1</td>
3999 </tr>
4000 </tbody>
4001</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4002
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4003<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4004<pre>
4005 &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]4006 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
4007 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4008</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4009</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4010<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4011<h4>
4012 <a name="i_or">'<tt>or</tt>' Instruction</a>
4013</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4014
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4015<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4016
4017<h5>Syntax:</h5>
4018<pre>
4019 &lt;result&gt; = or &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4020</pre>
4021
4022<h5>Overview:</h5>
4023<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive or of its
4024 two operands.</p>
4025
4026<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4027<p>The two arguments to the '<tt>or</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4028 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4029 values. Both arguments must have identical types.</p>
4030
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4031<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4032<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4033
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]4034<table border="1" cellspacing="0" cellpadding="4">
4035 <tbody>
4036 <tr>
4037 <td>In0</td>
4038 <td>In1</td>
4039 <td>Out</td>
4040 </tr>
4041 <tr>
4042 <td>0</td>
4043 <td>0</td>
4044 <td>0</td>
4045 </tr>
4046 <tr>
4047 <td>0</td>
4048 <td>1</td>
4049 <td>1</td>
4050 </tr>
4051 <tr>
4052 <td>1</td>
4053 <td>0</td>
4054 <td>1</td>
4055 </tr>
4056 <tr>
4057 <td>1</td>
4058 <td>1</td>
4059 <td>1</td>
4060 </tr>
4061 </tbody>
4062</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4063
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4064<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4065<pre>
4066 &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4067 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
4068 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4069</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4070
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4071</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4072
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4073<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4074<h4>
4075 <a name="i_xor">'<tt>xor</tt>' Instruction</a>
4076</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4077
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4078<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4079
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4080<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4081<pre>
4082 &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]4083</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4084
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4085<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4086<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive or of
4087 its two operands. The <tt>xor</tt> is used to implement the "one's
4088 complement" operation, which is the "~" operator in C.</p>
4089
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4090<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4091<p>The two arguments to the '<tt>xor</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4092 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4093 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4094
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4095<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4096<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4097
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]4098<table border="1" cellspacing="0" cellpadding="4">
4099 <tbody>
4100 <tr>
4101 <td>In0</td>
4102 <td>In1</td>
4103 <td>Out</td>
4104 </tr>
4105 <tr>
4106 <td>0</td>
4107 <td>0</td>
4108 <td>0</td>
4109 </tr>
4110 <tr>
4111 <td>0</td>
4112 <td>1</td>
4113 <td>1</td>
4114 </tr>
4115 <tr>
4116 <td>1</td>
4117 <td>0</td>
4118 <td>1</td>
4119 </tr>
4120 <tr>
4121 <td>1</td>
4122 <td>1</td>
4123 <td>0</td>
4124 </tr>
4125 </tbody>
4126</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4127
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4128<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4129<pre>
4130 &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4131 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
4132 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
4133 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4134</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4135
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4136</div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4137
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4138</div>
4139
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4140<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4141<h3>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4142 <a name="vectorops">Vector Operations</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4143</h3>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4144
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4145<div>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4146
4147<p>LLVM supports several instructions to represent vector operations in a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4148 target-independent manner. These instructions cover the element-access and
4149 vector-specific operations needed to process vectors effectively. While LLVM
4150 does directly support these vector operations, many sophisticated algorithms
4151 will want to use target-specific intrinsics to take full advantage of a
4152 specific target.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4153
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4154<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4155<h4>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4156 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4157</h4>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4158
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4159<div>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4160
4161<h5>Syntax:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4162<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4163 &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]4164</pre>
4165
4166<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4167<p>The '<tt>extractelement</tt>' instruction extracts a single scalar element
4168 from a vector at a specified index.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4169
4170
4171<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4172<p>The first operand of an '<tt>extractelement</tt>' instruction is a value
4173 of <a href="#t_vector">vector</a> type. The second operand is an index
4174 indicating the position from which to extract the element. The index may be
4175 a variable.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4176
4177<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4178<p>The result is a scalar of the same type as the element type of
4179 <tt>val</tt>. Its value is the value at position <tt>idx</tt> of
4180 <tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4181 results are undefined.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4182
4183<h5>Example:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4184<pre>
Gabor Greif03ab4dc2009-10-28 13:14:50 +0000[diff] [blame]4185 &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]4186</pre>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4187
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4188</div>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4189
4190<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4191<h4>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4192 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4193</h4>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4194
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4195<div>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4196
4197<h5>Syntax:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4198<pre>
Dan Gohman43ba0672008-05-12 23:38:42 +0000[diff] [blame]4199 &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]4200</pre>
4201
4202<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4203<p>The '<tt>insertelement</tt>' instruction inserts a scalar element into a
4204 vector at a specified index.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4205
4206<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4207<p>The first operand of an '<tt>insertelement</tt>' instruction is a value
4208 of <a href="#t_vector">vector</a> type. The second operand is a scalar value
4209 whose type must equal the element type of the first operand. The third
4210 operand is an index indicating the position at which to insert the value.
4211 The index may be a variable.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4212
4213<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4214<p>The result is a vector of the same type as <tt>val</tt>. Its element values
4215 are those of <tt>val</tt> except at position <tt>idx</tt>, where it gets the
4216 value <tt>elt</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4217 results are undefined.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4218
4219<h5>Example:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4220<pre>
Gabor Greif03ab4dc2009-10-28 13:14:50 +0000[diff] [blame]4221 &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]4222</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4223
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4224</div>
4225
4226<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4227<h4>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4228 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4229</h4>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4230
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4231<div>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4232
4233<h5>Syntax:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4234<pre>
Mon P Wang25f01062008-11-10 04:46:22 +0000[diff] [blame]4235 &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]4236</pre>
4237
4238<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4239<p>The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
4240 from two input vectors, returning a vector with the same element type as the
4241 input and length that is the same as the shuffle mask.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4242
4243<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4244<p>The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
4245 with types that match each other. The third argument is a shuffle mask whose
4246 element type is always 'i32'. The result of the instruction is a vector
4247 whose length is the same as the shuffle mask and whose element type is the
4248 same as the element type of the first two operands.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4249
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4250<p>The shuffle mask operand is required to be a constant vector with either
4251 constant integer or undef values.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4252
4253<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4254<p>The elements of the two input vectors are numbered from left to right across
4255 both of the vectors. The shuffle mask operand specifies, for each element of
4256 the result vector, which element of the two input vectors the result element
4257 gets. The element selector may be undef (meaning "don't care") and the
4258 second operand may be undef if performing a shuffle from only one vector.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4259
4260<h5>Example:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4261<pre>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4262 &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]4263 &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]4264 &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]4265 &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]4266 &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]4267 &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]4268 &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]4269 &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]4270</pre>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4271
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4272</div>
Tanya Lattnerb138bbe2006-04-14 19:24:33 +0000[diff] [blame]4273
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4274</div>
4275
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4276<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4277<h3>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4278 <a name="aggregateops">Aggregate Operations</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4279</h3>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4280
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4281<div>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4282
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4283<p>LLVM supports several instructions for working with
4284 <a href="#t_aggregate">aggregate</a> values.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4285
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4286<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4287<h4>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4288 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4289</h4>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4290
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4291<div>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4292
4293<h5>Syntax:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4294<pre>
4295 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
4296</pre>
4297
4298<h5>Overview:</h5>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4299<p>The '<tt>extractvalue</tt>' instruction extracts the value of a member field
4300 from an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4301
4302<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4303<p>The first operand of an '<tt>extractvalue</tt>' instruction is a value
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]4304 of <a href="#t_struct">struct</a> or
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4305 <a href="#t_array">array</a> type. The operands are constant indices to
4306 specify which value to extract in a similar manner as indices in a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4307 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
Frits van Bommel7cf63ac2010-12-05 20:54:38 +0000[diff] [blame]4308 <p>The major differences to <tt>getelementptr</tt> indexing are:</p>
4309 <ul>
4310 <li>Since the value being indexed is not a pointer, the first index is
4311 omitted and assumed to be zero.</li>
4312 <li>At least one index must be specified.</li>
4313 <li>Not only struct indices but also array indices must be in
4314 bounds.</li>
4315 </ul>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4316
4317<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4318<p>The result is the value at the position in the aggregate specified by the
4319 index operands.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4320
4321<h5>Example:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4322<pre>
Gabor Greif03ab4dc2009-10-28 13:14:50 +0000[diff] [blame]4323 &lt;result&gt; = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4324</pre>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4325
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4326</div>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4327
4328<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4329<h4>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4330 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4331</h4>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4332
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4333<div>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4334
4335<h5>Syntax:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4336<pre>
Chris Lattnerc2e85402011-05-22 07:18:08 +0000[diff] [blame]4337 &lt;result&gt; = insertvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;ty&gt; &lt;elt&gt;, &lt;idx&gt;{, <idx>}* <i>; yields &lt;aggregate type&gt;</i>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4338</pre>
4339
4340<h5>Overview:</h5>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4341<p>The '<tt>insertvalue</tt>' instruction inserts a value into a member field
4342 in an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4343
4344<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4345<p>The first operand of an '<tt>insertvalue</tt>' instruction is a value
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]4346 of <a href="#t_struct">struct</a> or
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4347 <a href="#t_array">array</a> type. The second operand is a first-class
4348 value to insert. The following operands are constant indices indicating
4349 the position at which to insert the value in a similar manner as indices in a
Frits van Bommel7cf63ac2010-12-05 20:54:38 +0000[diff] [blame]4350 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' instruction. The
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4351 value to insert must have the same type as the value identified by the
4352 indices.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4353
4354<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4355<p>The result is an aggregate of the same type as <tt>val</tt>. Its value is
4356 that of <tt>val</tt> except that the value at the position specified by the
4357 indices is that of <tt>elt</tt>.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4358
4359<h5>Example:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4360<pre>
Chris Lattnerc2e85402011-05-22 07:18:08 +0000[diff] [blame]4361 %agg1 = insertvalue {i32, float} undef, i32 1, 0 <i>; yields {i32 1, float undef}</i>
4362 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 <i>; yields {i32 1, float %val}</i>
4363 %agg3 = insertvalue {i32, {float}} %agg1, float %val, 1, 0 <i>; yields {i32 1, float %val}</i>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4364</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4365
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4366</div>
4367
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4368</div>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4369
4370<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4371<h3>
Chris Lattner6ab66722006-08-15 00:45:58 +0000[diff] [blame]4372 <a name="memoryops">Memory Access and Addressing Operations</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4373</h3>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4374
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4375<div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4376
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4377<p>A key design point of an SSA-based representation is how it represents
4378 memory. In LLVM, no memory locations are in SSA form, which makes things
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]4379 very simple. This section describes how to read, write, and allocate
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4380 memory in LLVM.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4381
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4382<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4383<h4>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4384 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4385</h4>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4386
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4387<div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4388
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4389<h5>Syntax:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4390<pre>
Dan Gohman2140a742010-05-28 01:14:11 +0000[diff] [blame]4391 &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]4392</pre>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4393
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4394<h5>Overview:</h5>
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]4395<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4396 currently executing function, to be automatically released when this function
4397 returns to its caller. The object is always allocated in the generic address
4398 space (address space zero).</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4399
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4400<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4401<p>The '<tt>alloca</tt>' instruction
4402 allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt> bytes of memory on the
4403 runtime stack, returning a pointer of the appropriate type to the program.
4404 If "NumElements" is specified, it is the number of elements allocated,
4405 otherwise "NumElements" is defaulted to be one. If a constant alignment is
4406 specified, the value result of the allocation is guaranteed to be aligned to
4407 at least that boundary. If not specified, or if zero, the target can choose
4408 to align the allocation on any convenient boundary compatible with the
4409 type.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4410
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4411<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4412
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4413<h5>Semantics:</h5>
Bill Wendling9ee6a312009-05-08 20:49:29 +0000[diff] [blame]4414<p>Memory is allocated; a pointer is returned. The operation is undefined if
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4415 there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
4416 memory is automatically released when the function returns. The
4417 '<tt>alloca</tt>' instruction is commonly used to represent automatic
4418 variables that must have an address available. When the function returns
4419 (either with the <tt><a href="#i_ret">ret</a></tt>
4420 or <tt><a href="#i_unwind">unwind</a></tt> instructions), the memory is
4421 reclaimed. Allocating zero bytes is legal, but the result is undefined.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4422
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4423<h5>Example:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4424<pre>
Dan Gohman7a5acb52009-01-04 23:49:44 +0000[diff] [blame]4425 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
4426 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
4427 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
4428 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4429</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4430
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4431</div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4432
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4433<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4434<h4>
4435 <a name="i_load">'<tt>load</tt>' Instruction</a>
4436</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4437
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4438<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4439
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4440<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4441<pre>
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4442 &lt;result&gt; = load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;]
4443 &lt;result&gt; = volatile load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;]
4444 !&lt;index&gt; = !{ i32 1 }
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4445</pre>
4446
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4447<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4448<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4449
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4450<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4451<p>The argument to the '<tt>load</tt>' instruction specifies the memory address
4452 from which to load. The pointer must point to
4453 a <a href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
4454 marked as <tt>volatile</tt>, then the optimizer is not allowed to modify the
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]4455 number or order of execution of this <tt>load</tt> with other <a
4456 href="#volatile">volatile operations</a>.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4457
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4458<p>The optional constant <tt>align</tt> argument specifies the alignment of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4459 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]4460 omitted <tt>align</tt> argument means that the operation has the preferential
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4461 alignment for the target. It is the responsibility of the code emitter to
4462 ensure that the alignment information is correct. Overestimating the
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4463 alignment results in undefined behavior. Underestimating the alignment may
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4464 produce less efficient code. An alignment of 1 is always safe.</p>
4465
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4466<p>The optional <tt>!nontemporal</tt> metadata must reference a single
4467 metatadata name &lt;index&gt; corresponding to a metadata node with
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]4468 one <tt>i32</tt> entry of value 1. The existence of
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4469 the <tt>!nontemporal</tt> metatadata on the instruction tells the optimizer
4470 and code generator that this load is not expected to be reused in the cache.
4471 The code generator may select special instructions to save cache bandwidth,
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]4472 such as the <tt>MOVNT</tt> instruction on x86.</p>
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4473
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4474<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4475<p>The location of memory pointed to is loaded. If the value being loaded is of
4476 scalar type then the number of bytes read does not exceed the minimum number
4477 of bytes needed to hold all bits of the type. For example, loading an
4478 <tt>i24</tt> reads at most three bytes. When loading a value of a type like
4479 <tt>i20</tt> with a size that is not an integral number of bytes, the result
4480 is undefined if the value was not originally written using a store of the
4481 same type.</p>
4482
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4483<h5>Examples:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4484<pre>
4485 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
4486 <a href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4487 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4488</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4489
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4490</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4491
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4492<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4493<h4>
4494 <a name="i_store">'<tt>store</tt>' Instruction</a>
4495</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4496
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4497<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4498
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4499<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4500<pre>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]4501 store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;] <i>; yields {void}</i>
4502 volatile store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;] <i>; yields {void}</i>
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4503</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4504
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4505<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4506<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4507
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4508<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4509<p>There are two arguments to the '<tt>store</tt>' instruction: a value to store
4510 and an address at which to store it. The type of the
4511 '<tt>&lt;pointer&gt;</tt>' operand must be a pointer to
4512 the <a href="#t_firstclass">first class</a> type of the
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]4513 '<tt>&lt;value&gt;</tt>' operand. If the <tt>store</tt> is marked as
4514 <tt>volatile</tt>, then the optimizer is not allowed to modify the number or
4515 order of execution of this <tt>store</tt> with other <a
4516 href="#volatile">volatile operations</a>.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4517
4518<p>The optional constant "align" argument specifies the alignment of the
4519 operation (that is, the alignment of the memory address). A value of 0 or an
4520 omitted "align" argument means that the operation has the preferential
4521 alignment for the target. It is the responsibility of the code emitter to
4522 ensure that the alignment information is correct. Overestimating the
4523 alignment results in an undefined behavior. Underestimating the alignment may
4524 produce less efficient code. An alignment of 1 is always safe.</p>
4525
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4526<p>The optional !nontemporal metadata must reference a single metatadata
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]4527 name &lt;index&gt; corresponding to a metadata node with one i32 entry of
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]4528 value 1. The existence of the !nontemporal metatadata on the
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4529 instruction tells the optimizer and code generator that this load is
4530 not expected to be reused in the cache. The code generator may
4531 select special instructions to save cache bandwidth, such as the
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]4532 MOVNT instruction on x86.</p>
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4533
4534
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]4535<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4536<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>' at the
4537 location specified by the '<tt>&lt;pointer&gt;</tt>' operand. If
4538 '<tt>&lt;value&gt;</tt>' is of scalar type then the number of bytes written
4539 does not exceed the minimum number of bytes needed to hold all bits of the
4540 type. For example, storing an <tt>i24</tt> writes at most three bytes. When
4541 writing a value of a type like <tt>i20</tt> with a size that is not an
4542 integral number of bytes, it is unspecified what happens to the extra bits
4543 that do not belong to the type, but they will typically be overwritten.</p>
4544
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4545<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4546<pre>
4547 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling8830ffe2007-10-22 05:10:05 +0000[diff] [blame]4548 store i32 3, i32* %ptr <i>; yields {void}</i>
4549 %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]4550</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4551
Reid Spencer443460a2006-11-09 21:15:49 +0000[diff] [blame]4552</div>
4553
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4554<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4555<h4>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4556 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4557</h4>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4558
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4559<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4560
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]4561<h5>Syntax:</h5>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4562<pre>
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]4563 &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]4564 &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]4565</pre>
4566
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]4567<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4568<p>The '<tt>getelementptr</tt>' instruction is used to get the address of a
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4569 subelement of an <a href="#t_aggregate">aggregate</a> data structure.
4570 It performs address calculation only and does not access memory.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4571
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]4572<h5>Arguments:</h5>
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]4573<p>The first argument is always a pointer, and forms the basis of the
Chris Lattnera40b9122009-07-29 06:44:13 +0000[diff] [blame]4574 calculation. The remaining arguments are indices that indicate which of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4575 elements of the aggregate object are indexed. The interpretation of each
4576 index is dependent on the type being indexed into. The first index always
4577 indexes the pointer value given as the first argument, the second index
4578 indexes a value of the type pointed to (not necessarily the value directly
4579 pointed to, since the first index can be non-zero), etc. The first type
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4580 indexed into must be a pointer value, subsequent types can be arrays,
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]4581 vectors, and structs. Note that subsequent types being indexed into
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4582 can never be pointers, since that would require loading the pointer before
4583 continuing calculation.</p>
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]4584
4585<p>The type of each index argument depends on the type it is indexing into.
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]4586 When indexing into a (optionally packed) structure, only <tt>i32</tt>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4587 integer <b>constants</b> are allowed. When indexing into an array, pointer
4588 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]4589 constant.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4590
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4591<p>For example, let's consider a C code fragment and how it gets compiled to
4592 LLVM:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4593
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]4594<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4595struct RT {
4596 char A;
Chris Lattnera446f1b2007-05-29 15:43:56 +0000[diff] [blame]4597 int B[10][20];
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4598 char C;
4599};
4600struct ST {
Chris Lattnera446f1b2007-05-29 15:43:56 +0000[diff] [blame]4601 int X;
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4602 double Y;
4603 struct RT Z;
4604};
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4605
Chris Lattnera446f1b2007-05-29 15:43:56 +0000[diff] [blame]4606int *foo(struct ST *s) {
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4607 return &amp;s[1].Z.B[5][13];
4608}
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4609</pre>
4610
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4611<p>The LLVM code generated by the GCC frontend is:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4612
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]4613<pre class="doc_code">
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]4614%RT = <a href="#namedtypes">type</a> { i8 , [10 x [20 x i32]], i8 }
4615%ST = <a href="#namedtypes">type</a> { i32, double, %RT }
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4616
Dan Gohman6b867702009-07-25 02:23:48 +0000[diff] [blame]4617define i32* @foo(%ST* %s) {
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4618entry:
4619 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
4620 ret i32* %reg
4621}
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4622</pre>
4623
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]4624<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4625<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]4626 type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
4627 }</tt>' type, a structure. The second index indexes into the third element
4628 of the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
4629 i8 }</tt>' type, another structure. The third index indexes into the second
4630 element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
4631 array. The two dimensions of the array are subscripted into, yielding an
4632 '<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a
4633 pointer to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4634
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4635<p>Note that it is perfectly legal to index partially through a structure,
4636 returning a pointer to an inner element. Because of this, the LLVM code for
4637 the given testcase is equivalent to:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4638
4639<pre>
Dan Gohman6b867702009-07-25 02:23:48 +0000[diff] [blame]4640 define i32* @foo(%ST* %s) {
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4641 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]4642 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
4643 %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]4644 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
4645 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
4646 ret i32* %t5
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4647 }
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]4648</pre>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +0000[diff] [blame]4649
Dan Gohman1639c392009-07-27 21:53:46 +0000[diff] [blame]4650<p>If the <tt>inbounds</tt> keyword is present, the result value of the
Dan Gohman57255802010-04-23 15:23:32 +0000[diff] [blame]4651 <tt>getelementptr</tt> is a <a href="#trapvalues">trap value</a> if the
4652 base pointer is not an <i>in bounds</i> address of an allocated object,
4653 or if any of the addresses that would be formed by successive addition of
4654 the offsets implied by the indices to the base address with infinitely
4655 precise arithmetic are not an <i>in bounds</i> address of that allocated
4656 object. The <i>in bounds</i> addresses for an allocated object are all
4657 the addresses that point into the object, plus the address one byte past
4658 the end.</p>
Dan Gohman1639c392009-07-27 21:53:46 +0000[diff] [blame]4659
4660<p>If the <tt>inbounds</tt> keyword is not present, the offsets are added to
4661 the base address with silently-wrapping two's complement arithmetic, and
4662 the result value of the <tt>getelementptr</tt> may be outside the object
4663 pointed to by the base pointer. The result value may not necessarily be
4664 used to access memory though, even if it happens to point into allocated
4665 storage. See the <a href="#pointeraliasing">Pointer Aliasing Rules</a>
4666 section for more information.</p>
4667
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4668<p>The getelementptr instruction is often confusing. For some more insight into
4669 how it works, see <a href="GetElementPtr.html">the getelementptr FAQ</a>.</p>
Chris Lattner6ab66722006-08-15 00:45:58 +0000[diff] [blame]4670
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]4671<h5>Example:</h5>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4672<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4673 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]4674 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
4675 <i>; yields i8*:vptr</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]4676 %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]4677 <i>; yields i8*:eptr</i>
4678 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Sanjiv Gupta0c155e62009-04-25 07:27:44 +0000[diff] [blame]4679 <i>; yields i32*:iptr</i>
Sanjiv Gupta77abea02009-04-24 16:38:13 +0000[diff] [blame]4680 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4681</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4682
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4683</div>
Reid Spencer443460a2006-11-09 21:15:49 +0000[diff] [blame]4684
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4685</div>
4686
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4687<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4688<h3>
4689 <a name="convertops">Conversion Operations</a>
4690</h3>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4691
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4692<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4693
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]4694<p>The instructions in this category are the conversion instructions (casting)
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4695 which all take a single operand and a type. They perform various bit
4696 conversions on the operand.</p>
4697
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]4698<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4699<h4>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4700 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4701</h4>
4702
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4703<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4704
4705<h5>Syntax:</h5>
4706<pre>
4707 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4708</pre>
4709
4710<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4711<p>The '<tt>trunc</tt>' instruction truncates its operand to the
4712 type <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4713
4714<h5>Arguments:</h5>
Nadav Rotem502f1b92011-02-24 21:01:34 +0000[diff] [blame]4715<p>The '<tt>trunc</tt>' instruction takes a value to trunc, and a type to trunc it to.
4716 Both types must be of <a href="#t_integer">integer</a> types, or vectors
4717 of the same number of integers.
4718 The bit size of the <tt>value</tt> must be larger than
4719 the bit size of the destination type, <tt>ty2</tt>.
4720 Equal sized types are not allowed.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4721
4722<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4723<p>The '<tt>trunc</tt>' instruction truncates the high order bits
4724 in <tt>value</tt> and converts the remaining bits to <tt>ty2</tt>. Since the
4725 source size must be larger than the destination size, <tt>trunc</tt> cannot
4726 be a <i>no-op cast</i>. It will always truncate bits.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4727
4728<h5>Example:</h5>
4729<pre>
Nadav Rotem502f1b92011-02-24 21:01:34 +0000[diff] [blame]4730 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
4731 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
4732 %Z = trunc i32 122 to i1 <i>; yields i1:false</i>
4733 %W = trunc &lt;2 x i16&gt; &lt;i16 8, i16 7&gt; to &lt;2 x i8&gt; <i>; yields &lt;i8 8, i8 7&gt;</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4734</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4735
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4736</div>
4737
4738<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4739<h4>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4740 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4741</h4>
4742
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4743<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4744
4745<h5>Syntax:</h5>
4746<pre>
4747 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4748</pre>
4749
4750<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4751<p>The '<tt>zext</tt>' instruction zero extends its operand to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4752 <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4753
4754
4755<h5>Arguments:</h5>
Nadav Rotem25f2ac92011-02-20 12:37:50 +0000[diff] [blame]4756<p>The '<tt>zext</tt>' instruction takes a value to cast, and a type to cast it to.
4757 Both types must be of <a href="#t_integer">integer</a> types, or vectors
4758 of the same number of integers.
4759 The bit size of the <tt>value</tt> must be smaller than
4760 the bit size of the destination type,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4761 <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4762
4763<h5>Semantics:</h5>
4764<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]4765 bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4766
Reid Spencer07c9c682007-01-12 15:46:11 +0000[diff] [blame]4767<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]4768
4769<h5>Example:</h5>
4770<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4771 %X = zext i32 257 to i64 <i>; yields i64:257</i>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]4772 %Y = zext i1 true to i32 <i>; yields i32:1</i>
Nadav Rotem25f2ac92011-02-20 12:37:50 +0000[diff] [blame]4773 %Z = zext &lt;2 x i16&gt; &lt;i16 8, i16 7&gt; to &lt;2 x i32&gt; <i>; yields &lt;i32 8, i32 7&gt;</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4774</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4775
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4776</div>
4777
4778<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4779<h4>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4780 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4781</h4>
4782
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4783<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4784
4785<h5>Syntax:</h5>
4786<pre>
4787 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4788</pre>
4789
4790<h5>Overview:</h5>
4791<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
4792
4793<h5>Arguments:</h5>
Nadav Rotem502f1b92011-02-24 21:01:34 +0000[diff] [blame]4794<p>The '<tt>sext</tt>' instruction takes a value to cast, and a type to cast it to.
4795 Both types must be of <a href="#t_integer">integer</a> types, or vectors
4796 of the same number of integers.
4797 The bit size of the <tt>value</tt> must be smaller than
4798 the bit size of the destination type,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4799 <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4800
4801<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4802<p>The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
4803 bit (highest order bit) of the <tt>value</tt> until it reaches the bit size
4804 of the type <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4805
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]4806<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]4807
4808<h5>Example:</h5>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4809<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4810 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]4811 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
Nadav Rotem502f1b92011-02-24 21:01:34 +0000[diff] [blame]4812 %Z = sext &lt;2 x i16&gt; &lt;i16 8, i16 7&gt; to &lt;2 x i32&gt; <i>; yields &lt;i32 8, i32 7&gt;</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4813</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4814
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4815</div>
4816
4817<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4818<h4>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4819 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4820</h4>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4821
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4822<div>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4823
4824<h5>Syntax:</h5>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4825<pre>
4826 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4827</pre>
4828
4829<h5>Overview:</h5>
4830<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4831 <tt>ty2</tt>.</p>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4832
4833<h5>Arguments:</h5>
4834<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4835 point</a> value to cast and a <a href="#t_floating">floating point</a> type
4836 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]4837 <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]4838 <i>no-op cast</i>.</p>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4839
4840<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4841<p>The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4842 <a href="#t_floating">floating point</a> type to a smaller
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4843 <a href="#t_floating">floating point</a> type. If the value cannot fit
4844 within the destination type, <tt>ty2</tt>, then the results are
4845 undefined.</p>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4846
4847<h5>Example:</h5>
4848<pre>
4849 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
4850 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
4851</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4852
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4853</div>
4854
4855<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4856<h4>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4857 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4858</h4>
4859
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4860<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4861
4862<h5>Syntax:</h5>
4863<pre>
4864 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4865</pre>
4866
4867<h5>Overview:</h5>
4868<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]4869 floating point value.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4870
4871<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4872<p>The '<tt>fpext</tt>' instruction takes a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4873 <a href="#t_floating">floating point</a> <tt>value</tt> to cast, and
4874 a <a href="#t_floating">floating point</a> type to cast it to. The source
4875 type must be smaller than the destination type.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4876
4877<h5>Semantics:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4878<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4879 <a href="#t_floating">floating point</a> type to a larger
4880 <a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
4881 used to make a <i>no-op cast</i> because it always changes bits. Use
4882 <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]4883
4884<h5>Example:</h5>
4885<pre>
Nick Lewycky9feca672011-03-31 18:20:19 +0000[diff] [blame]4886 %X = fpext float 3.125 to double <i>; yields double:3.125000e+00</i>
4887 %Y = fpext double %X to fp128 <i>; yields fp128:0xL00000000000000004000900000000000</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4888</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4889
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4890</div>
4891
4892<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4893<h4>
Reid Spencer2eadb532007-01-21 00:29:26 +0000[diff] [blame]4894 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4895</h4>
4896
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4897<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4898
4899<h5>Syntax:</h5>
4900<pre>
Reid Spencer753163d2007-07-31 14:40:14 +0000[diff] [blame]4901 &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]4902</pre>
4903
4904<h5>Overview:</h5>
Reid Spencer753163d2007-07-31 14:40:14 +0000[diff] [blame]4905<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4906 unsigned integer equivalent of type <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4907
4908<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4909<p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
4910 scalar or vector <a href="#t_floating">floating point</a> value, and a type
4911 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4912 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4913 vector integer type with the same number of elements as <tt>ty</tt></p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4914
4915<h5>Semantics:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4916<p>The '<tt>fptoui</tt>' instruction converts its
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4917 <a href="#t_floating">floating point</a> operand into the nearest (rounding
4918 towards zero) unsigned integer value. If the value cannot fit
4919 in <tt>ty2</tt>, the results are undefined.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4920
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4921<h5>Example:</h5>
4922<pre>
Reid Spencer753163d2007-07-31 14:40:14 +0000[diff] [blame]4923 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner5b95a172007-09-22 03:17:52 +0000[diff] [blame]4924 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Gabor Greiff50fd572009-10-28 09:21:30 +0000[diff] [blame]4925 %Z = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4926</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4927
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4928</div>
4929
4930<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4931<h4>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4932 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4933</h4>
4934
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4935<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4936
4937<h5>Syntax:</h5>
4938<pre>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4939 &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]4940</pre>
4941
4942<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4943<p>The '<tt>fptosi</tt>' instruction converts
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4944 <a href="#t_floating">floating point</a> <tt>value</tt> to
4945 type <tt>ty2</tt>.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]4946
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]4947<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4948<p>The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
4949 scalar or vector <a href="#t_floating">floating point</a> value, and a type
4950 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4951 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4952 vector integer type with the same number of elements as <tt>ty</tt></p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]4953
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]4954<h5>Semantics:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4955<p>The '<tt>fptosi</tt>' instruction converts its
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4956 <a href="#t_floating">floating point</a> operand into the nearest (rounding
4957 towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
4958 the results are undefined.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]4959
Chris Lattner70de6632001-07-09 00:26:23 +0000[diff] [blame]4960<h5>Example:</h5>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]4961<pre>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]4962 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner5b95a172007-09-22 03:17:52 +0000[diff] [blame]4963 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Gabor Greiff50fd572009-10-28 09:21:30 +0000[diff] [blame]4964 %Z = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4965</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4966
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4967</div>
4968
4969<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4970<h4>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4971 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4972</h4>
4973
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]4974<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4975
4976<h5>Syntax:</h5>
4977<pre>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4978 &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]4979</pre>
4980
4981<h5>Overview:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4982<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4983 integer and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4984
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4985<h5>Arguments:</h5>
Nate Begemand4d45c22007-11-17 03:58:34 +0000[diff] [blame]4986<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]4987 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
4988 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4989 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4990 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4991
4992<h5>Semantics:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4993<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4994 integer quantity and converts it to the corresponding floating point
4995 value. If the value cannot fit in the floating point value, the results are
4996 undefined.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4997
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4998<h5>Example:</h5>
4999<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5000 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]5001 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5002</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5003
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5004</div>
5005
5006<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5007<h4>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]5008 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5009</h4>
5010
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5011<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5012
5013<h5>Syntax:</h5>
5014<pre>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]5015 &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]5016</pre>
5017
5018<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5019<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed integer
5020 and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5021
5022<h5>Arguments:</h5>
Nate Begemand4d45c22007-11-17 03:58:34 +0000[diff] [blame]5023<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]5024 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
5025 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
5026 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
5027 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5028
5029<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5030<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed integer
5031 quantity and converts it to the corresponding floating point value. If the
5032 value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5033
5034<h5>Example:</h5>
5035<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5036 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]5037 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5038</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5039
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5040</div>
5041
5042<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5043<h4>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5044 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5045</h4>
5046
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5047<div>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5048
5049<h5>Syntax:</h5>
5050<pre>
5051 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5052</pre>
5053
5054<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5055<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
5056 the integer type <tt>ty2</tt>.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5057
5058<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5059<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
5060 must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
5061 <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]5062
5063<h5>Semantics:</h5>
5064<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5065 <tt>ty2</tt> by interpreting the pointer value as an integer and either
5066 truncating or zero extending that value to the size of the integer type. If
5067 <tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
5068 <tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
5069 are the same size, then nothing is done (<i>no-op cast</i>) other than a type
5070 change.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5071
5072<h5>Example:</h5>
5073<pre>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5074 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
5075 %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]5076</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5077
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5078</div>
5079
5080<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5081<h4>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5082 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5083</h4>
5084
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5085<div>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5086
5087<h5>Syntax:</h5>
5088<pre>
5089 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5090</pre>
5091
5092<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5093<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to a
5094 pointer type, <tt>ty2</tt>.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5095
5096<h5>Arguments:</h5>
Duncan Sands16f122e2007-03-30 12:22:09 +0000[diff] [blame]5097<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5098 value to cast, and a type to cast it to, which must be a
5099 <a href="#t_pointer">pointer</a> type.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5100
5101<h5>Semantics:</h5>
5102<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5103 <tt>ty2</tt> by applying either a zero extension or a truncation depending on
5104 the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
5105 size of a pointer then a truncation is done. If <tt>value</tt> is smaller
5106 than the size of a pointer then a zero extension is done. If they are the
5107 same size, nothing is done (<i>no-op cast</i>).</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5108
5109<h5>Example:</h5>
5110<pre>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5111 %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]5112 %Y = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
5113 %Z = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5114</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5115
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5116</div>
5117
5118<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5119<h4>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]5120 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5121</h4>
5122
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5123<div>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5124
5125<h5>Syntax:</h5>
5126<pre>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]5127 &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]5128</pre>
5129
5130<h5>Overview:</h5>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]5131<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5132 <tt>ty2</tt> without changing any bits.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5133
5134<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5135<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be a
5136 non-aggregate first class value, and a type to cast it to, which must also be
5137 a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes
5138 of <tt>value</tt> and the destination type, <tt>ty2</tt>, must be
5139 identical. If the source type is a pointer, the destination type must also be
5140 a pointer. This instruction supports bitwise conversion of vectors to
5141 integers and to vectors of other types (as long as they have the same
5142 size).</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5143
5144<h5>Semantics:</h5>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]5145<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5146 <tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
5147 this conversion. The conversion is done as if the <tt>value</tt> had been
5148 stored to memory and read back as type <tt>ty2</tt>. Pointer types may only
5149 be converted to other pointer types with this instruction. To convert
5150 pointers to other types, use the <a href="#i_inttoptr">inttoptr</a> or
5151 <a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5152
5153<h5>Example:</h5>
5154<pre>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5155 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5156 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5157 %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]5158</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5159
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5160</div>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5161
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5162</div>
5163
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5164<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5165<h3>
5166 <a name="otherops">Other Operations</a>
5167</h3>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5168
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5169<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5170
5171<p>The instructions in this category are the "miscellaneous" instructions, which
5172 defy better classification.</p>
5173
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5174<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5175<h4>
5176 <a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
5177</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5178
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5179<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5180
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5181<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5182<pre>
5183 &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]5184</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5185
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5186<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5187<p>The '<tt>icmp</tt>' instruction returns a boolean value or a vector of
5188 boolean values based on comparison of its two integer, integer vector, or
5189 pointer operands.</p>
5190
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5191<h5>Arguments:</h5>
5192<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5193 the condition code indicating the kind of comparison to perform. It is not a
5194 value, just a keyword. The possible condition code are:</p>
5195
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5196<ol>
5197 <li><tt>eq</tt>: equal</li>
5198 <li><tt>ne</tt>: not equal </li>
5199 <li><tt>ugt</tt>: unsigned greater than</li>
5200 <li><tt>uge</tt>: unsigned greater or equal</li>
5201 <li><tt>ult</tt>: unsigned less than</li>
5202 <li><tt>ule</tt>: unsigned less or equal</li>
5203 <li><tt>sgt</tt>: signed greater than</li>
5204 <li><tt>sge</tt>: signed greater or equal</li>
5205 <li><tt>slt</tt>: signed less than</li>
5206 <li><tt>sle</tt>: signed less or equal</li>
5207</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5208
Chris Lattnerc0f423a2007-01-15 01:54:13 +0000[diff] [blame]5209<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5210 <a href="#t_pointer">pointer</a> or integer <a href="#t_vector">vector</a>
5211 typed. They must also be identical types.</p>
5212
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5213<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5214<p>The '<tt>icmp</tt>' compares <tt>op1</tt> and <tt>op2</tt> according to the
5215 condition code given as <tt>cond</tt>. The comparison performed always yields
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]5216 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]5217 result, as follows:</p>
5218
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5219<ol>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5220 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5221 <tt>false</tt> otherwise. No sign interpretation is necessary or
5222 performed.</li>
5223
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5224 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5225 <tt>false</tt> otherwise. No sign interpretation is necessary or
5226 performed.</li>
5227
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5228 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5229 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5230
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5231 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5232 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5233 to <tt>op2</tt>.</li>
5234
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5235 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5236 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5237
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5238 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5239 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5240
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5241 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5242 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5243
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5244 <li><tt>sge</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5245 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5246 to <tt>op2</tt>.</li>
5247
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5248 <li><tt>slt</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5249 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5250
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5251 <li><tt>sle</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5252 <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]5253</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5254
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5255<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5256 values are compared as if they were integers.</p>
5257
5258<p>If the operands are integer vectors, then they are compared element by
5259 element. The result is an <tt>i1</tt> vector with the same number of elements
5260 as the values being compared. Otherwise, the result is an <tt>i1</tt>.</p>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5261
5262<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5263<pre>
5264 &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5265 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
5266 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
5267 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
5268 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
5269 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5270</pre>
Dan Gohmana5127ab2009-01-22 01:39:38 +0000[diff] [blame]5271
5272<p>Note that the code generator does not yet support vector types with
5273 the <tt>icmp</tt> instruction.</p>
5274
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5275</div>
5276
5277<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5278<h4>
5279 <a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
5280</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5281
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5282<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5283
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5284<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5285<pre>
5286 &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]5287</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5288
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5289<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5290<p>The '<tt>fcmp</tt>' instruction returns a boolean value or vector of boolean
5291 values based on comparison of its operands.</p>
5292
5293<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]5294(<a href="#t_integer"><tt>i1</tt></a>).</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5295
5296<p>If the operands are floating point vectors, then the result type is a vector
5297 of boolean with the same number of elements as the operands being
5298 compared.</p>
5299
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5300<h5>Arguments:</h5>
5301<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5302 the condition code indicating the kind of comparison to perform. It is not a
5303 value, just a keyword. The possible condition code are:</p>
5304
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5305<ol>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5306 <li><tt>false</tt>: no comparison, always returns false</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5307 <li><tt>oeq</tt>: ordered and equal</li>
5308 <li><tt>ogt</tt>: ordered and greater than </li>
5309 <li><tt>oge</tt>: ordered and greater than or equal</li>
5310 <li><tt>olt</tt>: ordered and less than </li>
5311 <li><tt>ole</tt>: ordered and less than or equal</li>
5312 <li><tt>one</tt>: ordered and not equal</li>
5313 <li><tt>ord</tt>: ordered (no nans)</li>
5314 <li><tt>ueq</tt>: unordered or equal</li>
5315 <li><tt>ugt</tt>: unordered or greater than </li>
5316 <li><tt>uge</tt>: unordered or greater than or equal</li>
5317 <li><tt>ult</tt>: unordered or less than </li>
5318 <li><tt>ule</tt>: unordered or less than or equal</li>
5319 <li><tt>une</tt>: unordered or not equal</li>
5320 <li><tt>uno</tt>: unordered (either nans)</li>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5321 <li><tt>true</tt>: no comparison, always returns true</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5322</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5323
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5324<p><i>Ordered</i> means that neither operand is a QNAN while
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5325 <i>unordered</i> means that either operand may be a QNAN.</p>
5326
5327<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be either
5328 a <a href="#t_floating">floating point</a> type or
5329 a <a href="#t_vector">vector</a> of floating point type. They must have
5330 identical types.</p>
5331
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5332<h5>Semantics:</h5>
Gabor Greif0f75ad02008-08-07 21:46:00 +0000[diff] [blame]5333<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5334 according to the condition code given as <tt>cond</tt>. If the operands are
5335 vectors, then the vectors are compared element by element. Each comparison
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]5336 performed always yields an <a href="#t_integer">i1</a> result, as
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5337 follows:</p>
5338
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5339<ol>
5340 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5341
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5342 <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]5343 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5344
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5345 <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]5346 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5347
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5348 <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]5349 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5350
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5351 <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]5352 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5353
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5354 <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]5355 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5356
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5357 <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]5358 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5359
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5360 <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]5361
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5362 <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]5363 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5364
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5365 <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]5366 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5367
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5368 <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]5369 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5370
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5371 <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]5372 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5373
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5374 <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]5375 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5376
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5377 <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]5378 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5379
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5380 <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]5381
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5382 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
5383</ol>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5384
5385<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5386<pre>
5387 &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]5388 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
5389 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
5390 &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]5391</pre>
Dan Gohmana5127ab2009-01-22 01:39:38 +0000[diff] [blame]5392
5393<p>Note that the code generator does not yet support vector types with
5394 the <tt>fcmp</tt> instruction.</p>
5395
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5396</div>
5397
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5398<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5399<h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5400 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5401</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5402
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5403<div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5404
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5405<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5406<pre>
5407 &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...
5408</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5409
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5410<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5411<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in the
5412 SSA graph representing the function.</p>
5413
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5414<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5415<p>The type of the incoming values is specified with the first type field. After
5416 this, the '<tt>phi</tt>' instruction takes a list of pairs as arguments, with
5417 one pair for each predecessor basic block of the current block. Only values
5418 of <a href="#t_firstclass">first class</a> type may be used as the value
5419 arguments to the PHI node. Only labels may be used as the label
5420 arguments.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5421
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5422<p>There must be no non-phi instructions between the start of a basic block and
5423 the PHI instructions: i.e. PHI instructions must be first in a basic
5424 block.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5425
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5426<p>For the purposes of the SSA form, the use of each incoming value is deemed to
5427 occur on the edge from the corresponding predecessor block to the current
5428 block (but after any definition of an '<tt>invoke</tt>' instruction's return
5429 value on the same edge).</p>
Jay Foad1a4eea52009-06-03 10:20:10 +0000[diff] [blame]5430
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5431<h5>Semantics:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5432<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5433 specified by the pair corresponding to the predecessor basic block that
5434 executed just prior to the current block.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5435
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5436<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5437<pre>
5438Loop: ; Infinite loop that counts from 0 on up...
5439 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
5440 %nextindvar = add i32 %indvar, 1
5441 br label %Loop
5442</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5443
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5444</div>
5445
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5446<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5447<h4>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5448 <a name="i_select">'<tt>select</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5449</h4>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5450
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5451<div>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5452
5453<h5>Syntax:</h5>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5454<pre>
Dan Gohmanc579d972008-09-09 01:02:47 +0000[diff] [blame]5455 &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>
5456
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]5457 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5458</pre>
5459
5460<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5461<p>The '<tt>select</tt>' instruction is used to choose one value based on a
5462 condition, without branching.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5463
5464
5465<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5466<p>The '<tt>select</tt>' instruction requires an 'i1' value or a vector of 'i1'
5467 values indicating the condition, and two values of the
5468 same <a href="#t_firstclass">first class</a> type. If the val1/val2 are
5469 vectors and the condition is a scalar, then entire vectors are selected, not
5470 individual elements.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5471
5472<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5473<p>If the condition is an i1 and it evaluates to 1, the instruction returns the
5474 first value argument; otherwise, it returns the second value argument.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5475
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5476<p>If the condition is a vector of i1, then the value arguments must be vectors
5477 of the same size, and the selection is done element by element.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5478
5479<h5>Example:</h5>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5480<pre>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]5481 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5482</pre>
Dan Gohmana5127ab2009-01-22 01:39:38 +0000[diff] [blame]5483
5484<p>Note that the code generator does not yet support conditions
5485 with vector type.</p>
5486
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5487</div>
5488
Robert Bocchinof72fdfe2006-01-15 20:48:27 +0000[diff] [blame]5489<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5490<h4>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5491 <a name="i_call">'<tt>call</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5492</h4>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5493
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5494<div>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5495
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5496<h5>Syntax:</h5>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5497<pre>
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]5498 &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]5499</pre>
5500
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5501<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5502<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5503
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5504<h5>Arguments:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5505<p>This instruction requires several arguments:</p>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5506
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]5507<ol>
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]5508 <li>The optional "tail" marker indicates that the callee function does not
5509 access any allocas or varargs in the caller. Note that calls may be
5510 marked "tail" even if they do not occur before
5511 a <a href="#i_ret"><tt>ret</tt></a> instruction. If the "tail" marker is
5512 present, the function call is eligible for tail call optimization,
5513 but <a href="CodeGenerator.html#tailcallopt">might not in fact be
Evan Cheng59676492010-03-08 21:05:02 +0000[diff] [blame]5514 optimized into a jump</a>. The code generator may optimize calls marked
5515 "tail" with either 1) automatic <a href="CodeGenerator.html#sibcallopt">
5516 sibling call optimization</a> when the caller and callee have
5517 matching signatures, or 2) forced tail call optimization when the
5518 following extra requirements are met:
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]5519 <ul>
5520 <li>Caller and callee both have the calling
5521 convention <tt>fastcc</tt>.</li>
5522 <li>The call is in tail position (ret immediately follows call and ret
5523 uses value of call or is void).</li>
5524 <li>Option <tt>-tailcallopt</tt> is enabled,
Dan Gohman6232f732010-03-02 01:08:11 +0000[diff] [blame]5525 or <code>llvm::GuaranteedTailCallOpt</code> is <code>true</code>.</li>
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]5526 <li><a href="CodeGenerator.html#tailcallopt">Platform specific
5527 constraints are met.</a></li>
5528 </ul>
5529 </li>
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]5530
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5531 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
5532 convention</a> the call should use. If none is specified, the call
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]5533 defaults to using C calling conventions. The calling convention of the
5534 call must match the calling convention of the target function, or else the
5535 behavior is undefined.</li>
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]5536
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5537 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
5538 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
5539 '<tt>inreg</tt>' attributes are valid here.</li>
5540
5541 <li>'<tt>ty</tt>': the type of the call instruction itself which is also the
5542 type of the return value. Functions that return no value are marked
5543 <tt><a href="#t_void">void</a></tt>.</li>
5544
5545 <li>'<tt>fnty</tt>': shall be the signature of the pointer to function value
5546 being invoked. The argument types must match the types implied by this
5547 signature. This type can be omitted if the function is not varargs and if
5548 the function type does not return a pointer to a function.</li>
5549
5550 <li>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
5551 be invoked. In most cases, this is a direct function invocation, but
5552 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
5553 to function value.</li>
5554
5555 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]5556 signature argument types and parameter attributes. All arguments must be
5557 of <a href="#t_firstclass">first class</a> type. If the function
5558 signature indicates the function accepts a variable number of arguments,
5559 the extra arguments can be specified.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5560
5561 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
5562 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
5563 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]5564</ol>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5565
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5566<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5567<p>The '<tt>call</tt>' instruction is used to cause control flow to transfer to
5568 a specified function, with its incoming arguments bound to the specified
5569 values. Upon a '<tt><a href="#i_ret">ret</a></tt>' instruction in the called
5570 function, control flow continues with the instruction after the function
5571 call, and the return value of the function is bound to the result
5572 argument.</p>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5573
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5574<h5>Example:</h5>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5575<pre>
Nick Lewyckya9b13d52007-09-08 13:57:50 +0000[diff] [blame]5576 %retval = call i32 @test(i32 %argc)
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]5577 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) <i>; yields i32</i>
Chris Lattnerfb7c88d2008-03-21 17:24:17 +0000[diff] [blame]5578 %X = tail call i32 @foo() <i>; yields i32</i>
5579 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
5580 call void %foo(i8 97 signext)
Devang Pateld6cff512008-03-10 20:49:15 +0000[diff] [blame]5581
5582 %struct.A = type { i32, i8 }
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]5583 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohmancc3132e2008-10-04 19:00:07 +0000[diff] [blame]5584 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
5585 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattner6cbe8e92008-10-08 06:26:11 +0000[diff] [blame]5586 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijmaneefa7df2008-10-07 10:03:45 +0000[diff] [blame]5587 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5588</pre>
5589
Dale Johannesen68f971b2009-09-24 18:38:21 +0000[diff] [blame]5590<p>llvm treats calls to some functions with names and arguments that match the
Dale Johannesen722212d2009-09-25 17:04:42 +0000[diff] [blame]5591standard C99 library as being the C99 library functions, and may perform
5592optimizations or generate code for them under that assumption. This is
5593something we'd like to change in the future to provide better support for
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]5594freestanding environments and non-C-based languages.</p>
Dale Johannesen68f971b2009-09-24 18:38:21 +0000[diff] [blame]5595
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5596</div>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5597
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5598<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5599<h4>
Chris Lattner33337472006-01-13 23:26:01 +0000[diff] [blame]5600 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5601</h4>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5602
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5603<div>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5604
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]5605<h5>Syntax:</h5>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5606<pre>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]5607 &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]5608</pre>
5609
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]5610<h5>Overview:</h5>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]5611<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5612 the "variable argument" area of a function call. It is used to implement the
5613 <tt>va_arg</tt> macro in C.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5614
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]5615<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5616<p>This instruction takes a <tt>va_list*</tt> value and the type of the
5617 argument. It returns a value of the specified argument type and increments
5618 the <tt>va_list</tt> to point to the next argument. The actual type
5619 of <tt>va_list</tt> is target specific.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5620
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]5621<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5622<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified type
5623 from the specified <tt>va_list</tt> and causes the <tt>va_list</tt> to point
5624 to the next argument. For more information, see the variable argument
5625 handling <a href="#int_varargs">Intrinsic Functions</a>.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5626
5627<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]5628 take a variable number of arguments, for example, the <tt>vfprintf</tt>
5629 function.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5630
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5631<p><tt>va_arg</tt> is an LLVM instruction instead of
5632 an <a href="#intrinsics">intrinsic function</a> because it takes a type as an
5633 argument.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5634
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]5635<h5>Example:</h5>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5636<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
5637
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5638<p>Note that the code generator does not yet fully support va_arg on many
5639 targets. Also, it does not currently support va_arg with aggregate types on
5640 any target.</p>
Dan Gohman3065b612009-01-12 23:12:39 +0000[diff] [blame]5641
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5642</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5643
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5644</div>
5645
5646</div>
5647
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5648<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5649<h2><a name="intrinsics">Intrinsic Functions</a></h2>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]5650<!-- *********************************************************************** -->
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5651
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5652<div>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5653
5654<p>LLVM supports the notion of an "intrinsic function". These functions have
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5655 well known names and semantics and are required to follow certain
5656 restrictions. Overall, these intrinsics represent an extension mechanism for
5657 the LLVM language that does not require changing all of the transformations
5658 in LLVM when adding to the language (or the bitcode reader/writer, the
5659 parser, etc...).</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5660
John Criswell88190562005-05-16 16:17:45 +0000[diff] [blame]5661<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5662 prefix is reserved in LLVM for intrinsic names; thus, function names may not
5663 begin with this prefix. Intrinsic functions must always be external
5664 functions: you cannot define the body of intrinsic functions. Intrinsic
5665 functions may only be used in call or invoke instructions: it is illegal to
5666 take the address of an intrinsic function. Additionally, because intrinsic
5667 functions are part of the LLVM language, it is required if any are added that
5668 they be documented here.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5669
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5670<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents a
5671 family of functions that perform the same operation but on different data
5672 types. Because LLVM can represent over 8 million different integer types,
5673 overloading is used commonly to allow an intrinsic function to operate on any
5674 integer type. One or more of the argument types or the result type can be
5675 overloaded to accept any integer type. Argument types may also be defined as
5676 exactly matching a previous argument's type or the result type. This allows
5677 an intrinsic function which accepts multiple arguments, but needs all of them
5678 to be of the same type, to only be overloaded with respect to a single
5679 argument or the result.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5680
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5681<p>Overloaded intrinsics will have the names of its overloaded argument types
5682 encoded into its function name, each preceded by a period. Only those types
5683 which are overloaded result in a name suffix. Arguments whose type is matched
5684 against another type do not. For example, the <tt>llvm.ctpop</tt> function
5685 can take an integer of any width and returns an integer of exactly the same
5686 integer width. This leads to a family of functions such as
5687 <tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29
5688 %val)</tt>. Only one type, the return type, is overloaded, and only one type
5689 suffix is required. Because the argument's type is matched against the return
5690 type, it does not require its own name suffix.</p>
Reid Spencer4eefaab2007-04-01 08:04:23 +0000[diff] [blame]5691
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5692<p>To learn how to add an intrinsic function, please see the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5693 <a href="ExtendingLLVM.html">Extending LLVM Guide</a>.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5694
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5695<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5696<h3>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5697 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5698</h3>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5699
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5700<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5701
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5702<p>Variable argument support is defined in LLVM with
5703 the <a href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
5704 intrinsic functions. These functions are related to the similarly named
5705 macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5706
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5707<p>All of these functions operate on arguments that use a target-specific value
5708 type "<tt>va_list</tt>". The LLVM assembly language reference manual does
5709 not define what this type is, so all transformations should be prepared to
5710 handle these functions regardless of the type used.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5711
Chris Lattner30b868d2006-05-15 17:26:46 +0000[diff] [blame]5712<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]5713 instruction and the variable argument handling intrinsic functions are
5714 used.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5715
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]5716<pre class="doc_code">
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]5717define i32 @test(i32 %X, ...) {
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5718 ; Initialize variable argument processing
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5719 %ap = alloca i8*
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]5720 %ap2 = bitcast i8** %ap to i8*
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]5721 call void @llvm.va_start(i8* %ap2)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5722
5723 ; Read a single integer argument
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5724 %tmp = va_arg i8** %ap, i32
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5725
5726 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5727 %aq = alloca i8*
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]5728 %aq2 = bitcast i8** %aq to i8*
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5729 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]5730 call void @llvm.va_end(i8* %aq2)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5731
5732 ; Stop processing of arguments.
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]5733 call void @llvm.va_end(i8* %ap2)
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5734 ret i32 %tmp
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5735}
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]5736
5737declare void @llvm.va_start(i8*)
5738declare void @llvm.va_copy(i8*, i8*)
5739declare void @llvm.va_end(i8*)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5740</pre>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5741
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5742<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5743<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5744 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5745</h4>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5746
5747
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5748<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5749
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5750<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5751<pre>
5752 declare void %llvm.va_start(i8* &lt;arglist&gt;)
5753</pre>
5754
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5755<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5756<p>The '<tt>llvm.va_start</tt>' intrinsic initializes <tt>*&lt;arglist&gt;</tt>
5757 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]5758
5759<h5>Arguments:</h5>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]5760<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]5761
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5762<h5>Semantics:</h5>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]5763<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]5764 macro available in C. In a target-dependent way, it initializes
5765 the <tt>va_list</tt> element to which the argument points, so that the next
5766 call to <tt>va_arg</tt> will produce the first variable argument passed to
5767 the function. Unlike the C <tt>va_start</tt> macro, this intrinsic does not
5768 need to know the last argument of the function as the compiler can figure
5769 that out.</p>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]5770
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5771</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5772
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5773<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5774<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5775 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5776</h4>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5777
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5778<div>
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]5779
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5780<h5>Syntax:</h5>
5781<pre>
5782 declare void @llvm.va_end(i8* &lt;arglist&gt;)
5783</pre>
5784
5785<h5>Overview:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5786<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5787 which has been initialized previously
5788 with <tt><a href="#int_va_start">llvm.va_start</a></tt>
5789 or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]5790
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5791<h5>Arguments:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5792<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]5793
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5794<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5795<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]5796 macro available in C. In a target-dependent way, it destroys
5797 the <tt>va_list</tt> element to which the argument points. Calls
5798 to <a href="#int_va_start"><tt>llvm.va_start</tt></a>
5799 and <a href="#int_va_copy"> <tt>llvm.va_copy</tt></a> must be matched exactly
5800 with calls to <tt>llvm.va_end</tt>.</p>
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]5801
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5802</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5803
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5804<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5805<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5806 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5807</h4>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5808
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5809<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5810
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5811<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5812<pre>
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]5813 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5814</pre>
5815
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5816<h5>Overview:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5817<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5818 from the source argument list to the destination argument list.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5819
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5820<h5>Arguments:</h5>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]5821<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]5822 The second argument is a pointer to a <tt>va_list</tt> element to copy
5823 from.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5824
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5825<h5>Semantics:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5826<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]5827 macro available in C. In a target-dependent way, it copies the
5828 source <tt>va_list</tt> element into the destination <tt>va_list</tt>
5829 element. This intrinsic is necessary because
5830 the <tt><a href="#int_va_start"> llvm.va_start</a></tt> intrinsic may be
5831 arbitrarily complex and require, for example, memory allocation.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5832
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5833</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5834
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5835</div>
5836
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5837<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5838<h3>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5839 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5840</h3>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5841
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5842<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5843
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5844<p>LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]5845Collection</a> (GC) requires the implementation and generation of these
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5846intrinsics. These intrinsics allow identification of <a href="#int_gcroot">GC
5847roots on the stack</a>, as well as garbage collector implementations that
5848require <a href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a>
5849barriers. Front-ends for type-safe garbage collected languages should generate
5850these intrinsics to make use of the LLVM garbage collectors. For more details,
5851see <a href="GarbageCollection.html">Accurate Garbage Collection with
5852LLVM</a>.</p>
Christopher Lamb55c6d4f2007-12-17 01:00:21 +0000[diff] [blame]5853
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5854<p>The garbage collection intrinsics only operate on objects in the generic
5855 address space (address space zero).</p>
Christopher Lamb55c6d4f2007-12-17 01:00:21 +0000[diff] [blame]5856
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5857<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5858<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5859 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5860</h4>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5861
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5862<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5863
5864<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5865<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]5866 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5867</pre>
5868
5869<h5>Overview:</h5>
John Criswelldfe6a862004-12-10 15:51:16 +0000[diff] [blame]5870<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]5871 the code generator, and allows some metadata to be associated with it.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5872
5873<h5>Arguments:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5874<p>The first argument specifies the address of a stack object that contains the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5875 root pointer. The second pointer (which must be either a constant or a
5876 global value address) contains the meta-data to be associated with the
5877 root.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5878
5879<h5>Semantics:</h5>
Chris Lattner851b7712008-04-24 05:59:56 +0000[diff] [blame]5880<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]5881 location. At compile-time, the code generator generates information to allow
5882 the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
5883 intrinsic may only be used in a function which <a href="#gc">specifies a GC
5884 algorithm</a>.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5885
5886</div>
5887
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5888<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5889<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5890 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5891</h4>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5892
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5893<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5894
5895<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5896<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]5897 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5898</pre>
5899
5900<h5>Overview:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5901<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5902 locations, allowing garbage collector implementations that require read
5903 barriers.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5904
5905<h5>Arguments:</h5>
Chris Lattnerf9228072006-03-14 20:02:51 +0000[diff] [blame]5906<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]5907 allocated from the garbage collector. The first object is a pointer to the
5908 start of the referenced object, if needed by the language runtime (otherwise
5909 null).</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5910
5911<h5>Semantics:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5912<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5913 instruction, but may be replaced with substantially more complex code by the
5914 garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
5915 may only be used in a function which <a href="#gc">specifies a GC
5916 algorithm</a>.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5917
5918</div>
5919
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5920<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5921<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5922 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5923</h4>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5924
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5925<div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5926
5927<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5928<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]5929 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5930</pre>
5931
5932<h5>Overview:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5933<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5934 locations, allowing garbage collector implementations that require write
5935 barriers (such as generational or reference counting collectors).</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5936
5937<h5>Arguments:</h5>
Chris Lattnerf9228072006-03-14 20:02:51 +0000[diff] [blame]5938<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]5939 object to store it to, and the third is the address of the field of Obj to
5940 store to. If the runtime does not require a pointer to the object, Obj may
5941 be null.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5942
5943<h5>Semantics:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5944<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5945 instruction, but may be replaced with substantially more complex code by the
5946 garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
5947 may only be used in a function which <a href="#gc">specifies a GC
5948 algorithm</a>.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5949
5950</div>
5951
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5952</div>
5953
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5954<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5955<h3>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5956 <a name="int_codegen">Code Generator Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5957</h3>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5958
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5959<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5960
5961<p>These intrinsics are provided by LLVM to expose special features that may
5962 only be implemented with code generator support.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5963
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5964<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5965<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5966 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5967</h4>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5968
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]5969<div>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5970
5971<h5>Syntax:</h5>
5972<pre>
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]5973 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5974</pre>
5975
5976<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5977<p>The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
5978 target-specific value indicating the return address of the current function
5979 or one of its callers.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5980
5981<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5982<p>The argument to this intrinsic indicates which function to return the address
5983 for. Zero indicates the calling function, one indicates its caller, etc.
5984 The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5985
5986<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5987<p>The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer
5988 indicating the return address of the specified call frame, or zero if it
5989 cannot be identified. The value returned by this intrinsic is likely to be
5990 incorrect or 0 for arguments other than zero, so it should only be used for
5991 debugging purposes.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5992
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5993<p>Note that calling this intrinsic does not prevent function inlining or other
5994 aggressive transformations, so the value returned may not be that of the
5995 obvious source-language caller.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5996
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5997</div>
5998
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5999<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6000<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6001 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6002</h4>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6003
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6004<div>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6005
6006<h5>Syntax:</h5>
6007<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6008 declare i8* @llvm.frameaddress(i32 &lt;level&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6009</pre>
6010
6011<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6012<p>The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
6013 target-specific frame pointer value for the specified stack frame.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6014
6015<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6016<p>The argument to this intrinsic indicates which function to return the frame
6017 pointer for. Zero indicates the calling function, one indicates its caller,
6018 etc. The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6019
6020<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6021<p>The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer
6022 indicating the frame address of the specified call frame, or zero if it
6023 cannot be identified. The value returned by this intrinsic is likely to be
6024 incorrect or 0 for arguments other than zero, so it should only be used for
6025 debugging purposes.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6026
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6027<p>Note that calling this intrinsic does not prevent function inlining or other
6028 aggressive transformations, so the value returned may not be that of the
6029 obvious source-language caller.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6030
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6031</div>
6032
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6033<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6034<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6035 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6036</h4>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6037
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6038<div>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6039
6040<h5>Syntax:</h5>
6041<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6042 declare i8* @llvm.stacksave()
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6043</pre>
6044
6045<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6046<p>The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state
6047 of the function stack, for use
6048 with <a href="#int_stackrestore"> <tt>llvm.stackrestore</tt></a>. This is
6049 useful for implementing language features like scoped automatic variable
6050 sized arrays in C99.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6051
6052<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6053<p>This intrinsic returns a opaque pointer value that can be passed
6054 to <a href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When
6055 an <tt>llvm.stackrestore</tt> intrinsic is executed with a value saved
6056 from <tt>llvm.stacksave</tt>, it effectively restores the state of the stack
6057 to the state it was in when the <tt>llvm.stacksave</tt> intrinsic executed.
6058 In practice, this pops any <a href="#i_alloca">alloca</a> blocks from the
6059 stack that were allocated after the <tt>llvm.stacksave</tt> was executed.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6060
6061</div>
6062
6063<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6064<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6065 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6066</h4>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6067
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6068<div>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6069
6070<h5>Syntax:</h5>
6071<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6072 declare void @llvm.stackrestore(i8* %ptr)
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6073</pre>
6074
6075<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6076<p>The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
6077 the function stack to the state it was in when the
6078 corresponding <a href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic
6079 executed. This is useful for implementing language features like scoped
6080 automatic variable sized arrays in C99.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6081
6082<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6083<p>See the description
6084 for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6085
6086</div>
6087
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6088<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6089<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6090 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6091</h4>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6092
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6093<div>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6094
6095<h5>Syntax:</h5>
6096<pre>
Bruno Cardoso Lopesdc9ff3a2011-06-14 04:58:37 +0000[diff] [blame]6097 declare void @llvm.prefetch(i8* &lt;address&gt;, i32 &lt;rw&gt;, i32 &lt;locality&gt;, i32 &lt;cache type&gt;)
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6098</pre>
6099
6100<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6101<p>The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to
6102 insert a prefetch instruction if supported; otherwise, it is a noop.
6103 Prefetches have no effect on the behavior of the program but can change its
6104 performance characteristics.</p>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6105
6106<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6107<p><tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the
6108 specifier determining if the fetch should be for a read (0) or write (1),
6109 and <tt>locality</tt> is a temporal locality specifier ranging from (0) - no
Bruno Cardoso Lopesdc9ff3a2011-06-14 04:58:37 +0000[diff] [blame]6110 locality, to (3) - extremely local keep in cache. The <tt>cache type</tt>
6111 specifies whether the prefetch is performed on the data (1) or instruction (0)
6112 cache. The <tt>rw</tt>, <tt>locality</tt> and <tt>cache type</tt> arguments
6113 must be constant integers.</p>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6114
6115<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6116<p>This intrinsic does not modify the behavior of the program. In particular,
6117 prefetches cannot trap and do not produce a value. On targets that support
6118 this intrinsic, the prefetch can provide hints to the processor cache for
6119 better performance.</p>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6120
6121</div>
6122
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6123<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6124<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6125 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6126</h4>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6127
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6128<div>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6129
6130<h5>Syntax:</h5>
6131<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]6132 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6133</pre>
6134
6135<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6136<p>The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program
6137 Counter (PC) in a region of code to simulators and other tools. The method
6138 is target specific, but it is expected that the marker will use exported
6139 symbols to transmit the PC of the marker. The marker makes no guarantees
6140 that it will remain with any specific instruction after optimizations. It is
6141 possible that the presence of a marker will inhibit optimizations. The
6142 intended use is to be inserted after optimizations to allow correlations of
6143 simulation runs.</p>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6144
6145<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6146<p><tt>id</tt> is a numerical id identifying the marker.</p>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6147
6148<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6149<p>This intrinsic does not modify the behavior of the program. Backends that do
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]6150 not support this intrinsic may ignore it.</p>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6151
6152</div>
6153
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6154<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6155<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6156 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6157</h4>
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6158
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6159<div>
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6160
6161<h5>Syntax:</h5>
6162<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]6163 declare i64 @llvm.readcyclecounter()
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6164</pre>
6165
6166<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6167<p>The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
6168 counter register (or similar low latency, high accuracy clocks) on those
6169 targets that support it. On X86, it should map to RDTSC. On Alpha, it
6170 should map to RPCC. As the backing counters overflow quickly (on the order
6171 of 9 seconds on alpha), this should only be used for small timings.</p>
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6172
6173<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6174<p>When directly supported, reading the cycle counter should not modify any
6175 memory. Implementations are allowed to either return a application specific
6176 value or a system wide value. On backends without support, this is lowered
6177 to a constant 0.</p>
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6178
6179</div>
6180
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6181</div>
6182
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6183<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6184<h3>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6185 <a name="int_libc">Standard C Library Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6186</h3>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6187
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6188<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6189
6190<p>LLVM provides intrinsics for a few important standard C library functions.
6191 These intrinsics allow source-language front-ends to pass information about
6192 the alignment of the pointer arguments to the code generator, providing
6193 opportunity for more efficient code generation.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6194
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6195<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6196<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6197 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6198</h4>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6199
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6200<div>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6201
6202<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6203<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]6204 integer bit width and for different address spaces. Not all targets support
6205 all bit widths however.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6206
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6207<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6208 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]6209 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6210 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]6211 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6212</pre>
6213
6214<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6215<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6216 source location to the destination location.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6217
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6218<p>Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6219 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6220 and the pointers can be in specified address spaces.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6221
6222<h5>Arguments:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6223
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6224<p>The first argument is a pointer to the destination, the second is a pointer
6225 to the source. The third argument is an integer argument specifying the
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6226 number of bytes to copy, the fourth argument is the alignment of the
6227 source and destination locations, and the fifth is a boolean indicating a
6228 volatile access.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6229
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]6230<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]6231 then the caller guarantees that both the source and destination pointers are
6232 aligned to that boundary.</p>
Chris Lattner4c67c482004-02-12 21:18:15 +0000[diff] [blame]6233
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]6234<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6235 <tt>llvm.memcpy</tt> call is a <a href="#volatile">volatile operation</a>.
6236 The detailed access behavior is not very cleanly specified and it is unwise
6237 to depend on it.</p>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6238
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6239<h5>Semantics:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6240
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6241<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6242 source location to the destination location, which are not allowed to
6243 overlap. It copies "len" bytes of memory over. If the argument is known to
6244 be aligned to some boundary, this can be specified as the fourth argument,
6245 otherwise it should be set to 0 or 1.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6246
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6247</div>
6248
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6249<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6250<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6251 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6252</h4>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6253
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6254<div>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6255
6256<h5>Syntax:</h5>
Chris Lattnerdd708342008-11-21 16:42:48 +0000[diff] [blame]6257<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]6258 width and for different address space. Not all targets support all bit
6259 widths however.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6260
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6261<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6262 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]6263 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6264 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]6265 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6266</pre>
6267
6268<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6269<p>The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the
6270 source location to the destination location. It is similar to the
6271 '<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to
6272 overlap.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6273
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6274<p>Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6275 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6276 and the pointers can be in specified address spaces.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6277
6278<h5>Arguments:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6279
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6280<p>The first argument is a pointer to the destination, the second is a pointer
6281 to the source. The third argument is an integer argument specifying the
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6282 number of bytes to copy, the fourth argument is the alignment of the
6283 source and destination locations, and the fifth is a boolean indicating a
6284 volatile access.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6285
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]6286<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]6287 then the caller guarantees that the source and destination pointers are
6288 aligned to that boundary.</p>
Chris Lattner4c67c482004-02-12 21:18:15 +0000[diff] [blame]6289
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]6290<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6291 <tt>llvm.memmove</tt> call is a <a href="#volatile">volatile operation</a>.
6292 The detailed access behavior is not very cleanly specified and it is unwise
6293 to depend on it.</p>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6294
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6295<h5>Semantics:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6296
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6297<p>The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the
6298 source location to the destination location, which may overlap. It copies
6299 "len" bytes of memory over. If the argument is known to be aligned to some
6300 boundary, this can be specified as the fourth argument, otherwise it should
6301 be set to 0 or 1.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6302
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6303</div>
6304
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6305<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6306<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6307 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6308</h4>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6309
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6310<div>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6311
6312<h5>Syntax:</h5>
Chris Lattnerdd708342008-11-21 16:42:48 +0000[diff] [blame]6313<p>This is an overloaded intrinsic. You can use llvm.memset on any integer bit
John Criswellad05ae42010-07-30 16:30:28 +0000[diff] [blame]6314 width and for different address spaces. However, not all targets support all
6315 bit widths.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6316
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6317<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6318 declare void @llvm.memset.p0i8.i32(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattner685db9d2010-04-08 00:54:34 +0000[diff] [blame]6319 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6320 declare void @llvm.memset.p0i8.i64(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattner685db9d2010-04-08 00:54:34 +0000[diff] [blame]6321 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6322</pre>
6323
6324<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6325<p>The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a
6326 particular byte value.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6327
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6328<p>Note that, unlike the standard libc function, the <tt>llvm.memset</tt>
John Criswellad05ae42010-07-30 16:30:28 +0000[diff] [blame]6329 intrinsic does not return a value and takes extra alignment/volatile
6330 arguments. Also, the destination can be in an arbitrary address space.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6331
6332<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6333<p>The first argument is a pointer to the destination to fill, the second is the
John Criswellad05ae42010-07-30 16:30:28 +0000[diff] [blame]6334 byte value with which to fill it, the third argument is an integer argument
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6335 specifying the number of bytes to fill, and the fourth argument is the known
John Criswellad05ae42010-07-30 16:30:28 +0000[diff] [blame]6336 alignment of the destination location.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6337
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]6338<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]6339 then the caller guarantees that the destination pointer is aligned to that
6340 boundary.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6341
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]6342<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6343 <tt>llvm.memset</tt> call is a <a href="#volatile">volatile operation</a>.
6344 The detailed access behavior is not very cleanly specified and it is unwise
6345 to depend on it.</p>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6346
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6347<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6348<p>The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting
6349 at the destination location. If the argument is known to be aligned to some
6350 boundary, this can be specified as the fourth argument, otherwise it should
6351 be set to 0 or 1.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6352
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6353</div>
6354
Chris Lattner3b4f4372004-06-11 02:28:03 +0000[diff] [blame]6355<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6356<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6357 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6358</h4>
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6359
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6360<div>
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6361
6362<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6363<p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
6364 floating point or vector of floating point type. Not all targets support all
6365 types however.</p>
6366
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6367<pre>
Dale Johannesendd89d272007-10-02 17:47:38 +0000[diff] [blame]6368 declare float @llvm.sqrt.f32(float %Val)
6369 declare double @llvm.sqrt.f64(double %Val)
6370 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
6371 declare fp128 @llvm.sqrt.f128(fp128 %Val)
6372 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6373</pre>
6374
6375<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6376<p>The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
6377 returning the same value as the libm '<tt>sqrt</tt>' functions would.
6378 Unlike <tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined
6379 behavior for negative numbers other than -0.0 (which allows for better
6380 optimization, because there is no need to worry about errno being
6381 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]6382
6383<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6384<p>The argument and return value are floating point numbers of the same
6385 type.</p>
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6386
6387<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6388<p>This function returns the sqrt of the specified operand if it is a
6389 nonnegative floating point number.</p>
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6390
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6391</div>
6392
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6393<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6394<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6395 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6396</h4>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6397
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6398<div>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6399
6400<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6401<p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
6402 floating point or vector of floating point type. Not all targets support all
6403 types however.</p>
6404
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6405<pre>
Dale Johannesendd89d272007-10-02 17:47:38 +0000[diff] [blame]6406 declare float @llvm.powi.f32(float %Val, i32 %power)
6407 declare double @llvm.powi.f64(double %Val, i32 %power)
6408 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
6409 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
6410 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6411</pre>
6412
6413<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6414<p>The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
6415 specified (positive or negative) power. The order of evaluation of
6416 multiplications is not defined. When a vector of floating point type is
6417 used, the second argument remains a scalar integer value.</p>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6418
6419<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6420<p>The second argument is an integer power, and the first is a value to raise to
6421 that power.</p>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6422
6423<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6424<p>This function returns the first value raised to the second power with an
6425 unspecified sequence of rounding operations.</p>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6426
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6427</div>
6428
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6429<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6430<h4>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6431 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6432</h4>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6433
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6434<div>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6435
6436<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6437<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
6438 floating point or vector of floating point type. Not all targets support all
6439 types however.</p>
6440
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6441<pre>
6442 declare float @llvm.sin.f32(float %Val)
6443 declare double @llvm.sin.f64(double %Val)
6444 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
6445 declare fp128 @llvm.sin.f128(fp128 %Val)
6446 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
6447</pre>
6448
6449<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6450<p>The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6451
6452<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6453<p>The argument and return value are floating point numbers of the same
6454 type.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6455
6456<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6457<p>This function returns the sine of the specified operand, returning the same
6458 values as the libm <tt>sin</tt> functions would, and handles error conditions
6459 in the same way.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6460
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6461</div>
6462
6463<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6464<h4>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6465 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6466</h4>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6467
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6468<div>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6469
6470<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6471<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
6472 floating point or vector of floating point type. Not all targets support all
6473 types however.</p>
6474
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6475<pre>
6476 declare float @llvm.cos.f32(float %Val)
6477 declare double @llvm.cos.f64(double %Val)
6478 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
6479 declare fp128 @llvm.cos.f128(fp128 %Val)
6480 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
6481</pre>
6482
6483<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6484<p>The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6485
6486<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6487<p>The argument and return value are floating point numbers of the same
6488 type.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6489
6490<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6491<p>This function returns the cosine of the specified operand, returning the same
6492 values as the libm <tt>cos</tt> functions would, and handles error conditions
6493 in the same way.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6494
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6495</div>
6496
6497<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6498<h4>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6499 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6500</h4>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6501
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6502<div>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6503
6504<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6505<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
6506 floating point or vector of floating point type. Not all targets support all
6507 types however.</p>
6508
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6509<pre>
6510 declare float @llvm.pow.f32(float %Val, float %Power)
6511 declare double @llvm.pow.f64(double %Val, double %Power)
6512 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
6513 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
6514 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
6515</pre>
6516
6517<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6518<p>The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
6519 specified (positive or negative) power.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6520
6521<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6522<p>The second argument is a floating point power, and the first is a value to
6523 raise to that power.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6524
6525<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6526<p>This function returns the first value raised to the second power, returning
6527 the same values as the libm <tt>pow</tt> functions would, and handles error
6528 conditions in the same way.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6529
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6530</div>
6531
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6532</div>
6533
Dan Gohman911fa902011-05-23 21:13:03 +0000[diff] [blame]6534<!-- _______________________________________________________________________ -->
6535<h4>
6536 <a name="int_exp">'<tt>llvm.exp.*</tt>' Intrinsic</a>
6537</h4>
6538
6539<div>
6540
6541<h5>Syntax:</h5>
6542<p>This is an overloaded intrinsic. You can use <tt>llvm.exp</tt> on any
6543 floating point or vector of floating point type. Not all targets support all
6544 types however.</p>
6545
6546<pre>
6547 declare float @llvm.exp.f32(float %Val)
6548 declare double @llvm.exp.f64(double %Val)
6549 declare x86_fp80 @llvm.exp.f80(x86_fp80 %Val)
6550 declare fp128 @llvm.exp.f128(fp128 %Val)
6551 declare ppc_fp128 @llvm.exp.ppcf128(ppc_fp128 %Val)
6552</pre>
6553
6554<h5>Overview:</h5>
6555<p>The '<tt>llvm.exp.*</tt>' intrinsics perform the exp function.</p>
6556
6557<h5>Arguments:</h5>
6558<p>The argument and return value are floating point numbers of the same
6559 type.</p>
6560
6561<h5>Semantics:</h5>
6562<p>This function returns the same values as the libm <tt>exp</tt> functions
6563 would, and handles error conditions in the same way.</p>
6564
6565</div>
6566
6567<!-- _______________________________________________________________________ -->
6568<h4>
6569 <a name="int_log">'<tt>llvm.log.*</tt>' Intrinsic</a>
6570</h4>
6571
6572<div>
6573
6574<h5>Syntax:</h5>
6575<p>This is an overloaded intrinsic. You can use <tt>llvm.log</tt> on any
6576 floating point or vector of floating point type. Not all targets support all
6577 types however.</p>
6578
6579<pre>
6580 declare float @llvm.log.f32(float %Val)
6581 declare double @llvm.log.f64(double %Val)
6582 declare x86_fp80 @llvm.log.f80(x86_fp80 %Val)
6583 declare fp128 @llvm.log.f128(fp128 %Val)
6584 declare ppc_fp128 @llvm.log.ppcf128(ppc_fp128 %Val)
6585</pre>
6586
6587<h5>Overview:</h5>
6588<p>The '<tt>llvm.log.*</tt>' intrinsics perform the log function.</p>
6589
6590<h5>Arguments:</h5>
6591<p>The argument and return value are floating point numbers of the same
6592 type.</p>
6593
6594<h5>Semantics:</h5>
6595<p>This function returns the same values as the libm <tt>log</tt> functions
6596 would, and handles error conditions in the same way.</p>
6597
Cameron Zwarichf03fa182011-07-08 21:39:21 +0000[diff] [blame]6598<h4>
6599 <a name="int_fma">'<tt>llvm.fma.*</tt>' Intrinsic</a>
6600</h4>
6601
6602<div>
6603
6604<h5>Syntax:</h5>
6605<p>This is an overloaded intrinsic. You can use <tt>llvm.fma</tt> on any
6606 floating point or vector of floating point type. Not all targets support all
6607 types however.</p>
6608
6609<pre>
6610 declare float @llvm.fma.f32(float %a, float %b, float %c)
6611 declare double @llvm.fma.f64(double %a, double %b, double %c)
6612 declare x86_fp80 @llvm.fma.f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c)
6613 declare fp128 @llvm.fma.f128(fp128 %a, fp128 %b, fp128 %c)
6614 declare ppc_fp128 @llvm.fma.ppcf128(ppc_fp128 %a, ppc_fp128 %b, ppc_fp128 %c)
6615</pre>
6616
6617<h5>Overview:</h5>
Cameron Zwaricha32fd212011-07-08 22:13:55 +0000[diff] [blame]6618<p>The '<tt>llvm.fma.*</tt>' intrinsics perform the fused multiply-add
Cameron Zwarichf03fa182011-07-08 21:39:21 +0000[diff] [blame]6619 operation.</p>
6620
6621<h5>Arguments:</h5>
6622<p>The argument and return value are floating point numbers of the same
6623 type.</p>
6624
6625<h5>Semantics:</h5>
6626<p>This function returns the same values as the libm <tt>fma</tt> functions
6627 would.</p>
6628
Dan Gohman911fa902011-05-23 21:13:03 +0000[diff] [blame]6629</div>
6630
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6631<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6632<h3>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6633 <a name="int_manip">Bit Manipulation Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6634</h3>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6635
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6636<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6637
6638<p>LLVM provides intrinsics for a few important bit manipulation operations.
6639 These allow efficient code generation for some algorithms.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6640
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6641<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6642<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6643 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6644</h4>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6645
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6646<div>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6647
6648<h5>Syntax:</h5>
Reid Spencer4eefaab2007-04-01 08:04:23 +0000[diff] [blame]6649<p>This is an overloaded intrinsic function. You can use bswap on any integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6650 type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
6651
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6652<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6653 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
6654 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
6655 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6656</pre>
6657
6658<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6659<p>The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
6660 values with an even number of bytes (positive multiple of 16 bits). These
6661 are useful for performing operations on data that is not in the target's
6662 native byte order.</p>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6663
6664<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6665<p>The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
6666 and low byte of the input i16 swapped. Similarly,
6667 the <tt>llvm.bswap.i32</tt> intrinsic returns an i32 value that has the four
6668 bytes of the input i32 swapped, so that if the input bytes are numbered 0, 1,
6669 2, 3 then the returned i32 will have its bytes in 3, 2, 1, 0 order.
6670 The <tt>llvm.bswap.i48</tt>, <tt>llvm.bswap.i64</tt> and other intrinsics
6671 extend this concept to additional even-byte lengths (6 bytes, 8 bytes and
6672 more, respectively).</p>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6673
6674</div>
6675
6676<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6677<h4>
Reid Spencerb4f9a6f2006-01-16 21:12:35 +0000[diff] [blame]6678 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6679</h4>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6680
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6681<div>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6682
6683<h5>Syntax:</h5>
Reid Spencer4eefaab2007-04-01 08:04:23 +0000[diff] [blame]6684<p>This is an overloaded intrinsic. You can use llvm.ctpop on any integer bit
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]6685 width, or on any vector with integer elements. Not all targets support all
6686 bit widths or vector types, however.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6687
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6688<pre>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6689 declare i8 @llvm.ctpop.i8(i8 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6690 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]6691 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6692 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
6693 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]6694 declare &lt;2 x i32&gt; @llvm.ctpop.v2i32(&lt;2 x i32&gt; &lt;src&gt;)
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6695</pre>
6696
6697<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6698<p>The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set
6699 in a value.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6700
6701<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6702<p>The only argument is the value to be counted. The argument may be of any
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]6703 integer type, or a vector with integer elements.
6704 The return type must match the argument type.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6705
6706<h5>Semantics:</h5>
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]6707<p>The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable, or within each
6708 element of a vector.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6709
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6710</div>
6711
6712<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6713<h4>
Chris Lattnerb748c672006-01-16 22:34:14 +0000[diff] [blame]6714 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6715</h4>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6716
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6717<div>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6718
6719<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6720<p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]6721 integer bit width, or any vector whose elements are integers. Not all
6722 targets support all bit widths or vector types, however.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6723
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6724<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6725 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
6726 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]6727 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6728 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
6729 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]6730 declare &lt;2 x i32&gt; @llvm.ctlz.v2i32(&lt;2 x i32&gt; &lt;src;gt)
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6731</pre>
6732
6733<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6734<p>The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
6735 leading zeros in a variable.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6736
6737<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6738<p>The only argument is the value to be counted. The argument may be of any
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]6739 integer type, or any vector type with integer element type.
6740 The return type must match the argument type.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6741
6742<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6743<p>The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant)
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]6744 zeros in a variable, or within each element of the vector if the operation
6745 is of vector type. If the src == 0 then the result is the size in bits of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6746 the type of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6747
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6748</div>
Chris Lattner3b4f4372004-06-11 02:28:03 +0000[diff] [blame]6749
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6750<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6751<h4>
Chris Lattnerb748c672006-01-16 22:34:14 +0000[diff] [blame]6752 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6753</h4>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6754
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6755<div>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6756
6757<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6758<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]6759 integer bit width, or any vector of integer elements. Not all targets
6760 support all bit widths or vector types, however.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6761
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6762<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6763 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
6764 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]6765 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6766 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
6767 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]6768 declase &lt;2 x i32&gt; @llvm.cttz.v2i32(&lt;2 x i32&gt; &lt;src&gt;)
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6769</pre>
6770
6771<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6772<p>The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
6773 trailing zeros.</p>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6774
6775<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6776<p>The only argument is the value to be counted. The argument may be of any
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]6777 integer type, or a vectory with integer element type.. The return type
6778 must match the argument type.</p>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6779
6780<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6781<p>The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant)
Owen Anderson2f37bdc2011-07-01 21:52:38 +0000[diff] [blame]6782 zeros in a variable, or within each element of a vector.
6783 If the src == 0 then the result is the size in bits of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6784 the type of src. For example, <tt>llvm.cttz(2) = 1</tt>.</p>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6785
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6786</div>
6787
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6788</div>
6789
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]6790<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6791<h3>
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]6792 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6793</h3>
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]6794
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6795<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6796
6797<p>LLVM provides intrinsics for some arithmetic with overflow operations.</p>
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]6798
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6799<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6800<h4>
6801 <a name="int_sadd_overflow">
6802 '<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics
6803 </a>
6804</h4>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6805
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6806<div>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6807
6808<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6809<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]6810 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6811
6812<pre>
6813 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
6814 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6815 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
6816</pre>
6817
6818<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6819<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6820 a signed addition of the two arguments, and indicate whether an overflow
6821 occurred during the signed summation.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6822
6823<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6824<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]6825 be of integer types of any bit width, but they must have the same bit
6826 width. The second element of the result structure must be of
6827 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6828 undergo signed addition.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6829
6830<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6831<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6832 a signed addition of the two variables. They return a structure &mdash; the
6833 first element of which is the signed summation, and the second element of
6834 which is a bit specifying if the signed summation resulted in an
6835 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6836
6837<h5>Examples:</h5>
6838<pre>
6839 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6840 %sum = extractvalue {i32, i1} %res, 0
6841 %obit = extractvalue {i32, i1} %res, 1
6842 br i1 %obit, label %overflow, label %normal
6843</pre>
6844
6845</div>
6846
6847<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6848<h4>
6849 <a name="int_uadd_overflow">
6850 '<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics
6851 </a>
6852</h4>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6853
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6854<div>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6855
6856<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6857<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]6858 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6859
6860<pre>
6861 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
6862 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6863 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
6864</pre>
6865
6866<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6867<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6868 an unsigned addition of the two arguments, and indicate whether a carry
6869 occurred during the unsigned summation.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6870
6871<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6872<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]6873 be of integer types of any bit width, but they must have the same bit
6874 width. The second element of the result structure must be of
6875 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6876 undergo unsigned addition.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6877
6878<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6879<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6880 an unsigned addition of the two arguments. They return a structure &mdash;
6881 the first element of which is the sum, and the second element of which is a
6882 bit specifying if the unsigned summation resulted in a carry.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6883
6884<h5>Examples:</h5>
6885<pre>
6886 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6887 %sum = extractvalue {i32, i1} %res, 0
6888 %obit = extractvalue {i32, i1} %res, 1
6889 br i1 %obit, label %carry, label %normal
6890</pre>
6891
6892</div>
6893
6894<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6895<h4>
6896 <a name="int_ssub_overflow">
6897 '<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics
6898 </a>
6899</h4>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6900
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6901<div>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6902
6903<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6904<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]6905 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6906
6907<pre>
6908 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
6909 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6910 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
6911</pre>
6912
6913<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6914<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6915 a signed subtraction of the two arguments, and indicate whether an overflow
6916 occurred during the signed subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6917
6918<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6919<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]6920 be of integer types of any bit width, but they must have the same bit
6921 width. The second element of the result structure must be of
6922 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6923 undergo signed subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6924
6925<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6926<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6927 a signed subtraction of the two arguments. They return a structure &mdash;
6928 the first element of which is the subtraction, and the second element of
6929 which is a bit specifying if the signed subtraction resulted in an
6930 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6931
6932<h5>Examples:</h5>
6933<pre>
6934 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6935 %sum = extractvalue {i32, i1} %res, 0
6936 %obit = extractvalue {i32, i1} %res, 1
6937 br i1 %obit, label %overflow, label %normal
6938</pre>
6939
6940</div>
6941
6942<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6943<h4>
6944 <a name="int_usub_overflow">
6945 '<tt>llvm.usub.with.overflow.*</tt>' Intrinsics
6946 </a>
6947</h4>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6948
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6949<div>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6950
6951<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6952<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]6953 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6954
6955<pre>
6956 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
6957 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6958 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
6959</pre>
6960
6961<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6962<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6963 an unsigned subtraction of the two arguments, and indicate whether an
6964 overflow occurred during the unsigned subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6965
6966<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6967<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]6968 be of integer types of any bit width, but they must have the same bit
6969 width. The second element of the result structure must be of
6970 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6971 undergo unsigned subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6972
6973<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6974<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6975 an unsigned subtraction of the two arguments. They return a structure &mdash;
6976 the first element of which is the subtraction, and the second element of
6977 which is a bit specifying if the unsigned subtraction resulted in an
6978 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6979
6980<h5>Examples:</h5>
6981<pre>
6982 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6983 %sum = extractvalue {i32, i1} %res, 0
6984 %obit = extractvalue {i32, i1} %res, 1
6985 br i1 %obit, label %overflow, label %normal
6986</pre>
6987
6988</div>
6989
6990<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6991<h4>
6992 <a name="int_smul_overflow">
6993 '<tt>llvm.smul.with.overflow.*</tt>' Intrinsics
6994 </a>
6995</h4>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6996
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]6997<div>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6998
6999<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7000<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]7001 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7002
7003<pre>
7004 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
7005 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
7006 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
7007</pre>
7008
7009<h5>Overview:</h5>
7010
7011<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7012 a signed multiplication of the two arguments, and indicate whether an
7013 overflow occurred during the signed multiplication.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7014
7015<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7016<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]7017 be of integer types of any bit width, but they must have the same bit
7018 width. The second element of the result structure must be of
7019 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7020 undergo signed multiplication.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7021
7022<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7023<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7024 a signed multiplication of the two arguments. They return a structure &mdash;
7025 the first element of which is the multiplication, and the second element of
7026 which is a bit specifying if the signed multiplication resulted in an
7027 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]7028
7029<h5>Examples:</h5>
7030<pre>
7031 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
7032 %sum = extractvalue {i32, i1} %res, 0
7033 %obit = extractvalue {i32, i1} %res, 1
7034 br i1 %obit, label %overflow, label %normal
7035</pre>
7036
Reid Spencer5bf54c82007-04-11 23:23:49 +0000[diff] [blame]7037</div>
7038
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7039<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7040<h4>
7041 <a name="int_umul_overflow">
7042 '<tt>llvm.umul.with.overflow.*</tt>' Intrinsics
7043 </a>
7044</h4>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7045
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7046<div>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7047
7048<h5>Syntax:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7049<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]7050 on any integer bit width.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7051
7052<pre>
7053 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
7054 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7055 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
7056</pre>
7057
7058<h5>Overview:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7059<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7060 a unsigned multiplication of the two arguments, and indicate whether an
7061 overflow occurred during the unsigned multiplication.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7062
7063<h5>Arguments:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7064<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]7065 be of integer types of any bit width, but they must have the same bit
7066 width. The second element of the result structure must be of
7067 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7068 undergo unsigned multiplication.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7069
7070<h5>Semantics:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7071<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7072 an unsigned multiplication of the two arguments. They return a structure
7073 &mdash; the first element of which is the multiplication, and the second
7074 element of which is a bit specifying if the unsigned multiplication resulted
7075 in an overflow.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]7076
7077<h5>Examples:</h5>
7078<pre>
7079 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7080 %sum = extractvalue {i32, i1} %res, 0
7081 %obit = extractvalue {i32, i1} %res, 1
7082 br i1 %obit, label %overflow, label %normal
7083</pre>
7084
7085</div>
7086
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7087</div>
7088
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]7089<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7090<h3>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7091 <a name="int_fp16">Half Precision Floating Point Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7092</h3>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7093
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7094<div>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7095
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]7096<p>Half precision floating point is a storage-only format. This means that it is
7097 a dense encoding (in memory) but does not support computation in the
7098 format.</p>
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]7099
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]7100<p>This means that code must first load the half-precision floating point
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]7101 value as an i16, then convert it to float with <a
7102 href="#int_convert_from_fp16"><tt>llvm.convert.from.fp16</tt></a>.
7103 Computation can then be performed on the float value (including extending to
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]7104 double etc). To store the value back to memory, it is first converted to
7105 float if needed, then converted to i16 with
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]7106 <a href="#int_convert_to_fp16"><tt>llvm.convert.to.fp16</tt></a>, then
7107 storing as an i16 value.</p>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7108
7109<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7110<h4>
7111 <a name="int_convert_to_fp16">
7112 '<tt>llvm.convert.to.fp16</tt>' Intrinsic
7113 </a>
7114</h4>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7115
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7116<div>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7117
7118<h5>Syntax:</h5>
7119<pre>
7120 declare i16 @llvm.convert.to.fp16(f32 %a)
7121</pre>
7122
7123<h5>Overview:</h5>
7124<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
7125 a conversion from single precision floating point format to half precision
7126 floating point format.</p>
7127
7128<h5>Arguments:</h5>
7129<p>The intrinsic function contains single argument - the value to be
7130 converted.</p>
7131
7132<h5>Semantics:</h5>
7133<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
7134 a conversion from single precision floating point format to half precision
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]7135 floating point format. The return value is an <tt>i16</tt> which
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]7136 contains the converted number.</p>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7137
7138<h5>Examples:</h5>
7139<pre>
7140 %res = call i16 @llvm.convert.to.fp16(f32 %a)
7141 store i16 %res, i16* @x, align 2
7142</pre>
7143
7144</div>
7145
7146<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7147<h4>
7148 <a name="int_convert_from_fp16">
7149 '<tt>llvm.convert.from.fp16</tt>' Intrinsic
7150 </a>
7151</h4>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7152
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7153<div>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7154
7155<h5>Syntax:</h5>
7156<pre>
7157 declare f32 @llvm.convert.from.fp16(i16 %a)
7158</pre>
7159
7160<h5>Overview:</h5>
7161<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs
7162 a conversion from half precision floating point format to single precision
7163 floating point format.</p>
7164
7165<h5>Arguments:</h5>
7166<p>The intrinsic function contains single argument - the value to be
7167 converted.</p>
7168
7169<h5>Semantics:</h5>
7170<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs a
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]7171 conversion from half single precision floating point format to single
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]7172 precision floating point format. The input half-float value is represented by
7173 an <tt>i16</tt> value.</p>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7174
7175<h5>Examples:</h5>
7176<pre>
7177 %a = load i16* @x, align 2
7178 %res = call f32 @llvm.convert.from.fp16(i16 %a)
7179</pre>
7180
7181</div>
7182
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7183</div>
7184
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7185<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7186<h3>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]7187 <a name="int_debugger">Debugger Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7188</h3>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]7189
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7190<div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]7191
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7192<p>The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt>
7193 prefix), are described in
7194 the <a href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source
7195 Level Debugging</a> document.</p>
7196
7197</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]7198
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]7199<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7200<h3>
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]7201 <a name="int_eh">Exception Handling Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7202</h3>
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]7203
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7204<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7205
7206<p>The LLVM exception handling intrinsics (which all start with
7207 <tt>llvm.eh.</tt> prefix), are described in
7208 the <a href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
7209 Handling</a> document.</p>
7210
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]7211</div>
7212
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7213<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7214<h3>
Duncan Sands86e01192007-09-11 14:10:23 +0000[diff] [blame]7215 <a name="int_trampoline">Trampoline Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7216</h3>
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7217
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7218<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7219
7220<p>This intrinsic makes it possible to excise one parameter, marked with
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]7221 the <a href="#nest"><tt>nest</tt></a> attribute, from a function.
7222 The result is a callable
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7223 function pointer lacking the nest parameter - the caller does not need to
7224 provide a value for it. Instead, the value to use is stored in advance in a
7225 "trampoline", a block of memory usually allocated on the stack, which also
7226 contains code to splice the nest value into the argument list. This is used
7227 to implement the GCC nested function address extension.</p>
7228
7229<p>For example, if the function is
7230 <tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
7231 pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as
7232 follows:</p>
7233
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]7234<pre class="doc_code">
Duncan Sands86e01192007-09-11 14:10:23 +0000[diff] [blame]7235 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
7236 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7237 %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]7238 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7239</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7240
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7241<p>The call <tt>%val = call i32 %fp(i32 %x, i32 %y)</tt> is then equivalent
7242 to <tt>%val = call i32 %f(i8* %nval, i32 %x, i32 %y)</tt>.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7243
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7244<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7245<h4>
7246 <a name="int_it">
7247 '<tt>llvm.init.trampoline</tt>' Intrinsic
7248 </a>
7249</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7250
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7251<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7252
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7253<h5>Syntax:</h5>
7254<pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7255 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]7256</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7257
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7258<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7259<p>This fills the memory pointed to by <tt>tramp</tt> with code and returns a
7260 function pointer suitable for executing it.</p>
7261
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7262<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7263<p>The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
7264 pointers. The <tt>tramp</tt> argument must point to a sufficiently large and
7265 sufficiently aligned block of memory; this memory is written to by the
7266 intrinsic. Note that the size and the alignment are target-specific - LLVM
7267 currently provides no portable way of determining them, so a front-end that
7268 generates this intrinsic needs to have some target-specific knowledge.
7269 The <tt>func</tt> argument must hold a function bitcast to
7270 an <tt>i8*</tt>.</p>
7271
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7272<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7273<p>The block of memory pointed to by <tt>tramp</tt> is filled with target
7274 dependent code, turning it into a function. A pointer to this function is
7275 returned, but needs to be bitcast to an <a href="#int_trampoline">appropriate
7276 function pointer type</a> before being called. The new function's signature
7277 is the same as that of <tt>func</tt> with any arguments marked with
7278 the <tt>nest</tt> attribute removed. At most one such <tt>nest</tt> argument
7279 is allowed, and it must be of pointer type. Calling the new function is
7280 equivalent to calling <tt>func</tt> with the same argument list, but
7281 with <tt>nval</tt> used for the missing <tt>nest</tt> argument. If, after
7282 calling <tt>llvm.init.trampoline</tt>, the memory pointed to
7283 by <tt>tramp</tt> is modified, then the effect of any later call to the
7284 returned function pointer is undefined.</p>
7285
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7286</div>
7287
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7288</div>
7289
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7290<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7291<h3>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7292 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7293</h3>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7294
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7295<div>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7296
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7297<p>These intrinsic functions expand the "universal IR" of LLVM to represent
7298 hardware constructs for atomic operations and memory synchronization. This
7299 provides an interface to the hardware, not an interface to the programmer. It
7300 is aimed at a low enough level to allow any programming models or APIs
7301 (Application Programming Interfaces) which need atomic behaviors to map
7302 cleanly onto it. It is also modeled primarily on hardware behavior. Just as
7303 hardware provides a "universal IR" for source languages, it also provides a
7304 starting point for developing a "universal" atomic operation and
7305 synchronization IR.</p>
7306
7307<p>These do <em>not</em> form an API such as high-level threading libraries,
7308 software transaction memory systems, atomic primitives, and intrinsic
7309 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
7310 application libraries. The hardware interface provided by LLVM should allow
7311 a clean implementation of all of these APIs and parallel programming models.
7312 No one model or paradigm should be selected above others unless the hardware
7313 itself ubiquitously does so.</p>
7314
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7315<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7316<h4>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7317 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7318</h4>
7319
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7320<div>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7321<h5>Syntax:</h5>
7322<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7323 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]7324</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7325
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7326<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7327<p>The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
7328 specific pairs of memory access types.</p>
7329
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7330<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7331<p>The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
7332 The first four arguments enables a specific barrier as listed below. The
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]7333 fifth argument specifies that the barrier applies to io or device or uncached
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7334 memory.</p>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7335
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7336<ul>
7337 <li><tt>ll</tt>: load-load barrier</li>
7338 <li><tt>ls</tt>: load-store barrier</li>
7339 <li><tt>sl</tt>: store-load barrier</li>
7340 <li><tt>ss</tt>: store-store barrier</li>
7341 <li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
7342</ul>
7343
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7344<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7345<p>This intrinsic causes the system to enforce some ordering constraints upon
7346 the loads and stores of the program. This barrier does not
7347 indicate <em>when</em> any events will occur, it only enforces
7348 an <em>order</em> in which they occur. For any of the specified pairs of load
7349 and store operations (f.ex. load-load, or store-load), all of the first
7350 operations preceding the barrier will complete before any of the second
7351 operations succeeding the barrier begin. Specifically the semantics for each
7352 pairing is as follows:</p>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7353
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7354<ul>
7355 <li><tt>ll</tt>: All loads before the barrier must complete before any load
7356 after the barrier begins.</li>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7357 <li><tt>ls</tt>: All loads before the barrier must complete before any
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7358 store after the barrier begins.</li>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7359 <li><tt>ss</tt>: All stores before the barrier must complete before any
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7360 store after the barrier begins.</li>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7361 <li><tt>sl</tt>: All stores before the barrier must complete before any
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7362 load after the barrier begins.</li>
7363</ul>
7364
7365<p>These semantics are applied with a logical "and" behavior when more than one
7366 is enabled in a single memory barrier intrinsic.</p>
7367
7368<p>Backends may implement stronger barriers than those requested when they do
7369 not support as fine grained a barrier as requested. Some architectures do
7370 not need all types of barriers and on such architectures, these become
7371 noops.</p>
7372
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7373<h5>Example:</h5>
7374<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7375%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7376%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7377 store i32 4, %ptr
7378
7379%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
Evan Cheng0ac49c62011-06-29 17:14:00 +0000[diff] [blame]7380 call void @llvm.memory.barrier(i1 false, i1 true, i1 false, i1 false, i1 true)
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7381 <i>; guarantee the above finishes</i>
7382 store i32 8, %ptr <i>; before this begins</i>
7383</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7384
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7385</div>
7386
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7387<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7388<h4>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7389 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7390</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7391
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7392<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7393
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7394<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7395<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
7396 any integer bit width and for different address spaces. Not all targets
7397 support all bit widths however.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7398
7399<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7400 declare i8 @llvm.atomic.cmp.swap.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt;)
7401 declare i16 @llvm.atomic.cmp.swap.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt;)
7402 declare i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt;)
7403 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]7404</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7405
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7406<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7407<p>This loads a value in memory and compares it to a given value. If they are
7408 equal, it stores a new value into the memory.</p>
7409
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7410<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7411<p>The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result
7412 as well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
7413 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
7414 this integer type. While any bit width integer may be used, targets may only
7415 lower representations they support in hardware.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7416
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7417<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7418<p>This entire intrinsic must be executed atomically. It first loads the value
7419 in memory pointed to by <tt>ptr</tt> and compares it with the
7420 value <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the
7421 memory. The loaded value is yielded in all cases. This provides the
7422 equivalent of an atomic compare-and-swap operation within the SSA
7423 framework.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7424
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7425<h5>Examples:</h5>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7426<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7427%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7428%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7429 store i32 4, %ptr
7430
7431%val1 = add i32 4, 4
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7432%result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %ptr, i32 4, %val1)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7433 <i>; yields {i32}:result1 = 4</i>
7434%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
7435%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
7436
7437%val2 = add i32 1, 1
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7438%result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %ptr, i32 5, %val2)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7439 <i>; yields {i32}:result2 = 8</i>
7440%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
7441
7442%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
7443</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7444
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7445</div>
7446
7447<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7448<h4>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7449 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7450</h4>
7451
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7452<div>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7453<h5>Syntax:</h5>
7454
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7455<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
7456 integer bit width. Not all targets support all bit widths however.</p>
7457
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7458<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7459 declare i8 @llvm.atomic.swap.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;val&gt;)
7460 declare i16 @llvm.atomic.swap.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;val&gt;)
7461 declare i32 @llvm.atomic.swap.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;val&gt;)
7462 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]7463</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7464
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7465<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7466<p>This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
7467 the value from memory. It then stores the value in <tt>val</tt> in the memory
7468 at <tt>ptr</tt>.</p>
7469
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7470<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7471<p>The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both
7472 the <tt>val</tt> argument and the result must be integers of the same bit
7473 width. The first argument, <tt>ptr</tt>, must be a pointer to a value of this
7474 integer type. The targets may only lower integer representations they
7475 support.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7476
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7477<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7478<p>This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
7479 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
7480 equivalent of an atomic swap operation within the SSA framework.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7481
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7482<h5>Examples:</h5>
7483<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7484%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7485%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7486 store i32 4, %ptr
7487
7488%val1 = add i32 4, 4
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7489%result1 = call i32 @llvm.atomic.swap.i32.p0i32(i32* %ptr, i32 %val1)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7490 <i>; yields {i32}:result1 = 4</i>
7491%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
7492%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
7493
7494%val2 = add i32 1, 1
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7495%result2 = call i32 @llvm.atomic.swap.i32.p0i32(i32* %ptr, i32 %val2)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7496 <i>; yields {i32}:result2 = 8</i>
7497
7498%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
7499%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
7500</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7501
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7502</div>
7503
7504<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7505<h4>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7506 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7507</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7508
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7509<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7510
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7511<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7512<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.add</tt> on
7513 any integer bit width. Not all targets support all bit widths however.</p>
7514
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7515<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7516 declare i8 @llvm.atomic.load.add.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7517 declare i16 @llvm.atomic.load.add.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7518 declare i32 @llvm.atomic.load.add.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7519 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]7520</pre>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7521
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7522<h5>Overview:</h5>
7523<p>This intrinsic adds <tt>delta</tt> to the value stored in memory
7524 at <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
7525
7526<h5>Arguments:</h5>
7527<p>The intrinsic takes two arguments, the first a pointer to an integer value
7528 and the second an integer value. The result is also an integer value. These
7529 integer types can have any bit width, but they must all have the same bit
7530 width. The targets may only lower integer representations they support.</p>
7531
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7532<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7533<p>This intrinsic does a series of operations atomically. It first loads the
7534 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
7535 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]7536
7537<h5>Examples:</h5>
7538<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7539%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7540%ptr = bitcast i8* %mallocP to i32*
7541 store i32 4, %ptr
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7542%result1 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 4)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7543 <i>; yields {i32}:result1 = 4</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7544%result2 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 2)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7545 <i>; yields {i32}:result2 = 8</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7546%result3 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 5)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7547 <i>; yields {i32}:result3 = 10</i>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7548%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7549</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7550
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7551</div>
7552
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7553<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7554<h4>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7555 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7556</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7557
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7558<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7559
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7560<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7561<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.sub</tt> on
7562 any integer bit width and for different address spaces. Not all targets
7563 support all bit widths however.</p>
7564
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7565<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7566 declare i8 @llvm.atomic.load.sub.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7567 declare i16 @llvm.atomic.load.sub.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7568 declare i32 @llvm.atomic.load.sub.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7569 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]7570</pre>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7571
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7572<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7573<p>This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7574 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
7575
7576<h5>Arguments:</h5>
7577<p>The intrinsic takes two arguments, the first a pointer to an integer value
7578 and the second an integer value. The result is also an integer value. These
7579 integer types can have any bit width, but they must all have the same bit
7580 width. The targets may only lower integer representations they support.</p>
7581
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7582<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7583<p>This intrinsic does a series of operations atomically. It first loads the
7584 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
7585 result to <tt>ptr</tt>. It yields the original value stored
7586 at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7587
7588<h5>Examples:</h5>
7589<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7590%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7591%ptr = bitcast i8* %mallocP to i32*
7592 store i32 8, %ptr
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7593%result1 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 4)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7594 <i>; yields {i32}:result1 = 8</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7595%result2 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 2)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7596 <i>; yields {i32}:result2 = 4</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7597%result3 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 5)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7598 <i>; yields {i32}:result3 = 2</i>
7599%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
7600</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7601
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7602</div>
7603
7604<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7605<h4>
7606 <a name="int_atomic_load_and">
7607 '<tt>llvm.atomic.load.and.*</tt>' Intrinsic
7608 </a>
7609 <br>
7610 <a name="int_atomic_load_nand">
7611 '<tt>llvm.atomic.load.nand.*</tt>' Intrinsic
7612 </a>
7613 <br>
7614 <a name="int_atomic_load_or">
7615 '<tt>llvm.atomic.load.or.*</tt>' Intrinsic
7616 </a>
7617 <br>
7618 <a name="int_atomic_load_xor">
7619 '<tt>llvm.atomic.load.xor.*</tt>' Intrinsic
7620 </a>
7621</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7622
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7623<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7624
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7625<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7626<p>These are overloaded intrinsics. You can
7627 use <tt>llvm.atomic.load_and</tt>, <tt>llvm.atomic.load_nand</tt>,
7628 <tt>llvm.atomic.load_or</tt>, and <tt>llvm.atomic.load_xor</tt> on any integer
7629 bit width and for different address spaces. Not all targets support all bit
7630 widths however.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7631
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7632<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7633 declare i8 @llvm.atomic.load.and.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7634 declare i16 @llvm.atomic.load.and.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7635 declare i32 @llvm.atomic.load.and.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7636 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]7637</pre>
7638
7639<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7640 declare i8 @llvm.atomic.load.or.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7641 declare i16 @llvm.atomic.load.or.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7642 declare i32 @llvm.atomic.load.or.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7643 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]7644</pre>
7645
7646<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7647 declare i8 @llvm.atomic.load.nand.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7648 declare i16 @llvm.atomic.load.nand.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7649 declare i32 @llvm.atomic.load.nand.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7650 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]7651</pre>
7652
7653<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7654 declare i8 @llvm.atomic.load.xor.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7655 declare i16 @llvm.atomic.load.xor.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7656 declare i32 @llvm.atomic.load.xor.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7657 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]7658</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7659
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7660<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7661<p>These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
7662 the value stored in memory at <tt>ptr</tt>. It yields the original value
7663 at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7664
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7665<h5>Arguments:</h5>
7666<p>These intrinsics take two arguments, the first a pointer to an integer value
7667 and the second an integer value. The result is also an integer value. These
7668 integer types can have any bit width, but they must all have the same bit
7669 width. The targets may only lower integer representations they support.</p>
7670
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7671<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7672<p>These intrinsics does a series of operations atomically. They first load the
7673 value stored at <tt>ptr</tt>. They then do the bitwise
7674 operation <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the
7675 original value stored at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7676
7677<h5>Examples:</h5>
7678<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7679%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7680%ptr = bitcast i8* %mallocP to i32*
7681 store i32 0x0F0F, %ptr
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7682%result0 = call i32 @llvm.atomic.load.nand.i32.p0i32(i32* %ptr, i32 0xFF)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7683 <i>; yields {i32}:result0 = 0x0F0F</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7684%result1 = call i32 @llvm.atomic.load.and.i32.p0i32(i32* %ptr, i32 0xFF)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7685 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7686%result2 = call i32 @llvm.atomic.load.or.i32.p0i32(i32* %ptr, i32 0F)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7687 <i>; yields {i32}:result2 = 0xF0</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7688%result3 = call i32 @llvm.atomic.load.xor.i32.p0i32(i32* %ptr, i32 0F)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7689 <i>; yields {i32}:result3 = FF</i>
7690%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
7691</pre>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7692
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7693</div>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7694
7695<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7696<h4>
7697 <a name="int_atomic_load_max">
7698 '<tt>llvm.atomic.load.max.*</tt>' Intrinsic
7699 </a>
7700 <br>
7701 <a name="int_atomic_load_min">
7702 '<tt>llvm.atomic.load.min.*</tt>' Intrinsic
7703 </a>
7704 <br>
7705 <a name="int_atomic_load_umax">
7706 '<tt>llvm.atomic.load.umax.*</tt>' Intrinsic
7707 </a>
7708 <br>
7709 <a name="int_atomic_load_umin">
7710 '<tt>llvm.atomic.load.umin.*</tt>' Intrinsic
7711 </a>
7712</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7713
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7714<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7715
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7716<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7717<p>These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
7718 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
7719 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
7720 address spaces. Not all targets support all bit widths however.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7721
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7722<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7723 declare i8 @llvm.atomic.load.max.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7724 declare i16 @llvm.atomic.load.max.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7725 declare i32 @llvm.atomic.load.max.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7726 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]7727</pre>
7728
7729<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7730 declare i8 @llvm.atomic.load.min.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7731 declare i16 @llvm.atomic.load.min.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7732 declare i32 @llvm.atomic.load.min.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7733 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]7734</pre>
7735
7736<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7737 declare i8 @llvm.atomic.load.umax.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7738 declare i16 @llvm.atomic.load.umax.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7739 declare i32 @llvm.atomic.load.umax.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7740 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]7741</pre>
7742
7743<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7744 declare i8 @llvm.atomic.load.umin.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7745 declare i16 @llvm.atomic.load.umin.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7746 declare i32 @llvm.atomic.load.umin.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7747 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]7748</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7749
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7750<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7751<p>These intrinsics takes the signed or unsigned minimum or maximum of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7752 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
7753 original value at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7754
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7755<h5>Arguments:</h5>
7756<p>These intrinsics take two arguments, the first a pointer to an integer value
7757 and the second an integer value. The result is also an integer value. These
7758 integer types can have any bit width, but they must all have the same bit
7759 width. The targets may only lower integer representations they support.</p>
7760
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7761<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7762<p>These intrinsics does a series of operations atomically. They first load the
7763 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or
7764 max <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They
7765 yield the original value stored at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7766
7767<h5>Examples:</h5>
7768<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7769%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7770%ptr = bitcast i8* %mallocP to i32*
7771 store i32 7, %ptr
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7772%result0 = call i32 @llvm.atomic.load.min.i32.p0i32(i32* %ptr, i32 -2)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7773 <i>; yields {i32}:result0 = 7</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7774%result1 = call i32 @llvm.atomic.load.max.i32.p0i32(i32* %ptr, i32 8)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7775 <i>; yields {i32}:result1 = -2</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7776%result2 = call i32 @llvm.atomic.load.umin.i32.p0i32(i32* %ptr, i32 10)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7777 <i>; yields {i32}:result2 = 8</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7778%result3 = call i32 @llvm.atomic.load.umax.i32.p0i32(i32* %ptr, i32 30)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7779 <i>; yields {i32}:result3 = 8</i>
7780%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
7781</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7782
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7783</div>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7784
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7785</div>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7786
7787<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7788<h3>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7789 <a name="int_memorymarkers">Memory Use Markers</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7790</h3>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7791
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7792<div>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7793
7794<p>This class of intrinsics exists to information about the lifetime of memory
7795 objects and ranges where variables are immutable.</p>
7796
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7797<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7798<h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7799 <a name="int_lifetime_start">'<tt>llvm.lifetime.start</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7800</h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7801
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7802<div>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7803
7804<h5>Syntax:</h5>
7805<pre>
7806 declare void @llvm.lifetime.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7807</pre>
7808
7809<h5>Overview:</h5>
7810<p>The '<tt>llvm.lifetime.start</tt>' intrinsic specifies the start of a memory
7811 object's lifetime.</p>
7812
7813<h5>Arguments:</h5>
Nick Lewycky9bc89042009-10-13 07:57:33 +0000[diff] [blame]7814<p>The first argument is a constant integer representing the size of the
7815 object, or -1 if it is variable sized. The second argument is a pointer to
7816 the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7817
7818<h5>Semantics:</h5>
7819<p>This intrinsic indicates that before this point in the code, the value of the
7820 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]7821 never be used and has an undefined value. A load from the pointer that
7822 precedes this intrinsic can be replaced with
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7823 <tt>'<a href="#undefvalues">undef</a>'</tt>.</p>
7824
7825</div>
7826
7827<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7828<h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7829 <a name="int_lifetime_end">'<tt>llvm.lifetime.end</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7830</h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7831
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7832<div>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7833
7834<h5>Syntax:</h5>
7835<pre>
7836 declare void @llvm.lifetime.end(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7837</pre>
7838
7839<h5>Overview:</h5>
7840<p>The '<tt>llvm.lifetime.end</tt>' intrinsic specifies the end of a memory
7841 object's lifetime.</p>
7842
7843<h5>Arguments:</h5>
Nick Lewycky9bc89042009-10-13 07:57:33 +0000[diff] [blame]7844<p>The first argument is a constant integer representing the size of the
7845 object, or -1 if it is variable sized. The second argument is a pointer to
7846 the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7847
7848<h5>Semantics:</h5>
7849<p>This intrinsic indicates that after this point in the code, the value of the
7850 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
7851 never be used and has an undefined value. Any stores into the memory object
7852 following this intrinsic may be removed as dead.
7853
7854</div>
7855
7856<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7857<h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7858 <a name="int_invariant_start">'<tt>llvm.invariant.start</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7859</h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7860
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7861<div>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7862
7863<h5>Syntax:</h5>
7864<pre>
Nick Lewycky2965d3e2010-11-30 04:13:41 +0000[diff] [blame]7865 declare {}* @llvm.invariant.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7866</pre>
7867
7868<h5>Overview:</h5>
7869<p>The '<tt>llvm.invariant.start</tt>' intrinsic specifies that the contents of
7870 a memory object will not change.</p>
7871
7872<h5>Arguments:</h5>
Nick Lewycky9bc89042009-10-13 07:57:33 +0000[diff] [blame]7873<p>The first argument is a constant integer representing the size of the
7874 object, or -1 if it is variable sized. The second argument is a pointer to
7875 the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7876
7877<h5>Semantics:</h5>
7878<p>This intrinsic indicates that until an <tt>llvm.invariant.end</tt> that uses
7879 the return value, the referenced memory location is constant and
7880 unchanging.</p>
7881
7882</div>
7883
7884<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7885<h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7886 <a name="int_invariant_end">'<tt>llvm.invariant.end</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7887</h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7888
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7889<div>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7890
7891<h5>Syntax:</h5>
7892<pre>
7893 declare void @llvm.invariant.end({}* &lt;start&gt;, i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7894</pre>
7895
7896<h5>Overview:</h5>
7897<p>The '<tt>llvm.invariant.end</tt>' intrinsic specifies that the contents of
7898 a memory object are mutable.</p>
7899
7900<h5>Arguments:</h5>
7901<p>The first argument is the matching <tt>llvm.invariant.start</tt> intrinsic.
Nick Lewycky9bc89042009-10-13 07:57:33 +0000[diff] [blame]7902 The second argument is a constant integer representing the size of the
7903 object, or -1 if it is variable sized and the third argument is a pointer
7904 to the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7905
7906<h5>Semantics:</h5>
7907<p>This intrinsic indicates that the memory is mutable again.</p>
7908
7909</div>
7910
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7911</div>
7912
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7913<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7914<h3>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7915 <a name="int_general">General Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7916</h3>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7917
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7918<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7919
7920<p>This class of intrinsics is designed to be generic and has no specific
7921 purpose.</p>
7922
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7923<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7924<h4>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7925 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7926</h4>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7927
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7928<div>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7929
7930<h5>Syntax:</h5>
7931<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7932 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]7933</pre>
7934
7935<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7936<p>The '<tt>llvm.var.annotation</tt>' intrinsic.</p>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7937
7938<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7939<p>The first argument is a pointer to a value, the second is a pointer to a
7940 global string, the third is a pointer to a global string which is the source
7941 file name, and the last argument is the line number.</p>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7942
7943<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7944<p>This intrinsic allows annotation of local variables with arbitrary strings.
7945 This can be useful for special purpose optimizations that want to look for
7946 these annotations. These have no other defined use, they are ignored by code
7947 generation and optimization.</p>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7948
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7949</div>
7950
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7951<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7952<h4>
Tanya Lattner0186a652007-09-21 23:57:59 +0000[diff] [blame]7953 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7954</h4>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7955
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7956<div>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7957
7958<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7959<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
7960 any integer bit width.</p>
7961
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7962<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7963 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7964 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7965 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7966 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7967 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]7968</pre>
7969
7970<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7971<p>The '<tt>llvm.annotation</tt>' intrinsic.</p>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7972
7973<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7974<p>The first argument is an integer value (result of some expression), the
7975 second is a pointer to a global string, the third is a pointer to a global
7976 string which is the source file name, and the last argument is the line
7977 number. It returns the value of the first argument.</p>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7978
7979<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7980<p>This intrinsic allows annotations to be put on arbitrary expressions with
7981 arbitrary strings. This can be useful for special purpose optimizations that
7982 want to look for these annotations. These have no other defined use, they
7983 are ignored by code generation and optimization.</p>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7984
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7985</div>
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]7986
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]7987<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7988<h4>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]7989 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7990</h4>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]7991
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]7992<div>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]7993
7994<h5>Syntax:</h5>
7995<pre>
7996 declare void @llvm.trap()
7997</pre>
7998
7999<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8000<p>The '<tt>llvm.trap</tt>' intrinsic.</p>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]8001
8002<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8003<p>None.</p>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]8004
8005<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8006<p>This intrinsics is lowered to the target dependent trap instruction. If the
8007 target does not have a trap instruction, this intrinsic will be lowered to
8008 the call of the <tt>abort()</tt> function.</p>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]8009
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]8010</div>
8011
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]8012<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8013<h4>
Misha Brukman50de2b22008-11-22 23:55:29 +0000[diff] [blame]8014 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8015</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8016
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8017<div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8018
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]8019<h5>Syntax:</h5>
8020<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8021 declare void @llvm.stackprotector(i8* &lt;guard&gt;, i8** &lt;slot&gt;)
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]8022</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8023
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]8024<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8025<p>The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and
8026 stores it onto the stack at <tt>slot</tt>. The stack slot is adjusted to
8027 ensure that it is placed on the stack before local variables.</p>
8028
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]8029<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8030<p>The <tt>llvm.stackprotector</tt> intrinsic requires two pointer
8031 arguments. The first argument is the value loaded from the stack
8032 guard <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt>
8033 that has enough space to hold the value of the guard.</p>
8034
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]8035<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8036<p>This intrinsic causes the prologue/epilogue inserter to force the position of
8037 the <tt>AllocaInst</tt> stack slot to be before local variables on the
8038 stack. This is to ensure that if a local variable on the stack is
8039 overwritten, it will destroy the value of the guard. When the function exits,
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]8040 the guard on the stack is checked against the original guard. If they are
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]8041 different, then the program aborts by calling the <tt>__stack_chk_fail()</tt>
8042 function.</p>
8043
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]8044</div>
8045
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8046<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8047<h4>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8048 <a name="int_objectsize">'<tt>llvm.objectsize</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]8049</h4>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8050
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8051<div>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8052
8053<h5>Syntax:</h5>
8054<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]8055 declare i32 @llvm.objectsize.i32(i8* &lt;object&gt;, i1 &lt;type&gt;)
8056 declare i64 @llvm.objectsize.i64(i8* &lt;object&gt;, i1 &lt;type&gt;)
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8057</pre>
8058
8059<h5>Overview:</h5>
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]8060<p>The <tt>llvm.objectsize</tt> intrinsic is designed to provide information to
8061 the optimizers to determine at compile time whether a) an operation (like
8062 memcpy) will overflow a buffer that corresponds to an object, or b) that a
8063 runtime check for overflow isn't necessary. An object in this context means
8064 an allocation of a specific class, structure, array, or other object.</p>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8065
8066<h5>Arguments:</h5>
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]8067<p>The <tt>llvm.objectsize</tt> intrinsic takes two arguments. The first
Eric Christopher31e39bd2009-12-23 00:29:49 +0000[diff] [blame]8068 argument is a pointer to or into the <tt>object</tt>. The second argument
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]8069 is a boolean 0 or 1. This argument determines whether you want the
8070 maximum (0) or minimum (1) bytes remaining. This needs to be a literal 0 or
Eric Christopher31e39bd2009-12-23 00:29:49 +0000[diff] [blame]8071 1, variables are not allowed.</p>
8072
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8073<h5>Semantics:</h5>
8074<p>The <tt>llvm.objectsize</tt> intrinsic is lowered to either a constant
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]8075 representing the size of the object concerned, or <tt>i32/i64 -1 or 0</tt>,
8076 depending on the <tt>type</tt> argument, if the size cannot be determined at
8077 compile time.</p>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]8078
8079</div>
8080
NAKAMURA Takumiaa3d6242011-04-23 00:30:22 +0000[diff] [blame]8081</div>
8082
8083</div>
8084
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]8085<!-- *********************************************************************** -->
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]8086<hr>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]8087<address>
8088 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman86242e12008-12-11 17:34:48 +0000[diff] [blame]8089 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]8090 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman86242e12008-12-11 17:34:48 +0000[diff] [blame]8091 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]8092
8093 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
NAKAMURA Takumica46f5a2011-04-09 02:13:37 +0000[diff] [blame]8094 <a href="http://llvm.org/">The LLVM Compiler Infrastructure</a><br>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]8095 Last modified: $Date$
8096</address>
Chris Lattnerb8f816e2008-01-04 04:33:49 +0000[diff] [blame]8097
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]8098</body>
8099</html>