Gitiles
Code Review Sign In
gerrit-public.fairphone.software / toolchain / llvm-project / 45e6c195d7f48a9507edb4888db2964f1cd928d7 / . / llvm / docs / LangRef.html
blob: 2121fc67408e8b6b923861acb4f2d378317286ef [file] [log] [blame]
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3<html>
4<head>
5 <title>LLVM Assembly Language Reference Manual</title>
Reid Spencercb84e432004-08-26 20:44:00 +0000[diff] [blame]6 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
7 <meta name="author" content="Chris Lattner">
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]8 <meta name="description"
Reid Spencercb84e432004-08-26 20:44:00 +0000[diff] [blame]9 content="LLVM Assembly Language Reference Manual.">
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]10 <link rel="stylesheet" href="llvm.css" type="text/css">
11</head>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]12
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]13<body>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]14
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]15<div class="doc_title"> LLVM Language Reference Manual </div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]16<ol>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]17 <li><a href="#abstract">Abstract</a></li>
18 <li><a href="#introduction">Introduction</a></li>
19 <li><a href="#identifiers">Identifiers</a></li>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]20 <li><a href="#highlevel">High Level Structure</a>
21 <ol>
22 <li><a href="#modulestructure">Module Structure</a></li>
Bill Wendlinga3c6f6b2009-07-20 01:03:30 +0000[diff] [blame]23 <li><a href="#linkage">Linkage Types</a>
24 <ol>
Bill Wendling8693ef82009-07-20 02:41:50 +0000[diff] [blame]25 <li><a href="#linkage_private">'<tt>private</tt>' Linkage</a></li>
26 <li><a href="#linkage_linker_private">'<tt>linker_private</tt>' Linkage</a></li>
Bill Wendling03bcd6e2010-07-01 21:55:59 +0000[diff] [blame]27 <li><a href="#linkage_linker_private_weak">'<tt>linker_private_weak</tt>' Linkage</a></li>
Bill 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>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]56 </ol>
57 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]58 <li><a href="#typesystem">Type System</a>
59 <ol>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]60 <li><a href="#t_classifications">Type Classifications</a></li>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]61 <li><a href="#t_primitive">Primitive Types</a>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]62 <ol>
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]63 <li><a href="#t_integer">Integer Type</a></li>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]64 <li><a href="#t_floating">Floating Point Types</a></li>
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]65 <li><a href="#t_x86mmx">X86mmx Type</a></li>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]66 <li><a href="#t_void">Void Type</a></li>
67 <li><a href="#t_label">Label Type</a></li>
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]68 <li><a href="#t_metadata">Metadata Type</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]69 </ol>
70 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]71 <li><a href="#t_derived">Derived Types</a>
72 <ol>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]73 <li><a href="#t_aggregate">Aggregate Types</a>
74 <ol>
75 <li><a href="#t_array">Array Type</a></li>
76 <li><a href="#t_struct">Structure Type</a></li>
77 <li><a href="#t_pstruct">Packed Structure Type</a></li>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]78 <li><a href="#t_vector">Vector Type</a></li>
79 </ol>
80 </li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]81 <li><a href="#t_function">Function Type</a></li>
82 <li><a href="#t_pointer">Pointer Type</a></li>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]83 <li><a href="#t_opaque">Opaque Type</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]84 </ol>
85 </li>
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]86 <li><a href="#t_uprefs">Type Up-references</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]87 </ol>
88 </li>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]89 <li><a href="#constants">Constants</a>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]90 <ol>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]91 <li><a href="#simpleconstants">Simple Constants</a></li>
Chris Lattner361bfcd2009-02-28 18:32:25 +0000[diff] [blame]92 <li><a href="#complexconstants">Complex Constants</a></li>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]93 <li><a href="#globalconstants">Global Variable and Function Addresses</a></li>
94 <li><a href="#undefvalues">Undefined Values</a></li>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]95 <li><a href="#trapvalues">Trap Values</a></li>
Chris Lattner2bfd3202009-10-27 21:19:13 +0000[diff] [blame]96 <li><a href="#blockaddress">Addresses of Basic Blocks</a></li>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]97 <li><a href="#constantexprs">Constant Expressions</a></li>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]98 </ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]99 </li>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]100 <li><a href="#othervalues">Other Values</a>
101 <ol>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]102 <li><a href="#inlineasm">Inline Assembler Expressions</a></li>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]103 <li><a href="#metadata">Metadata Nodes and Metadata Strings</a></li>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]104 </ol>
105 </li>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]106 <li><a href="#intrinsic_globals">Intrinsic Global Variables</a>
107 <ol>
108 <li><a href="#intg_used">The '<tt>llvm.used</tt>' Global Variable</a></li>
Chris Lattner58f9bb22009-07-20 06:14:25 +0000[diff] [blame]109 <li><a href="#intg_compiler_used">The '<tt>llvm.compiler.used</tt>'
110 Global Variable</a></li>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]111 <li><a href="#intg_global_ctors">The '<tt>llvm.global_ctors</tt>'
112 Global Variable</a></li>
113 <li><a href="#intg_global_dtors">The '<tt>llvm.global_dtors</tt>'
114 Global Variable</a></li>
115 </ol>
116 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]117 <li><a href="#instref">Instruction Reference</a>
118 <ol>
119 <li><a href="#terminators">Terminator Instructions</a>
120 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]121 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
122 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]123 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]124 <li><a href="#i_indirectbr">'<tt>indirectbr</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]125 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]126 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]127 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]128 </ol>
129 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]130 <li><a href="#binaryops">Binary Operations</a>
131 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]132 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]133 <li><a href="#i_fadd">'<tt>fadd</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]134 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]135 <li><a href="#i_fsub">'<tt>fsub</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]136 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]137 <li><a href="#i_fmul">'<tt>fmul</tt>' Instruction</a></li>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]138 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
139 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
140 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]141 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
142 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
143 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]144 </ol>
145 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]146 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
147 <ol>
Reid Spencer2ab01932007-02-02 13:57:07 +0000[diff] [blame]148 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
149 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
150 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]151 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]152 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]153 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]154 </ol>
155 </li>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]156 <li><a href="#vectorops">Vector Operations</a>
157 <ol>
158 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
159 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
160 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]161 </ol>
162 </li>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]163 <li><a href="#aggregateops">Aggregate Operations</a>
164 <ol>
165 <li><a href="#i_extractvalue">'<tt>extractvalue</tt>' Instruction</a></li>
166 <li><a href="#i_insertvalue">'<tt>insertvalue</tt>' Instruction</a></li>
167 </ol>
168 </li>
Chris Lattner6ab66722006-08-15 00:45:58 +0000[diff] [blame]169 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]170 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]171 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
Robert Bocchino820bc75b2006-02-17 21:18:08 +0000[diff] [blame]172 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
173 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
174 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]175 </ol>
176 </li>
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]177 <li><a href="#convertops">Conversion Operations</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]178 <ol>
179 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
180 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
181 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
182 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
183 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]184 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
185 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
186 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
187 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]188 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
189 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]190 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]191 </ol>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]192 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]193 <li><a href="#otherops">Other Operations</a>
194 <ol>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]195 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
196 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]197 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]198 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]199 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Chris Lattner33337472006-01-13 23:26:01 +0000[diff] [blame]200 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]201 </ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]202 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]203 </ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]204 </li>
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]205 <li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]206 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]207 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
208 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]209 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
210 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
211 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]212 </ol>
213 </li>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]214 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
215 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]216 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
217 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
218 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]219 </ol>
220 </li>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]221 <li><a href="#int_codegen">Code Generator Intrinsics</a>
222 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]223 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
224 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
225 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
226 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
227 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
228 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
Dan Gohmane58f7b32010-05-26 21:56:15 +0000[diff] [blame]229 <li><a href="#int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswellaa1c3c12004-04-09 16:43:20 +0000[diff] [blame]230 </ol>
231 </li>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]232 <li><a href="#int_libc">Standard C Library Intrinsics</a>
233 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]234 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
235 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
236 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
237 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
238 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]239 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
240 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
241 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]242 </ol>
243 </li>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]244 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]245 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]246 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattnerb748c672006-01-16 22:34:14 +0000[diff] [blame]247 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
248 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
249 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]250 </ol>
251 </li>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]252 <li><a href="#int_overflow">Arithmetic with Overflow Intrinsics</a>
253 <ol>
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]254 <li><a href="#int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt> Intrinsics</a></li>
255 <li><a href="#int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt> Intrinsics</a></li>
256 <li><a href="#int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt> Intrinsics</a></li>
257 <li><a href="#int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt> Intrinsics</a></li>
258 <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]259 <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]260 </ol>
261 </li>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]262 <li><a href="#int_fp16">Half Precision Floating Point Intrinsics</a>
263 <ol>
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]264 <li><a href="#int_convert_to_fp16">'<tt>llvm.convert.to.fp16</tt>' Intrinsic</a></li>
265 <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]266 </ol>
267 </li>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]268 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]269 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sands86e01192007-09-11 14:10:23 +0000[diff] [blame]270 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]271 <ol>
272 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]273 </ol>
274 </li>
Bill Wendlingf85850f2008-11-18 22:10:53 +0000[diff] [blame]275 <li><a href="#int_atomics">Atomic intrinsics</a>
276 <ol>
277 <li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
278 <li><a href="#int_atomic_cmp_swap"><tt>llvm.atomic.cmp.swap</tt></a></li>
279 <li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
280 <li><a href="#int_atomic_load_add"><tt>llvm.atomic.load.add</tt></a></li>
281 <li><a href="#int_atomic_load_sub"><tt>llvm.atomic.load.sub</tt></a></li>
282 <li><a href="#int_atomic_load_and"><tt>llvm.atomic.load.and</tt></a></li>
283 <li><a href="#int_atomic_load_nand"><tt>llvm.atomic.load.nand</tt></a></li>
284 <li><a href="#int_atomic_load_or"><tt>llvm.atomic.load.or</tt></a></li>
285 <li><a href="#int_atomic_load_xor"><tt>llvm.atomic.load.xor</tt></a></li>
286 <li><a href="#int_atomic_load_max"><tt>llvm.atomic.load.max</tt></a></li>
287 <li><a href="#int_atomic_load_min"><tt>llvm.atomic.load.min</tt></a></li>
288 <li><a href="#int_atomic_load_umax"><tt>llvm.atomic.load.umax</tt></a></li>
289 <li><a href="#int_atomic_load_umin"><tt>llvm.atomic.load.umin</tt></a></li>
290 </ol>
291 </li>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]292 <li><a href="#int_memorymarkers">Memory Use Markers</a>
293 <ol>
294 <li><a href="#int_lifetime_start"><tt>llvm.lifetime.start</tt></a></li>
295 <li><a href="#int_lifetime_end"><tt>llvm.lifetime.end</tt></a></li>
296 <li><a href="#int_invariant_start"><tt>llvm.invariant.start</tt></a></li>
297 <li><a href="#int_invariant_end"><tt>llvm.invariant.end</tt></a></li>
298 </ol>
299 </li>
Reid Spencer5b2cb0f2007-07-20 19:59:11 +0000[diff] [blame]300 <li><a href="#int_general">General intrinsics</a>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]301 <ol>
Reid Spencer5b2cb0f2007-07-20 19:59:11 +0000[diff] [blame]302 <li><a href="#int_var_annotation">
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]303 '<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]304 <li><a href="#int_annotation">
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]305 '<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]306 <li><a href="#int_trap">
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]307 '<tt>llvm.trap</tt>' Intrinsic</a></li>
308 <li><a href="#int_stackprotector">
309 '<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]310 <li><a href="#int_objectsize">
311 '<tt>llvm.objectsize</tt>' Intrinsic</a></li>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]312 </ol>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]313 </li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]314 </ol>
315 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]316</ol>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]317
318<div class="doc_author">
319 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
320 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]321</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]322
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]323<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]324<div class="doc_section"> <a name="abstract">Abstract </a></div>
325<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]326
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]327<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]328
329<p>This document is a reference manual for the LLVM assembly language. LLVM is
330 a Static Single Assignment (SSA) based representation that provides type
331 safety, low-level operations, flexibility, and the capability of representing
332 'all' high-level languages cleanly. It is the common code representation
333 used throughout all phases of the LLVM compilation strategy.</p>
334
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]335</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]336
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]337<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]338<div class="doc_section"> <a name="introduction">Introduction</a> </div>
339<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]340
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]341<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]342
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]343<p>The LLVM code representation is designed to be used in three different forms:
344 as an in-memory compiler IR, as an on-disk bitcode representation (suitable
345 for fast loading by a Just-In-Time compiler), and as a human readable
346 assembly language representation. This allows LLVM to provide a powerful
347 intermediate representation for efficient compiler transformations and
348 analysis, while providing a natural means to debug and visualize the
349 transformations. The three different forms of LLVM are all equivalent. This
350 document describes the human readable representation and notation.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]351
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]352<p>The LLVM representation aims to be light-weight and low-level while being
353 expressive, typed, and extensible at the same time. It aims to be a
354 "universal IR" of sorts, by being at a low enough level that high-level ideas
355 may be cleanly mapped to it (similar to how microprocessors are "universal
356 IR's", allowing many source languages to be mapped to them). By providing
357 type information, LLVM can be used as the target of optimizations: for
358 example, through pointer analysis, it can be proven that a C automatic
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]359 variable is never accessed outside of the current function, allowing it to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]360 be promoted to a simple SSA value instead of a memory location.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]361
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]362</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]363
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]364<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]365<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]366
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]367<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]368
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]369<p>It is important to note that this document describes 'well formed' LLVM
370 assembly language. There is a difference between what the parser accepts and
371 what is considered 'well formed'. For example, the following instruction is
372 syntactically okay, but not well formed:</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]373
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]374<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]375%x = <a href="#i_add">add</a> i32 1, %x
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]376</pre>
377
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]378<p>because the definition of <tt>%x</tt> does not dominate all of its uses. The
379 LLVM infrastructure provides a verification pass that may be used to verify
380 that an LLVM module is well formed. This pass is automatically run by the
381 parser after parsing input assembly and by the optimizer before it outputs
382 bitcode. The violations pointed out by the verifier pass indicate bugs in
383 transformation passes or input to the parser.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]384
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]385</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]386
Chris Lattner87a3dbe2007-10-03 17:34:29 +0000[diff] [blame]387<!-- Describe the typesetting conventions here. -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]388
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]389<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]390<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]391<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]392
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]393<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]394
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]395<p>LLVM identifiers come in two basic types: global and local. Global
396 identifiers (functions, global variables) begin with the <tt>'@'</tt>
397 character. Local identifiers (register names, types) begin with
398 the <tt>'%'</tt> character. Additionally, there are three different formats
399 for identifiers, for different purposes:</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]400
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]401<ol>
Reid Spencerb23b65f2007-08-07 14:34:28 +0000[diff] [blame]402 <li>Named values are represented as a string of characters with their prefix.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]403 For example, <tt>%foo</tt>, <tt>@DivisionByZero</tt>,
404 <tt>%a.really.long.identifier</tt>. The actual regular expression used is
405 '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'. Identifiers which require
406 other characters in their names can be surrounded with quotes. Special
407 characters may be escaped using <tt>"\xx"</tt> where <tt>xx</tt> is the
408 ASCII code for the character in hexadecimal. In this way, any character
409 can be used in a name value, even quotes themselves.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]410
Reid Spencerb23b65f2007-08-07 14:34:28 +0000[diff] [blame]411 <li>Unnamed values are represented as an unsigned numeric value with their
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]412 prefix. For example, <tt>%12</tt>, <tt>@2</tt>, <tt>%44</tt>.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]413
Reid Spencer8f08d802004-12-09 18:02:53 +0000[diff] [blame]414 <li>Constants, which are described in a <a href="#constants">section about
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]415 constants</a>, below.</li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]416</ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]417
Reid Spencerb23b65f2007-08-07 14:34:28 +0000[diff] [blame]418<p>LLVM requires that values start with a prefix for two reasons: Compilers
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]419 don't need to worry about name clashes with reserved words, and the set of
420 reserved words may be expanded in the future without penalty. Additionally,
421 unnamed identifiers allow a compiler to quickly come up with a temporary
422 variable without having to avoid symbol table conflicts.</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]423
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]424<p>Reserved words in LLVM are very similar to reserved words in other
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]425 languages. There are keywords for different opcodes
426 ('<tt><a href="#i_add">add</a></tt>',
427 '<tt><a href="#i_bitcast">bitcast</a></tt>',
428 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names
429 ('<tt><a href="#t_void">void</a></tt>',
430 '<tt><a href="#t_primitive">i32</a></tt>', etc...), and others. These
431 reserved words cannot conflict with variable names, because none of them
432 start with a prefix character (<tt>'%'</tt> or <tt>'@'</tt>).</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]433
434<p>Here is an example of LLVM code to multiply the integer variable
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]435 '<tt>%X</tt>' by 8:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]436
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]437<p>The easy way:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]438
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]439<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]440%result = <a href="#i_mul">mul</a> i32 %X, 8
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]441</pre>
442
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]443<p>After strength reduction:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]444
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]445<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]446%result = <a href="#i_shl">shl</a> i32 %X, i8 3
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]447</pre>
448
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]449<p>And the hard way:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]450
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]451<pre class="doc_code">
Gabor Greifbd0328f2009-10-28 13:05:07 +0000[diff] [blame]452%0 = <a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
453%1 = <a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]454%result = <a href="#i_add">add</a> i32 %1, %1
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]455</pre>
456
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]457<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several important
458 lexical features of LLVM:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]459
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]460<ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]461 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]462 line.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]463
464 <li>Unnamed temporaries are created when the result of a computation is not
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]465 assigned to a named value.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]466
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]467 <li>Unnamed temporaries are numbered sequentially</li>
468</ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]469
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]470<p>It also shows a convention that we follow in this document. When
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]471 demonstrating instructions, we will follow an instruction with a comment that
472 defines the type and name of value produced. Comments are shown in italic
473 text.</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]474
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]475</div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]476
477<!-- *********************************************************************** -->
478<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
479<!-- *********************************************************************** -->
480
481<!-- ======================================================================= -->
482<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
483</div>
484
485<div class="doc_text">
486
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]487<p>LLVM programs are composed of "Module"s, each of which is a translation unit
488 of the input programs. Each module consists of functions, global variables,
489 and symbol table entries. Modules may be combined together with the LLVM
490 linker, which merges function (and global variable) definitions, resolves
491 forward declarations, and merges symbol table entries. Here is an example of
492 the "hello world" module:</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]493
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]494<pre class="doc_code">
Chris Lattner54a7be72010-08-17 17:13:42 +0000[diff] [blame]495<i>; Declare the string constant as a global constant.</i>&nbsp;
496<a href="#identifiers">@.LC0</a> = <a href="#linkage_internal">internal</a> <a href="#globalvars">constant</a> <a href="#t_array">[13 x i8]</a> c"hello world\0A\00" <i>; [13 x i8]*</i>&nbsp;
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]497
Chris Lattner54a7be72010-08-17 17:13:42 +0000[diff] [blame]498<i>; External declaration of the puts function</i>&nbsp;
499<a href="#functionstructure">declare</a> i32 @puts(i8*) <i>; i32 (i8*)* </i>&nbsp;
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]500
501<i>; Definition of main function</i>
Chris Lattner54a7be72010-08-17 17:13:42 +0000[diff] [blame]502define i32 @main() { <i>; i32()* </i>&nbsp;
503 <i>; Convert [13 x i8]* to i8 *...</i>&nbsp;
504 %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]505
Chris Lattner54a7be72010-08-17 17:13:42 +0000[diff] [blame]506 <i>; Call puts function to write out the string to stdout.</i>&nbsp;
507 <a href="#i_call">call</a> i32 @puts(i8* %cast210) <i>; i32</i>&nbsp;
508 <a href="#i_ret">ret</a> i32 0&nbsp;
509}
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]510
511<i>; Named metadata</i>
512!1 = metadata !{i32 41}
513!foo = !{!1, null}
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]514</pre>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]515
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]516<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]517 "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>" function,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]518 a <a href="#functionstructure">function definition</a> for
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]519 "<tt>main</tt>" and <a href="#namedmetadatastructure">named metadata</a>
520 "<tt>foo"</tt>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]521
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]522<p>In general, a module is made up of a list of global values, where both
523 functions and global variables are global values. Global values are
524 represented by a pointer to a memory location (in this case, a pointer to an
525 array of char, and a pointer to a function), and have one of the
526 following <a href="#linkage">linkage types</a>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]527
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]528</div>
529
530<!-- ======================================================================= -->
531<div class="doc_subsection">
532 <a name="linkage">Linkage Types</a>
533</div>
534
535<div class="doc_text">
536
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]537<p>All Global Variables and Functions have one of the following types of
538 linkage:</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]539
540<dl>
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]541 <dt><tt><b><a name="linkage_private">private</a></b></tt></dt>
Bill Wendling03bcd6e2010-07-01 21:55:59 +0000[diff] [blame]542 <dd>Global values with "<tt>private</tt>" linkage are only directly accessible
543 by objects in the current module. In particular, linking code into a
544 module with an private global value may cause the private to be renamed as
545 necessary to avoid collisions. Because the symbol is private to the
546 module, all references can be updated. This doesn't show up in any symbol
547 table in the object file.</dd>
Rafael Espindola6de96a12009-01-15 20:18:42 +0000[diff] [blame]548
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]549 <dt><tt><b><a name="linkage_linker_private">linker_private</a></b></tt></dt>
Bill Wendling03bcd6e2010-07-01 21:55:59 +0000[diff] [blame]550 <dd>Similar to <tt>private</tt>, but the symbol is passed through the
551 assembler and evaluated by the linker. Unlike normal strong symbols, they
552 are removed by the linker from the final linked image (executable or
553 dynamic library).</dd>
554
555 <dt><tt><b><a name="linkage_linker_private_weak">linker_private_weak</a></b></tt></dt>
556 <dd>Similar to "<tt>linker_private</tt>", but the symbol is weak. Note that
557 <tt>linker_private_weak</tt> symbols are subject to coalescing by the
558 linker. The symbols are removed by the linker from the final linked image
559 (executable or dynamic library).</dd>
Bill Wendlinga3c6f6b2009-07-20 01:03:30 +0000[diff] [blame]560
Bill Wendling578ee402010-08-20 22:05:50 +0000[diff] [blame]561 <dt><tt><b><a name="linkage_linker_private_weak_def_auto">linker_private_weak_def_auto</a></b></tt></dt>
562 <dd>Similar to "<tt>linker_private_weak</tt>", but it's known that the address
563 of the object is not taken. For instance, functions that had an inline
564 definition, but the compiler decided not to inline it. Note,
565 unlike <tt>linker_private</tt> and <tt>linker_private_weak</tt>,
566 <tt>linker_private_weak_def_auto</tt> may have only <tt>default</tt>
567 visibility. The symbols are removed by the linker from the final linked
568 image (executable or dynamic library).</dd>
569
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]570 <dt><tt><b><a name="linkage_internal">internal</a></b></tt></dt>
Bill Wendling36321712010-06-29 22:34:52 +0000[diff] [blame]571 <dd>Similar to private, but the value shows as a local symbol
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]572 (<tt>STB_LOCAL</tt> in the case of ELF) in the object file. This
573 corresponds to the notion of the '<tt>static</tt>' keyword in C.</dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]574
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]575 <dt><tt><b><a name="linkage_available_externally">available_externally</a></b></tt></dt>
Chris Lattner184f1be2009-04-13 05:44:34 +0000[diff] [blame]576 <dd>Globals with "<tt>available_externally</tt>" linkage are never emitted
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]577 into the object file corresponding to the LLVM module. They exist to
578 allow inlining and other optimizations to take place given knowledge of
579 the definition of the global, which is known to be somewhere outside the
580 module. Globals with <tt>available_externally</tt> linkage are allowed to
581 be discarded at will, and are otherwise the same as <tt>linkonce_odr</tt>.
582 This linkage type is only allowed on definitions, not declarations.</dd>
Chris Lattner184f1be2009-04-13 05:44:34 +0000[diff] [blame]583
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]584 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt></dt>
Chris Lattnere20b4702007-01-14 06:51:48 +0000[diff] [blame]585 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
Chris Lattner0de4caa2010-01-09 19:15:14 +0000[diff] [blame]586 the same name when linkage occurs. This can be used to implement
587 some forms of inline functions, templates, or other code which must be
588 generated in each translation unit that uses it, but where the body may
589 be overridden with a more definitive definition later. Unreferenced
590 <tt>linkonce</tt> globals are allowed to be discarded. Note that
591 <tt>linkonce</tt> linkage does not actually allow the optimizer to
592 inline the body of this function into callers because it doesn't know if
593 this definition of the function is the definitive definition within the
594 program or whether it will be overridden by a stronger definition.
595 To enable inlining and other optimizations, use "<tt>linkonce_odr</tt>"
596 linkage.</dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]597
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]598 <dt><tt><b><a name="linkage_weak">weak</a></b></tt></dt>
Chris Lattnerd0554882009-08-05 05:21:07 +0000[diff] [blame]599 <dd>"<tt>weak</tt>" linkage has the same merging semantics as
600 <tt>linkonce</tt> linkage, except that unreferenced globals with
601 <tt>weak</tt> linkage may not be discarded. This is used for globals that
602 are declared "weak" in C source code.</dd>
603
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]604 <dt><tt><b><a name="linkage_common">common</a></b></tt></dt>
Chris Lattnerd0554882009-08-05 05:21:07 +0000[diff] [blame]605 <dd>"<tt>common</tt>" linkage is most similar to "<tt>weak</tt>" linkage, but
606 they are used for tentative definitions in C, such as "<tt>int X;</tt>" at
607 global scope.
608 Symbols with "<tt>common</tt>" linkage are merged in the same way as
609 <tt>weak symbols</tt>, and they may not be deleted if unreferenced.
Chris Lattner0aff0b22009-08-05 05:41:44 +0000[diff] [blame]610 <tt>common</tt> symbols may not have an explicit section,
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]611 must have a zero initializer, and may not be marked '<a
Chris Lattner0aff0b22009-08-05 05:41:44 +0000[diff] [blame]612 href="#globalvars"><tt>constant</tt></a>'. Functions and aliases may not
613 have common linkage.</dd>
Chris Lattnerd0554882009-08-05 05:21:07 +0000[diff] [blame]614
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]615
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]616 <dt><tt><b><a name="linkage_appending">appending</a></b></tt></dt>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]617 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]618 pointer to array type. When two global variables with appending linkage
619 are linked together, the two global arrays are appended together. This is
620 the LLVM, typesafe, equivalent of having the system linker append together
621 "sections" with identical names when .o files are linked.</dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]622
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]623 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]624 <dd>The semantics of this linkage follow the ELF object file model: the symbol
625 is weak until linked, if not linked, the symbol becomes null instead of
626 being an undefined reference.</dd>
Anton Korobeynikova0554d92007-01-12 19:20:47 +0000[diff] [blame]627
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]628 <dt><tt><b><a name="linkage_linkonce_odr">linkonce_odr</a></b></tt></dt>
629 <dt><tt><b><a name="linkage_weak_odr">weak_odr</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]630 <dd>Some languages allow differing globals to be merged, such as two functions
631 with different semantics. Other languages, such as <tt>C++</tt>, ensure
Bill Wendling03bcd6e2010-07-01 21:55:59 +0000[diff] [blame]632 that only equivalent globals are ever merged (the "one definition rule"
633 &mdash; "ODR"). Such languages can use the <tt>linkonce_odr</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]634 and <tt>weak_odr</tt> linkage types to indicate that the global will only
635 be merged with equivalent globals. These linkage types are otherwise the
636 same as their non-<tt>odr</tt> versions.</dd>
Duncan Sands12da8ce2009-03-07 15:45:40 +0000[diff] [blame]637
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]638 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]639 <dd>If none of the above identifiers are used, the global is externally
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]640 visible, meaning that it participates in linkage and can be used to
641 resolve external symbol references.</dd>
Reid Spencer7972c472007-04-11 23:49:50 +0000[diff] [blame]642</dl>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]643
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]644<p>The next two types of linkage are targeted for Microsoft Windows platform
645 only. They are designed to support importing (exporting) symbols from (to)
646 DLLs (Dynamic Link Libraries).</p>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]647
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]648<dl>
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]649 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt></dt>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]650 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]651 or variable via a global pointer to a pointer that is set up by the DLL
652 exporting the symbol. On Microsoft Windows targets, the pointer name is
653 formed by combining <code>__imp_</code> and the function or variable
654 name.</dd>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]655
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]656 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt></dt>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]657 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]658 pointer to a pointer in a DLL, so that it can be referenced with the
659 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
660 name is formed by combining <code>__imp_</code> and the function or
661 variable name.</dd>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]662</dl>
663
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]664<p>For example, since the "<tt>.LC0</tt>" variable is defined to be internal, if
665 another module defined a "<tt>.LC0</tt>" variable and was linked with this
666 one, one of the two would be renamed, preventing a collision. Since
667 "<tt>main</tt>" and "<tt>puts</tt>" are external (i.e., lacking any linkage
668 declarations), they are accessible outside of the current module.</p>
669
670<p>It is illegal for a function <i>declaration</i> to have any linkage type
671 other than "externally visible", <tt>dllimport</tt>
672 or <tt>extern_weak</tt>.</p>
673
Duncan Sands12da8ce2009-03-07 15:45:40 +0000[diff] [blame]674<p>Aliases can have only <tt>external</tt>, <tt>internal</tt>, <tt>weak</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]675 or <tt>weak_odr</tt> linkages.</p>
676
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]677</div>
678
679<!-- ======================================================================= -->
680<div class="doc_subsection">
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]681 <a name="callingconv">Calling Conventions</a>
682</div>
683
684<div class="doc_text">
685
686<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]687 and <a href="#i_invoke">invokes</a> can all have an optional calling
688 convention specified for the call. The calling convention of any pair of
689 dynamic caller/callee must match, or the behavior of the program is
690 undefined. The following calling conventions are supported by LLVM, and more
691 may be added in the future:</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]692
693<dl>
694 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]695 <dd>This calling convention (the default if no other calling convention is
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]696 specified) matches the target C calling conventions. This calling
697 convention supports varargs function calls and tolerates some mismatch in
698 the declared prototype and implemented declaration of the function (as
699 does normal C).</dd>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]700
701 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]702 <dd>This calling convention attempts to make calls as fast as possible
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]703 (e.g. by passing things in registers). This calling convention allows the
704 target to use whatever tricks it wants to produce fast code for the
705 target, without having to conform to an externally specified ABI
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]706 (Application Binary Interface).
707 <a href="CodeGenerator.html#tailcallopt">Tail calls can only be optimized
Chris Lattnera179e4d2010-03-11 00:22:57 +0000[diff] [blame]708 when this or the GHC convention is used.</a> This calling convention
709 does not support varargs and requires the prototype of all callees to
710 exactly match the prototype of the function definition.</dd>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]711
712 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]713 <dd>This calling convention attempts to make code in the caller as efficient
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]714 as possible under the assumption that the call is not commonly executed.
715 As such, these calls often preserve all registers so that the call does
716 not break any live ranges in the caller side. This calling convention
717 does not support varargs and requires the prototype of all callees to
718 exactly match the prototype of the function definition.</dd>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]719
Chris Lattnera179e4d2010-03-11 00:22:57 +0000[diff] [blame]720 <dt><b>"<tt>cc <em>10</em></tt>" - GHC convention</b>:</dt>
721 <dd>This calling convention has been implemented specifically for use by the
722 <a href="http://www.haskell.org/ghc">Glasgow Haskell Compiler (GHC)</a>.
723 It passes everything in registers, going to extremes to achieve this by
724 disabling callee save registers. This calling convention should not be
725 used lightly but only for specific situations such as an alternative to
726 the <em>register pinning</em> performance technique often used when
727 implementing functional programming languages.At the moment only X86
728 supports this convention and it has the following limitations:
729 <ul>
730 <li>On <em>X86-32</em> only supports up to 4 bit type parameters. No
731 floating point types are supported.</li>
732 <li>On <em>X86-64</em> only supports up to 10 bit type parameters and
733 6 floating point parameters.</li>
734 </ul>
735 This calling convention supports
736 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> but
737 requires both the caller and callee are using it.
738 </dd>
739
Chris Lattner573f64e2005-05-07 01:46:40 +0000[diff] [blame]740 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]741 <dd>Any calling convention may be specified by number, allowing
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]742 target-specific calling conventions to be used. Target specific calling
743 conventions start at 64.</dd>
Chris Lattner573f64e2005-05-07 01:46:40 +0000[diff] [blame]744</dl>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]745
746<p>More calling conventions can be added/defined on an as-needed basis, to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]747 support Pascal conventions or any other well-known target-independent
748 convention.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]749
750</div>
751
752<!-- ======================================================================= -->
753<div class="doc_subsection">
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]754 <a name="visibility">Visibility Styles</a>
755</div>
756
757<div class="doc_text">
758
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]759<p>All Global Variables and Functions have one of the following visibility
760 styles:</p>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]761
762<dl>
763 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]764 <dd>On targets that use the ELF object file format, default visibility means
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]765 that the declaration is visible to other modules and, in shared libraries,
766 means that the declared entity may be overridden. On Darwin, default
767 visibility means that the declaration is visible to other modules. Default
768 visibility corresponds to "external linkage" in the language.</dd>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]769
770 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]771 <dd>Two declarations of an object with hidden visibility refer to the same
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]772 object if they are in the same shared object. Usually, hidden visibility
773 indicates that the symbol will not be placed into the dynamic symbol
774 table, so no other module (executable or shared library) can reference it
775 directly.</dd>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]776
Anton Korobeynikov39f3cff2007-04-29 18:35:00 +0000[diff] [blame]777 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
Anton Korobeynikov39f3cff2007-04-29 18:35:00 +0000[diff] [blame]778 <dd>On ELF, protected visibility indicates that the symbol will be placed in
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]779 the dynamic symbol table, but that references within the defining module
780 will bind to the local symbol. That is, the symbol cannot be overridden by
781 another module.</dd>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]782</dl>
783
784</div>
785
786<!-- ======================================================================= -->
787<div class="doc_subsection">
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]788 <a name="namedtypes">Named Types</a>
789</div>
790
791<div class="doc_text">
792
793<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]794 it easier to read the IR and make the IR more condensed (particularly when
795 recursive types are involved). An example of a name specification is:</p>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]796
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]797<pre class="doc_code">
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]798%mytype = type { %mytype*, i32 }
799</pre>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]800
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]801<p>You may give a name to any <a href="#typesystem">type</a> except
Chris Lattner249b9762010-08-17 23:26:04 +0000[diff] [blame]802 "<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]803 is expected with the syntax "%mytype".</p>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]804
805<p>Note that type names are aliases for the structural type that they indicate,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]806 and that you can therefore specify multiple names for the same type. This
807 often leads to confusing behavior when dumping out a .ll file. Since LLVM IR
808 uses structural typing, the name is not part of the type. When printing out
809 LLVM IR, the printer will pick <em>one name</em> to render all types of a
810 particular shape. This means that if you have code where two different
811 source types end up having the same LLVM type, that the dumper will sometimes
812 print the "wrong" or unexpected type. This is an important design point and
813 isn't going to change.</p>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]814
815</div>
816
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]817<!-- ======================================================================= -->
818<div class="doc_subsection">
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]819 <a name="globalvars">Global Variables</a>
820</div>
821
822<div class="doc_text">
823
Chris Lattner5d5aede2005-02-12 19:30:21 +0000[diff] [blame]824<p>Global variables define regions of memory allocated at compilation time
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]825 instead of run-time. Global variables may optionally be initialized, may
826 have an explicit section to be placed in, and may have an optional explicit
827 alignment specified. A variable may be defined as "thread_local", which
828 means that it will not be shared by threads (each thread will have a
829 separated copy of the variable). A variable may be defined as a global
830 "constant," which indicates that the contents of the variable
831 will <b>never</b> be modified (enabling better optimization, allowing the
832 global data to be placed in the read-only section of an executable, etc).
833 Note that variables that need runtime initialization cannot be marked
834 "constant" as there is a store to the variable.</p>
Chris Lattner5d5aede2005-02-12 19:30:21 +0000[diff] [blame]835
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]836<p>LLVM explicitly allows <em>declarations</em> of global variables to be marked
837 constant, even if the final definition of the global is not. This capability
838 can be used to enable slightly better optimization of the program, but
839 requires the language definition to guarantee that optimizations based on the
840 'constantness' are valid for the translation units that do not include the
841 definition.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]842
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]843<p>As SSA values, global variables define pointer values that are in scope
844 (i.e. they dominate) all basic blocks in the program. Global variables
845 always define a pointer to their "content" type because they describe a
846 region of memory, and all memory objects in LLVM are accessed through
847 pointers.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]848
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame^]849<p>Global variables can be marked with <tt>unnamed_addr</tt> which indicates
850 that the address is not significant, only the content. Constants marked
851 like this can be merged if they have the same content.</p>
852
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]853<p>A global variable may be declared to reside in a target-specific numbered
854 address space. For targets that support them, address spaces may affect how
855 optimizations are performed and/or what target instructions are used to
856 access the variable. The default address space is zero. The address space
857 qualifier must precede any other attributes.</p>
Christopher Lamb308121c2007-12-11 09:31:00 +0000[diff] [blame]858
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]859<p>LLVM allows an explicit section to be specified for globals. If the target
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]860 supports it, it will emit globals to the section specified.</p>
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]861
Chris Lattner78e00bc2010-04-28 00:13:42 +0000[diff] [blame]862<p>An explicit alignment may be specified for a global, which must be a power
863 of 2. If not present, or if the alignment is set to zero, the alignment of
864 the global is set by the target to whatever it feels convenient. If an
865 explicit alignment is specified, the global is forced to have exactly that
Chris Lattner4bd85e42010-04-28 00:31:12 +0000[diff] [blame]866 alignment. Targets and optimizers are not allowed to over-align the global
867 if the global has an assigned section. In this case, the extra alignment
868 could be observable: for example, code could assume that the globals are
869 densely packed in their section and try to iterate over them as an array,
870 alignment padding would break this iteration.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]871
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]872<p>For example, the following defines a global in a numbered address space with
873 an initializer, section, and alignment:</p>
Chris Lattner5760c502007-01-14 00:27:09 +0000[diff] [blame]874
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]875<pre class="doc_code">
Dan Gohmanaaa679b2009-01-11 00:40:00 +0000[diff] [blame]876@G = addrspace(5) constant float 1.0, section "foo", align 4
Chris Lattner5760c502007-01-14 00:27:09 +0000[diff] [blame]877</pre>
878
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]879</div>
880
881
882<!-- ======================================================================= -->
883<div class="doc_subsection">
884 <a name="functionstructure">Functions</a>
885</div>
886
887<div class="doc_text">
888
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]889<p>LLVM function definitions consist of the "<tt>define</tt>" keyword, an
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]890 optional <a href="#linkage">linkage type</a>, an optional
891 <a href="#visibility">visibility style</a>, an optional
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame^]892 <a href="#callingconv">calling convention</a>,
893 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]894 <a href="#paramattrs">parameter attribute</a> for the return type, a function
895 name, a (possibly empty) argument list (each with optional
896 <a href="#paramattrs">parameter attributes</a>), optional
897 <a href="#fnattrs">function attributes</a>, an optional section, an optional
898 alignment, an optional <a href="#gc">garbage collector name</a>, an opening
899 curly brace, a list of basic blocks, and a closing curly brace.</p>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]900
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]901<p>LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
902 optional <a href="#linkage">linkage type</a>, an optional
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]903 <a href="#visibility">visibility style</a>, an optional
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame^]904 <a href="#callingconv">calling convention</a>,
905 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]906 <a href="#paramattrs">parameter attribute</a> for the return type, a function
907 name, a possibly empty list of arguments, an optional alignment, and an
908 optional <a href="#gc">garbage collector name</a>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]909
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]910<p>A function definition contains a list of basic blocks, forming the CFG
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]911 (Control Flow Graph) for the function. Each basic block may optionally start
912 with a label (giving the basic block a symbol table entry), contains a list
913 of instructions, and ends with a <a href="#terminators">terminator</a>
914 instruction (such as a branch or function return).</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]915
Chris Lattnera59fb102007-06-08 16:52:14 +0000[diff] [blame]916<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]917 executed on entrance to the function, and it is not allowed to have
918 predecessor basic blocks (i.e. there can not be any branches to the entry
919 block of a function). Because the block can have no predecessors, it also
920 cannot have any <a href="#i_phi">PHI nodes</a>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]921
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]922<p>LLVM allows an explicit section to be specified for functions. If the target
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]923 supports it, it will emit functions to the section specified.</p>
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]924
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]925<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]926 the alignment is set to zero, the alignment of the function is set by the
927 target to whatever it feels convenient. If an explicit alignment is
928 specified, the function is forced to have at least that much alignment. All
929 alignments must be a power of 2.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]930
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame^]931<p>If the <tt>unnamed_addr</tt> attribute is given, the address is know to not
932 be significant and two identical functions can be merged</p>.
933
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]934<h5>Syntax:</h5>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]935<pre class="doc_code">
Chris Lattner0ae02092008-10-13 16:55:18 +0000[diff] [blame]936define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]937 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
938 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
939 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
940 [<a href="#gc">gc</a>] { ... }
941</pre>
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]942
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]943</div>
944
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]945<!-- ======================================================================= -->
946<div class="doc_subsection">
947 <a name="aliasstructure">Aliases</a>
948</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]949
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]950<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]951
952<p>Aliases act as "second name" for the aliasee value (which can be either
953 function, global variable, another alias or bitcast of global value). Aliases
954 may have an optional <a href="#linkage">linkage type</a>, and an
955 optional <a href="#visibility">visibility style</a>.</p>
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]956
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]957<h5>Syntax:</h5>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]958<pre class="doc_code">
Duncan Sands7e99a942008-09-12 20:48:21 +0000[diff] [blame]959@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Bill Wendling2d8b9a82007-05-29 09:42:13 +0000[diff] [blame]960</pre>
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]961
962</div>
963
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]964<!-- ======================================================================= -->
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]965<div class="doc_subsection">
966 <a name="namedmetadatastructure">Named Metadata</a>
967</div>
968
969<div class="doc_text">
970
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]971<p>Named metadata is a collection of metadata. <a href="#metadata">Metadata
Dan Gohman093cb792010-07-21 18:54:18 +0000[diff] [blame]972 nodes</a> (but not metadata strings) are the only valid operands for
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]973 a named metadata.</p>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]974
975<h5>Syntax:</h5>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]976<pre class="doc_code">
Dan Gohman093cb792010-07-21 18:54:18 +0000[diff] [blame]977; Some unnamed metadata nodes, which are referenced by the named metadata.
978!0 = metadata !{metadata !"zero"}
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]979!1 = metadata !{metadata !"one"}
Dan Gohman093cb792010-07-21 18:54:18 +0000[diff] [blame]980!2 = metadata !{metadata !"two"}
Dan Gohman58cd65f2010-07-13 19:48:13 +0000[diff] [blame]981; A named metadata.
Dan Gohman093cb792010-07-21 18:54:18 +0000[diff] [blame]982!name = !{!0, !1, !2}
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]983</pre>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]984
985</div>
986
987<!-- ======================================================================= -->
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]988<div class="doc_subsection"><a name="paramattrs">Parameter Attributes</a></div>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]989
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]990<div class="doc_text">
991
992<p>The return type and each parameter of a function type may have a set of
993 <i>parameter attributes</i> associated with them. Parameter attributes are
994 used to communicate additional information about the result or parameters of
995 a function. Parameter attributes are considered to be part of the function,
996 not of the function type, so functions with different parameter attributes
997 can have the same function type.</p>
998
999<p>Parameter attributes are simple keywords that follow the type specified. If
1000 multiple parameter attributes are needed, they are space separated. For
1001 example:</p>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]1002
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1003<pre class="doc_code">
Nick Lewyckydac78d82009-02-15 23:06:14 +0000[diff] [blame]1004declare i32 @printf(i8* noalias nocapture, ...)
Chris Lattnerd2597d72008-10-04 18:33:34 +0000[diff] [blame]1005declare i32 @atoi(i8 zeroext)
1006declare signext i8 @returns_signed_char()
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]1007</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]1008
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1009<p>Note that any attributes for the function result (<tt>nounwind</tt>,
1010 <tt>readonly</tt>) come immediately after the argument list.</p>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1011
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1012<p>Currently, only the following parameter attributes are defined:</p>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000[diff] [blame]1013
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1014<dl>
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1015 <dt><tt><b>zeroext</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1016 <dd>This indicates to the code generator that the parameter or return value
1017 should be zero-extended to a 32-bit value by the caller (for a parameter)
1018 or the callee (for a return value).</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000[diff] [blame]1019
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1020 <dt><tt><b>signext</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1021 <dd>This indicates to the code generator that the parameter or return value
1022 should be sign-extended to a 32-bit value by the caller (for a parameter)
1023 or the callee (for a return value).</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000[diff] [blame]1024
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1025 <dt><tt><b>inreg</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1026 <dd>This indicates that this parameter or return value should be treated in a
1027 special target-dependent fashion during while emitting code for a function
1028 call or return (usually, by putting it in a register as opposed to memory,
1029 though some targets use it to distinguish between two different kinds of
1030 registers). Use of this attribute is target-specific.</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000[diff] [blame]1031
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1032 <dt><tt><b><a name="byval">byval</a></b></tt></dt>
Chris Lattnerd78dbee2010-11-20 23:49:06 +0000[diff] [blame]1033 <dd><p>This indicates that the pointer parameter should really be passed by
1034 value to the function. The attribute implies that a hidden copy of the
1035 pointee
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1036 is made between the caller and the callee, so the callee is unable to
1037 modify the value in the callee. This attribute is only valid on LLVM
1038 pointer arguments. It is generally used to pass structs and arrays by
1039 value, but is also valid on pointers to scalars. The copy is considered
1040 to belong to the caller not the callee (for example,
1041 <tt><a href="#readonly">readonly</a></tt> functions should not write to
1042 <tt>byval</tt> parameters). This is not a valid attribute for return
Chris Lattnerd78dbee2010-11-20 23:49:06 +0000[diff] [blame]1043 values.</p>
1044
1045 <p>The byval attribute also supports specifying an alignment with
1046 the align attribute. It indicates the alignment of the stack slot to
1047 form and the known alignment of the pointer specified to the call site. If
1048 the alignment is not specified, then the code generator makes a
1049 target-specific assumption.</p></dd>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1050
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]1051 <dt><tt><b><a name="sret">sret</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1052 <dd>This indicates that the pointer parameter specifies the address of a
1053 structure that is the return value of the function in the source program.
1054 This pointer must be guaranteed by the caller to be valid: loads and
1055 stores to the structure may be assumed by the callee to not to trap. This
1056 may only be applied to the first parameter. This is not a valid attribute
1057 for return values. </dd>
1058
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]1059 <dt><tt><b><a name="noalias">noalias</a></b></tt></dt>
Dan Gohmandf12d082010-07-02 18:41:32 +0000[diff] [blame]1060 <dd>This indicates that pointer values
1061 <a href="#pointeraliasing"><i>based</i></a> on the argument or return
Dan Gohmande256292010-07-02 23:46:54 +0000[diff] [blame]1062 value do not alias pointer values which are not <i>based</i> on it,
1063 ignoring certain "irrelevant" dependencies.
1064 For a call to the parent function, dependencies between memory
1065 references from before or after the call and from those during the call
1066 are "irrelevant" to the <tt>noalias</tt> keyword for the arguments and
1067 return value used in that call.
Dan Gohmandf12d082010-07-02 18:41:32 +0000[diff] [blame]1068 The caller shares the responsibility with the callee for ensuring that
1069 these requirements are met.
1070 For further details, please see the discussion of the NoAlias response in
Dan Gohman6c858db2010-07-06 15:26:33 +0000[diff] [blame]1071 <a href="AliasAnalysis.html#MustMayNo">alias analysis</a>.<br>
1072<br>
John McCall72ed8902010-07-06 21:07:14 +0000[diff] [blame]1073 Note that this definition of <tt>noalias</tt> is intentionally
1074 similar to the definition of <tt>restrict</tt> in C99 for function
Chris Lattner5eff9ca2010-07-06 20:51:35 +0000[diff] [blame]1075 arguments, though it is slightly weaker.
Dan Gohman6c858db2010-07-06 15:26:33 +0000[diff] [blame]1076<br>
1077 For function return values, C99's <tt>restrict</tt> is not meaningful,
1078 while LLVM's <tt>noalias</tt> is.
1079 </dd>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1080
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]1081 <dt><tt><b><a name="nocapture">nocapture</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1082 <dd>This indicates that the callee does not make any copies of the pointer
1083 that outlive the callee itself. This is not a valid attribute for return
1084 values.</dd>
1085
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]1086 <dt><tt><b><a name="nest">nest</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1087 <dd>This indicates that the pointer parameter can be excised using the
1088 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
1089 attribute for return values.</dd>
1090</dl>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1091
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1092</div>
1093
1094<!-- ======================================================================= -->
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1095<div class="doc_subsection">
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1096 <a name="gc">Garbage Collector Names</a>
1097</div>
1098
1099<div class="doc_text">
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1100
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1101<p>Each function may specify a garbage collector name, which is simply a
1102 string:</p>
1103
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1104<pre class="doc_code">
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1105define void @f() gc "name" { ... }
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1106</pre>
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1107
1108<p>The compiler declares the supported values of <i>name</i>. Specifying a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1109 collector which will cause the compiler to alter its output in order to
1110 support the named garbage collection algorithm.</p>
1111
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1112</div>
1113
1114<!-- ======================================================================= -->
1115<div class="doc_subsection">
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1116 <a name="fnattrs">Function Attributes</a>
Devang Patelcaacdba2008-09-04 23:05:13 +0000[diff] [blame]1117</div>
1118
1119<div class="doc_text">
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1120
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1121<p>Function attributes are set to communicate additional information about a
1122 function. Function attributes are considered to be part of the function, not
1123 of the function type, so functions with different parameter attributes can
1124 have the same function type.</p>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1125
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1126<p>Function attributes are simple keywords that follow the type specified. If
1127 multiple attributes are needed, they are space separated. For example:</p>
Devang Patelcaacdba2008-09-04 23:05:13 +0000[diff] [blame]1128
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1129<pre class="doc_code">
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1130define void @f() noinline { ... }
1131define void @f() alwaysinline { ... }
1132define void @f() alwaysinline optsize { ... }
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1133define void @f() optsize { ... }
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1134</pre>
Devang Patelcaacdba2008-09-04 23:05:13 +0000[diff] [blame]1135
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1136<dl>
Charles Davisbe5557e2010-02-12 00:31:15 +0000[diff] [blame]1137 <dt><tt><b>alignstack(&lt;<em>n</em>&gt;)</b></tt></dt>
1138 <dd>This attribute indicates that, when emitting the prologue and epilogue,
1139 the backend should forcibly align the stack pointer. Specify the
1140 desired alignment, which must be a power of two, in parentheses.
1141
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1142 <dt><tt><b>alwaysinline</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1143 <dd>This attribute indicates that the inliner should attempt to inline this
1144 function into callers whenever possible, ignoring any active inlining size
1145 threshold for this caller.</dd>
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1146
Charles Davis22fe1862010-10-25 15:37:09 +0000[diff] [blame]1147 <dt><tt><b>hotpatch</b></tt></dt>
Charles Davis1b2d3722010-10-25 16:29:03 +0000[diff] [blame]1148 <dd>This attribute indicates that the function should be 'hotpatchable',
Charles Davis74205252010-10-25 19:07:39 +0000[diff] [blame]1149 meaning the function can be patched and/or hooked even while it is
1150 loaded into memory. On x86, the function prologue will be preceded
1151 by six bytes of padding and will begin with a two-byte instruction.
1152 Most of the functions in the Windows system DLLs in Windows XP SP2 or
1153 higher were compiled in this fashion.</dd>
Charles Davis22fe1862010-10-25 15:37:09 +0000[diff] [blame]1154
Jakob Stoklund Olesen74bb06c2010-02-06 01:16:28 +0000[diff] [blame]1155 <dt><tt><b>inlinehint</b></tt></dt>
1156 <dd>This attribute indicates that the source code contained a hint that inlining
1157 this function is desirable (such as the "inline" keyword in C/C++). It
1158 is just a hint; it imposes no requirements on the inliner.</dd>
1159
Nick Lewycky14b58da2010-07-06 18:24:09 +0000[diff] [blame]1160 <dt><tt><b>naked</b></tt></dt>
1161 <dd>This attribute disables prologue / epilogue emission for the function.
1162 This can have very system-specific consequences.</dd>
1163
1164 <dt><tt><b>noimplicitfloat</b></tt></dt>
1165 <dd>This attributes disables implicit floating point instructions.</dd>
1166
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1167 <dt><tt><b>noinline</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1168 <dd>This attribute indicates that the inliner should never inline this
1169 function in any situation. This attribute may not be used together with
1170 the <tt>alwaysinline</tt> attribute.</dd>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1171
Nick Lewycky14b58da2010-07-06 18:24:09 +0000[diff] [blame]1172 <dt><tt><b>noredzone</b></tt></dt>
1173 <dd>This attribute indicates that the code generator should not use a red
1174 zone, even if the target-specific ABI normally permits it.</dd>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1175
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1176 <dt><tt><b>noreturn</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1177 <dd>This function attribute indicates that the function never returns
1178 normally. This produces undefined behavior at runtime if the function
1179 ever does dynamically return.</dd>
Bill Wendlinga8130172008-11-13 01:02:51 +0000[diff] [blame]1180
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1181 <dt><tt><b>nounwind</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1182 <dd>This function attribute indicates that the function never returns with an
1183 unwind or exceptional control flow. If the function does unwind, its
1184 runtime behavior is undefined.</dd>
Bill Wendling0f5541e2008-11-26 19:07:40 +0000[diff] [blame]1185
Nick Lewycky14b58da2010-07-06 18:24:09 +0000[diff] [blame]1186 <dt><tt><b>optsize</b></tt></dt>
1187 <dd>This attribute suggests that optimization passes and code generator passes
1188 make choices that keep the code size of this function low, and otherwise
1189 do optimizations specifically to reduce code size.</dd>
1190
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1191 <dt><tt><b>readnone</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1192 <dd>This attribute indicates that the function computes its result (or decides
1193 to unwind an exception) based strictly on its arguments, without
1194 dereferencing any pointer arguments or otherwise accessing any mutable
1195 state (e.g. memory, control registers, etc) visible to caller functions.
1196 It does not write through any pointer arguments
1197 (including <tt><a href="#byval">byval</a></tt> arguments) and never
1198 changes any state visible to callers. This means that it cannot unwind
1199 exceptions by calling the <tt>C++</tt> exception throwing methods, but
1200 could use the <tt>unwind</tt> instruction.</dd>
Devang Patel310fd4a2009-06-12 19:45:19 +0000[diff] [blame]1201
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1202 <dt><tt><b><a name="readonly">readonly</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1203 <dd>This attribute indicates that the function does not write through any
1204 pointer arguments (including <tt><a href="#byval">byval</a></tt>
1205 arguments) or otherwise modify any state (e.g. memory, control registers,
1206 etc) visible to caller functions. It may dereference pointer arguments
1207 and read state that may be set in the caller. A readonly function always
1208 returns the same value (or unwinds an exception identically) when called
1209 with the same set of arguments and global state. It cannot unwind an
1210 exception by calling the <tt>C++</tt> exception throwing methods, but may
1211 use the <tt>unwind</tt> instruction.</dd>
Anton Korobeynikovc8ce7b082009-07-17 18:07:26 +0000[diff] [blame]1212
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1213 <dt><tt><b><a name="ssp">ssp</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1214 <dd>This attribute indicates that the function should emit a stack smashing
1215 protector. It is in the form of a "canary"&mdash;a random value placed on
1216 the stack before the local variables that's checked upon return from the
1217 function to see if it has been overwritten. A heuristic is used to
1218 determine if a function needs stack protectors or not.<br>
1219<br>
1220 If a function that has an <tt>ssp</tt> attribute is inlined into a
1221 function that doesn't have an <tt>ssp</tt> attribute, then the resulting
1222 function will have an <tt>ssp</tt> attribute.</dd>
1223
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1224 <dt><tt><b>sspreq</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1225 <dd>This attribute indicates that the function should <em>always</em> emit a
1226 stack smashing protector. This overrides
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1227 the <tt><a href="#ssp">ssp</a></tt> function attribute.<br>
1228<br>
1229 If a function that has an <tt>sspreq</tt> attribute is inlined into a
1230 function that doesn't have an <tt>sspreq</tt> attribute or which has
1231 an <tt>ssp</tt> attribute, then the resulting function will have
1232 an <tt>sspreq</tt> attribute.</dd>
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1233</dl>
1234
Devang Patelcaacdba2008-09-04 23:05:13 +0000[diff] [blame]1235</div>
1236
1237<!-- ======================================================================= -->
1238<div class="doc_subsection">
Chris Lattner93564892006-04-08 04:40:53 +0000[diff] [blame]1239 <a name="moduleasm">Module-Level Inline Assembly</a>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1240</div>
1241
1242<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1243
1244<p>Modules may contain "module-level inline asm" blocks, which corresponds to
1245 the GCC "file scope inline asm" blocks. These blocks are internally
1246 concatenated by LLVM and treated as a single unit, but may be separated in
1247 the <tt>.ll</tt> file if desired. The syntax is very simple:</p>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1248
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1249<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]1250module asm "inline asm code goes here"
1251module asm "more can go here"
1252</pre>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1253
1254<p>The strings can contain any character by escaping non-printable characters.
1255 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]1256 for the number.</p>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1257
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1258<p>The inline asm code is simply printed to the machine code .s file when
1259 assembly code is generated.</p>
1260
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1261</div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1262
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1263<!-- ======================================================================= -->
1264<div class="doc_subsection">
1265 <a name="datalayout">Data Layout</a>
1266</div>
1267
1268<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1269
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1270<p>A module may specify a target specific data layout string that specifies how
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1271 data is to be laid out in memory. The syntax for the data layout is
1272 simply:</p>
1273
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1274<pre class="doc_code">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1275target datalayout = "<i>layout specification</i>"
1276</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1277
1278<p>The <i>layout specification</i> consists of a list of specifications
1279 separated by the minus sign character ('-'). Each specification starts with
1280 a letter and may include other information after the letter to define some
1281 aspect of the data layout. The specifications accepted are as follows:</p>
1282
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1283<dl>
1284 <dt><tt>E</tt></dt>
1285 <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]1286 bits with the most significance have the lowest address location.</dd>
1287
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1288 <dt><tt>e</tt></dt>
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]1289 <dd>Specifies that the target lays out data in little-endian form. That is,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1290 the bits with the least significance have the lowest address
1291 location.</dd>
1292
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1293 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1294 <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]1295 <i>preferred</i> alignments. All sizes are in bits. Specifying
1296 the <i>pref</i> alignment is optional. If omitted, the
1297 preceding <tt>:</tt> should be omitted too.</dd>
1298
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1299 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1300 <dd>This specifies the alignment for an integer type of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1301 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1302
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1303 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1304 <dd>This specifies the alignment for a vector type of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1305 <i>size</i>.</dd>
1306
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1307 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1308 <dd>This specifies the alignment for a floating point type of a given bit
Dale Johannesence522852010-05-28 18:54:47 +0000[diff] [blame]1309 <i>size</i>. Only values of <i>size</i> that are supported by the target
1310 will work. 32 (float) and 64 (double) are supported on all targets;
1311 80 or 128 (different flavors of long double) are also supported on some
1312 targets.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1313
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1314 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1315 <dd>This specifies the alignment for an aggregate type of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1316 <i>size</i>.</dd>
1317
Daniel Dunbar7921a592009-06-08 22:17:53 +0000[diff] [blame]1318 <dt><tt>s<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1319 <dd>This specifies the alignment for a stack object of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1320 <i>size</i>.</dd>
Chris Lattnera381eff2009-11-07 09:35:34 +0000[diff] [blame]1321
1322 <dt><tt>n<i>size1</i>:<i>size2</i>:<i>size3</i>...</tt></dt>
1323 <dd>This specifies a set of native integer widths for the target CPU
1324 in bits. For example, it might contain "n32" for 32-bit PowerPC,
1325 "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]1326 this set are considered to support most general arithmetic
Chris Lattnera381eff2009-11-07 09:35:34 +0000[diff] [blame]1327 operations efficiently.</dd>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1328</dl>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1329
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1330<p>When constructing the data layout for a given target, LLVM starts with a
Dan Gohman61110ae2010-04-28 00:36:01 +0000[diff] [blame]1331 default set of specifications which are then (possibly) overridden by the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1332 specifications in the <tt>datalayout</tt> keyword. The default specifications
1333 are given in this list:</p>
1334
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1335<ul>
1336 <li><tt>E</tt> - big endian</li>
Dan Gohman8ad777d2010-02-23 02:44:03 +0000[diff] [blame]1337 <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]1338 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1339 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1340 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1341 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]1342 <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]1343 alignment of 64-bits</li>
1344 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1345 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1346 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1347 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1348 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
Daniel Dunbar7921a592009-06-08 22:17:53 +0000[diff] [blame]1349 <li><tt>s0:64:64</tt> - stack objects are 64-bit aligned</li>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1350</ul>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1351
1352<p>When LLVM is determining the alignment for a given type, it uses the
1353 following rules:</p>
1354
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1355<ol>
1356 <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]1357 specification is used.</li>
1358
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1359 <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]1360 smallest integer type that is larger than the bitwidth of the sought type
1361 is used. If none of the specifications are larger than the bitwidth then
1362 the the largest integer type is used. For example, given the default
1363 specifications above, the i7 type will use the alignment of i8 (next
1364 largest) while both i65 and i256 will use the alignment of i64 (largest
1365 specified).</li>
1366
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1367 <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]1368 largest vector type that is smaller than the sought vector type will be
1369 used as a fall back. This happens because &lt;128 x double&gt; can be
1370 implemented in terms of 64 &lt;2 x double&gt;, for example.</li>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1371</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1372
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1373</div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1374
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1375<!-- ======================================================================= -->
1376<div class="doc_subsection">
1377 <a name="pointeraliasing">Pointer Aliasing Rules</a>
1378</div>
1379
1380<div class="doc_text">
1381
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1382<p>Any memory access must be done through a pointer value associated
Andreas Bolkae39f0332009-07-27 20:37:10 +0000[diff] [blame]1383with an address range of the memory access, otherwise the behavior
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1384is undefined. Pointer values are associated with address ranges
1385according to the following rules:</p>
1386
1387<ul>
Dan Gohmandf12d082010-07-02 18:41:32 +0000[diff] [blame]1388 <li>A pointer value is associated with the addresses associated with
1389 any value it is <i>based</i> on.
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1390 <li>An address of a global variable is associated with the address
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1391 range of the variable's storage.</li>
1392 <li>The result value of an allocation instruction is associated with
1393 the address range of the allocated storage.</li>
1394 <li>A null pointer in the default address-space is associated with
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1395 no address.</li>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1396 <li>An integer constant other than zero or a pointer value returned
1397 from a function not defined within LLVM may be associated with address
1398 ranges allocated through mechanisms other than those provided by
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1399 LLVM. Such ranges shall not overlap with any ranges of addresses
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1400 allocated by mechanisms provided by LLVM.</li>
Dan Gohmandf12d082010-07-02 18:41:32 +0000[diff] [blame]1401</ul>
1402
1403<p>A pointer value is <i>based</i> on another pointer value according
1404 to the following rules:</p>
1405
1406<ul>
1407 <li>A pointer value formed from a
1408 <tt><a href="#i_getelementptr">getelementptr</a></tt> operation
1409 is <i>based</i> on the first operand of the <tt>getelementptr</tt>.</li>
1410 <li>The result value of a
1411 <tt><a href="#i_bitcast">bitcast</a></tt> is <i>based</i> on the operand
1412 of the <tt>bitcast</tt>.</li>
1413 <li>A pointer value formed by an
1414 <tt><a href="#i_inttoptr">inttoptr</a></tt> is <i>based</i> on all
1415 pointer values that contribute (directly or indirectly) to the
1416 computation of the pointer's value.</li>
1417 <li>The "<i>based</i> on" relationship is transitive.</li>
1418</ul>
1419
1420<p>Note that this definition of <i>"based"</i> is intentionally
1421 similar to the definition of <i>"based"</i> in C99, though it is
1422 slightly weaker.</p>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1423
1424<p>LLVM IR does not associate types with memory. The result type of a
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1425<tt><a href="#i_load">load</a></tt> merely indicates the size and
1426alignment of the memory from which to load, as well as the
Dan Gohman4eb47192010-06-17 19:23:50 +0000[diff] [blame]1427interpretation of the value. The first operand type of a
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1428<tt><a href="#i_store">store</a></tt> similarly only indicates the size
1429and alignment of the store.</p>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1430
1431<p>Consequently, type-based alias analysis, aka TBAA, aka
1432<tt>-fstrict-aliasing</tt>, is not applicable to general unadorned
1433LLVM IR. <a href="#metadata">Metadata</a> may be used to encode
1434additional information which specialized optimization passes may use
1435to implement type-based alias analysis.</p>
1436
1437</div>
1438
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]1439<!-- ======================================================================= -->
1440<div class="doc_subsection">
1441 <a name="volatile">Volatile Memory Accesses</a>
1442</div>
1443
1444<div class="doc_text">
1445
1446<p>Certain memory accesses, such as <a href="#i_load"><tt>load</tt></a>s, <a
1447href="#i_store"><tt>store</tt></a>s, and <a
1448href="#int_memcpy"><tt>llvm.memcpy</tt></a>s may be marked <tt>volatile</tt>.
1449The optimizers must not change the number of volatile operations or change their
1450order of execution relative to other volatile operations. The optimizers
1451<i>may</i> change the order of volatile operations relative to non-volatile
1452operations. This is not Java's "volatile" and has no cross-thread
1453synchronization behavior.</p>
1454
1455</div>
1456
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1457<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1458<div class="doc_section"> <a name="typesystem">Type System</a> </div>
1459<!-- *********************************************************************** -->
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1460
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1461<div class="doc_text">
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1462
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1463<p>The LLVM type system is one of the most important features of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1464 intermediate representation. Being typed enables a number of optimizations
1465 to be performed on the intermediate representation directly, without having
1466 to do extra analyses on the side before the transformation. A strong type
1467 system makes it easier to read the generated code and enables novel analyses
1468 and transformations that are not feasible to perform on normal three address
1469 code representations.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1470
1471</div>
1472
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1473<!-- ======================================================================= -->
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1474<div class="doc_subsection"> <a name="t_classifications">Type
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1475Classifications</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1476
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1477<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1478
1479<p>The types fall into a few useful classifications:</p>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]1480
1481<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1482 <tbody>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1483 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1484 <tr>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1485 <td><a href="#t_integer">integer</a></td>
Reid Spencer138249b2007-05-16 18:44:01 +0000[diff] [blame]1486 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1487 </tr>
1488 <tr>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1489 <td><a href="#t_floating">floating point</a></td>
1490 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1491 </tr>
1492 <tr>
1493 <td><a name="t_firstclass">first class</a></td>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1494 <td><a href="#t_integer">integer</a>,
1495 <a href="#t_floating">floating point</a>,
1496 <a href="#t_pointer">pointer</a>,
Dan Gohman08783a882008-06-18 18:42:13 +0000[diff] [blame]1497 <a href="#t_vector">vector</a>,
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]1498 <a href="#t_struct">structure</a>,
1499 <a href="#t_array">array</a>,
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1500 <a href="#t_label">label</a>,
1501 <a href="#t_metadata">metadata</a>.
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1502 </td>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1503 </tr>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1504 <tr>
1505 <td><a href="#t_primitive">primitive</a></td>
1506 <td><a href="#t_label">label</a>,
1507 <a href="#t_void">void</a>,
Tobias Grosser4c8c95b2010-12-28 20:29:31 +0000[diff] [blame]1508 <a href="#t_integer">integer</a>,
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1509 <a href="#t_floating">floating point</a>,
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]1510 <a href="#t_x86mmx">x86mmx</a>,
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1511 <a href="#t_metadata">metadata</a>.</td>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1512 </tr>
1513 <tr>
1514 <td><a href="#t_derived">derived</a></td>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1515 <td><a href="#t_array">array</a>,
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1516 <a href="#t_function">function</a>,
1517 <a href="#t_pointer">pointer</a>,
1518 <a href="#t_struct">structure</a>,
1519 <a href="#t_pstruct">packed structure</a>,
1520 <a href="#t_vector">vector</a>,
1521 <a href="#t_opaque">opaque</a>.
Dan Gohman93bf60d2008-10-14 16:32:04 +0000[diff] [blame]1522 </td>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1523 </tr>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1524 </tbody>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1525</table>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]1526
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1527<p>The <a href="#t_firstclass">first class</a> types are perhaps the most
1528 important. Values of these types are the only ones which can be produced by
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1529 instructions.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1530
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1531</div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1532
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1533<!-- ======================================================================= -->
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1534<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Chris Lattner43542b32008-01-04 04:34:14 +0000[diff] [blame]1535
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1536<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1537
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1538<p>The primitive types are the fundamental building blocks of the LLVM
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1539 system.</p>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1540
Chris Lattner43542b32008-01-04 04:34:14 +0000[diff] [blame]1541</div>
1542
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1543<!-- _______________________________________________________________________ -->
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1544<div class="doc_subsubsection"> <a name="t_integer">Integer Type</a> </div>
1545
1546<div class="doc_text">
1547
1548<h5>Overview:</h5>
1549<p>The integer type is a very simple type that simply specifies an arbitrary
1550 bit width for the integer type desired. Any bit width from 1 bit to
1551 2<sup>23</sup>-1 (about 8 million) can be specified.</p>
1552
1553<h5>Syntax:</h5>
1554<pre>
1555 iN
1556</pre>
1557
1558<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1559 value.</p>
1560
1561<h5>Examples:</h5>
1562<table class="layout">
1563 <tr class="layout">
1564 <td class="left"><tt>i1</tt></td>
1565 <td class="left">a single-bit integer.</td>
1566 </tr>
1567 <tr class="layout">
1568 <td class="left"><tt>i32</tt></td>
1569 <td class="left">a 32-bit integer.</td>
1570 </tr>
1571 <tr class="layout">
1572 <td class="left"><tt>i1942652</tt></td>
1573 <td class="left">a really big integer of over 1 million bits.</td>
1574 </tr>
1575</table>
1576
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1577</div>
1578
1579<!-- _______________________________________________________________________ -->
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1580<div class="doc_subsubsection"> <a name="t_floating">Floating Point Types</a> </div>
1581
1582<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1583
1584<table>
1585 <tbody>
1586 <tr><th>Type</th><th>Description</th></tr>
1587 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1588 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1589 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1590 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1591 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1592 </tbody>
1593</table>
1594
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1595</div>
1596
1597<!-- _______________________________________________________________________ -->
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]1598<div class="doc_subsubsection"> <a name="t_x86mmx">X86mmx Type</a> </div>
1599
1600<div class="doc_text">
1601
1602<h5>Overview:</h5>
1603<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>
1604
1605<h5>Syntax:</h5>
1606<pre>
Dale Johannesenb1f0ff12010-10-01 01:07:02 +0000[diff] [blame]1607 x86mmx
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]1608</pre>
1609
1610</div>
1611
1612<!-- _______________________________________________________________________ -->
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1613<div class="doc_subsubsection"> <a name="t_void">Void Type</a> </div>
1614
1615<div class="doc_text">
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1616
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1617<h5>Overview:</h5>
1618<p>The void type does not represent any value and has no size.</p>
1619
1620<h5>Syntax:</h5>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1621<pre>
1622 void
1623</pre>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1624
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1625</div>
1626
1627<!-- _______________________________________________________________________ -->
1628<div class="doc_subsubsection"> <a name="t_label">Label Type</a> </div>
1629
1630<div class="doc_text">
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1631
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1632<h5>Overview:</h5>
1633<p>The label type represents code labels.</p>
1634
1635<h5>Syntax:</h5>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1636<pre>
1637 label
1638</pre>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1639
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1640</div>
1641
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1642<!-- _______________________________________________________________________ -->
1643<div class="doc_subsubsection"> <a name="t_metadata">Metadata Type</a> </div>
1644
1645<div class="doc_text">
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1646
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1647<h5>Overview:</h5>
Nick Lewycky93e06a52009-09-27 23:27:42 +0000[diff] [blame]1648<p>The metadata type represents embedded metadata. No derived types may be
1649 created from metadata except for <a href="#t_function">function</a>
1650 arguments.
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1651
1652<h5>Syntax:</h5>
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1653<pre>
1654 metadata
1655</pre>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1656
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1657</div>
1658
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1659
1660<!-- ======================================================================= -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1661<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1662
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1663<div class="doc_text">
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1664
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1665<p>The real power in LLVM comes from the derived types in the system. This is
1666 what allows a programmer to represent arrays, functions, pointers, and other
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1667 useful types. Each of these types contain one or more element types which
1668 may be a primitive type, or another derived type. For example, it is
1669 possible to have a two dimensional array, using an array as the element type
1670 of another array.</p>
Dan Gohman142ccc02009-01-24 15:58:40 +0000[diff] [blame]1671
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1672
1673</div>
1674
1675<!-- _______________________________________________________________________ -->
1676<div class="doc_subsubsection"> <a name="t_aggregate">Aggregate Types</a> </div>
1677
1678<div class="doc_text">
1679
1680<p>Aggregate Types are a subset of derived types that can contain multiple
1681 member types. <a href="#t_array">Arrays</a>,
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]1682 <a href="#t_struct">structs</a>, and <a href="#t_vector">vectors</a> are
1683 aggregate types.</p>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1684
1685</div>
1686
Reid Spencer138249b2007-05-16 18:44:01 +0000[diff] [blame]1687<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1688<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1689
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1690<div class="doc_text">
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1691
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1692<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1693<p>The array type is a very simple derived type that arranges elements
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1694 sequentially in memory. The array type requires a size (number of elements)
1695 and an underlying data type.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1696
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]1697<h5>Syntax:</h5>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1698<pre>
1699 [&lt;# elements&gt; x &lt;elementtype&gt;]
1700</pre>
1701
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1702<p>The number of elements is a constant integer value; <tt>elementtype</tt> may
1703 be any type with a size.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1704
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]1705<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1706<table class="layout">
1707 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1708 <td class="left"><tt>[40 x i32]</tt></td>
1709 <td class="left">Array of 40 32-bit integer values.</td>
1710 </tr>
1711 <tr class="layout">
1712 <td class="left"><tt>[41 x i32]</tt></td>
1713 <td class="left">Array of 41 32-bit integer values.</td>
1714 </tr>
1715 <tr class="layout">
1716 <td class="left"><tt>[4 x i8]</tt></td>
1717 <td class="left">Array of 4 8-bit integer values.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1718 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1719</table>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1720<p>Here are some examples of multidimensional arrays:</p>
1721<table class="layout">
1722 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1723 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1724 <td class="left">3x4 array of 32-bit integer values.</td>
1725 </tr>
1726 <tr class="layout">
1727 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1728 <td class="left">12x10 array of single precision floating point values.</td>
1729 </tr>
1730 <tr class="layout">
1731 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1732 <td class="left">2x3x4 array of 16-bit integer values.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1733 </tr>
1734</table>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +0000[diff] [blame]1735
Dan Gohmanc74bc282009-11-09 19:01:53 +0000[diff] [blame]1736<p>There is no restriction on indexing beyond the end of the array implied by
1737 a static type (though there are restrictions on indexing beyond the bounds
1738 of an allocated object in some cases). This means that single-dimension
1739 'variable sized array' addressing can be implemented in LLVM with a zero
1740 length array type. An implementation of 'pascal style arrays' in LLVM could
1741 use the type "<tt>{ i32, [0 x float]}</tt>", for example.</p>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +0000[diff] [blame]1742
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1743</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1744
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1745<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1746<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1747
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1748<div class="doc_text">
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]1749
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1750<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1751<p>The function type can be thought of as a function signature. It consists of
1752 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]1753 function type is a first class type or a void type.</p>
Devang Pateld6cff512008-03-10 20:49:15 +0000[diff] [blame]1754
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1755<h5>Syntax:</h5>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]1756<pre>
Nick Lewycky14d1ccc2009-09-27 07:55:32 +0000[diff] [blame]1757 &lt;returntype&gt; (&lt;parameter list&gt;)
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]1758</pre>
1759
John Criswell4c0cf7f2005-10-24 16:17:18 +0000[diff] [blame]1760<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]1761 specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1762 which indicates that the function takes a variable number of arguments.
1763 Variable argument functions can access their arguments with
1764 the <a href="#int_varargs">variable argument handling intrinsic</a>
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]1765 functions. '<tt>&lt;returntype&gt;</tt>' is any type except
Nick Lewycky93e06a52009-09-27 23:27:42 +0000[diff] [blame]1766 <a href="#t_label">label</a>.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]1767
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1768<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1769<table class="layout">
1770 <tr class="layout">
Reid Spencer58c08712006-12-31 07:18:34 +0000[diff] [blame]1771 <td class="left"><tt>i32 (i32)</tt></td>
1772 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1773 </td>
Reid Spencer58c08712006-12-31 07:18:34 +0000[diff] [blame]1774 </tr><tr class="layout">
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]1775 <td class="left"><tt>float&nbsp;(i16,&nbsp;i32&nbsp;*)&nbsp;*
Reid Spencer655dcc62006-12-31 07:20:23 +0000[diff] [blame]1776 </tt></td>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1777 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]1778 an <tt>i16</tt> and a <a href="#t_pointer">pointer</a> to <tt>i32</tt>,
1779 returning <tt>float</tt>.
Reid Spencer58c08712006-12-31 07:18:34 +0000[diff] [blame]1780 </td>
1781 </tr><tr class="layout">
1782 <td class="left"><tt>i32 (i8*, ...)</tt></td>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1783 <td class="left">A vararg function that takes at least one
1784 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1785 which returns an integer. This is the signature for <tt>printf</tt> in
Reid Spencer58c08712006-12-31 07:18:34 +0000[diff] [blame]1786 LLVM.
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1787 </td>
Devang Patele3dfc1c2008-03-24 05:35:41 +0000[diff] [blame]1788 </tr><tr class="layout">
1789 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Nick Lewycky14d1ccc2009-09-27 07:55:32 +0000[diff] [blame]1790 <td class="left">A function taking an <tt>i32</tt>, returning a
1791 <a href="#t_struct">structure</a> containing two <tt>i32</tt> values
Devang Patele3dfc1c2008-03-24 05:35:41 +0000[diff] [blame]1792 </td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1793 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1794</table>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]1795
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1796</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1797
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1798<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1799<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1800
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1801<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1802
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1803<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1804<p>The structure type is used to represent a collection of data members together
1805 in memory. The packing of the field types is defined to match the ABI of the
1806 underlying processor. The elements of a structure may be any type that has a
1807 size.</p>
1808
Jeffrey Yasskinf991bbb2010-01-11 19:19:26 +0000[diff] [blame]1809<p>Structures in memory are accessed using '<tt><a href="#i_load">load</a></tt>'
1810 and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field
1811 with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
1812 Structures in registers are accessed using the
1813 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' and
1814 '<tt><a href="#i_insertvalue">insertvalue</a></tt>' instructions.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1815<h5>Syntax:</h5>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1816<pre>
1817 { &lt;type list&gt; }
1818</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1819
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1820<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1821<table class="layout">
1822 <tr class="layout">
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]1823 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1824 <td class="left">A triple of three <tt>i32</tt> values</td>
1825 </tr><tr class="layout">
1826 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1827 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1828 second element is a <a href="#t_pointer">pointer</a> to a
1829 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1830 an <tt>i32</tt>.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1831 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1832</table>
Dan Gohman142ccc02009-01-24 15:58:40 +0000[diff] [blame]1833
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1834</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1835
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1836<!-- _______________________________________________________________________ -->
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]1837<div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
1838</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1839
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]1840<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1841
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]1842<h5>Overview:</h5>
1843<p>The packed structure type is used to represent a collection of data members
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1844 together in memory. There is no padding between fields. Further, the
1845 alignment of a packed structure is 1 byte. The elements of a packed
1846 structure may be any type that has a size.</p>
1847
1848<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt> and
1849 '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field with
1850 the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
1851
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]1852<h5>Syntax:</h5>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1853<pre>
1854 &lt; { &lt;type list&gt; } &gt;
1855</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1856
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]1857<h5>Examples:</h5>
1858<table class="layout">
1859 <tr class="layout">
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]1860 <td class="left"><tt>&lt; { i32, i32, i32 } &gt;</tt></td>
1861 <td class="left">A triple of three <tt>i32</tt> values</td>
1862 </tr><tr class="layout">
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1863 <td class="left">
1864<tt>&lt;&nbsp;{&nbsp;float,&nbsp;i32&nbsp;(i32)*&nbsp;}&nbsp;&gt;</tt></td>
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]1865 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1866 second element is a <a href="#t_pointer">pointer</a> to a
1867 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1868 an <tt>i32</tt>.</td>
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]1869 </tr>
1870</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1871
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]1872</div>
1873
1874<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1875<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
Chris Lattner4a67c912009-02-08 19:53:29 +0000[diff] [blame]1876
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1877<div class="doc_text">
1878
1879<h5>Overview:</h5>
Dan Gohman88481112010-02-25 16:50:07 +0000[diff] [blame]1880<p>The pointer type is used to specify memory locations.
1881 Pointers are commonly used to reference objects in memory.</p>
1882
1883<p>Pointer types may have an optional address space attribute defining the
1884 numbered address space where the pointed-to object resides. The default
1885 address space is number zero. The semantics of non-zero address
1886 spaces are target-specific.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1887
1888<p>Note that LLVM does not permit pointers to void (<tt>void*</tt>) nor does it
1889 permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</tt> instead.</p>
Chris Lattner4a67c912009-02-08 19:53:29 +0000[diff] [blame]1890
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]1891<h5>Syntax:</h5>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1892<pre>
1893 &lt;type&gt; *
1894</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1895
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]1896<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1897<table class="layout">
1898 <tr class="layout">
Dan Gohman623806e2009-01-04 23:44:43 +0000[diff] [blame]1899 <td class="left"><tt>[4 x i32]*</tt></td>
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1900 <td class="left">A <a href="#t_pointer">pointer</a> to <a
1901 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
1902 </tr>
1903 <tr class="layout">
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]1904 <td class="left"><tt>i32 (i32*) *</tt></td>
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1905 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1906 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1907 <tt>i32</tt>.</td>
1908 </tr>
1909 <tr class="layout">
1910 <td class="left"><tt>i32 addrspace(5)*</tt></td>
1911 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
1912 that resides in address space #5.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1913 </tr>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1914</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1915
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1916</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1917
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]1918<!-- _______________________________________________________________________ -->
Reid Spencer404a3252007-02-15 03:07:05 +0000[diff] [blame]1919<div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1920
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1921<div class="doc_text">
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1922
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]1923<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1924<p>A vector type is a simple derived type that represents a vector of elements.
1925 Vector types are used when multiple primitive data are operated in parallel
1926 using a single instruction (SIMD). A vector type requires a size (number of
Duncan Sands31c0e0e2009-11-27 13:38:03 +0000[diff] [blame]1927 elements) and an underlying primitive data type. Vector types are considered
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1928 <a href="#t_firstclass">first class</a>.</p>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1929
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]1930<h5>Syntax:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1931<pre>
1932 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1933</pre>
1934
Chris Lattnerf11031a2010-10-10 18:20:35 +0000[diff] [blame]1935<p>The number of elements is a constant integer value larger than 0; elementtype
1936 may be any integer or floating point type. Vectors of size zero are not
1937 allowed, and pointers are not allowed as the element type.</p>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1938
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]1939<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1940<table class="layout">
1941 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1942 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
1943 <td class="left">Vector of 4 32-bit integer values.</td>
1944 </tr>
1945 <tr class="layout">
1946 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
1947 <td class="left">Vector of 8 32-bit floating-point values.</td>
1948 </tr>
1949 <tr class="layout">
1950 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
1951 <td class="left">Vector of 2 64-bit integer values.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1952 </tr>
1953</table>
Dan Gohman142ccc02009-01-24 15:58:40 +0000[diff] [blame]1954
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1955</div>
1956
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1957<!-- _______________________________________________________________________ -->
1958<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1959<div class="doc_text">
1960
1961<h5>Overview:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1962<p>Opaque types are used to represent unknown types in the system. This
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1963 corresponds (for example) to the C notion of a forward declared structure
1964 type. In LLVM, opaque types can eventually be resolved to any type (not just
1965 a structure type).</p>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1966
1967<h5>Syntax:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1968<pre>
1969 opaque
1970</pre>
1971
1972<h5>Examples:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1973<table class="layout">
1974 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1975 <td class="left"><tt>opaque</tt></td>
1976 <td class="left">An opaque type.</td>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1977 </tr>
1978</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1979
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1980</div>
1981
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]1982<!-- ======================================================================= -->
1983<div class="doc_subsection">
1984 <a name="t_uprefs">Type Up-references</a>
1985</div>
1986
1987<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1988
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]1989<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1990<p>An "up reference" allows you to refer to a lexically enclosing type without
1991 requiring it to have a name. For instance, a structure declaration may
1992 contain a pointer to any of the types it is lexically a member of. Example
1993 of up references (with their equivalent as named type declarations)
1994 include:</p>
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]1995
1996<pre>
Chris Lattnerbf1d5452009-02-09 10:00:56 +0000[diff] [blame]1997 { \2 * } %x = type { %x* }
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]1998 { \2 }* %y = type { %y }*
1999 \1* %z = type %z*
2000</pre>
2001
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2002<p>An up reference is needed by the asmprinter for printing out cyclic types
2003 when there is no declared name for a type in the cycle. Because the
2004 asmprinter does not want to print out an infinite type string, it needs a
2005 syntax to handle recursive types that have no names (all names are optional
2006 in llvm IR).</p>
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]2007
2008<h5>Syntax:</h5>
2009<pre>
2010 \&lt;level&gt;
2011</pre>
2012
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2013<p>The level is the count of the lexical type that is being referred to.</p>
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]2014
2015<h5>Examples:</h5>
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]2016<table class="layout">
2017 <tr class="layout">
2018 <td class="left"><tt>\1*</tt></td>
2019 <td class="left">Self-referential pointer.</td>
2020 </tr>
2021 <tr class="layout">
2022 <td class="left"><tt>{ { \3*, i8 }, i32 }</tt></td>
2023 <td class="left">Recursive structure where the upref refers to the out-most
2024 structure.</td>
2025 </tr>
2026</table>
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]2027
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2028</div>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]2029
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2030<!-- *********************************************************************** -->
2031<div class="doc_section"> <a name="constants">Constants</a> </div>
2032<!-- *********************************************************************** -->
2033
2034<div class="doc_text">
2035
2036<p>LLVM has several different basic types of constants. This section describes
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2037 them all and their syntax.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2038
2039</div>
2040
2041<!-- ======================================================================= -->
Reid Spencer8f08d802004-12-09 18:02:53 +0000[diff] [blame]2042<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2043
2044<div class="doc_text">
2045
2046<dl>
2047 <dt><b>Boolean constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2048 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]2049 constants of the <tt><a href="#t_integer">i1</a></tt> type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2050
2051 <dt><b>Integer constants</b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2052 <dd>Standard integers (such as '4') are constants of
2053 the <a href="#t_integer">integer</a> type. Negative numbers may be used
2054 with integer types.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2055
2056 <dt><b>Floating point constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2057 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2058 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
2059 notation (see below). The assembler requires the exact decimal value of a
2060 floating-point constant. For example, the assembler accepts 1.25 but
2061 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
2062 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2063
2064 <dt><b>Null pointer constants</b></dt>
John Criswelldfe6a862004-12-10 15:51:16 +0000[diff] [blame]2065 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2066 and must be of <a href="#t_pointer">pointer type</a>.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2067</dl>
2068
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2069<p>The one non-intuitive notation for constants is the hexadecimal form of
2070 floating point constants. For example, the form '<tt>double
2071 0x432ff973cafa8000</tt>' is equivalent to (but harder to read than)
2072 '<tt>double 4.5e+15</tt>'. The only time hexadecimal floating point
2073 constants are required (and the only time that they are generated by the
2074 disassembler) is when a floating point constant must be emitted but it cannot
2075 be represented as a decimal floating point number in a reasonable number of
2076 digits. For example, NaN's, infinities, and other special values are
2077 represented in their IEEE hexadecimal format so that assembly and disassembly
2078 do not cause any bits to change in the constants.</p>
2079
Dale Johannesencd4a3012009-02-11 22:14:51 +0000[diff] [blame]2080<p>When using the hexadecimal form, constants of types float and double are
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2081 represented using the 16-digit form shown above (which matches the IEEE754
2082 representation for double); float values must, however, be exactly
2083 representable as IEE754 single precision. Hexadecimal format is always used
2084 for long double, and there are three forms of long double. The 80-bit format
2085 used by x86 is represented as <tt>0xK</tt> followed by 20 hexadecimal digits.
2086 The 128-bit format used by PowerPC (two adjacent doubles) is represented
2087 by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit format
2088 is represented by <tt>0xL</tt> followed by 32 hexadecimal digits; no
2089 currently supported target uses this format. Long doubles will only work if
2090 they match the long double format on your target. All hexadecimal formats
2091 are big-endian (sign bit at the left).</p>
2092
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]2093<p>There are no constants of type x86mmx.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2094</div>
2095
2096<!-- ======================================================================= -->
Chris Lattner361bfcd2009-02-28 18:32:25 +0000[diff] [blame]2097<div class="doc_subsection">
Bill Wendling972b7202009-07-20 02:32:41 +0000[diff] [blame]2098<a name="aggregateconstants"></a> <!-- old anchor -->
2099<a name="complexconstants">Complex Constants</a>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2100</div>
2101
2102<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2103
Chris Lattner361bfcd2009-02-28 18:32:25 +0000[diff] [blame]2104<p>Complex constants are a (potentially recursive) combination of simple
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2105 constants and smaller complex constants.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2106
2107<dl>
2108 <dt><b>Structure constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2109 <dd>Structure constants are represented with notation similar to structure
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2110 type definitions (a comma separated list of elements, surrounded by braces
2111 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
2112 where "<tt>@G</tt>" is declared as "<tt>@G = external global i32</tt>".
2113 Structure constants must have <a href="#t_struct">structure type</a>, and
2114 the number and types of elements must match those specified by the
2115 type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2116
2117 <dt><b>Array constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2118 <dd>Array constants are represented with notation similar to array type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2119 definitions (a comma separated list of elements, surrounded by square
2120 brackets (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74
2121 ]</tt>". Array constants must have <a href="#t_array">array type</a>, and
2122 the number and types of elements must match those specified by the
2123 type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2124
Reid Spencer404a3252007-02-15 03:07:05 +0000[diff] [blame]2125 <dt><b>Vector constants</b></dt>
Reid Spencer404a3252007-02-15 03:07:05 +0000[diff] [blame]2126 <dd>Vector constants are represented with notation similar to vector type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2127 definitions (a comma separated list of elements, surrounded by
2128 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32
2129 42, i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must
2130 have <a href="#t_vector">vector type</a>, and the number and types of
2131 elements must match those specified by the type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2132
2133 <dt><b>Zero initialization</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2134 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]2135 value to zero of <em>any</em> type, including scalar and
2136 <a href="#t_aggregate">aggregate</a> types.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2137 This is often used to avoid having to print large zero initializers
2138 (e.g. for large arrays) and is always exactly equivalent to using explicit
2139 zero initializers.</dd>
Nick Lewycky49f89192009-04-04 07:22:01 +0000[diff] [blame]2140
2141 <dt><b>Metadata node</b></dt>
Nick Lewycky8e2c4f42009-05-30 16:08:30 +0000[diff] [blame]2142 <dd>A metadata node is a structure-like constant with
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2143 <a href="#t_metadata">metadata type</a>. For example: "<tt>metadata !{
2144 i32 0, metadata !"test" }</tt>". Unlike other constants that are meant to
2145 be interpreted as part of the instruction stream, metadata is a place to
2146 attach additional information such as debug info.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2147</dl>
2148
2149</div>
2150
2151<!-- ======================================================================= -->
2152<div class="doc_subsection">
2153 <a name="globalconstants">Global Variable and Function Addresses</a>
2154</div>
2155
2156<div class="doc_text">
2157
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2158<p>The addresses of <a href="#globalvars">global variables</a>
2159 and <a href="#functionstructure">functions</a> are always implicitly valid
2160 (link-time) constants. These constants are explicitly referenced when
2161 the <a href="#identifiers">identifier for the global</a> is used and always
2162 have <a href="#t_pointer">pointer</a> type. For example, the following is a
2163 legal LLVM file:</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2164
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2165<pre class="doc_code">
Chris Lattner00538a12007-06-06 18:28:13 +0000[diff] [blame]2166@X = global i32 17
2167@Y = global i32 42
2168@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2169</pre>
2170
2171</div>
2172
2173<!-- ======================================================================= -->
Reid Spencer641f5c92004-12-09 18:13:12 +0000[diff] [blame]2174<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2175<div class="doc_text">
2176
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2177<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]2178 indicates that the user of the value may receive an unspecified bit-pattern.
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2179 Undefined values may be of any type (other than '<tt>label</tt>'
2180 or '<tt>void</tt>') and be used anywhere a constant is permitted.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2181
Chris Lattner92ada5d2009-09-11 01:49:31 +0000[diff] [blame]2182<p>Undefined values are useful because they indicate to the compiler that the
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2183 program is well defined no matter what value is used. This gives the
2184 compiler more freedom to optimize. Here are some examples of (potentially
2185 surprising) transformations that are valid (in pseudo IR):</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2186
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2187
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2188<pre class="doc_code">
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2189 %A = add %X, undef
2190 %B = sub %X, undef
2191 %C = xor %X, undef
2192Safe:
2193 %A = undef
2194 %B = undef
2195 %C = undef
2196</pre>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2197
2198<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]2199 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]2200
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2201<pre class="doc_code">
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2202 %A = or %X, undef
2203 %B = and %X, undef
2204Safe:
2205 %A = -1
2206 %B = 0
2207Unsafe:
2208 %A = undef
2209 %B = undef
2210</pre>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2211
2212<p>These logical operations have bits that are not always affected by the input.
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2213 For example, if <tt>%X</tt> has a zero bit, then the output of the
2214 '<tt>and</tt>' operation will always be a zero for that bit, no matter what
2215 the corresponding bit from the '<tt>undef</tt>' is. As such, it is unsafe to
2216 optimize or assume that the result of the '<tt>and</tt>' is '<tt>undef</tt>'.
2217 However, it is safe to assume that all bits of the '<tt>undef</tt>' could be
2218 0, and optimize the '<tt>and</tt>' to 0. Likewise, it is safe to assume that
2219 all the bits of the '<tt>undef</tt>' operand to the '<tt>or</tt>' could be
2220 set, allowing the '<tt>or</tt>' to be folded to -1.</p>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2221
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2222<pre class="doc_code">
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2223 %A = select undef, %X, %Y
2224 %B = select undef, 42, %Y
2225 %C = select %X, %Y, undef
2226Safe:
2227 %A = %X (or %Y)
2228 %B = 42 (or %Y)
2229 %C = %Y
2230Unsafe:
2231 %A = undef
2232 %B = undef
2233 %C = undef
2234</pre>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2235
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2236<p>This set of examples shows that undefined '<tt>select</tt>' (and conditional
2237 branch) conditions can go <em>either way</em>, but they have to come from one
2238 of the two operands. In the <tt>%A</tt> example, if <tt>%X</tt> and
2239 <tt>%Y</tt> were both known to have a clear low bit, then <tt>%A</tt> would
2240 have to have a cleared low bit. However, in the <tt>%C</tt> example, the
2241 optimizer is allowed to assume that the '<tt>undef</tt>' operand could be the
2242 same as <tt>%Y</tt>, allowing the whole '<tt>select</tt>' to be
2243 eliminated.</p>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2244
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2245<pre class="doc_code">
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2246 %A = xor undef, undef
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2247
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2248 %B = undef
2249 %C = xor %B, %B
2250
2251 %D = undef
2252 %E = icmp lt %D, 4
2253 %F = icmp gte %D, 4
2254
2255Safe:
2256 %A = undef
2257 %B = undef
2258 %C = undef
2259 %D = undef
2260 %E = undef
2261 %F = undef
2262</pre>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2263
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2264<p>This example points out that two '<tt>undef</tt>' operands are not
2265 necessarily the same. This can be surprising to people (and also matches C
2266 semantics) where they assume that "<tt>X^X</tt>" is always zero, even
2267 if <tt>X</tt> is undefined. This isn't true for a number of reasons, but the
2268 short answer is that an '<tt>undef</tt>' "variable" can arbitrarily change
2269 its value over its "live range". This is true because the variable doesn't
2270 actually <em>have a live range</em>. Instead, the value is logically read
2271 from arbitrary registers that happen to be around when needed, so the value
2272 is not necessarily consistent over time. In fact, <tt>%A</tt> and <tt>%C</tt>
2273 need to have the same semantics or the core LLVM "replace all uses with"
2274 concept would not hold.</p>
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2275
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2276<pre class="doc_code">
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2277 %A = fdiv undef, %X
2278 %B = fdiv %X, undef
2279Safe:
2280 %A = undef
2281b: unreachable
2282</pre>
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2283
2284<p>These examples show the crucial difference between an <em>undefined
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2285 value</em> and <em>undefined behavior</em>. An undefined value (like
2286 '<tt>undef</tt>') is allowed to have an arbitrary bit-pattern. This means that
2287 the <tt>%A</tt> operation can be constant folded to '<tt>undef</tt>', because
2288 the '<tt>undef</tt>' could be an SNaN, and <tt>fdiv</tt> is not (currently)
2289 defined on SNaN's. However, in the second example, we can make a more
2290 aggressive assumption: because the <tt>undef</tt> is allowed to be an
2291 arbitrary value, we are allowed to assume that it could be zero. Since a
2292 divide by zero has <em>undefined behavior</em>, we are allowed to assume that
2293 the operation does not execute at all. This allows us to delete the divide and
2294 all code after it. Because the undefined operation "can't happen", the
2295 optimizer can assume that it occurs in dead code.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2296
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2297<pre class="doc_code">
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2298a: store undef -> %X
2299b: store %X -> undef
2300Safe:
2301a: &lt;deleted&gt;
2302b: unreachable
2303</pre>
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2304
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2305<p>These examples reiterate the <tt>fdiv</tt> example: a store <em>of</em> an
2306 undefined value can be assumed to not have any effect; we can assume that the
2307 value is overwritten with bits that happen to match what was already there.
2308 However, a store <em>to</em> an undefined location could clobber arbitrary
2309 memory, therefore, it has undefined behavior.</p>
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2310
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2311</div>
2312
2313<!-- ======================================================================= -->
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2314<div class="doc_subsection"><a name="trapvalues">Trap Values</a></div>
2315<div class="doc_text">
2316
Dan Gohmanb2a709b2010-04-26 20:21:21 +0000[diff] [blame]2317<p>Trap values are similar to <a href="#undefvalues">undef values</a>, however
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2318 instead of representing an unspecified bit pattern, they represent the
2319 fact that an instruction or constant expression which cannot evoke side
2320 effects has nevertheless detected a condition which results in undefined
Dan Gohmanb2a709b2010-04-26 20:21:21 +0000[diff] [blame]2321 behavior.</p>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2322
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2323<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]2324 only exist when produced by operations such as
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2325 <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]2326
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2327<p>Trap value behavior is defined in terms of value <i>dependence</i>:</p>
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2328
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2329<ul>
2330<li>Values other than <a href="#i_phi"><tt>phi</tt></a> nodes depend on
2331 their operands.</li>
2332
2333<li><a href="#i_phi"><tt>Phi</tt></a> nodes depend on the operand corresponding
2334 to their dynamic predecessor basic block.</li>
2335
2336<li>Function arguments depend on the corresponding actual argument values in
2337 the dynamic callers of their functions.</li>
2338
2339<li><a href="#i_call"><tt>Call</tt></a> instructions depend on the
2340 <a href="#i_ret"><tt>ret</tt></a> instructions that dynamically transfer
2341 control back to them.</li>
2342
Dan Gohman7292a752010-05-03 14:55:22 +0000[diff] [blame]2343<li><a href="#i_invoke"><tt>Invoke</tt></a> instructions depend on the
2344 <a href="#i_ret"><tt>ret</tt></a>, <a href="#i_unwind"><tt>unwind</tt></a>,
2345 or exception-throwing call instructions that dynamically transfer control
2346 back to them.</li>
2347
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2348<li>Non-volatile loads and stores depend on the most recent stores to all of the
2349 referenced memory addresses, following the order in the IR
2350 (including loads and stores implied by intrinsics such as
2351 <a href="#int_memcpy"><tt>@llvm.memcpy</tt></a>.)</li>
2352
Dan Gohman3513ea52010-05-03 14:59:34 +0000[diff] [blame]2353<!-- TODO: In the case of multiple threads, this only applies if the store
2354 "happens-before" the load or store. -->
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2355
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2356<!-- TODO: floating-point exception state -->
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2357
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2358<li>An instruction with externally visible side effects depends on the most
2359 recent preceding instruction with externally visible side effects, following
Dan Gohman6c858db2010-07-06 15:26:33 +0000[diff] [blame]2360 the order in the IR. (This includes
2361 <a href="#volatile">volatile operations</a>.)</li>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2362
Dan Gohman7292a752010-05-03 14:55:22 +0000[diff] [blame]2363<li>An instruction <i>control-depends</i> on a
2364 <a href="#terminators">terminator instruction</a>
2365 if the terminator instruction has multiple successors and the instruction
2366 is always executed when control transfers to one of the successors, and
2367 may not be executed when control is transfered to another.</li>
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2368
2369<li>Dependence is transitive.</li>
2370
2371</ul>
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2372
2373<p>Whenever a trap value is generated, all values which depend on it evaluate
2374 to trap. If they have side effects, the evoke their side effects as if each
2375 operand with a trap value were undef. If they have externally-visible side
2376 effects, the behavior is undefined.</p>
2377
2378<p>Here are some examples:</p>
Dan Gohman48a25882010-04-26 20:54:53 +0000[diff] [blame]2379
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2380<pre class="doc_code">
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2381entry:
2382 %trap = sub nuw i32 0, 1 ; Results in a trap value.
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2383 %still_trap = and i32 %trap, 0 ; Whereas (and i32 undef, 0) would return 0.
2384 %trap_yet_again = getelementptr i32* @h, i32 %still_trap
2385 store i32 0, i32* %trap_yet_again ; undefined behavior
2386
2387 store i32 %trap, i32* @g ; Trap value conceptually stored to memory.
2388 %trap2 = load i32* @g ; Returns a trap value, not just undef.
2389
2390 volatile store i32 %trap, i32* @g ; External observation; undefined behavior.
2391
2392 %narrowaddr = bitcast i32* @g to i16*
2393 %wideaddr = bitcast i32* @g to i64*
2394 %trap3 = load 16* %narrowaddr ; Returns a trap value.
2395 %trap4 = load i64* %widaddr ; Returns a trap value.
2396
2397 %cmp = icmp i32 slt %trap, 0 ; Returns a trap value.
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2398 %br i1 %cmp, %true, %end ; Branch to either destination.
2399
2400true:
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2401 volatile store i32 0, i32* @g ; This is control-dependent on %cmp, so
2402 ; it has undefined behavior.
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2403 br label %end
2404
2405end:
2406 %p = phi i32 [ 0, %entry ], [ 1, %true ]
2407 ; Both edges into this PHI are
2408 ; control-dependent on %cmp, so this
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2409 ; always results in a trap value.
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2410
2411 volatile store i32 0, i32* @g ; %end is control-equivalent to %entry
2412 ; so this is defined (ignoring earlier
2413 ; undefined behavior in this example).
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2414</pre>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2415
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2416</div>
2417
2418<!-- ======================================================================= -->
Chris Lattner2bfd3202009-10-27 21:19:13 +0000[diff] [blame]2419<div class="doc_subsection"><a name="blockaddress">Addresses of Basic
2420 Blocks</a></div>
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2421<div class="doc_text">
2422
Chris Lattneraa99c942009-11-01 01:27:45 +0000[diff] [blame]2423<p><b><tt>blockaddress(@function, %block)</tt></b></p>
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2424
2425<p>The '<tt>blockaddress</tt>' constant computes the address of the specified
Chris Lattner5c5f0ac2009-10-27 21:49:40 +0000[diff] [blame]2426 basic block in the specified function, and always has an i8* type. Taking
Chris Lattneraa99c942009-11-01 01:27:45 +0000[diff] [blame]2427 the address of the entry block is illegal.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2428
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2429<p>This value only has defined behavior when used as an operand to the
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2430 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>' instruction, or for
2431 comparisons against null. Pointer equality tests between labels addresses
2432 results in undefined behavior &mdash; though, again, comparison against null
2433 is ok, and no label is equal to the null pointer. This may be passed around
2434 as an opaque pointer sized value as long as the bits are not inspected. This
2435 allows <tt>ptrtoint</tt> and arithmetic to be performed on these values so
2436 long as the original value is reconstituted before the <tt>indirectbr</tt>
2437 instruction.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2438
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2439<p>Finally, some targets may provide defined semantics when using the value as
2440 the operand to an inline assembly, but that is target specific.</p>
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2441
2442</div>
2443
2444
2445<!-- ======================================================================= -->
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2446<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
2447</div>
2448
2449<div class="doc_text">
2450
2451<p>Constant expressions are used to allow expressions involving other constants
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2452 to be used as constants. Constant expressions may be of
2453 any <a href="#t_firstclass">first class</a> type and may involve any LLVM
2454 operation that does not have side effects (e.g. load and call are not
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2455 supported). The following is the syntax for constant expressions:</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2456
2457<dl>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2458 <dt><b><tt>trunc (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2459 <dd>Truncate a constant to another type. The bit size of CST must be larger
2460 than the bit size of TYPE. Both types must be integers.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2461
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2462 <dt><b><tt>zext (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2463 <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]2464 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2465
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2466 <dt><b><tt>sext (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2467 <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]2468 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2469
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2470 <dt><b><tt>fptrunc (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2471 <dd>Truncate a floating point constant to another floating point type. The
2472 size of CST must be larger than the size of TYPE. Both types must be
2473 floating point.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2474
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2475 <dt><b><tt>fpext (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2476 <dd>Floating point extend a constant to another type. The size of CST must be
2477 smaller or equal to the size of TYPE. Both types must be floating
2478 point.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2479
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2480 <dt><b><tt>fptoui (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2481 <dd>Convert a floating point constant to the corresponding unsigned integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2482 constant. TYPE must be a scalar or vector integer type. CST must be of
2483 scalar or vector floating point type. Both CST and TYPE must be scalars,
2484 or vectors of the same number of elements. If the value won't fit in the
2485 integer type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2486
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2487 <dt><b><tt>fptosi (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2488 <dd>Convert a floating point constant to the corresponding signed integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2489 constant. TYPE must be a scalar or vector integer type. CST must be of
2490 scalar or vector floating point type. Both CST and TYPE must be scalars,
2491 or vectors of the same number of elements. If the value won't fit in the
2492 integer type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2493
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2494 <dt><b><tt>uitofp (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2495 <dd>Convert an unsigned integer constant to the corresponding floating point
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2496 constant. TYPE must be a scalar or vector floating point type. CST must be
2497 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2498 vectors of the same number of elements. If the value won't fit in the
2499 floating point type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2500
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2501 <dt><b><tt>sitofp (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2502 <dd>Convert a signed integer constant to the corresponding floating point
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2503 constant. TYPE must be a scalar or vector floating point type. CST must be
2504 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2505 vectors of the same number of elements. If the value won't fit in the
2506 floating point type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2507
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2508 <dt><b><tt>ptrtoint (CST to TYPE)</tt></b></dt>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]2509 <dd>Convert a pointer typed constant to the corresponding integer constant
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2510 <tt>TYPE</tt> must be an integer type. <tt>CST</tt> must be of pointer
2511 type. The <tt>CST</tt> value is zero extended, truncated, or unchanged to
2512 make it fit in <tt>TYPE</tt>.</dd>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]2513
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2514 <dt><b><tt>inttoptr (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2515 <dd>Convert a integer constant to a pointer constant. TYPE must be a pointer
2516 type. CST must be of integer type. The CST value is zero extended,
2517 truncated, or unchanged to make it fit in a pointer size. This one is
2518 <i>really</i> dangerous!</dd>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]2519
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2520 <dt><b><tt>bitcast (CST to TYPE)</tt></b></dt>
Chris Lattner789dee32009-02-28 18:27:03 +0000[diff] [blame]2521 <dd>Convert a constant, CST, to another TYPE. The constraints of the operands
2522 are the same as those for the <a href="#i_bitcast">bitcast
2523 instruction</a>.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2524
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2525 <dt><b><tt>getelementptr (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
2526 <dt><b><tt>getelementptr inbounds (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2527 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2528 constants. As with the <a href="#i_getelementptr">getelementptr</a>
2529 instruction, the index list may have zero or more indexes, which are
2530 required to make sense for the type of "CSTPTR".</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2531
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2532 <dt><b><tt>select (COND, VAL1, VAL2)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2533 <dd>Perform the <a href="#i_select">select operation</a> on constants.</dd>
Reid Spencer9965ee72006-12-04 19:23:19 +0000[diff] [blame]2534
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2535 <dt><b><tt>icmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer9965ee72006-12-04 19:23:19 +0000[diff] [blame]2536 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
2537
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2538 <dt><b><tt>fcmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer9965ee72006-12-04 19:23:19 +0000[diff] [blame]2539 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
Robert Bocchino7e97a6d2006-01-10 19:31:34 +0000[diff] [blame]2540
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2541 <dt><b><tt>extractelement (VAL, IDX)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2542 <dd>Perform the <a href="#i_extractelement">extractelement operation</a> on
2543 constants.</dd>
Robert Bocchino7e97a6d2006-01-10 19:31:34 +0000[diff] [blame]2544
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2545 <dt><b><tt>insertelement (VAL, ELT, IDX)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2546 <dd>Perform the <a href="#i_insertelement">insertelement operation</a> on
2547 constants.</dd>
Chris Lattner016a0e52006-04-08 00:13:41 +0000[diff] [blame]2548
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2549 <dt><b><tt>shufflevector (VEC1, VEC2, IDXMASK)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2550 <dd>Perform the <a href="#i_shufflevector">shufflevector operation</a> on
2551 constants.</dd>
Chris Lattner016a0e52006-04-08 00:13:41 +0000[diff] [blame]2552
Nick Lewycky9ab9a7f2010-05-29 06:44:15 +0000[diff] [blame]2553 <dt><b><tt>extractvalue (VAL, IDX0, IDX1, ...)</tt></b></dt>
2554 <dd>Perform the <a href="#i_extractvalue">extractvalue operation</a> on
2555 constants. The index list is interpreted in a similar manner as indices in
2556 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2557 index value must be specified.</dd>
2558
2559 <dt><b><tt>insertvalue (VAL, ELT, IDX0, IDX1, ...)</tt></b></dt>
2560 <dd>Perform the <a href="#i_insertvalue">insertvalue operation</a> on
2561 constants. The index list is interpreted in a similar manner as indices in
2562 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2563 index value must be specified.</dd>
2564
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2565 <dt><b><tt>OPCODE (LHS, RHS)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2566 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
2567 be any of the <a href="#binaryops">binary</a>
2568 or <a href="#bitwiseops">bitwise binary</a> operations. The constraints
2569 on operands are the same as those for the corresponding instruction
2570 (e.g. no bitwise operations on floating point values are allowed).</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2571</dl>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2572
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2573</div>
Chris Lattnerb1652612004-03-08 16:49:10 +0000[diff] [blame]2574
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2575<!-- *********************************************************************** -->
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2576<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
2577<!-- *********************************************************************** -->
2578
2579<!-- ======================================================================= -->
2580<div class="doc_subsection">
2581<a name="inlineasm">Inline Assembler Expressions</a>
2582</div>
2583
2584<div class="doc_text">
2585
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2586<p>LLVM supports inline assembler expressions (as opposed
2587 to <a href="#moduleasm"> Module-Level Inline Assembly</a>) through the use of
2588 a special value. This value represents the inline assembler as a string
2589 (containing the instructions to emit), a list of operand constraints (stored
Dale Johannesen63c94fe2009-10-13 21:56:55 +0000[diff] [blame]2590 as a string), a flag that indicates whether or not the inline asm
Dale Johannesen1cfb9582009-10-21 23:28:00 +0000[diff] [blame]2591 expression has side effects, and a flag indicating whether the function
2592 containing the asm needs to align its stack conservatively. An example
2593 inline assembler expression is:</p>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2594
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2595<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2596i32 (i32) asm "bswap $0", "=r,r"
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2597</pre>
2598
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2599<p>Inline assembler expressions may <b>only</b> be used as the callee operand of
2600 a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we
2601 have:</p>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2602
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2603<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2604%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]2605</pre>
2606
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2607<p>Inline asms with side effects not visible in the constraint list must be
2608 marked as having side effects. This is done through the use of the
2609 '<tt>sideeffect</tt>' keyword, like so:</p>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2610
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2611<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2612call void asm sideeffect "eieio", ""()
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2613</pre>
2614
Dale Johannesen1cfb9582009-10-21 23:28:00 +0000[diff] [blame]2615<p>In some cases inline asms will contain code that will not work unless the
2616 stack is aligned in some way, such as calls or SSE instructions on x86,
2617 yet will not contain code that does that alignment within the asm.
2618 The compiler should make conservative assumptions about what the asm might
2619 contain and should generate its usual stack alignment code in the prologue
2620 if the '<tt>alignstack</tt>' keyword is present:</p>
Dale Johannesen63c94fe2009-10-13 21:56:55 +0000[diff] [blame]2621
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2622<pre class="doc_code">
Dale Johannesen1cfb9582009-10-21 23:28:00 +0000[diff] [blame]2623call void asm alignstack "eieio", ""()
Dale Johannesen63c94fe2009-10-13 21:56:55 +0000[diff] [blame]2624</pre>
Dale Johannesen63c94fe2009-10-13 21:56:55 +0000[diff] [blame]2625
2626<p>If both keywords appear the '<tt>sideeffect</tt>' keyword must come
2627 first.</p>
2628
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2629<p>TODO: The format of the asm and constraints string still need to be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2630 documented here. Constraints on what can be done (e.g. duplication, moving,
2631 etc need to be documented). This is probably best done by reference to
2632 another document that covers inline asm from a holistic perspective.</p>
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2633</div>
2634
2635<div class="doc_subsubsection">
2636<a name="inlineasm_md">Inline Asm Metadata</a>
2637</div>
2638
2639<div class="doc_text">
2640
2641<p>The call instructions that wrap inline asm nodes may have a "!srcloc" MDNode
Chris Lattner79ffdc72010-11-17 08:20:42 +0000[diff] [blame]2642 attached to it that contains a list of constant integers. If present, the
2643 code generator will use the integer as the location cookie value when report
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2644 errors through the LLVMContext error reporting mechanisms. This allows a
Dan Gohman61110ae2010-04-28 00:36:01 +0000[diff] [blame]2645 front-end to correlate backend errors that occur with inline asm back to the
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2646 source code that produced it. For example:</p>
2647
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2648<pre class="doc_code">
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2649call void asm sideeffect "something bad", ""()<b>, !srcloc !42</b>
2650...
2651!42 = !{ i32 1234567 }
2652</pre>
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2653
2654<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]2655 IR. If the MDNode contains multiple constants, the code generator will use
2656 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]2657
2658</div>
2659
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]2660<!-- ======================================================================= -->
2661<div class="doc_subsection"><a name="metadata">Metadata Nodes and Metadata
2662 Strings</a>
2663</div>
2664
2665<div class="doc_text">
2666
2667<p>LLVM IR allows metadata to be attached to instructions in the program that
2668 can convey extra information about the code to the optimizers and code
2669 generator. One example application of metadata is source-level debug
2670 information. There are two metadata primitives: strings and nodes. All
2671 metadata has the <tt>metadata</tt> type and is identified in syntax by a
2672 preceding exclamation point ('<tt>!</tt>').</p>
2673
2674<p>A metadata string is a string surrounded by double quotes. It can contain
2675 any character by escaping non-printable characters with "\xx" where "xx" is
2676 the two digit hex code. For example: "<tt>!"test\00"</tt>".</p>
2677
2678<p>Metadata nodes are represented with notation similar to structure constants
2679 (a comma separated list of elements, surrounded by braces and preceded by an
2680 exclamation point). For example: "<tt>!{ metadata !"test\00", i32
2681 10}</tt>". Metadata nodes can have any values as their operand.</p>
2682
2683<p>A <a href="#namedmetadatastructure">named metadata</a> is a collection of
2684 metadata nodes, which can be looked up in the module symbol table. For
2685 example: "<tt>!foo = metadata !{!4, !3}</tt>".
2686
Devang Patel9984bd62010-03-04 23:44:48 +0000[diff] [blame]2687<p>Metadata can be used as function arguments. Here <tt>llvm.dbg.value</tt>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2688 function is using two metadata arguments.</p>
Devang Patel9984bd62010-03-04 23:44:48 +0000[diff] [blame]2689
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2690 <pre class="doc_code">
Devang Patel9984bd62010-03-04 23:44:48 +0000[diff] [blame]2691 call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2692 </pre>
Devang Patel9984bd62010-03-04 23:44:48 +0000[diff] [blame]2693
2694<p>Metadata can be attached with an instruction. Here metadata <tt>!21</tt> is
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2695 attached with <tt>add</tt> instruction using <tt>!dbg</tt> identifier.</p>
Devang Patel9984bd62010-03-04 23:44:48 +0000[diff] [blame]2696
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2697 <pre class="doc_code">
Devang Patel9984bd62010-03-04 23:44:48 +0000[diff] [blame]2698 %indvar.next = add i64 %indvar, 1, !dbg !21
2699 </pre>
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]2700</div>
2701
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2702
2703<!-- *********************************************************************** -->
2704<div class="doc_section">
2705 <a name="intrinsic_globals">Intrinsic Global Variables</a>
2706</div>
2707<!-- *********************************************************************** -->
2708
2709<p>LLVM has a number of "magic" global variables that contain data that affect
2710code generation or other IR semantics. These are documented here. All globals
Chris Lattner58f9bb22009-07-20 06:14:25 +0000[diff] [blame]2711of this sort should have a section specified as "<tt>llvm.metadata</tt>". This
2712section and all globals that start with "<tt>llvm.</tt>" are reserved for use
2713by LLVM.</p>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2714
2715<!-- ======================================================================= -->
2716<div class="doc_subsection">
2717<a name="intg_used">The '<tt>llvm.used</tt>' Global Variable</a>
2718</div>
2719
2720<div class="doc_text">
2721
2722<p>The <tt>@llvm.used</tt> global is an array with i8* element type which has <a
2723href="#linkage_appending">appending linkage</a>. This array contains a list of
2724pointers to global variables and functions which may optionally have a pointer
2725cast formed of bitcast or getelementptr. For example, a legal use of it is:</p>
2726
2727<pre>
2728 @X = global i8 4
2729 @Y = global i32 123
2730
2731 @llvm.used = appending global [2 x i8*] [
2732 i8* @X,
2733 i8* bitcast (i32* @Y to i8*)
2734 ], section "llvm.metadata"
2735</pre>
2736
2737<p>If a global variable appears in the <tt>@llvm.used</tt> list, then the
2738compiler, assembler, and linker are required to treat the symbol as if there is
2739a reference to the global that it cannot see. For example, if a variable has
2740internal linkage and no references other than that from the <tt>@llvm.used</tt>
2741list, it cannot be deleted. This is commonly used to represent references from
2742inline asms and other things the compiler cannot "see", and corresponds to
2743"attribute((used))" in GNU C.</p>
2744
2745<p>On some targets, the code generator must emit a directive to the assembler or
2746object file to prevent the assembler and linker from molesting the symbol.</p>
2747
2748</div>
2749
2750<!-- ======================================================================= -->
2751<div class="doc_subsection">
Chris Lattner58f9bb22009-07-20 06:14:25 +0000[diff] [blame]2752<a name="intg_compiler_used">The '<tt>llvm.compiler.used</tt>' Global Variable</a>
2753</div>
2754
2755<div class="doc_text">
2756
2757<p>The <tt>@llvm.compiler.used</tt> directive is the same as the
2758<tt>@llvm.used</tt> directive, except that it only prevents the compiler from
2759touching the symbol. On targets that support it, this allows an intelligent
2760linker to optimize references to the symbol without being impeded as it would be
2761by <tt>@llvm.used</tt>.</p>
2762
2763<p>This is a rare construct that should only be used in rare circumstances, and
2764should not be exposed to source languages.</p>
2765
2766</div>
2767
2768<!-- ======================================================================= -->
2769<div class="doc_subsection">
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2770<a name="intg_global_ctors">The '<tt>llvm.global_ctors</tt>' Global Variable</a>
2771</div>
2772
2773<div class="doc_text">
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]2774<pre>
2775%0 = type { i32, void ()* }
David Chisnallb492b812010-04-30 19:27:35 +0000[diff] [blame]2776@llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor }]
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]2777</pre>
2778<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.
2779</p>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2780
2781</div>
2782
2783<!-- ======================================================================= -->
2784<div class="doc_subsection">
2785<a name="intg_global_dtors">The '<tt>llvm.global_dtors</tt>' Global Variable</a>
2786</div>
2787
2788<div class="doc_text">
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]2789<pre>
2790%0 = type { i32, void ()* }
David Chisnallb492b812010-04-30 19:27:35 +0000[diff] [blame]2791@llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor }]
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]2792</pre>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2793
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]2794<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.
2795</p>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2796
2797</div>
2798
2799
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2800<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2801<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
2802<!-- *********************************************************************** -->
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2803
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2804<div class="doc_text">
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2805
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2806<p>The LLVM instruction set consists of several different classifications of
2807 instructions: <a href="#terminators">terminator
2808 instructions</a>, <a href="#binaryops">binary instructions</a>,
2809 <a href="#bitwiseops">bitwise binary instructions</a>,
2810 <a href="#memoryops">memory instructions</a>, and
2811 <a href="#otherops">other instructions</a>.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2812
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2813</div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2814
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2815<!-- ======================================================================= -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2816<div class="doc_subsection"> <a name="terminators">Terminator
2817Instructions</a> </div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2818
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2819<div class="doc_text">
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2820
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2821<p>As mentioned <a href="#functionstructure">previously</a>, every basic block
2822 in a program ends with a "Terminator" instruction, which indicates which
2823 block should be executed after the current block is finished. These
2824 terminator instructions typically yield a '<tt>void</tt>' value: they produce
2825 control flow, not values (the one exception being the
2826 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
2827
Duncan Sands626b0242010-04-15 20:35:54 +0000[diff] [blame]2828<p>There are seven different terminator instructions: the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2829 '<a href="#i_ret"><tt>ret</tt></a>' instruction, the
2830 '<a href="#i_br"><tt>br</tt></a>' instruction, the
2831 '<a href="#i_switch"><tt>switch</tt></a>' instruction, the
Bill Wendling33fef7e2009-11-02 00:25:26 +0000[diff] [blame]2832 '<a href="#i_indirectbr">'<tt>indirectbr</tt></a>' Instruction, the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2833 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the
2834 '<a href="#i_unwind"><tt>unwind</tt></a>' instruction, and the
2835 '<a href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2836
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2837</div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2838
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2839<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2840<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
2841Instruction</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2842
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2843<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2844
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2845<h5>Syntax:</h5>
Dan Gohmancc3132e2008-10-04 19:00:07 +0000[diff] [blame]2846<pre>
2847 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]2848 ret void <i>; Return from void function</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2849</pre>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]2850
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2851<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2852<p>The '<tt>ret</tt>' instruction is used to return control flow (and optionally
2853 a value) from a function back to the caller.</p>
2854
2855<p>There are two forms of the '<tt>ret</tt>' instruction: one that returns a
2856 value and then causes control flow, and one that just causes control flow to
2857 occur.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]2858
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2859<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2860<p>The '<tt>ret</tt>' instruction optionally accepts a single argument, the
2861 return value. The type of the return value must be a
2862 '<a href="#t_firstclass">first class</a>' type.</p>
Dan Gohmancc3132e2008-10-04 19:00:07 +0000[diff] [blame]2863
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2864<p>A function is not <a href="#wellformed">well formed</a> if it it has a
2865 non-void return type and contains a '<tt>ret</tt>' instruction with no return
2866 value or a return value with a type that does not match its type, or if it
2867 has a void return type and contains a '<tt>ret</tt>' instruction with a
2868 return value.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]2869
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2870<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2871<p>When the '<tt>ret</tt>' instruction is executed, control flow returns back to
2872 the calling function's context. If the caller is a
2873 "<a href="#i_call"><tt>call</tt></a>" instruction, execution continues at the
2874 instruction after the call. If the caller was an
2875 "<a href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues at
2876 the beginning of the "normal" destination block. If the instruction returns
2877 a value, that value shall set the call or invoke instruction's return
2878 value.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]2879
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2880<h5>Example:</h5>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]2881<pre>
2882 ret i32 5 <i>; Return an integer value of 5</i>
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]2883 ret void <i>; Return from a void function</i>
Bill Wendling050ee8f2009-02-28 22:12:54 +0000[diff] [blame]2884 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]2885</pre>
Dan Gohman3065b612009-01-12 23:12:39 +0000[diff] [blame]2886
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2887</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2888<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2889<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2890
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2891<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2892
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2893<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2894<pre>
2895 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]2896</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2897
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2898<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2899<p>The '<tt>br</tt>' instruction is used to cause control flow to transfer to a
2900 different basic block in the current function. There are two forms of this
2901 instruction, corresponding to a conditional branch and an unconditional
2902 branch.</p>
2903
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2904<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2905<p>The conditional branch form of the '<tt>br</tt>' instruction takes a single
2906 '<tt>i1</tt>' value and two '<tt>label</tt>' values. The unconditional form
2907 of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>' value as a
2908 target.</p>
2909
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2910<h5>Semantics:</h5>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]2911<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2912 argument is evaluated. If the value is <tt>true</tt>, control flows to the
2913 '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
2914 control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
2915
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2916<h5>Example:</h5>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]2917<pre>
2918Test:
2919 %cond = <a href="#i_icmp">icmp</a> eq i32 %a, %b
2920 br i1 %cond, label %IfEqual, label %IfUnequal
2921IfEqual:
2922 <a href="#i_ret">ret</a> i32 1
2923IfUnequal:
2924 <a href="#i_ret">ret</a> i32 0
2925</pre>
2926
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2927</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2928
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2929<!-- _______________________________________________________________________ -->
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]2930<div class="doc_subsubsection">
2931 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
2932</div>
2933
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2934<div class="doc_text">
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]2935
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2936<h5>Syntax:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]2937<pre>
2938 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
2939</pre>
2940
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2941<h5>Overview:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]2942<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]2943 several different places. It is a generalization of the '<tt>br</tt>'
2944 instruction, allowing a branch to occur to one of many possible
2945 destinations.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]2946
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2947<h5>Arguments:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]2948<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2949 comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination,
2950 and an array of pairs of comparison value constants and '<tt>label</tt>'s.
2951 The table is not allowed to contain duplicate constant entries.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]2952
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2953<h5>Semantics:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2954<p>The <tt>switch</tt> instruction specifies a table of values and
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2955 destinations. When the '<tt>switch</tt>' instruction is executed, this table
2956 is searched for the given value. If the value is found, control flow is
Benjamin Kramer0f420382009-10-12 14:46:08 +0000[diff] [blame]2957 transferred to the corresponding destination; otherwise, control flow is
2958 transferred to the default destination.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2959
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]2960<h5>Implementation:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]2961<p>Depending on properties of the target machine and the particular
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2962 <tt>switch</tt> instruction, this instruction may be code generated in
2963 different ways. For example, it could be generated as a series of chained
2964 conditional branches or with a lookup table.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]2965
2966<h5>Example:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]2967<pre>
2968 <i>; Emulate a conditional br instruction</i>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]2969 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Dan Gohman623806e2009-01-04 23:44:43 +0000[diff] [blame]2970 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]2971
2972 <i>; Emulate an unconditional br instruction</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]2973 switch i32 0, label %dest [ ]
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]2974
2975 <i>; Implement a jump table:</i>
Dan Gohman623806e2009-01-04 23:44:43 +0000[diff] [blame]2976 switch i32 %val, label %otherwise [ i32 0, label %onzero
2977 i32 1, label %onone
2978 i32 2, label %ontwo ]
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2979</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2980
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2981</div>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]2982
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]2983
2984<!-- _______________________________________________________________________ -->
2985<div class="doc_subsubsection">
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]2986 <a name="i_indirectbr">'<tt>indirectbr</tt>' Instruction</a>
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]2987</div>
2988
2989<div class="doc_text">
2990
2991<h5>Syntax:</h5>
2992<pre>
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]2993 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]2994</pre>
2995
2996<h5>Overview:</h5>
2997
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]2998<p>The '<tt>indirectbr</tt>' instruction implements an indirect branch to a label
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]2999 within the current function, whose address is specified by
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]3000 "<tt>address</tt>". Address must be derived from a <a
3001 href="#blockaddress">blockaddress</a> constant.</p>
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3002
3003<h5>Arguments:</h5>
3004
3005<p>The '<tt>address</tt>' argument is the address of the label to jump to. The
3006 rest of the arguments indicate the full set of possible destinations that the
3007 address may point to. Blocks are allowed to occur multiple times in the
3008 destination list, though this isn't particularly useful.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3009
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3010<p>This destination list is required so that dataflow analysis has an accurate
3011 understanding of the CFG.</p>
3012
3013<h5>Semantics:</h5>
3014
3015<p>Control transfers to the block specified in the address argument. All
3016 possible destination blocks must be listed in the label list, otherwise this
3017 instruction has undefined behavior. This implies that jumps to labels
3018 defined in other functions have undefined behavior as well.</p>
3019
3020<h5>Implementation:</h5>
3021
3022<p>This is typically implemented with a jump through a register.</p>
3023
3024<h5>Example:</h5>
3025<pre>
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]3026 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3027</pre>
3028
3029</div>
3030
3031
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3032<!-- _______________________________________________________________________ -->
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3033<div class="doc_subsubsection">
3034 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
3035</div>
3036
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3037<div class="doc_text">
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3038
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3039<h5>Syntax:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3040<pre>
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]3041 &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]3042 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3043</pre>
3044
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]3045<h5>Overview:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3046<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3047 function, with the possibility of control flow transfer to either the
3048 '<tt>normal</tt>' label or the '<tt>exception</tt>' label. If the callee
3049 function returns with the "<tt><a href="#i_ret">ret</a></tt>" instruction,
3050 control flow will return to the "normal" label. If the callee (or any
3051 indirect callees) returns with the "<a href="#i_unwind"><tt>unwind</tt></a>"
3052 instruction, control is interrupted and continued at the dynamically nearest
3053 "exception" label.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3054
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3055<h5>Arguments:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3056<p>This instruction requires several arguments:</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3057
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3058<ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3059 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
3060 convention</a> the call should use. If none is specified, the call
3061 defaults to using C calling conventions.</li>
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]3062
3063 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3064 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
3065 '<tt>inreg</tt>' attributes are valid here.</li>
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]3066
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3067 <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]3068 function value being invoked. In most cases, this is a direct function
3069 invocation, but indirect <tt>invoke</tt>s are just as possible, branching
3070 off an arbitrary pointer to function value.</li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3071
3072 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3073 function to be invoked. </li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3074
3075 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]3076 signature argument types and parameter attributes. All arguments must be
3077 of <a href="#t_firstclass">first class</a> type. If the function
3078 signature indicates the function accepts a variable number of arguments,
3079 the extra arguments can be specified.</li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3080
3081 <li>'<tt>normal label</tt>': the label reached when the called function
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3082 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3083
3084 <li>'<tt>exception label</tt>': the label reached when a callee returns with
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3085 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3086
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]3087 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3088 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
3089 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3090</ol>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3091
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3092<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3093<p>This instruction is designed to operate as a standard
3094 '<tt><a href="#i_call">call</a></tt>' instruction in most regards. The
3095 primary difference is that it establishes an association with a label, which
3096 is used by the runtime library to unwind the stack.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3097
3098<p>This instruction is used in languages with destructors to ensure that proper
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3099 cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
3100 exception. Additionally, this is important for implementation of
3101 '<tt>catch</tt>' clauses in high-level languages that support them.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3102
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3103<p>For the purposes of the SSA form, the definition of the value returned by the
3104 '<tt>invoke</tt>' instruction is deemed to occur on the edge from the current
3105 block to the "normal" label. If the callee unwinds then no return value is
3106 available.</p>
Dan Gohman9069d892009-05-22 21:47:08 +0000[diff] [blame]3107
Chris Lattner97257f82010-01-15 18:08:37 +0000[diff] [blame]3108<p>Note that the code generator does not yet completely support unwind, and
3109that the invoke/unwind semantics are likely to change in future versions.</p>
3110
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3111<h5>Example:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3112<pre>
Nick Lewycky084ab472008-03-16 07:18:12 +0000[diff] [blame]3113 %retval = invoke i32 @Test(i32 15) to label %Continue
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]3114 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewycky084ab472008-03-16 07:18:12 +0000[diff] [blame]3115 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]3116 unwind label %TestCleanup <i>; {i32}:retval set</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3117</pre>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3118
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3119</div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3120
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]3121<!-- _______________________________________________________________________ -->
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3122
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3123<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
3124Instruction</a> </div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3125
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3126<div class="doc_text">
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3127
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]3128<h5>Syntax:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3129<pre>
3130 unwind
3131</pre>
3132
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]3133<h5>Overview:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3134<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3135 at the first callee in the dynamic call stack which used
3136 an <a href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call.
3137 This is primarily used to implement exception handling.</p>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3138
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]3139<h5>Semantics:</h5>
Chris Lattnerfe8519c2008-04-19 21:01:16 +0000[diff] [blame]3140<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3141 immediately halt. The dynamic call stack is then searched for the
3142 first <a href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack.
3143 Once found, execution continues at the "exceptional" destination block
3144 specified by the <tt>invoke</tt> instruction. If there is no <tt>invoke</tt>
3145 instruction in the dynamic call chain, undefined behavior results.</p>
3146
Chris Lattner97257f82010-01-15 18:08:37 +0000[diff] [blame]3147<p>Note that the code generator does not yet completely support unwind, and
3148that the invoke/unwind semantics are likely to change in future versions.</p>
3149
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3150</div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3151
3152<!-- _______________________________________________________________________ -->
3153
3154<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
3155Instruction</a> </div>
3156
3157<div class="doc_text">
3158
3159<h5>Syntax:</h5>
3160<pre>
3161 unreachable
3162</pre>
3163
3164<h5>Overview:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3165<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3166 instruction is used to inform the optimizer that a particular portion of the
3167 code is not reachable. This can be used to indicate that the code after a
3168 no-return function cannot be reached, and other facts.</p>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3169
3170<h5>Semantics:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3171<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3172
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3173</div>
3174
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3175<!-- ======================================================================= -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3176<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3177
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3178<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3179
3180<p>Binary operators are used to do most of the computation in a program. They
3181 require two operands of the same type, execute an operation on them, and
3182 produce a single value. The operands might represent multiple data, as is
3183 the case with the <a href="#t_vector">vector</a> data type. The result value
3184 has the same type as its operands.</p>
3185
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3186<p>There are several different binary operators:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3187
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3188</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3189
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3190<!-- _______________________________________________________________________ -->
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3191<div class="doc_subsubsection">
3192 <a name="i_add">'<tt>add</tt>' Instruction</a>
3193</div>
3194
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3195<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3196
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3197<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3198<pre>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3199 &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]3200 &lt;result&gt; = add nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3201 &lt;result&gt; = add nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3202 &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]3203</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3204
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3205<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3206<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3207
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3208<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3209<p>The two arguments to the '<tt>add</tt>' instruction must
3210 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3211 integer values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3212
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3213<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3214<p>The value produced is the integer sum of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3215
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3216<p>If the sum has unsigned overflow, the result returned is the mathematical
3217 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]3218
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3219<p>Because LLVM integers use a two's complement representation, this instruction
3220 is appropriate for both signed and unsigned integers.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3221
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3222<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3223 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3224 <tt>nsw</tt> keywords are present, the result value of the <tt>add</tt>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]3225 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3226 respectively, occurs.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3227
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3228<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3229<pre>
3230 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3231</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3232
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3233</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3234
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3235<!-- _______________________________________________________________________ -->
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3236<div class="doc_subsubsection">
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3237 <a name="i_fadd">'<tt>fadd</tt>' Instruction</a>
3238</div>
3239
3240<div class="doc_text">
3241
3242<h5>Syntax:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3243<pre>
3244 &lt;result&gt; = fadd &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3245</pre>
3246
3247<h5>Overview:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3248<p>The '<tt>fadd</tt>' instruction returns the sum of its two operands.</p>
3249
3250<h5>Arguments:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3251<p>The two arguments to the '<tt>fadd</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3252 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3253 floating point values. Both arguments must have identical types.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3254
3255<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3256<p>The value produced is the floating point sum of the two operands.</p>
3257
3258<h5>Example:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3259<pre>
3260 &lt;result&gt; = fadd float 4.0, %var <i>; yields {float}:result = 4.0 + %var</i>
3261</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3262
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3263</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3264
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3265<!-- _______________________________________________________________________ -->
3266<div class="doc_subsubsection">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3267 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
3268</div>
3269
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3270<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3271
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3272<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3273<pre>
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3274 &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]3275 &lt;result&gt; = sub nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3276 &lt;result&gt; = sub nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3277 &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]3278</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3279
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3280<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3281<p>The '<tt>sub</tt>' instruction returns the difference of its two
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3282 operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3283
3284<p>Note that the '<tt>sub</tt>' instruction is used to represent the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3285 '<tt>neg</tt>' instruction present in most other intermediate
3286 representations.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3287
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3288<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3289<p>The two arguments to the '<tt>sub</tt>' instruction must
3290 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3291 integer values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3292
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3293<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3294<p>The value produced is the integer difference of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3295
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3296<p>If the difference has unsigned overflow, the result returned is the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3297 mathematical result modulo 2<sup>n</sup>, where n is the bit width of the
3298 result.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3299
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3300<p>Because LLVM integers use a two's complement representation, this instruction
3301 is appropriate for both signed and unsigned integers.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3302
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3303<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3304 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3305 <tt>nsw</tt> keywords are present, the result value of the <tt>sub</tt>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]3306 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3307 respectively, occurs.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3308
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3309<h5>Example:</h5>
Bill Wendling2d8b9a82007-05-29 09:42:13 +0000[diff] [blame]3310<pre>
3311 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]3312 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3313</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3314
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3315</div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3316
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3317<!-- _______________________________________________________________________ -->
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3318<div class="doc_subsubsection">
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3319 <a name="i_fsub">'<tt>fsub</tt>' Instruction</a>
3320</div>
3321
3322<div class="doc_text">
3323
3324<h5>Syntax:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3325<pre>
3326 &lt;result&gt; = fsub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3327</pre>
3328
3329<h5>Overview:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3330<p>The '<tt>fsub</tt>' instruction returns the difference of its two
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3331 operands.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3332
3333<p>Note that the '<tt>fsub</tt>' instruction is used to represent the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3334 '<tt>fneg</tt>' instruction present in most other intermediate
3335 representations.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3336
3337<h5>Arguments:</h5>
Bill Wendling972b7202009-07-20 02:32:41 +0000[diff] [blame]3338<p>The two arguments to the '<tt>fsub</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3339 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3340 floating point values. Both arguments must have identical types.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3341
3342<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3343<p>The value produced is the floating point difference of the two operands.</p>
3344
3345<h5>Example:</h5>
3346<pre>
3347 &lt;result&gt; = fsub float 4.0, %var <i>; yields {float}:result = 4.0 - %var</i>
3348 &lt;result&gt; = fsub float -0.0, %val <i>; yields {float}:result = -%var</i>
3349</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3350
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3351</div>
3352
3353<!-- _______________________________________________________________________ -->
3354<div class="doc_subsubsection">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3355 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
3356</div>
3357
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3358<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3359
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3360<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3361<pre>
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3362 &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]3363 &lt;result&gt; = mul nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3364 &lt;result&gt; = mul nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3365 &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]3366</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3367
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3368<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3369<p>The '<tt>mul</tt>' instruction returns the product of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3370
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3371<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3372<p>The two arguments to the '<tt>mul</tt>' instruction must
3373 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3374 integer values. Both arguments must have identical types.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3375
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3376<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3377<p>The value produced is the integer product of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3378
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3379<p>If the result of the multiplication has unsigned overflow, the result
3380 returned is the mathematical result modulo 2<sup>n</sup>, where n is the bit
3381 width of the result.</p>
3382
3383<p>Because LLVM integers use a two's complement representation, and the result
3384 is the same width as the operands, this instruction returns the correct
3385 result for both signed and unsigned integers. If a full product
3386 (e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands should
3387 be sign-extended or zero-extended as appropriate to the width of the full
3388 product.</p>
3389
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3390<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3391 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3392 <tt>nsw</tt> keywords are present, the result value of the <tt>mul</tt>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]3393 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3394 respectively, occurs.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3395
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3396<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3397<pre>
3398 &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3399</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3400
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3401</div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3402
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3403<!-- _______________________________________________________________________ -->
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3404<div class="doc_subsubsection">
3405 <a name="i_fmul">'<tt>fmul</tt>' Instruction</a>
3406</div>
3407
3408<div class="doc_text">
3409
3410<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3411<pre>
3412 &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]3413</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3414
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3415<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3416<p>The '<tt>fmul</tt>' instruction returns the product of its two operands.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3417
3418<h5>Arguments:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3419<p>The two arguments to the '<tt>fmul</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3420 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3421 floating point values. Both arguments must have identical types.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3422
3423<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3424<p>The value produced is the floating point product of the two operands.</p>
3425
3426<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3427<pre>
3428 &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]3429</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3430
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3431</div>
3432
3433<!-- _______________________________________________________________________ -->
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3434<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
3435</a></div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3436
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3437<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3438
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3439<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3440<pre>
3441 &lt;result&gt; = udiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3442</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3443
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3444<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3445<p>The '<tt>udiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3446
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3447<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3448<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3449 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3450 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3451
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3452<h5>Semantics:</h5>
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3453<p>The value produced is the unsigned integer quotient of the two operands.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3454
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3455<p>Note that unsigned integer division and signed integer division are distinct
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3456 operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
3457
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3458<p>Division by zero leads to undefined behavior.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3459
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3460<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3461<pre>
3462 &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3463</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3464
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3465</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3466
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3467<!-- _______________________________________________________________________ -->
3468<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
3469</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3470
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3471<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3472
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3473<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3474<pre>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3475 &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]3476 &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]3477</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3478
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3479<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3480<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3481
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3482<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3483<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3484 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3485 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3486
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3487<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3488<p>The value produced is the signed integer quotient of the two operands rounded
3489 towards zero.</p>
3490
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3491<p>Note that signed integer division and unsigned integer division are distinct
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3492 operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
3493
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3494<p>Division by zero leads to undefined behavior. Overflow also leads to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3495 undefined behavior; this is a rare case, but can occur, for example, by doing
3496 a 32-bit division of -2147483648 by -1.</p>
3497
Dan Gohman71dfd782009-07-22 00:04:19 +0000[diff] [blame]3498<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohman57255802010-04-23 15:23:32 +0000[diff] [blame]3499 <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]3500 be rounded.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3501
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3502<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3503<pre>
3504 &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3505</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3506
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3507</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3508
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3509<!-- _______________________________________________________________________ -->
3510<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3511Instruction</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3512
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3513<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3514
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3515<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3516<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +0000[diff] [blame]3517 &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]3518</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3519
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3520<h5>Overview:</h5>
3521<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3522
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3523<h5>Arguments:</h5>
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]3524<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3525 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3526 floating point values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3527
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3528<h5>Semantics:</h5>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3529<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3530
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3531<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3532<pre>
3533 &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]3534</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3535
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3536</div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3537
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3538<!-- _______________________________________________________________________ -->
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3539<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
3540</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3541
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3542<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3543
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3544<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3545<pre>
3546 &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]3547</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3548
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3549<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3550<p>The '<tt>urem</tt>' instruction returns the remainder from the unsigned
3551 division of its two arguments.</p>
3552
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3553<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3554<p>The two arguments to the '<tt>urem</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3555 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3556 values. Both arguments must have identical types.</p>
3557
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3558<h5>Semantics:</h5>
3559<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3560 This instruction always performs an unsigned division to get the
3561 remainder.</p>
3562
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3563<p>Note that unsigned integer remainder and signed integer remainder are
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3564 distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
3565
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3566<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3567
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3568<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3569<pre>
3570 &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3571</pre>
3572
3573</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3574
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3575<!-- _______________________________________________________________________ -->
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3576<div class="doc_subsubsection">
3577 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
3578</div>
3579
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3580<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3581
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3582<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3583<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +0000[diff] [blame]3584 &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]3585</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3586
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3587<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3588<p>The '<tt>srem</tt>' instruction returns the remainder from the signed
3589 division of its two operands. This instruction can also take
3590 <a href="#t_vector">vector</a> versions of the values in which case the
3591 elements must be integers.</p>
Chris Lattnerb8f816e2008-01-04 04:33:49 +0000[diff] [blame]3592
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3593<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3594<p>The two arguments to the '<tt>srem</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3595 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3596 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3597
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3598<h5>Semantics:</h5>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3599<p>This instruction returns the <i>remainder</i> of a division (where the result
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3600 has the same sign as the dividend, <tt>op1</tt>), not the <i>modulo</i>
3601 operator (where the result has the same sign as the divisor, <tt>op2</tt>) of
3602 a value. For more information about the difference,
3603 see <a href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
3604 Math Forum</a>. For a table of how this is implemented in various languages,
3605 please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
3606 Wikipedia: modulo operation</a>.</p>
3607
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3608<p>Note that signed integer remainder and unsigned integer remainder are
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3609 distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
3610
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3611<p>Taking the remainder of a division by zero leads to undefined behavior.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3612 Overflow also leads to undefined behavior; this is a rare case, but can
3613 occur, for example, by taking the remainder of a 32-bit division of
3614 -2147483648 by -1. (The remainder doesn't actually overflow, but this rule
3615 lets srem be implemented using instructions that return both the result of
3616 the division and the remainder.)</p>
3617
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3618<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3619<pre>
3620 &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3621</pre>
3622
3623</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3624
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3625<!-- _______________________________________________________________________ -->
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3626<div class="doc_subsubsection">
3627 <a name="i_frem">'<tt>frem</tt>' Instruction</a> </div>
3628
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3629<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3630
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3631<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3632<pre>
3633 &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]3634</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3635
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3636<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3637<p>The '<tt>frem</tt>' instruction returns the remainder from the division of
3638 its two operands.</p>
3639
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3640<h5>Arguments:</h5>
3641<p>The two arguments to the '<tt>frem</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3642 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3643 floating point values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3644
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3645<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3646<p>This instruction returns the <i>remainder</i> of a division. The remainder
3647 has the same sign as the dividend.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3648
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3649<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3650<pre>
3651 &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]3652</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3653
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3654</div>
Robert Bocchino820bc75b2006-02-17 21:18:08 +0000[diff] [blame]3655
Reid Spencer2ab01932007-02-02 13:57:07 +0000[diff] [blame]3656<!-- ======================================================================= -->
3657<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
3658Operations</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3659
Reid Spencer2ab01932007-02-02 13:57:07 +0000[diff] [blame]3660<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3661
3662<p>Bitwise binary operators are used to do various forms of bit-twiddling in a
3663 program. They are generally very efficient instructions and can commonly be
3664 strength reduced from other instructions. They require two operands of the
3665 same type, execute an operation on them, and produce a single value. The
3666 resulting value is the same type as its operands.</p>
3667
Reid Spencer2ab01932007-02-02 13:57:07 +0000[diff] [blame]3668</div>
3669
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3670<!-- _______________________________________________________________________ -->
3671<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
3672Instruction</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3673
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3674<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3675
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3676<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3677<pre>
3678 &lt;result&gt; = shl &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3679</pre>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3680
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3681<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3682<p>The '<tt>shl</tt>' instruction returns the first operand shifted to the left
3683 a specified number of bits.</p>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3684
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3685<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3686<p>Both arguments to the '<tt>shl</tt>' instruction must be the
3687 same <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3688 integer type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3689
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3690<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3691<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod
3692 2<sup>n</sup>, where <tt>n</tt> is the width of the result. If <tt>op2</tt>
3693 is (statically or dynamically) negative or equal to or larger than the number
3694 of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3695 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3696 shift amount in <tt>op2</tt>.</p>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3697
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3698<h5>Example:</h5>
3699<pre>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3700 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
3701 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
3702 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3703 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Mon P Wang4dd832d2008-12-09 05:46:39 +0000[diff] [blame]3704 &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]3705</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3706
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3707</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3708
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3709<!-- _______________________________________________________________________ -->
3710<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
3711Instruction</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3712
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3713<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3714
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3715<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3716<pre>
3717 &lt;result&gt; = lshr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3718</pre>
3719
3720<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3721<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
3722 operand shifted to the right a specified number of bits with zero fill.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3723
3724<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3725<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
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 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3728
3729<h5>Semantics:</h5>
3730<p>This instruction always performs a logical shift right operation. The most
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3731 significant bits of the result will be filled with zero bits after the shift.
3732 If <tt>op2</tt> is (statically or dynamically) equal to or larger than the
3733 number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3734 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3735 shift amount in <tt>op2</tt>.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3736
3737<h5>Example:</h5>
3738<pre>
3739 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
3740 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
3741 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
3742 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3743 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Mon P Wang4dd832d2008-12-09 05:46:39 +0000[diff] [blame]3744 &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]3745</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3746
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3747</div>
3748
Reid Spencer2ab01932007-02-02 13:57:07 +0000[diff] [blame]3749<!-- _______________________________________________________________________ -->
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3750<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
3751Instruction</a> </div>
3752<div class="doc_text">
3753
3754<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3755<pre>
3756 &lt;result&gt; = ashr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3757</pre>
3758
3759<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3760<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
3761 operand shifted to the right a specified number of bits with sign
3762 extension.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3763
3764<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3765<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3766 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3767 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3768
3769<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3770<p>This instruction always performs an arithmetic shift right operation, The
3771 most significant bits of the result will be filled with the sign bit
3772 of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
3773 larger than the number of bits in <tt>op1</tt>, the result is undefined. If
3774 the arguments are vectors, each vector element of <tt>op1</tt> is shifted by
3775 the corresponding shift amount in <tt>op2</tt>.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3776
3777<h5>Example:</h5>
3778<pre>
3779 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
3780 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
3781 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
3782 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3783 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Mon P Wang4dd832d2008-12-09 05:46:39 +0000[diff] [blame]3784 &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]3785</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3786
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3787</div>
3788
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3789<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3790<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
3791Instruction</a> </div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3792
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3793<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3794
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3795<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3796<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +0000[diff] [blame]3797 &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]3798</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3799
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3800<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3801<p>The '<tt>and</tt>' instruction returns the bitwise logical and of its two
3802 operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3803
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3804<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3805<p>The two arguments to the '<tt>and</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3806 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3807 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3808
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3809<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3810<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3811
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3812<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3813 <tbody>
3814 <tr>
3815 <td>In0</td>
3816 <td>In1</td>
3817 <td>Out</td>
3818 </tr>
3819 <tr>
3820 <td>0</td>
3821 <td>0</td>
3822 <td>0</td>
3823 </tr>
3824 <tr>
3825 <td>0</td>
3826 <td>1</td>
3827 <td>0</td>
3828 </tr>
3829 <tr>
3830 <td>1</td>
3831 <td>0</td>
3832 <td>0</td>
3833 </tr>
3834 <tr>
3835 <td>1</td>
3836 <td>1</td>
3837 <td>1</td>
3838 </tr>
3839 </tbody>
3840</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3841
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3842<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3843<pre>
3844 &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]3845 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
3846 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3847</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3848</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3849<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3850<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3851
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3852<div class="doc_text">
3853
3854<h5>Syntax:</h5>
3855<pre>
3856 &lt;result&gt; = or &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3857</pre>
3858
3859<h5>Overview:</h5>
3860<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive or of its
3861 two operands.</p>
3862
3863<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3864<p>The two arguments to the '<tt>or</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3865 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3866 values. Both arguments must have identical types.</p>
3867
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3868<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3869<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3870
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3871<table border="1" cellspacing="0" cellpadding="4">
3872 <tbody>
3873 <tr>
3874 <td>In0</td>
3875 <td>In1</td>
3876 <td>Out</td>
3877 </tr>
3878 <tr>
3879 <td>0</td>
3880 <td>0</td>
3881 <td>0</td>
3882 </tr>
3883 <tr>
3884 <td>0</td>
3885 <td>1</td>
3886 <td>1</td>
3887 </tr>
3888 <tr>
3889 <td>1</td>
3890 <td>0</td>
3891 <td>1</td>
3892 </tr>
3893 <tr>
3894 <td>1</td>
3895 <td>1</td>
3896 <td>1</td>
3897 </tr>
3898 </tbody>
3899</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3900
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3901<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3902<pre>
3903 &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]3904 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
3905 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3906</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3907
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3908</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3909
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3910<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3911<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
3912Instruction</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3913
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3914<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3915
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3916<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3917<pre>
3918 &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]3919</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3920
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3921<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3922<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive or of
3923 its two operands. The <tt>xor</tt> is used to implement the "one's
3924 complement" operation, which is the "~" operator in C.</p>
3925
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3926<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3927<p>The two arguments to the '<tt>xor</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3928 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3929 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3930
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3931<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3932<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3933
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3934<table border="1" cellspacing="0" cellpadding="4">
3935 <tbody>
3936 <tr>
3937 <td>In0</td>
3938 <td>In1</td>
3939 <td>Out</td>
3940 </tr>
3941 <tr>
3942 <td>0</td>
3943 <td>0</td>
3944 <td>0</td>
3945 </tr>
3946 <tr>
3947 <td>0</td>
3948 <td>1</td>
3949 <td>1</td>
3950 </tr>
3951 <tr>
3952 <td>1</td>
3953 <td>0</td>
3954 <td>1</td>
3955 </tr>
3956 <tr>
3957 <td>1</td>
3958 <td>1</td>
3959 <td>0</td>
3960 </tr>
3961 </tbody>
3962</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3963
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3964<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3965<pre>
3966 &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]3967 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
3968 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
3969 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3970</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3971
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3972</div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]3973
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3974<!-- ======================================================================= -->
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3975<div class="doc_subsection">
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]3976 <a name="vectorops">Vector Operations</a>
3977</div>
3978
3979<div class="doc_text">
3980
3981<p>LLVM supports several instructions to represent vector operations in a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3982 target-independent manner. These instructions cover the element-access and
3983 vector-specific operations needed to process vectors effectively. While LLVM
3984 does directly support these vector operations, many sophisticated algorithms
3985 will want to use target-specific intrinsics to take full advantage of a
3986 specific target.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]3987
3988</div>
3989
3990<!-- _______________________________________________________________________ -->
3991<div class="doc_subsubsection">
3992 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
3993</div>
3994
3995<div class="doc_text">
3996
3997<h5>Syntax:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]3998<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]3999 &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]4000</pre>
4001
4002<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4003<p>The '<tt>extractelement</tt>' instruction extracts a single scalar element
4004 from a vector at a specified index.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4005
4006
4007<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4008<p>The first operand of an '<tt>extractelement</tt>' instruction is a value
4009 of <a href="#t_vector">vector</a> type. The second operand is an index
4010 indicating the position from which to extract the element. The index may be
4011 a variable.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4012
4013<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4014<p>The result is a scalar of the same type as the element type of
4015 <tt>val</tt>. Its value is the value at position <tt>idx</tt> of
4016 <tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4017 results are undefined.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4018
4019<h5>Example:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4020<pre>
Gabor Greif03ab4dc2009-10-28 13:14:50 +0000[diff] [blame]4021 &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]4022</pre>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4023
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4024</div>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4025
4026<!-- _______________________________________________________________________ -->
4027<div class="doc_subsubsection">
4028 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
4029</div>
4030
4031<div class="doc_text">
4032
4033<h5>Syntax:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4034<pre>
Dan Gohman43ba0672008-05-12 23:38:42 +0000[diff] [blame]4035 &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]4036</pre>
4037
4038<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4039<p>The '<tt>insertelement</tt>' instruction inserts a scalar element into a
4040 vector at a specified index.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4041
4042<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4043<p>The first operand of an '<tt>insertelement</tt>' instruction is a value
4044 of <a href="#t_vector">vector</a> type. The second operand is a scalar value
4045 whose type must equal the element type of the first operand. The third
4046 operand is an index indicating the position at which to insert the value.
4047 The index may be a variable.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4048
4049<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4050<p>The result is a vector of the same type as <tt>val</tt>. Its element values
4051 are those of <tt>val</tt> except at position <tt>idx</tt>, where it gets the
4052 value <tt>elt</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4053 results are undefined.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4054
4055<h5>Example:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4056<pre>
Gabor Greif03ab4dc2009-10-28 13:14:50 +0000[diff] [blame]4057 &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]4058</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4059
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4060</div>
4061
4062<!-- _______________________________________________________________________ -->
4063<div class="doc_subsubsection">
4064 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
4065</div>
4066
4067<div class="doc_text">
4068
4069<h5>Syntax:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4070<pre>
Mon P Wang25f01062008-11-10 04:46:22 +0000[diff] [blame]4071 &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]4072</pre>
4073
4074<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4075<p>The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
4076 from two input vectors, returning a vector with the same element type as the
4077 input and length that is the same as the shuffle mask.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4078
4079<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4080<p>The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
4081 with types that match each other. The third argument is a shuffle mask whose
4082 element type is always 'i32'. The result of the instruction is a vector
4083 whose length is the same as the shuffle mask and whose element type is the
4084 same as the element type of the first two operands.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4085
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4086<p>The shuffle mask operand is required to be a constant vector with either
4087 constant integer or undef values.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4088
4089<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4090<p>The elements of the two input vectors are numbered from left to right across
4091 both of the vectors. The shuffle mask operand specifies, for each element of
4092 the result vector, which element of the two input vectors the result element
4093 gets. The element selector may be undef (meaning "don't care") and the
4094 second operand may be undef if performing a shuffle from only one vector.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4095
4096<h5>Example:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4097<pre>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4098 &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]4099 &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]4100 &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]4101 &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]4102 &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]4103 &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]4104 &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]4105 &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]4106</pre>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4107
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4108</div>
Tanya Lattnerb138bbe2006-04-14 19:24:33 +0000[diff] [blame]4109
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4110<!-- ======================================================================= -->
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4111<div class="doc_subsection">
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4112 <a name="aggregateops">Aggregate Operations</a>
4113</div>
4114
4115<div class="doc_text">
4116
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4117<p>LLVM supports several instructions for working with
4118 <a href="#t_aggregate">aggregate</a> values.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4119
4120</div>
4121
4122<!-- _______________________________________________________________________ -->
4123<div class="doc_subsubsection">
4124 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
4125</div>
4126
4127<div class="doc_text">
4128
4129<h5>Syntax:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4130<pre>
4131 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
4132</pre>
4133
4134<h5>Overview:</h5>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4135<p>The '<tt>extractvalue</tt>' instruction extracts the value of a member field
4136 from an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4137
4138<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4139<p>The first operand of an '<tt>extractvalue</tt>' instruction is a value
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]4140 of <a href="#t_struct">struct</a> or
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4141 <a href="#t_array">array</a> type. The operands are constant indices to
4142 specify which value to extract in a similar manner as indices in a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4143 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
Frits van Bommel7cf63ac2010-12-05 20:54:38 +0000[diff] [blame]4144 <p>The major differences to <tt>getelementptr</tt> indexing are:</p>
4145 <ul>
4146 <li>Since the value being indexed is not a pointer, the first index is
4147 omitted and assumed to be zero.</li>
4148 <li>At least one index must be specified.</li>
4149 <li>Not only struct indices but also array indices must be in
4150 bounds.</li>
4151 </ul>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4152
4153<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4154<p>The result is the value at the position in the aggregate specified by the
4155 index operands.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4156
4157<h5>Example:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4158<pre>
Gabor Greif03ab4dc2009-10-28 13:14:50 +0000[diff] [blame]4159 &lt;result&gt; = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4160</pre>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4161
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4162</div>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4163
4164<!-- _______________________________________________________________________ -->
4165<div class="doc_subsubsection">
4166 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
4167</div>
4168
4169<div class="doc_text">
4170
4171<h5>Syntax:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4172<pre>
Jeffrey Yasskinf991bbb2010-01-11 19:19:26 +0000[diff] [blame]4173 &lt;result&gt; = insertvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;ty&gt; &lt;elt&gt;, &lt;idx&gt; <i>; yields &lt;aggregate type&gt;</i>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4174</pre>
4175
4176<h5>Overview:</h5>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4177<p>The '<tt>insertvalue</tt>' instruction inserts a value into a member field
4178 in an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4179
4180<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4181<p>The first operand of an '<tt>insertvalue</tt>' instruction is a value
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]4182 of <a href="#t_struct">struct</a> or
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4183 <a href="#t_array">array</a> type. The second operand is a first-class
4184 value to insert. The following operands are constant indices indicating
4185 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]4186 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' instruction. The
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4187 value to insert must have the same type as the value identified by the
4188 indices.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4189
4190<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4191<p>The result is an aggregate of the same type as <tt>val</tt>. Its value is
4192 that of <tt>val</tt> except that the value at the position specified by the
4193 indices is that of <tt>elt</tt>.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4194
4195<h5>Example:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4196<pre>
Jeffrey Yasskinf991bbb2010-01-11 19:19:26 +0000[diff] [blame]4197 %agg1 = insertvalue {i32, float} undef, i32 1, 0 <i>; yields {i32 1, float undef}</i>
4198 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 <i>; yields {i32 1, float %val}</i>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4199</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4200
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4201</div>
4202
4203
4204<!-- ======================================================================= -->
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4205<div class="doc_subsection">
Chris Lattner6ab66722006-08-15 00:45:58 +0000[diff] [blame]4206 <a name="memoryops">Memory Access and Addressing Operations</a>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4207</div>
4208
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4209<div class="doc_text">
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4210
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4211<p>A key design point of an SSA-based representation is how it represents
4212 memory. In LLVM, no memory locations are in SSA form, which makes things
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]4213 very simple. This section describes how to read, write, and allocate
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4214 memory in LLVM.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4215
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4216</div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4217
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4218<!-- _______________________________________________________________________ -->
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4219<div class="doc_subsubsection">
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4220 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
4221</div>
4222
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4223<div class="doc_text">
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4224
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4225<h5>Syntax:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4226<pre>
Dan Gohman2140a742010-05-28 01:14:11 +0000[diff] [blame]4227 &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]4228</pre>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4229
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4230<h5>Overview:</h5>
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]4231<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4232 currently executing function, to be automatically released when this function
4233 returns to its caller. The object is always allocated in the generic address
4234 space (address space zero).</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4235
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4236<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4237<p>The '<tt>alloca</tt>' instruction
4238 allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt> bytes of memory on the
4239 runtime stack, returning a pointer of the appropriate type to the program.
4240 If "NumElements" is specified, it is the number of elements allocated,
4241 otherwise "NumElements" is defaulted to be one. If a constant alignment is
4242 specified, the value result of the allocation is guaranteed to be aligned to
4243 at least that boundary. If not specified, or if zero, the target can choose
4244 to align the allocation on any convenient boundary compatible with the
4245 type.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4246
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4247<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4248
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4249<h5>Semantics:</h5>
Bill Wendling9ee6a312009-05-08 20:49:29 +0000[diff] [blame]4250<p>Memory is allocated; a pointer is returned. The operation is undefined if
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4251 there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
4252 memory is automatically released when the function returns. The
4253 '<tt>alloca</tt>' instruction is commonly used to represent automatic
4254 variables that must have an address available. When the function returns
4255 (either with the <tt><a href="#i_ret">ret</a></tt>
4256 or <tt><a href="#i_unwind">unwind</a></tt> instructions), the memory is
4257 reclaimed. Allocating zero bytes is legal, but the result is undefined.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4258
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4259<h5>Example:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4260<pre>
Dan Gohman7a5acb52009-01-04 23:49:44 +0000[diff] [blame]4261 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
4262 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
4263 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
4264 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4265</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4266
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4267</div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4268
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4269<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]4270<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
4271Instruction</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4272
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4273<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4274
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4275<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4276<pre>
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4277 &lt;result&gt; = load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;]
4278 &lt;result&gt; = volatile load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;]
4279 !&lt;index&gt; = !{ i32 1 }
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4280</pre>
4281
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4282<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4283<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4284
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4285<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4286<p>The argument to the '<tt>load</tt>' instruction specifies the memory address
4287 from which to load. The pointer must point to
4288 a <a href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
4289 marked as <tt>volatile</tt>, then the optimizer is not allowed to modify the
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]4290 number or order of execution of this <tt>load</tt> with other <a
4291 href="#volatile">volatile operations</a>.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4292
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4293<p>The optional constant <tt>align</tt> argument specifies the alignment of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4294 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]4295 omitted <tt>align</tt> argument means that the operation has the preferential
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4296 alignment for the target. It is the responsibility of the code emitter to
4297 ensure that the alignment information is correct. Overestimating the
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4298 alignment results in undefined behavior. Underestimating the alignment may
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4299 produce less efficient code. An alignment of 1 is always safe.</p>
4300
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4301<p>The optional <tt>!nontemporal</tt> metadata must reference a single
4302 metatadata name &lt;index&gt; corresponding to a metadata node with
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]4303 one <tt>i32</tt> entry of value 1. The existence of
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4304 the <tt>!nontemporal</tt> metatadata on the instruction tells the optimizer
4305 and code generator that this load is not expected to be reused in the cache.
4306 The code generator may select special instructions to save cache bandwidth,
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]4307 such as the <tt>MOVNT</tt> instruction on x86.</p>
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4308
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4309<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4310<p>The location of memory pointed to is loaded. If the value being loaded is of
4311 scalar type then the number of bytes read does not exceed the minimum number
4312 of bytes needed to hold all bits of the type. For example, loading an
4313 <tt>i24</tt> reads at most three bytes. When loading a value of a type like
4314 <tt>i20</tt> with a size that is not an integral number of bytes, the result
4315 is undefined if the value was not originally written using a store of the
4316 same type.</p>
4317
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4318<h5>Examples:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4319<pre>
4320 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
4321 <a href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4322 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4323</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4324
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4325</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4326
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4327<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]4328<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
4329Instruction</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4330
Reid Spencera89fb182006-11-09 21:18:01 +0000[diff] [blame]4331<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4332
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4333<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4334<pre>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]4335 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>
4336 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]4337</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4338
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4339<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4340<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4341
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4342<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4343<p>There are two arguments to the '<tt>store</tt>' instruction: a value to store
4344 and an address at which to store it. The type of the
4345 '<tt>&lt;pointer&gt;</tt>' operand must be a pointer to
4346 the <a href="#t_firstclass">first class</a> type of the
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]4347 '<tt>&lt;value&gt;</tt>' operand. If the <tt>store</tt> is marked as
4348 <tt>volatile</tt>, then the optimizer is not allowed to modify the number or
4349 order of execution of this <tt>store</tt> with other <a
4350 href="#volatile">volatile operations</a>.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4351
4352<p>The optional constant "align" argument specifies the alignment of the
4353 operation (that is, the alignment of the memory address). A value of 0 or an
4354 omitted "align" argument means that the operation has the preferential
4355 alignment for the target. It is the responsibility of the code emitter to
4356 ensure that the alignment information is correct. Overestimating the
4357 alignment results in an undefined behavior. Underestimating the alignment may
4358 produce less efficient code. An alignment of 1 is always safe.</p>
4359
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4360<p>The optional !nontemporal metadata must reference a single metatadata
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]4361 name &lt;index&gt; corresponding to a metadata node with one i32 entry of
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]4362 value 1. The existence of the !nontemporal metatadata on the
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4363 instruction tells the optimizer and code generator that this load is
4364 not expected to be reused in the cache. The code generator may
4365 select special instructions to save cache bandwidth, such as the
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]4366 MOVNT instruction on x86.</p>
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4367
4368
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]4369<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4370<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>' at the
4371 location specified by the '<tt>&lt;pointer&gt;</tt>' operand. If
4372 '<tt>&lt;value&gt;</tt>' is of scalar type then the number of bytes written
4373 does not exceed the minimum number of bytes needed to hold all bits of the
4374 type. For example, storing an <tt>i24</tt> writes at most three bytes. When
4375 writing a value of a type like <tt>i20</tt> with a size that is not an
4376 integral number of bytes, it is unspecified what happens to the extra bits
4377 that do not belong to the type, but they will typically be overwritten.</p>
4378
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4379<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4380<pre>
4381 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling8830ffe2007-10-22 05:10:05 +0000[diff] [blame]4382 store i32 3, i32* %ptr <i>; yields {void}</i>
4383 %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]4384</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4385
Reid Spencer443460a2006-11-09 21:15:49 +0000[diff] [blame]4386</div>
4387
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4388<!-- _______________________________________________________________________ -->
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4389<div class="doc_subsubsection">
4390 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
4391</div>
4392
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4393<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4394
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]4395<h5>Syntax:</h5>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4396<pre>
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]4397 &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]4398 &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]4399</pre>
4400
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]4401<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4402<p>The '<tt>getelementptr</tt>' instruction is used to get the address of a
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4403 subelement of an <a href="#t_aggregate">aggregate</a> data structure.
4404 It performs address calculation only and does not access memory.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4405
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]4406<h5>Arguments:</h5>
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]4407<p>The first argument is always a pointer, and forms the basis of the
Chris Lattnera40b9122009-07-29 06:44:13 +0000[diff] [blame]4408 calculation. The remaining arguments are indices that indicate which of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4409 elements of the aggregate object are indexed. The interpretation of each
4410 index is dependent on the type being indexed into. The first index always
4411 indexes the pointer value given as the first argument, the second index
4412 indexes a value of the type pointed to (not necessarily the value directly
4413 pointed to, since the first index can be non-zero), etc. The first type
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4414 indexed into must be a pointer value, subsequent types can be arrays,
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]4415 vectors, and structs. Note that subsequent types being indexed into
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4416 can never be pointers, since that would require loading the pointer before
4417 continuing calculation.</p>
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]4418
4419<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]4420 When indexing into a (optionally packed) structure, only <tt>i32</tt>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4421 integer <b>constants</b> are allowed. When indexing into an array, pointer
4422 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]4423 constant.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4424
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4425<p>For example, let's consider a C code fragment and how it gets compiled to
4426 LLVM:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4427
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]4428<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4429struct RT {
4430 char A;
Chris Lattnera446f1b2007-05-29 15:43:56 +0000[diff] [blame]4431 int B[10][20];
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4432 char C;
4433};
4434struct ST {
Chris Lattnera446f1b2007-05-29 15:43:56 +0000[diff] [blame]4435 int X;
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4436 double Y;
4437 struct RT Z;
4438};
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4439
Chris Lattnera446f1b2007-05-29 15:43:56 +0000[diff] [blame]4440int *foo(struct ST *s) {
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4441 return &amp;s[1].Z.B[5][13];
4442}
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4443</pre>
4444
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4445<p>The LLVM code generated by the GCC frontend is:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4446
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]4447<pre class="doc_code">
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]4448%RT = <a href="#namedtypes">type</a> { i8 , [10 x [20 x i32]], i8 }
4449%ST = <a href="#namedtypes">type</a> { i32, double, %RT }
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4450
Dan Gohman6b867702009-07-25 02:23:48 +0000[diff] [blame]4451define i32* @foo(%ST* %s) {
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4452entry:
4453 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
4454 ret i32* %reg
4455}
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4456</pre>
4457
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]4458<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4459<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]4460 type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
4461 }</tt>' type, a structure. The second index indexes into the third element
4462 of the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
4463 i8 }</tt>' type, another structure. The third index indexes into the second
4464 element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
4465 array. The two dimensions of the array are subscripted into, yielding an
4466 '<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a
4467 pointer to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4468
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4469<p>Note that it is perfectly legal to index partially through a structure,
4470 returning a pointer to an inner element. Because of this, the LLVM code for
4471 the given testcase is equivalent to:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4472
4473<pre>
Dan Gohman6b867702009-07-25 02:23:48 +0000[diff] [blame]4474 define i32* @foo(%ST* %s) {
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4475 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]4476 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
4477 %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]4478 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
4479 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
4480 ret i32* %t5
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4481 }
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]4482</pre>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +0000[diff] [blame]4483
Dan Gohman1639c392009-07-27 21:53:46 +0000[diff] [blame]4484<p>If the <tt>inbounds</tt> keyword is present, the result value of the
Dan Gohman57255802010-04-23 15:23:32 +0000[diff] [blame]4485 <tt>getelementptr</tt> is a <a href="#trapvalues">trap value</a> if the
4486 base pointer is not an <i>in bounds</i> address of an allocated object,
4487 or if any of the addresses that would be formed by successive addition of
4488 the offsets implied by the indices to the base address with infinitely
4489 precise arithmetic are not an <i>in bounds</i> address of that allocated
4490 object. The <i>in bounds</i> addresses for an allocated object are all
4491 the addresses that point into the object, plus the address one byte past
4492 the end.</p>
Dan Gohman1639c392009-07-27 21:53:46 +0000[diff] [blame]4493
4494<p>If the <tt>inbounds</tt> keyword is not present, the offsets are added to
4495 the base address with silently-wrapping two's complement arithmetic, and
4496 the result value of the <tt>getelementptr</tt> may be outside the object
4497 pointed to by the base pointer. The result value may not necessarily be
4498 used to access memory though, even if it happens to point into allocated
4499 storage. See the <a href="#pointeraliasing">Pointer Aliasing Rules</a>
4500 section for more information.</p>
4501
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4502<p>The getelementptr instruction is often confusing. For some more insight into
4503 how it works, see <a href="GetElementPtr.html">the getelementptr FAQ</a>.</p>
Chris Lattner6ab66722006-08-15 00:45:58 +0000[diff] [blame]4504
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]4505<h5>Example:</h5>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4506<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4507 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]4508 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
4509 <i>; yields i8*:vptr</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]4510 %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]4511 <i>; yields i8*:eptr</i>
4512 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Sanjiv Gupta0c155e62009-04-25 07:27:44 +0000[diff] [blame]4513 <i>; yields i32*:iptr</i>
Sanjiv Gupta77abea02009-04-24 16:38:13 +0000[diff] [blame]4514 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4515</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4516
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4517</div>
Reid Spencer443460a2006-11-09 21:15:49 +0000[diff] [blame]4518
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4519<!-- ======================================================================= -->
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]4520<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4521</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4522
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4523<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4524
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]4525<p>The instructions in this category are the conversion instructions (casting)
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4526 which all take a single operand and a type. They perform various bit
4527 conversions on the operand.</p>
4528
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4529</div>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]4530
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]4531<!-- _______________________________________________________________________ -->
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]4532<div class="doc_subsubsection">
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4533 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
4534</div>
4535<div class="doc_text">
4536
4537<h5>Syntax:</h5>
4538<pre>
4539 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4540</pre>
4541
4542<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4543<p>The '<tt>trunc</tt>' instruction truncates its operand to the
4544 type <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4545
4546<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4547<p>The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
4548 be an <a href="#t_integer">integer</a> type, and a type that specifies the
4549 size and type of the result, which must be
4550 an <a href="#t_integer">integer</a> type. The bit size of <tt>value</tt> must
4551 be larger than the bit size of <tt>ty2</tt>. Equal sized types are not
4552 allowed.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4553
4554<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4555<p>The '<tt>trunc</tt>' instruction truncates the high order bits
4556 in <tt>value</tt> and converts the remaining bits to <tt>ty2</tt>. Since the
4557 source size must be larger than the destination size, <tt>trunc</tt> cannot
4558 be a <i>no-op cast</i>. It will always truncate bits.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4559
4560<h5>Example:</h5>
4561<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4562 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]4563 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
Gabor Greiff50fd572009-10-28 09:21:30 +0000[diff] [blame]4564 %Z = trunc i32 122 to i1 <i>; yields i1:false</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4565</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4566
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4567</div>
4568
4569<!-- _______________________________________________________________________ -->
4570<div class="doc_subsubsection">
4571 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
4572</div>
4573<div class="doc_text">
4574
4575<h5>Syntax:</h5>
4576<pre>
4577 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4578</pre>
4579
4580<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4581<p>The '<tt>zext</tt>' instruction zero extends its operand to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4582 <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4583
4584
4585<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4586<p>The '<tt>zext</tt>' instruction takes a value to cast, which must be of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4587 <a href="#t_integer">integer</a> type, and a type to cast it to, which must
4588 also be of <a href="#t_integer">integer</a> type. The bit size of the
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4589 <tt>value</tt> must be smaller than the bit size of the destination type,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4590 <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4591
4592<h5>Semantics:</h5>
4593<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]4594 bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4595
Reid Spencer07c9c682007-01-12 15:46:11 +0000[diff] [blame]4596<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]4597
4598<h5>Example:</h5>
4599<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4600 %X = zext i32 257 to i64 <i>; yields i64:257</i>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]4601 %Y = zext i1 true to i32 <i>; yields i32:1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4602</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4603
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4604</div>
4605
4606<!-- _______________________________________________________________________ -->
4607<div class="doc_subsubsection">
4608 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
4609</div>
4610<div class="doc_text">
4611
4612<h5>Syntax:</h5>
4613<pre>
4614 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4615</pre>
4616
4617<h5>Overview:</h5>
4618<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
4619
4620<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4621<p>The '<tt>sext</tt>' instruction takes a value to cast, which must be of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4622 <a href="#t_integer">integer</a> type, and a type to cast it to, which must
4623 also be of <a href="#t_integer">integer</a> type. The bit size of the
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4624 <tt>value</tt> must be smaller than the bit size of the destination type,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4625 <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4626
4627<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4628<p>The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
4629 bit (highest order bit) of the <tt>value</tt> until it reaches the bit size
4630 of the type <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4631
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]4632<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]4633
4634<h5>Example:</h5>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4635<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4636 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]4637 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4638</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4639
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4640</div>
4641
4642<!-- _______________________________________________________________________ -->
4643<div class="doc_subsubsection">
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4644 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
4645</div>
4646
4647<div class="doc_text">
4648
4649<h5>Syntax:</h5>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4650<pre>
4651 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4652</pre>
4653
4654<h5>Overview:</h5>
4655<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4656 <tt>ty2</tt>.</p>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4657
4658<h5>Arguments:</h5>
4659<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4660 point</a> value to cast and a <a href="#t_floating">floating point</a> type
4661 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]4662 <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]4663 <i>no-op cast</i>.</p>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4664
4665<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4666<p>The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4667 <a href="#t_floating">floating point</a> type to a smaller
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4668 <a href="#t_floating">floating point</a> type. If the value cannot fit
4669 within the destination type, <tt>ty2</tt>, then the results are
4670 undefined.</p>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4671
4672<h5>Example:</h5>
4673<pre>
4674 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
4675 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
4676</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4677
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4678</div>
4679
4680<!-- _______________________________________________________________________ -->
4681<div class="doc_subsubsection">
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4682 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
4683</div>
4684<div class="doc_text">
4685
4686<h5>Syntax:</h5>
4687<pre>
4688 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4689</pre>
4690
4691<h5>Overview:</h5>
4692<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]4693 floating point value.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4694
4695<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4696<p>The '<tt>fpext</tt>' instruction takes a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4697 <a href="#t_floating">floating point</a> <tt>value</tt> to cast, and
4698 a <a href="#t_floating">floating point</a> type to cast it to. The source
4699 type must be smaller than the destination type.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4700
4701<h5>Semantics:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4702<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4703 <a href="#t_floating">floating point</a> type to a larger
4704 <a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
4705 used to make a <i>no-op cast</i> because it always changes bits. Use
4706 <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]4707
4708<h5>Example:</h5>
4709<pre>
4710 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
4711 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
4712</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4713
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4714</div>
4715
4716<!-- _______________________________________________________________________ -->
4717<div class="doc_subsubsection">
Reid Spencer2eadb532007-01-21 00:29:26 +0000[diff] [blame]4718 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4719</div>
4720<div class="doc_text">
4721
4722<h5>Syntax:</h5>
4723<pre>
Reid Spencer753163d2007-07-31 14:40:14 +0000[diff] [blame]4724 &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]4725</pre>
4726
4727<h5>Overview:</h5>
Reid Spencer753163d2007-07-31 14:40:14 +0000[diff] [blame]4728<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4729 unsigned integer equivalent of type <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4730
4731<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4732<p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
4733 scalar or vector <a href="#t_floating">floating point</a> value, and a type
4734 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4735 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4736 vector integer type with the same number of elements as <tt>ty</tt></p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4737
4738<h5>Semantics:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4739<p>The '<tt>fptoui</tt>' instruction converts its
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4740 <a href="#t_floating">floating point</a> operand into the nearest (rounding
4741 towards zero) unsigned integer value. If the value cannot fit
4742 in <tt>ty2</tt>, the results are undefined.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4743
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4744<h5>Example:</h5>
4745<pre>
Reid Spencer753163d2007-07-31 14:40:14 +0000[diff] [blame]4746 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner5b95a172007-09-22 03:17:52 +0000[diff] [blame]4747 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Gabor Greiff50fd572009-10-28 09:21:30 +0000[diff] [blame]4748 %Z = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4749</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4750
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4751</div>
4752
4753<!-- _______________________________________________________________________ -->
4754<div class="doc_subsubsection">
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4755 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4756</div>
4757<div class="doc_text">
4758
4759<h5>Syntax:</h5>
4760<pre>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4761 &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]4762</pre>
4763
4764<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4765<p>The '<tt>fptosi</tt>' instruction converts
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4766 <a href="#t_floating">floating point</a> <tt>value</tt> to
4767 type <tt>ty2</tt>.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]4768
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]4769<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4770<p>The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
4771 scalar or vector <a href="#t_floating">floating point</a> value, and a type
4772 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4773 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4774 vector integer type with the same number of elements as <tt>ty</tt></p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]4775
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]4776<h5>Semantics:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4777<p>The '<tt>fptosi</tt>' instruction converts its
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4778 <a href="#t_floating">floating point</a> operand into the nearest (rounding
4779 towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
4780 the results are undefined.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]4781
Chris Lattner70de6632001-07-09 00:26:23 +0000[diff] [blame]4782<h5>Example:</h5>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]4783<pre>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]4784 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner5b95a172007-09-22 03:17:52 +0000[diff] [blame]4785 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Gabor Greiff50fd572009-10-28 09:21:30 +0000[diff] [blame]4786 %Z = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4787</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4788
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4789</div>
4790
4791<!-- _______________________________________________________________________ -->
4792<div class="doc_subsubsection">
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4793 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4794</div>
4795<div class="doc_text">
4796
4797<h5>Syntax:</h5>
4798<pre>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4799 &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]4800</pre>
4801
4802<h5>Overview:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4803<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4804 integer and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4805
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4806<h5>Arguments:</h5>
Nate Begemand4d45c22007-11-17 03:58:34 +0000[diff] [blame]4807<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]4808 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
4809 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4810 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4811 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4812
4813<h5>Semantics:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4814<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4815 integer quantity and converts it to the corresponding floating point
4816 value. If the value cannot fit in the floating point value, the results are
4817 undefined.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4818
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4819<h5>Example:</h5>
4820<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4821 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]4822 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4823</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4824
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4825</div>
4826
4827<!-- _______________________________________________________________________ -->
4828<div class="doc_subsubsection">
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4829 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4830</div>
4831<div class="doc_text">
4832
4833<h5>Syntax:</h5>
4834<pre>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4835 &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]4836</pre>
4837
4838<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4839<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed integer
4840 and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4841
4842<h5>Arguments:</h5>
Nate Begemand4d45c22007-11-17 03:58:34 +0000[diff] [blame]4843<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]4844 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
4845 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4846 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4847 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4848
4849<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4850<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed integer
4851 quantity and converts it to the corresponding floating point value. If the
4852 value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4853
4854<h5>Example:</h5>
4855<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4856 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]4857 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4858</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4859
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4860</div>
4861
4862<!-- _______________________________________________________________________ -->
4863<div class="doc_subsubsection">
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]4864 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
4865</div>
4866<div class="doc_text">
4867
4868<h5>Syntax:</h5>
4869<pre>
4870 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4871</pre>
4872
4873<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4874<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
4875 the integer type <tt>ty2</tt>.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]4876
4877<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4878<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
4879 must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
4880 <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]4881
4882<h5>Semantics:</h5>
4883<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4884 <tt>ty2</tt> by interpreting the pointer value as an integer and either
4885 truncating or zero extending that value to the size of the integer type. If
4886 <tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
4887 <tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
4888 are the same size, then nothing is done (<i>no-op cast</i>) other than a type
4889 change.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]4890
4891<h5>Example:</h5>
4892<pre>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]4893 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
4894 %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]4895</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4896
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]4897</div>
4898
4899<!-- _______________________________________________________________________ -->
4900<div class="doc_subsubsection">
4901 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
4902</div>
4903<div class="doc_text">
4904
4905<h5>Syntax:</h5>
4906<pre>
4907 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4908</pre>
4909
4910<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4911<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to a
4912 pointer type, <tt>ty2</tt>.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]4913
4914<h5>Arguments:</h5>
Duncan Sands16f122e2007-03-30 12:22:09 +0000[diff] [blame]4915<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4916 value to cast, and a type to cast it to, which must be a
4917 <a href="#t_pointer">pointer</a> type.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]4918
4919<h5>Semantics:</h5>
4920<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4921 <tt>ty2</tt> by applying either a zero extension or a truncation depending on
4922 the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
4923 size of a pointer then a truncation is done. If <tt>value</tt> is smaller
4924 than the size of a pointer then a zero extension is done. If they are the
4925 same size, nothing is done (<i>no-op cast</i>).</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]4926
4927<h5>Example:</h5>
4928<pre>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]4929 %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]4930 %Y = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
4931 %Z = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]4932</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4933
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]4934</div>
4935
4936<!-- _______________________________________________________________________ -->
4937<div class="doc_subsubsection">
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]4938 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4939</div>
4940<div class="doc_text">
4941
4942<h5>Syntax:</h5>
4943<pre>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]4944 &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]4945</pre>
4946
4947<h5>Overview:</h5>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]4948<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4949 <tt>ty2</tt> without changing any bits.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4950
4951<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4952<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be a
4953 non-aggregate first class value, and a type to cast it to, which must also be
4954 a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes
4955 of <tt>value</tt> and the destination type, <tt>ty2</tt>, must be
4956 identical. If the source type is a pointer, the destination type must also be
4957 a pointer. This instruction supports bitwise conversion of vectors to
4958 integers and to vectors of other types (as long as they have the same
4959 size).</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4960
4961<h5>Semantics:</h5>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]4962<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4963 <tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
4964 this conversion. The conversion is done as if the <tt>value</tt> had been
4965 stored to memory and read back as type <tt>ty2</tt>. Pointer types may only
4966 be converted to other pointer types with this instruction. To convert
4967 pointers to other types, use the <a href="#i_inttoptr">inttoptr</a> or
4968 <a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4969
4970<h5>Example:</h5>
4971<pre>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]4972 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4973 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4974 %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]4975</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4976
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4977</div>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]4978
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]4979<!-- ======================================================================= -->
4980<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4981
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]4982<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4983
4984<p>The instructions in this category are the "miscellaneous" instructions, which
4985 defy better classification.</p>
4986
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]4987</div>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]4988
4989<!-- _______________________________________________________________________ -->
4990<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
4991</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4992
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]4993<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4994
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]4995<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4996<pre>
4997 &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]4998</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4999
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5000<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5001<p>The '<tt>icmp</tt>' instruction returns a boolean value or a vector of
5002 boolean values based on comparison of its two integer, integer vector, or
5003 pointer operands.</p>
5004
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5005<h5>Arguments:</h5>
5006<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5007 the condition code indicating the kind of comparison to perform. It is not a
5008 value, just a keyword. The possible condition code are:</p>
5009
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5010<ol>
5011 <li><tt>eq</tt>: equal</li>
5012 <li><tt>ne</tt>: not equal </li>
5013 <li><tt>ugt</tt>: unsigned greater than</li>
5014 <li><tt>uge</tt>: unsigned greater or equal</li>
5015 <li><tt>ult</tt>: unsigned less than</li>
5016 <li><tt>ule</tt>: unsigned less or equal</li>
5017 <li><tt>sgt</tt>: signed greater than</li>
5018 <li><tt>sge</tt>: signed greater or equal</li>
5019 <li><tt>slt</tt>: signed less than</li>
5020 <li><tt>sle</tt>: signed less or equal</li>
5021</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5022
Chris Lattnerc0f423a2007-01-15 01:54:13 +0000[diff] [blame]5023<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5024 <a href="#t_pointer">pointer</a> or integer <a href="#t_vector">vector</a>
5025 typed. They must also be identical types.</p>
5026
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5027<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5028<p>The '<tt>icmp</tt>' compares <tt>op1</tt> and <tt>op2</tt> according to the
5029 condition code given as <tt>cond</tt>. The comparison performed always yields
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]5030 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]5031 result, as follows:</p>
5032
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5033<ol>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5034 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5035 <tt>false</tt> otherwise. No sign interpretation is necessary or
5036 performed.</li>
5037
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5038 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5039 <tt>false</tt> otherwise. No sign interpretation is necessary or
5040 performed.</li>
5041
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5042 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5043 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5044
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5045 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5046 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5047 to <tt>op2</tt>.</li>
5048
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5049 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5050 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5051
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5052 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5053 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5054
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5055 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5056 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5057
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5058 <li><tt>sge</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5059 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5060 to <tt>op2</tt>.</li>
5061
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5062 <li><tt>slt</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5063 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5064
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5065 <li><tt>sle</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5066 <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]5067</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5068
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5069<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5070 values are compared as if they were integers.</p>
5071
5072<p>If the operands are integer vectors, then they are compared element by
5073 element. The result is an <tt>i1</tt> vector with the same number of elements
5074 as the values being compared. Otherwise, the result is an <tt>i1</tt>.</p>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5075
5076<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5077<pre>
5078 &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5079 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
5080 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
5081 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
5082 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
5083 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5084</pre>
Dan Gohmana5127ab2009-01-22 01:39:38 +0000[diff] [blame]5085
5086<p>Note that the code generator does not yet support vector types with
5087 the <tt>icmp</tt> instruction.</p>
5088
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5089</div>
5090
5091<!-- _______________________________________________________________________ -->
5092<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
5093</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5094
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5095<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5096
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5097<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5098<pre>
5099 &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]5100</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5101
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5102<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5103<p>The '<tt>fcmp</tt>' instruction returns a boolean value or vector of boolean
5104 values based on comparison of its operands.</p>
5105
5106<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]5107(<a href="#t_integer"><tt>i1</tt></a>).</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5108
5109<p>If the operands are floating point vectors, then the result type is a vector
5110 of boolean with the same number of elements as the operands being
5111 compared.</p>
5112
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5113<h5>Arguments:</h5>
5114<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5115 the condition code indicating the kind of comparison to perform. It is not a
5116 value, just a keyword. The possible condition code are:</p>
5117
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5118<ol>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5119 <li><tt>false</tt>: no comparison, always returns false</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5120 <li><tt>oeq</tt>: ordered and equal</li>
5121 <li><tt>ogt</tt>: ordered and greater than </li>
5122 <li><tt>oge</tt>: ordered and greater than or equal</li>
5123 <li><tt>olt</tt>: ordered and less than </li>
5124 <li><tt>ole</tt>: ordered and less than or equal</li>
5125 <li><tt>one</tt>: ordered and not equal</li>
5126 <li><tt>ord</tt>: ordered (no nans)</li>
5127 <li><tt>ueq</tt>: unordered or equal</li>
5128 <li><tt>ugt</tt>: unordered or greater than </li>
5129 <li><tt>uge</tt>: unordered or greater than or equal</li>
5130 <li><tt>ult</tt>: unordered or less than </li>
5131 <li><tt>ule</tt>: unordered or less than or equal</li>
5132 <li><tt>une</tt>: unordered or not equal</li>
5133 <li><tt>uno</tt>: unordered (either nans)</li>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5134 <li><tt>true</tt>: no comparison, always returns true</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5135</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5136
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5137<p><i>Ordered</i> means that neither operand is a QNAN while
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5138 <i>unordered</i> means that either operand may be a QNAN.</p>
5139
5140<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be either
5141 a <a href="#t_floating">floating point</a> type or
5142 a <a href="#t_vector">vector</a> of floating point type. They must have
5143 identical types.</p>
5144
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5145<h5>Semantics:</h5>
Gabor Greif0f75ad02008-08-07 21:46:00 +0000[diff] [blame]5146<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5147 according to the condition code given as <tt>cond</tt>. If the operands are
5148 vectors, then the vectors are compared element by element. Each comparison
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]5149 performed always yields an <a href="#t_integer">i1</a> result, as
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5150 follows:</p>
5151
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5152<ol>
5153 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5154
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5155 <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]5156 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5157
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5158 <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]5159 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5160
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5161 <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]5162 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5163
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5164 <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]5165 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5166
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5167 <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]5168 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5169
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5170 <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]5171 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5172
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5173 <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]5174
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5175 <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]5176 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5177
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5178 <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]5179 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5180
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5181 <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]5182 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5183
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5184 <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]5185 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5186
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5187 <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]5188 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5189
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5190 <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]5191 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5192
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5193 <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]5194
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5195 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
5196</ol>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5197
5198<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5199<pre>
5200 &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]5201 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
5202 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
5203 &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]5204</pre>
Dan Gohmana5127ab2009-01-22 01:39:38 +0000[diff] [blame]5205
5206<p>Note that the code generator does not yet support vector types with
5207 the <tt>fcmp</tt> instruction.</p>
5208
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5209</div>
5210
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5211<!-- _______________________________________________________________________ -->
Nate Begemand2195702008-05-12 19:01:56 +0000[diff] [blame]5212<div class="doc_subsubsection">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5213 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
5214</div>
5215
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5216<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5217
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5218<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5219<pre>
5220 &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...
5221</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5222
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5223<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5224<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in the
5225 SSA graph representing the function.</p>
5226
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5227<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5228<p>The type of the incoming values is specified with the first type field. After
5229 this, the '<tt>phi</tt>' instruction takes a list of pairs as arguments, with
5230 one pair for each predecessor basic block of the current block. Only values
5231 of <a href="#t_firstclass">first class</a> type may be used as the value
5232 arguments to the PHI node. Only labels may be used as the label
5233 arguments.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5234
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5235<p>There must be no non-phi instructions between the start of a basic block and
5236 the PHI instructions: i.e. PHI instructions must be first in a basic
5237 block.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5238
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5239<p>For the purposes of the SSA form, the use of each incoming value is deemed to
5240 occur on the edge from the corresponding predecessor block to the current
5241 block (but after any definition of an '<tt>invoke</tt>' instruction's return
5242 value on the same edge).</p>
Jay Foad1a4eea52009-06-03 10:20:10 +0000[diff] [blame]5243
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5244<h5>Semantics:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5245<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5246 specified by the pair corresponding to the predecessor basic block that
5247 executed just prior to the current block.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5248
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5249<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5250<pre>
5251Loop: ; Infinite loop that counts from 0 on up...
5252 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
5253 %nextindvar = add i32 %indvar, 1
5254 br label %Loop
5255</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5256
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5257</div>
5258
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5259<!-- _______________________________________________________________________ -->
5260<div class="doc_subsubsection">
5261 <a name="i_select">'<tt>select</tt>' Instruction</a>
5262</div>
5263
5264<div class="doc_text">
5265
5266<h5>Syntax:</h5>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5267<pre>
Dan Gohmanc579d972008-09-09 01:02:47 +0000[diff] [blame]5268 &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>
5269
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]5270 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5271</pre>
5272
5273<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5274<p>The '<tt>select</tt>' instruction is used to choose one value based on a
5275 condition, without branching.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5276
5277
5278<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5279<p>The '<tt>select</tt>' instruction requires an 'i1' value or a vector of 'i1'
5280 values indicating the condition, and two values of the
5281 same <a href="#t_firstclass">first class</a> type. If the val1/val2 are
5282 vectors and the condition is a scalar, then entire vectors are selected, not
5283 individual elements.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5284
5285<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5286<p>If the condition is an i1 and it evaluates to 1, the instruction returns the
5287 first value argument; otherwise, it returns the second value argument.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5288
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5289<p>If the condition is a vector of i1, then the value arguments must be vectors
5290 of the same size, and the selection is done element by element.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5291
5292<h5>Example:</h5>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5293<pre>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]5294 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5295</pre>
Dan Gohmana5127ab2009-01-22 01:39:38 +0000[diff] [blame]5296
5297<p>Note that the code generator does not yet support conditions
5298 with vector type.</p>
5299
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5300</div>
5301
Robert Bocchinof72fdfe2006-01-15 20:48:27 +0000[diff] [blame]5302<!-- _______________________________________________________________________ -->
5303<div class="doc_subsubsection">
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5304 <a name="i_call">'<tt>call</tt>' Instruction</a>
5305</div>
5306
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5307<div class="doc_text">
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5308
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5309<h5>Syntax:</h5>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5310<pre>
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]5311 &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]5312</pre>
5313
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5314<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5315<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5316
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5317<h5>Arguments:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5318<p>This instruction requires several arguments:</p>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5319
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]5320<ol>
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]5321 <li>The optional "tail" marker indicates that the callee function does not
5322 access any allocas or varargs in the caller. Note that calls may be
5323 marked "tail" even if they do not occur before
5324 a <a href="#i_ret"><tt>ret</tt></a> instruction. If the "tail" marker is
5325 present, the function call is eligible for tail call optimization,
5326 but <a href="CodeGenerator.html#tailcallopt">might not in fact be
Evan Cheng59676492010-03-08 21:05:02 +0000[diff] [blame]5327 optimized into a jump</a>. The code generator may optimize calls marked
5328 "tail" with either 1) automatic <a href="CodeGenerator.html#sibcallopt">
5329 sibling call optimization</a> when the caller and callee have
5330 matching signatures, or 2) forced tail call optimization when the
5331 following extra requirements are met:
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]5332 <ul>
5333 <li>Caller and callee both have the calling
5334 convention <tt>fastcc</tt>.</li>
5335 <li>The call is in tail position (ret immediately follows call and ret
5336 uses value of call or is void).</li>
5337 <li>Option <tt>-tailcallopt</tt> is enabled,
Dan Gohman6232f732010-03-02 01:08:11 +0000[diff] [blame]5338 or <code>llvm::GuaranteedTailCallOpt</code> is <code>true</code>.</li>
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]5339 <li><a href="CodeGenerator.html#tailcallopt">Platform specific
5340 constraints are met.</a></li>
5341 </ul>
5342 </li>
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]5343
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5344 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
5345 convention</a> the call should use. If none is specified, the call
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]5346 defaults to using C calling conventions. The calling convention of the
5347 call must match the calling convention of the target function, or else the
5348 behavior is undefined.</li>
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]5349
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5350 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
5351 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
5352 '<tt>inreg</tt>' attributes are valid here.</li>
5353
5354 <li>'<tt>ty</tt>': the type of the call instruction itself which is also the
5355 type of the return value. Functions that return no value are marked
5356 <tt><a href="#t_void">void</a></tt>.</li>
5357
5358 <li>'<tt>fnty</tt>': shall be the signature of the pointer to function value
5359 being invoked. The argument types must match the types implied by this
5360 signature. This type can be omitted if the function is not varargs and if
5361 the function type does not return a pointer to a function.</li>
5362
5363 <li>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
5364 be invoked. In most cases, this is a direct function invocation, but
5365 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
5366 to function value.</li>
5367
5368 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]5369 signature argument types and parameter attributes. All arguments must be
5370 of <a href="#t_firstclass">first class</a> type. If the function
5371 signature indicates the function accepts a variable number of arguments,
5372 the extra arguments can be specified.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5373
5374 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
5375 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
5376 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]5377</ol>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5378
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5379<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5380<p>The '<tt>call</tt>' instruction is used to cause control flow to transfer to
5381 a specified function, with its incoming arguments bound to the specified
5382 values. Upon a '<tt><a href="#i_ret">ret</a></tt>' instruction in the called
5383 function, control flow continues with the instruction after the function
5384 call, and the return value of the function is bound to the result
5385 argument.</p>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5386
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5387<h5>Example:</h5>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5388<pre>
Nick Lewyckya9b13d52007-09-08 13:57:50 +0000[diff] [blame]5389 %retval = call i32 @test(i32 %argc)
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]5390 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) <i>; yields i32</i>
Chris Lattnerfb7c88d2008-03-21 17:24:17 +0000[diff] [blame]5391 %X = tail call i32 @foo() <i>; yields i32</i>
5392 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
5393 call void %foo(i8 97 signext)
Devang Pateld6cff512008-03-10 20:49:15 +0000[diff] [blame]5394
5395 %struct.A = type { i32, i8 }
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]5396 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohmancc3132e2008-10-04 19:00:07 +0000[diff] [blame]5397 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
5398 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattner6cbe8e92008-10-08 06:26:11 +0000[diff] [blame]5399 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijmaneefa7df2008-10-07 10:03:45 +0000[diff] [blame]5400 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5401</pre>
5402
Dale Johannesen68f971b2009-09-24 18:38:21 +0000[diff] [blame]5403<p>llvm treats calls to some functions with names and arguments that match the
Dale Johannesen722212d2009-09-25 17:04:42 +0000[diff] [blame]5404standard C99 library as being the C99 library functions, and may perform
5405optimizations or generate code for them under that assumption. This is
5406something we'd like to change in the future to provide better support for
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]5407freestanding environments and non-C-based languages.</p>
Dale Johannesen68f971b2009-09-24 18:38:21 +0000[diff] [blame]5408
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5409</div>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5410
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5411<!-- _______________________________________________________________________ -->
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5412<div class="doc_subsubsection">
Chris Lattner33337472006-01-13 23:26:01 +0000[diff] [blame]5413 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5414</div>
5415
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5416<div class="doc_text">
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5417
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]5418<h5>Syntax:</h5>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5419<pre>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]5420 &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]5421</pre>
5422
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]5423<h5>Overview:</h5>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]5424<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5425 the "variable argument" area of a function call. It is used to implement the
5426 <tt>va_arg</tt> macro in C.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5427
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]5428<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5429<p>This instruction takes a <tt>va_list*</tt> value and the type of the
5430 argument. It returns a value of the specified argument type and increments
5431 the <tt>va_list</tt> to point to the next argument. The actual type
5432 of <tt>va_list</tt> is target specific.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5433
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]5434<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5435<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified type
5436 from the specified <tt>va_list</tt> and causes the <tt>va_list</tt> to point
5437 to the next argument. For more information, see the variable argument
5438 handling <a href="#int_varargs">Intrinsic Functions</a>.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5439
5440<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]5441 take a variable number of arguments, for example, the <tt>vfprintf</tt>
5442 function.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5443
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5444<p><tt>va_arg</tt> is an LLVM instruction instead of
5445 an <a href="#intrinsics">intrinsic function</a> because it takes a type as an
5446 argument.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5447
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]5448<h5>Example:</h5>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5449<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
5450
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5451<p>Note that the code generator does not yet fully support va_arg on many
5452 targets. Also, it does not currently support va_arg with aggregate types on
5453 any target.</p>
Dan Gohman3065b612009-01-12 23:12:39 +0000[diff] [blame]5454
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5455</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5456
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5457<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]5458<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
5459<!-- *********************************************************************** -->
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5460
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5461<div class="doc_text">
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5462
5463<p>LLVM supports the notion of an "intrinsic function". These functions have
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5464 well known names and semantics and are required to follow certain
5465 restrictions. Overall, these intrinsics represent an extension mechanism for
5466 the LLVM language that does not require changing all of the transformations
5467 in LLVM when adding to the language (or the bitcode reader/writer, the
5468 parser, etc...).</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5469
John Criswell88190562005-05-16 16:17:45 +0000[diff] [blame]5470<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5471 prefix is reserved in LLVM for intrinsic names; thus, function names may not
5472 begin with this prefix. Intrinsic functions must always be external
5473 functions: you cannot define the body of intrinsic functions. Intrinsic
5474 functions may only be used in call or invoke instructions: it is illegal to
5475 take the address of an intrinsic function. Additionally, because intrinsic
5476 functions are part of the LLVM language, it is required if any are added that
5477 they be documented here.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5478
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5479<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents a
5480 family of functions that perform the same operation but on different data
5481 types. Because LLVM can represent over 8 million different integer types,
5482 overloading is used commonly to allow an intrinsic function to operate on any
5483 integer type. One or more of the argument types or the result type can be
5484 overloaded to accept any integer type. Argument types may also be defined as
5485 exactly matching a previous argument's type or the result type. This allows
5486 an intrinsic function which accepts multiple arguments, but needs all of them
5487 to be of the same type, to only be overloaded with respect to a single
5488 argument or the result.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5489
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5490<p>Overloaded intrinsics will have the names of its overloaded argument types
5491 encoded into its function name, each preceded by a period. Only those types
5492 which are overloaded result in a name suffix. Arguments whose type is matched
5493 against another type do not. For example, the <tt>llvm.ctpop</tt> function
5494 can take an integer of any width and returns an integer of exactly the same
5495 integer width. This leads to a family of functions such as
5496 <tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29
5497 %val)</tt>. Only one type, the return type, is overloaded, and only one type
5498 suffix is required. Because the argument's type is matched against the return
5499 type, it does not require its own name suffix.</p>
Reid Spencer4eefaab2007-04-01 08:04:23 +0000[diff] [blame]5500
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5501<p>To learn how to add an intrinsic function, please see the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5502 <a href="ExtendingLLVM.html">Extending LLVM Guide</a>.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5503
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5504</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5505
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5506<!-- ======================================================================= -->
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5507<div class="doc_subsection">
5508 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
5509</div>
5510
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5511<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5512
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5513<p>Variable argument support is defined in LLVM with
5514 the <a href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
5515 intrinsic functions. These functions are related to the similarly named
5516 macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5517
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5518<p>All of these functions operate on arguments that use a target-specific value
5519 type "<tt>va_list</tt>". The LLVM assembly language reference manual does
5520 not define what this type is, so all transformations should be prepared to
5521 handle these functions regardless of the type used.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5522
Chris Lattner30b868d2006-05-15 17:26:46 +0000[diff] [blame]5523<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]5524 instruction and the variable argument handling intrinsic functions are
5525 used.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5526
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]5527<pre class="doc_code">
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]5528define i32 @test(i32 %X, ...) {
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5529 ; Initialize variable argument processing
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5530 %ap = alloca i8*
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]5531 %ap2 = bitcast i8** %ap to i8*
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]5532 call void @llvm.va_start(i8* %ap2)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5533
5534 ; Read a single integer argument
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5535 %tmp = va_arg i8** %ap, i32
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5536
5537 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5538 %aq = alloca i8*
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]5539 %aq2 = bitcast i8** %aq to i8*
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5540 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]5541 call void @llvm.va_end(i8* %aq2)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5542
5543 ; Stop processing of arguments.
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]5544 call void @llvm.va_end(i8* %ap2)
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5545 ret i32 %tmp
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5546}
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]5547
5548declare void @llvm.va_start(i8*)
5549declare void @llvm.va_copy(i8*, i8*)
5550declare void @llvm.va_end(i8*)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5551</pre>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5552
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]5553</div>
5554
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5555<!-- _______________________________________________________________________ -->
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5556<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5557 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5558</div>
5559
5560
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5561<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5562
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5563<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5564<pre>
5565 declare void %llvm.va_start(i8* &lt;arglist&gt;)
5566</pre>
5567
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5568<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5569<p>The '<tt>llvm.va_start</tt>' intrinsic initializes <tt>*&lt;arglist&gt;</tt>
5570 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]5571
5572<h5>Arguments:</h5>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]5573<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]5574
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5575<h5>Semantics:</h5>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]5576<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]5577 macro available in C. In a target-dependent way, it initializes
5578 the <tt>va_list</tt> element to which the argument points, so that the next
5579 call to <tt>va_arg</tt> will produce the first variable argument passed to
5580 the function. Unlike the C <tt>va_start</tt> macro, this intrinsic does not
5581 need to know the last argument of the function as the compiler can figure
5582 that out.</p>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]5583
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5584</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5585
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5586<!-- _______________________________________________________________________ -->
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5587<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5588 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5589</div>
5590
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5591<div class="doc_text">
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]5592
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5593<h5>Syntax:</h5>
5594<pre>
5595 declare void @llvm.va_end(i8* &lt;arglist&gt;)
5596</pre>
5597
5598<h5>Overview:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5599<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5600 which has been initialized previously
5601 with <tt><a href="#int_va_start">llvm.va_start</a></tt>
5602 or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]5603
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5604<h5>Arguments:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5605<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]5606
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5607<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5608<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]5609 macro available in C. In a target-dependent way, it destroys
5610 the <tt>va_list</tt> element to which the argument points. Calls
5611 to <a href="#int_va_start"><tt>llvm.va_start</tt></a>
5612 and <a href="#int_va_copy"> <tt>llvm.va_copy</tt></a> must be matched exactly
5613 with calls to <tt>llvm.va_end</tt>.</p>
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]5614
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5615</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5616
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5617<!-- _______________________________________________________________________ -->
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5618<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5619 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5620</div>
5621
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5622<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5623
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5624<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5625<pre>
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]5626 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5627</pre>
5628
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5629<h5>Overview:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5630<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5631 from the source argument list to the destination argument list.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5632
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5633<h5>Arguments:</h5>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]5634<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]5635 The second argument is a pointer to a <tt>va_list</tt> element to copy
5636 from.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5637
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5638<h5>Semantics:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5639<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]5640 macro available in C. In a target-dependent way, it copies the
5641 source <tt>va_list</tt> element into the destination <tt>va_list</tt>
5642 element. This intrinsic is necessary because
5643 the <tt><a href="#int_va_start"> llvm.va_start</a></tt> intrinsic may be
5644 arbitrarily complex and require, for example, memory allocation.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5645
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5646</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5647
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5648<!-- ======================================================================= -->
5649<div class="doc_subsection">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5650 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
5651</div>
5652
5653<div class="doc_text">
5654
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5655<p>LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]5656Collection</a> (GC) requires the implementation and generation of these
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5657intrinsics. These intrinsics allow identification of <a href="#int_gcroot">GC
5658roots on the stack</a>, as well as garbage collector implementations that
5659require <a href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a>
5660barriers. Front-ends for type-safe garbage collected languages should generate
5661these intrinsics to make use of the LLVM garbage collectors. For more details,
5662see <a href="GarbageCollection.html">Accurate Garbage Collection with
5663LLVM</a>.</p>
Christopher Lamb55c6d4f2007-12-17 01:00:21 +0000[diff] [blame]5664
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5665<p>The garbage collection intrinsics only operate on objects in the generic
5666 address space (address space zero).</p>
Christopher Lamb55c6d4f2007-12-17 01:00:21 +0000[diff] [blame]5667
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5668</div>
5669
5670<!-- _______________________________________________________________________ -->
5671<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5672 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5673</div>
5674
5675<div class="doc_text">
5676
5677<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5678<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]5679 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5680</pre>
5681
5682<h5>Overview:</h5>
John Criswelldfe6a862004-12-10 15:51:16 +0000[diff] [blame]5683<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]5684 the code generator, and allows some metadata to be associated with it.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5685
5686<h5>Arguments:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5687<p>The first argument specifies the address of a stack object that contains the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5688 root pointer. The second pointer (which must be either a constant or a
5689 global value address) contains the meta-data to be associated with the
5690 root.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5691
5692<h5>Semantics:</h5>
Chris Lattner851b7712008-04-24 05:59:56 +0000[diff] [blame]5693<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]5694 location. At compile-time, the code generator generates information to allow
5695 the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
5696 intrinsic may only be used in a function which <a href="#gc">specifies a GC
5697 algorithm</a>.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5698
5699</div>
5700
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5701<!-- _______________________________________________________________________ -->
5702<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5703 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5704</div>
5705
5706<div class="doc_text">
5707
5708<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5709<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]5710 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5711</pre>
5712
5713<h5>Overview:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5714<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5715 locations, allowing garbage collector implementations that require read
5716 barriers.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5717
5718<h5>Arguments:</h5>
Chris Lattnerf9228072006-03-14 20:02:51 +0000[diff] [blame]5719<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]5720 allocated from the garbage collector. The first object is a pointer to the
5721 start of the referenced object, if needed by the language runtime (otherwise
5722 null).</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5723
5724<h5>Semantics:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5725<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5726 instruction, but may be replaced with substantially more complex code by the
5727 garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
5728 may only be used in a function which <a href="#gc">specifies a GC
5729 algorithm</a>.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5730
5731</div>
5732
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5733<!-- _______________________________________________________________________ -->
5734<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5735 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5736</div>
5737
5738<div class="doc_text">
5739
5740<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5741<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]5742 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5743</pre>
5744
5745<h5>Overview:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5746<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5747 locations, allowing garbage collector implementations that require write
5748 barriers (such as generational or reference counting collectors).</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5749
5750<h5>Arguments:</h5>
Chris Lattnerf9228072006-03-14 20:02:51 +0000[diff] [blame]5751<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]5752 object to store it to, and the third is the address of the field of Obj to
5753 store to. If the runtime does not require a pointer to the object, Obj may
5754 be null.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5755
5756<h5>Semantics:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5757<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5758 instruction, but may be replaced with substantially more complex code by the
5759 garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
5760 may only be used in a function which <a href="#gc">specifies a GC
5761 algorithm</a>.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5762
5763</div>
5764
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5765<!-- ======================================================================= -->
5766<div class="doc_subsection">
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5767 <a name="int_codegen">Code Generator Intrinsics</a>
5768</div>
5769
5770<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5771
5772<p>These intrinsics are provided by LLVM to expose special features that may
5773 only be implemented with code generator support.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5774
5775</div>
5776
5777<!-- _______________________________________________________________________ -->
5778<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5779 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5780</div>
5781
5782<div class="doc_text">
5783
5784<h5>Syntax:</h5>
5785<pre>
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]5786 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5787</pre>
5788
5789<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5790<p>The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
5791 target-specific value indicating the return address of the current function
5792 or one of its callers.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5793
5794<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5795<p>The argument to this intrinsic indicates which function to return the address
5796 for. Zero indicates the calling function, one indicates its caller, etc.
5797 The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5798
5799<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5800<p>The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer
5801 indicating the return address of the specified call frame, or zero if it
5802 cannot be identified. The value returned by this intrinsic is likely to be
5803 incorrect or 0 for arguments other than zero, so it should only be used for
5804 debugging purposes.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5805
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5806<p>Note that calling this intrinsic does not prevent function inlining or other
5807 aggressive transformations, so the value returned may not be that of the
5808 obvious source-language caller.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5809
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5810</div>
5811
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5812<!-- _______________________________________________________________________ -->
5813<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5814 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5815</div>
5816
5817<div class="doc_text">
5818
5819<h5>Syntax:</h5>
5820<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]5821 declare i8* @llvm.frameaddress(i32 &lt;level&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5822</pre>
5823
5824<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5825<p>The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
5826 target-specific frame pointer value for the specified stack frame.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5827
5828<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5829<p>The argument to this intrinsic indicates which function to return the frame
5830 pointer for. Zero indicates the calling function, one indicates its caller,
5831 etc. The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5832
5833<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5834<p>The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer
5835 indicating the frame address of the specified call frame, or zero if it
5836 cannot be identified. The value returned by this intrinsic is likely to be
5837 incorrect or 0 for arguments other than zero, so it should only be used for
5838 debugging purposes.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5839
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5840<p>Note that calling this intrinsic does not prevent function inlining or other
5841 aggressive transformations, so the value returned may not be that of the
5842 obvious source-language caller.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5843
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5844</div>
5845
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]5846<!-- _______________________________________________________________________ -->
5847<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5848 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]5849</div>
5850
5851<div class="doc_text">
5852
5853<h5>Syntax:</h5>
5854<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]5855 declare i8* @llvm.stacksave()
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]5856</pre>
5857
5858<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5859<p>The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state
5860 of the function stack, for use
5861 with <a href="#int_stackrestore"> <tt>llvm.stackrestore</tt></a>. This is
5862 useful for implementing language features like scoped automatic variable
5863 sized arrays in C99.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]5864
5865<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5866<p>This intrinsic returns a opaque pointer value that can be passed
5867 to <a href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When
5868 an <tt>llvm.stackrestore</tt> intrinsic is executed with a value saved
5869 from <tt>llvm.stacksave</tt>, it effectively restores the state of the stack
5870 to the state it was in when the <tt>llvm.stacksave</tt> intrinsic executed.
5871 In practice, this pops any <a href="#i_alloca">alloca</a> blocks from the
5872 stack that were allocated after the <tt>llvm.stacksave</tt> was executed.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]5873
5874</div>
5875
5876<!-- _______________________________________________________________________ -->
5877<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5878 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]5879</div>
5880
5881<div class="doc_text">
5882
5883<h5>Syntax:</h5>
5884<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]5885 declare void @llvm.stackrestore(i8* %ptr)
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]5886</pre>
5887
5888<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5889<p>The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
5890 the function stack to the state it was in when the
5891 corresponding <a href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic
5892 executed. This is useful for implementing language features like scoped
5893 automatic variable sized arrays in C99.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]5894
5895<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5896<p>See the description
5897 for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]5898
5899</div>
5900
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]5901<!-- _______________________________________________________________________ -->
5902<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5903 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]5904</div>
5905
5906<div class="doc_text">
5907
5908<h5>Syntax:</h5>
5909<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]5910 declare void @llvm.prefetch(i8* &lt;address&gt;, i32 &lt;rw&gt;, i32 &lt;locality&gt;)
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]5911</pre>
5912
5913<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5914<p>The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to
5915 insert a prefetch instruction if supported; otherwise, it is a noop.
5916 Prefetches have no effect on the behavior of the program but can change its
5917 performance characteristics.</p>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]5918
5919<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5920<p><tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the
5921 specifier determining if the fetch should be for a read (0) or write (1),
5922 and <tt>locality</tt> is a temporal locality specifier ranging from (0) - no
5923 locality, to (3) - extremely local keep in cache. The <tt>rw</tt>
5924 and <tt>locality</tt> arguments must be constant integers.</p>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]5925
5926<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5927<p>This intrinsic does not modify the behavior of the program. In particular,
5928 prefetches cannot trap and do not produce a value. On targets that support
5929 this intrinsic, the prefetch can provide hints to the processor cache for
5930 better performance.</p>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]5931
5932</div>
5933
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]5934<!-- _______________________________________________________________________ -->
5935<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5936 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]5937</div>
5938
5939<div class="doc_text">
5940
5941<h5>Syntax:</h5>
5942<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]5943 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]5944</pre>
5945
5946<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5947<p>The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program
5948 Counter (PC) in a region of code to simulators and other tools. The method
5949 is target specific, but it is expected that the marker will use exported
5950 symbols to transmit the PC of the marker. The marker makes no guarantees
5951 that it will remain with any specific instruction after optimizations. It is
5952 possible that the presence of a marker will inhibit optimizations. The
5953 intended use is to be inserted after optimizations to allow correlations of
5954 simulation runs.</p>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]5955
5956<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5957<p><tt>id</tt> is a numerical id identifying the marker.</p>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]5958
5959<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5960<p>This intrinsic does not modify the behavior of the program. Backends that do
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]5961 not support this intrinsic may ignore it.</p>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]5962
5963</div>
5964
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]5965<!-- _______________________________________________________________________ -->
5966<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5967 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]5968</div>
5969
5970<div class="doc_text">
5971
5972<h5>Syntax:</h5>
5973<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]5974 declare i64 @llvm.readcyclecounter()
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]5975</pre>
5976
5977<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5978<p>The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
5979 counter register (or similar low latency, high accuracy clocks) on those
5980 targets that support it. On X86, it should map to RDTSC. On Alpha, it
5981 should map to RPCC. As the backing counters overflow quickly (on the order
5982 of 9 seconds on alpha), this should only be used for small timings.</p>
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]5983
5984<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5985<p>When directly supported, reading the cycle counter should not modify any
5986 memory. Implementations are allowed to either return a application specific
5987 value or a system wide value. On backends without support, this is lowered
5988 to a constant 0.</p>
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]5989
5990</div>
5991
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5992<!-- ======================================================================= -->
5993<div class="doc_subsection">
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5994 <a name="int_libc">Standard C Library Intrinsics</a>
5995</div>
5996
5997<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5998
5999<p>LLVM provides intrinsics for a few important standard C library functions.
6000 These intrinsics allow source-language front-ends to pass information about
6001 the alignment of the pointer arguments to the code generator, providing
6002 opportunity for more efficient code generation.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6003
6004</div>
6005
6006<!-- _______________________________________________________________________ -->
6007<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6008 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6009</div>
6010
6011<div class="doc_text">
6012
6013<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6014<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]6015 integer bit width and for different address spaces. Not all targets support
6016 all bit widths however.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6017
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6018<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6019 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]6020 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6021 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]6022 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6023</pre>
6024
6025<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6026<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6027 source location to the destination location.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6028
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6029<p>Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6030 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6031 and the pointers can be in specified address spaces.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6032
6033<h5>Arguments:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6034
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6035<p>The first argument is a pointer to the destination, the second is a pointer
6036 to the source. The third argument is an integer argument specifying the
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6037 number of bytes to copy, the fourth argument is the alignment of the
6038 source and destination locations, and the fifth is a boolean indicating a
6039 volatile access.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6040
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]6041<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]6042 then the caller guarantees that both the source and destination pointers are
6043 aligned to that boundary.</p>
Chris Lattner4c67c482004-02-12 21:18:15 +0000[diff] [blame]6044
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]6045<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6046 <tt>llvm.memcpy</tt> call is a <a href="#volatile">volatile operation</a>.
6047 The detailed access behavior is not very cleanly specified and it is unwise
6048 to depend on it.</p>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6049
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6050<h5>Semantics:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6051
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6052<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6053 source location to the destination location, which are not allowed to
6054 overlap. It copies "len" bytes of memory over. If the argument is known to
6055 be aligned to some boundary, this can be specified as the fourth argument,
6056 otherwise it should be set to 0 or 1.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6057
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6058</div>
6059
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6060<!-- _______________________________________________________________________ -->
6061<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6062 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6063</div>
6064
6065<div class="doc_text">
6066
6067<h5>Syntax:</h5>
Chris Lattnerdd708342008-11-21 16:42:48 +0000[diff] [blame]6068<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]6069 width and for different address space. Not all targets support all bit
6070 widths however.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6071
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6072<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6073 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]6074 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6075 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]6076 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6077</pre>
6078
6079<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6080<p>The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the
6081 source location to the destination location. It is similar to the
6082 '<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to
6083 overlap.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6084
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6085<p>Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6086 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6087 and the pointers can be in specified address spaces.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6088
6089<h5>Arguments:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6090
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6091<p>The first argument is a pointer to the destination, the second is a pointer
6092 to the source. The third argument is an integer argument specifying the
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6093 number of bytes to copy, the fourth argument is the alignment of the
6094 source and destination locations, and the fifth is a boolean indicating a
6095 volatile access.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6096
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]6097<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]6098 then the caller guarantees that the source and destination pointers are
6099 aligned to that boundary.</p>
Chris Lattner4c67c482004-02-12 21:18:15 +0000[diff] [blame]6100
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]6101<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6102 <tt>llvm.memmove</tt> call is a <a href="#volatile">volatile operation</a>.
6103 The detailed access behavior is not very cleanly specified and it is unwise
6104 to depend on it.</p>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6105
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6106<h5>Semantics:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6107
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6108<p>The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the
6109 source location to the destination location, which may overlap. It copies
6110 "len" bytes of memory over. If the argument is known to be aligned to some
6111 boundary, this can be specified as the fourth argument, otherwise it should
6112 be set to 0 or 1.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6113
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6114</div>
6115
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6116<!-- _______________________________________________________________________ -->
6117<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6118 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6119</div>
6120
6121<div class="doc_text">
6122
6123<h5>Syntax:</h5>
Chris Lattnerdd708342008-11-21 16:42:48 +0000[diff] [blame]6124<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]6125 width and for different address spaces. However, not all targets support all
6126 bit widths.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6127
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6128<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6129 declare void @llvm.memset.p0i8.i32(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattner685db9d2010-04-08 00:54:34 +0000[diff] [blame]6130 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6131 declare void @llvm.memset.p0i8.i64(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattner685db9d2010-04-08 00:54:34 +0000[diff] [blame]6132 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +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.memset.*</tt>' intrinsics fill a block of memory with a
6137 particular byte value.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6138
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6139<p>Note that, unlike the standard libc function, the <tt>llvm.memset</tt>
John Criswellad05ae42010-07-30 16:30:28 +0000[diff] [blame]6140 intrinsic does not return a value and takes extra alignment/volatile
6141 arguments. Also, the destination can be in an arbitrary address space.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6142
6143<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6144<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]6145 byte value with which to fill it, the third argument is an integer argument
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6146 specifying the number of bytes to fill, and the fourth argument is the known
John Criswellad05ae42010-07-30 16:30:28 +0000[diff] [blame]6147 alignment of the destination location.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6148
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]6149<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]6150 then the caller guarantees that the destination pointer is aligned to that
6151 boundary.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6152
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]6153<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6154 <tt>llvm.memset</tt> call is a <a href="#volatile">volatile operation</a>.
6155 The detailed access behavior is not very cleanly specified and it is unwise
6156 to depend on it.</p>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6157
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6158<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6159<p>The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting
6160 at the destination location. If the argument is known to be aligned to some
6161 boundary, this can be specified as the fourth argument, otherwise it should
6162 be set to 0 or 1.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6163
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6164</div>
6165
Chris Lattner3b4f4372004-06-11 02:28:03 +0000[diff] [blame]6166<!-- _______________________________________________________________________ -->
6167<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6168 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6169</div>
6170
6171<div class="doc_text">
6172
6173<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6174<p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
6175 floating point or vector of floating point type. Not all targets support all
6176 types however.</p>
6177
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6178<pre>
Dale Johannesendd89d272007-10-02 17:47:38 +0000[diff] [blame]6179 declare float @llvm.sqrt.f32(float %Val)
6180 declare double @llvm.sqrt.f64(double %Val)
6181 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
6182 declare fp128 @llvm.sqrt.f128(fp128 %Val)
6183 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6184</pre>
6185
6186<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6187<p>The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
6188 returning the same value as the libm '<tt>sqrt</tt>' functions would.
6189 Unlike <tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined
6190 behavior for negative numbers other than -0.0 (which allows for better
6191 optimization, because there is no need to worry about errno being
6192 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]6193
6194<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6195<p>The argument and return value are floating point numbers of the same
6196 type.</p>
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6197
6198<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6199<p>This function returns the sqrt of the specified operand if it is a
6200 nonnegative floating point number.</p>
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6201
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6202</div>
6203
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6204<!-- _______________________________________________________________________ -->
6205<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6206 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6207</div>
6208
6209<div class="doc_text">
6210
6211<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6212<p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
6213 floating point or vector of floating point type. Not all targets support all
6214 types however.</p>
6215
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6216<pre>
Dale Johannesendd89d272007-10-02 17:47:38 +0000[diff] [blame]6217 declare float @llvm.powi.f32(float %Val, i32 %power)
6218 declare double @llvm.powi.f64(double %Val, i32 %power)
6219 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
6220 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
6221 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6222</pre>
6223
6224<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6225<p>The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
6226 specified (positive or negative) power. The order of evaluation of
6227 multiplications is not defined. When a vector of floating point type is
6228 used, the second argument remains a scalar integer value.</p>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6229
6230<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6231<p>The second argument is an integer power, and the first is a value to raise to
6232 that power.</p>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6233
6234<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6235<p>This function returns the first value raised to the second power with an
6236 unspecified sequence of rounding operations.</p>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6237
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6238</div>
6239
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6240<!-- _______________________________________________________________________ -->
6241<div class="doc_subsubsection">
6242 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
6243</div>
6244
6245<div class="doc_text">
6246
6247<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6248<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
6249 floating point or vector of floating point type. Not all targets support all
6250 types however.</p>
6251
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6252<pre>
6253 declare float @llvm.sin.f32(float %Val)
6254 declare double @llvm.sin.f64(double %Val)
6255 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
6256 declare fp128 @llvm.sin.f128(fp128 %Val)
6257 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
6258</pre>
6259
6260<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6261<p>The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6262
6263<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6264<p>The argument and return value are floating point numbers of the same
6265 type.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6266
6267<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6268<p>This function returns the sine of the specified operand, returning the same
6269 values as the libm <tt>sin</tt> functions would, and handles error conditions
6270 in the same way.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6271
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6272</div>
6273
6274<!-- _______________________________________________________________________ -->
6275<div class="doc_subsubsection">
6276 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
6277</div>
6278
6279<div class="doc_text">
6280
6281<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6282<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
6283 floating point or vector of floating point type. Not all targets support all
6284 types however.</p>
6285
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6286<pre>
6287 declare float @llvm.cos.f32(float %Val)
6288 declare double @llvm.cos.f64(double %Val)
6289 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
6290 declare fp128 @llvm.cos.f128(fp128 %Val)
6291 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
6292</pre>
6293
6294<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6295<p>The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6296
6297<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6298<p>The argument and return value are floating point numbers of the same
6299 type.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6300
6301<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6302<p>This function returns the cosine of the specified operand, returning the same
6303 values as the libm <tt>cos</tt> functions would, and handles error conditions
6304 in the same way.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6305
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6306</div>
6307
6308<!-- _______________________________________________________________________ -->
6309<div class="doc_subsubsection">
6310 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
6311</div>
6312
6313<div class="doc_text">
6314
6315<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6316<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
6317 floating point or vector of floating point type. Not all targets support all
6318 types however.</p>
6319
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6320<pre>
6321 declare float @llvm.pow.f32(float %Val, float %Power)
6322 declare double @llvm.pow.f64(double %Val, double %Power)
6323 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
6324 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
6325 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
6326</pre>
6327
6328<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6329<p>The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
6330 specified (positive or negative) power.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6331
6332<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6333<p>The second argument is a floating point power, and the first is a value to
6334 raise to that power.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6335
6336<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6337<p>This function returns the first value raised to the second power, returning
6338 the same values as the libm <tt>pow</tt> functions would, and handles error
6339 conditions in the same way.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6340
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6341</div>
6342
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6343<!-- ======================================================================= -->
6344<div class="doc_subsection">
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6345 <a name="int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6346</div>
6347
6348<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6349
6350<p>LLVM provides intrinsics for a few important bit manipulation operations.
6351 These allow efficient code generation for some algorithms.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6352
6353</div>
6354
6355<!-- _______________________________________________________________________ -->
6356<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6357 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6358</div>
6359
6360<div class="doc_text">
6361
6362<h5>Syntax:</h5>
Reid Spencer4eefaab2007-04-01 08:04:23 +0000[diff] [blame]6363<p>This is an overloaded intrinsic function. You can use bswap on any integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6364 type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
6365
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6366<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6367 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
6368 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
6369 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6370</pre>
6371
6372<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6373<p>The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
6374 values with an even number of bytes (positive multiple of 16 bits). These
6375 are useful for performing operations on data that is not in the target's
6376 native byte order.</p>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6377
6378<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6379<p>The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
6380 and low byte of the input i16 swapped. Similarly,
6381 the <tt>llvm.bswap.i32</tt> intrinsic returns an i32 value that has the four
6382 bytes of the input i32 swapped, so that if the input bytes are numbered 0, 1,
6383 2, 3 then the returned i32 will have its bytes in 3, 2, 1, 0 order.
6384 The <tt>llvm.bswap.i48</tt>, <tt>llvm.bswap.i64</tt> and other intrinsics
6385 extend this concept to additional even-byte lengths (6 bytes, 8 bytes and
6386 more, respectively).</p>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6387
6388</div>
6389
6390<!-- _______________________________________________________________________ -->
6391<div class="doc_subsubsection">
Reid Spencerb4f9a6f2006-01-16 21:12:35 +0000[diff] [blame]6392 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6393</div>
6394
6395<div class="doc_text">
6396
6397<h5>Syntax:</h5>
Reid Spencer4eefaab2007-04-01 08:04:23 +0000[diff] [blame]6398<p>This is an overloaded intrinsic. You can use llvm.ctpop on any integer bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6399 width. Not all targets support all bit widths however.</p>
6400
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6401<pre>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6402 declare i8 @llvm.ctpop.i8(i8 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6403 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]6404 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6405 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
6406 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6407</pre>
6408
6409<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6410<p>The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set
6411 in a value.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6412
6413<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6414<p>The only argument is the value to be counted. The argument may be of any
6415 integer type. The return type must match the argument type.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6416
6417<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6418<p>The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6419
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6420</div>
6421
6422<!-- _______________________________________________________________________ -->
6423<div class="doc_subsubsection">
Chris Lattnerb748c672006-01-16 22:34:14 +0000[diff] [blame]6424 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6425</div>
6426
6427<div class="doc_text">
6428
6429<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6430<p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
6431 integer bit width. Not all targets support all bit widths however.</p>
6432
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6433<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6434 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
6435 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]6436 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6437 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
6438 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6439</pre>
6440
6441<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6442<p>The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
6443 leading zeros in a variable.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6444
6445<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6446<p>The only argument is the value to be counted. The argument may be of any
6447 integer type. The return type must match the argument type.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6448
6449<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6450<p>The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant)
6451 zeros in a variable. If the src == 0 then the result is the size in bits of
6452 the type of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6453
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6454</div>
Chris Lattner3b4f4372004-06-11 02:28:03 +0000[diff] [blame]6455
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6456<!-- _______________________________________________________________________ -->
6457<div class="doc_subsubsection">
Chris Lattnerb748c672006-01-16 22:34:14 +0000[diff] [blame]6458 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6459</div>
6460
6461<div class="doc_text">
6462
6463<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6464<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
6465 integer bit width. Not all targets support all bit widths however.</p>
6466
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6467<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6468 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
6469 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]6470 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6471 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
6472 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6473</pre>
6474
6475<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6476<p>The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
6477 trailing zeros.</p>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6478
6479<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6480<p>The only argument is the value to be counted. The argument may be of any
6481 integer type. The return type must match the argument type.</p>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6482
6483<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6484<p>The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant)
6485 zeros in a variable. If the src == 0 then the result is the size in bits of
6486 the type of src. For example, <tt>llvm.cttz(2) = 1</tt>.</p>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6487
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6488</div>
6489
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]6490<!-- ======================================================================= -->
6491<div class="doc_subsection">
6492 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
6493</div>
6494
6495<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6496
6497<p>LLVM provides intrinsics for some arithmetic with overflow operations.</p>
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]6498
6499</div>
6500
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6501<!-- _______________________________________________________________________ -->
6502<div class="doc_subsubsection">
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]6503 <a name="int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics</a>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6504</div>
6505
6506<div class="doc_text">
6507
6508<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6509<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]6510 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6511
6512<pre>
6513 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
6514 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6515 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
6516</pre>
6517
6518<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6519<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6520 a signed addition of the two arguments, and indicate whether an overflow
6521 occurred during the signed summation.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6522
6523<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6524<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]6525 be of integer types of any bit width, but they must have the same bit
6526 width. The second element of the result structure must be of
6527 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6528 undergo signed addition.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6529
6530<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6531<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6532 a signed addition of the two variables. They return a structure &mdash; the
6533 first element of which is the signed summation, and the second element of
6534 which is a bit specifying if the signed summation resulted in an
6535 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6536
6537<h5>Examples:</h5>
6538<pre>
6539 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6540 %sum = extractvalue {i32, i1} %res, 0
6541 %obit = extractvalue {i32, i1} %res, 1
6542 br i1 %obit, label %overflow, label %normal
6543</pre>
6544
6545</div>
6546
6547<!-- _______________________________________________________________________ -->
6548<div class="doc_subsubsection">
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]6549 <a name="int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics</a>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6550</div>
6551
6552<div class="doc_text">
6553
6554<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6555<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]6556 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6557
6558<pre>
6559 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
6560 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6561 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
6562</pre>
6563
6564<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6565<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6566 an unsigned addition of the two arguments, and indicate whether a carry
6567 occurred during the unsigned summation.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6568
6569<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6570<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]6571 be of integer types of any bit width, but they must have the same bit
6572 width. The second element of the result structure must be of
6573 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6574 undergo unsigned addition.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6575
6576<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6577<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6578 an unsigned addition of the two arguments. They return a structure &mdash;
6579 the first element of which is the sum, and the second element of which is a
6580 bit specifying if the unsigned summation resulted in a carry.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6581
6582<h5>Examples:</h5>
6583<pre>
6584 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6585 %sum = extractvalue {i32, i1} %res, 0
6586 %obit = extractvalue {i32, i1} %res, 1
6587 br i1 %obit, label %carry, label %normal
6588</pre>
6589
6590</div>
6591
6592<!-- _______________________________________________________________________ -->
6593<div class="doc_subsubsection">
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]6594 <a name="int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics</a>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6595</div>
6596
6597<div class="doc_text">
6598
6599<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6600<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]6601 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6602
6603<pre>
6604 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
6605 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6606 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
6607</pre>
6608
6609<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6610<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6611 a signed subtraction of the two arguments, and indicate whether an overflow
6612 occurred during the signed subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6613
6614<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6615<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]6616 be of integer types of any bit width, but they must have the same bit
6617 width. The second element of the result structure must be of
6618 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6619 undergo signed subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6620
6621<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6622<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6623 a signed subtraction of the two arguments. They return a structure &mdash;
6624 the first element of which is the subtraction, and the second element of
6625 which is a bit specifying if the signed subtraction resulted in an
6626 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6627
6628<h5>Examples:</h5>
6629<pre>
6630 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6631 %sum = extractvalue {i32, i1} %res, 0
6632 %obit = extractvalue {i32, i1} %res, 1
6633 br i1 %obit, label %overflow, label %normal
6634</pre>
6635
6636</div>
6637
6638<!-- _______________________________________________________________________ -->
6639<div class="doc_subsubsection">
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]6640 <a name="int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt>' Intrinsics</a>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6641</div>
6642
6643<div class="doc_text">
6644
6645<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6646<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]6647 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6648
6649<pre>
6650 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
6651 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6652 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
6653</pre>
6654
6655<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6656<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6657 an unsigned subtraction of the two arguments, and indicate whether an
6658 overflow occurred during the unsigned subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6659
6660<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6661<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]6662 be of integer types of any bit width, but they must have the same bit
6663 width. The second element of the result structure must be of
6664 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6665 undergo unsigned subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6666
6667<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6668<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6669 an unsigned subtraction of the two arguments. They return a structure &mdash;
6670 the first element of which is the subtraction, and the second element of
6671 which is a bit specifying if the unsigned subtraction resulted in an
6672 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6673
6674<h5>Examples:</h5>
6675<pre>
6676 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6677 %sum = extractvalue {i32, i1} %res, 0
6678 %obit = extractvalue {i32, i1} %res, 1
6679 br i1 %obit, label %overflow, label %normal
6680</pre>
6681
6682</div>
6683
6684<!-- _______________________________________________________________________ -->
6685<div class="doc_subsubsection">
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]6686 <a name="int_smul_overflow">'<tt>llvm.smul.with.overflow.*</tt>' Intrinsics</a>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6687</div>
6688
6689<div class="doc_text">
6690
6691<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6692<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]6693 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6694
6695<pre>
6696 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
6697 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6698 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
6699</pre>
6700
6701<h5>Overview:</h5>
6702
6703<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6704 a signed multiplication of the two arguments, and indicate whether an
6705 overflow occurred during the signed multiplication.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6706
6707<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6708<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]6709 be of integer types of any bit width, but they must have the same bit
6710 width. The second element of the result structure must be of
6711 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6712 undergo signed multiplication.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6713
6714<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6715<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6716 a signed multiplication of the two arguments. They return a structure &mdash;
6717 the first element of which is the multiplication, and the second element of
6718 which is a bit specifying if the signed multiplication resulted in an
6719 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6720
6721<h5>Examples:</h5>
6722<pre>
6723 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6724 %sum = extractvalue {i32, i1} %res, 0
6725 %obit = extractvalue {i32, i1} %res, 1
6726 br i1 %obit, label %overflow, label %normal
6727</pre>
6728
Reid Spencer5bf54c82007-04-11 23:23:49 +0000[diff] [blame]6729</div>
6730
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6731<!-- _______________________________________________________________________ -->
6732<div class="doc_subsubsection">
6733 <a name="int_umul_overflow">'<tt>llvm.umul.with.overflow.*</tt>' Intrinsics</a>
6734</div>
6735
6736<div class="doc_text">
6737
6738<h5>Syntax:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6739<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]6740 on any integer bit width.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6741
6742<pre>
6743 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
6744 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6745 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
6746</pre>
6747
6748<h5>Overview:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6749<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6750 a unsigned multiplication of the two arguments, and indicate whether an
6751 overflow occurred during the unsigned multiplication.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6752
6753<h5>Arguments:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6754<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]6755 be of integer types of any bit width, but they must have the same bit
6756 width. The second element of the result structure must be of
6757 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6758 undergo unsigned multiplication.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6759
6760<h5>Semantics:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6761<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6762 an unsigned multiplication of the two arguments. They return a structure
6763 &mdash; the first element of which is the multiplication, and the second
6764 element of which is a bit specifying if the unsigned multiplication resulted
6765 in an overflow.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6766
6767<h5>Examples:</h5>
6768<pre>
6769 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6770 %sum = extractvalue {i32, i1} %res, 0
6771 %obit = extractvalue {i32, i1} %res, 1
6772 br i1 %obit, label %overflow, label %normal
6773</pre>
6774
6775</div>
6776
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6777<!-- ======================================================================= -->
6778<div class="doc_subsection">
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]6779 <a name="int_fp16">Half Precision Floating Point Intrinsics</a>
6780</div>
6781
6782<div class="doc_text">
6783
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]6784<p>Half precision floating point is a storage-only format. This means that it is
6785 a dense encoding (in memory) but does not support computation in the
6786 format.</p>
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]6787
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]6788<p>This means that code must first load the half-precision floating point
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]6789 value as an i16, then convert it to float with <a
6790 href="#int_convert_from_fp16"><tt>llvm.convert.from.fp16</tt></a>.
6791 Computation can then be performed on the float value (including extending to
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]6792 double etc). To store the value back to memory, it is first converted to
6793 float if needed, then converted to i16 with
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]6794 <a href="#int_convert_to_fp16"><tt>llvm.convert.to.fp16</tt></a>, then
6795 storing as an i16 value.</p>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]6796</div>
6797
6798<!-- _______________________________________________________________________ -->
6799<div class="doc_subsubsection">
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]6800 <a name="int_convert_to_fp16">'<tt>llvm.convert.to.fp16</tt>' Intrinsic</a>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]6801</div>
6802
6803<div class="doc_text">
6804
6805<h5>Syntax:</h5>
6806<pre>
6807 declare i16 @llvm.convert.to.fp16(f32 %a)
6808</pre>
6809
6810<h5>Overview:</h5>
6811<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
6812 a conversion from single precision floating point format to half precision
6813 floating point format.</p>
6814
6815<h5>Arguments:</h5>
6816<p>The intrinsic function contains single argument - the value to be
6817 converted.</p>
6818
6819<h5>Semantics:</h5>
6820<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
6821 a conversion from single precision floating point format to half precision
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]6822 floating point format. The return value is an <tt>i16</tt> which
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]6823 contains the converted number.</p>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]6824
6825<h5>Examples:</h5>
6826<pre>
6827 %res = call i16 @llvm.convert.to.fp16(f32 %a)
6828 store i16 %res, i16* @x, align 2
6829</pre>
6830
6831</div>
6832
6833<!-- _______________________________________________________________________ -->
6834<div class="doc_subsubsection">
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]6835 <a name="int_convert_from_fp16">'<tt>llvm.convert.from.fp16</tt>' Intrinsic</a>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]6836</div>
6837
6838<div class="doc_text">
6839
6840<h5>Syntax:</h5>
6841<pre>
6842 declare f32 @llvm.convert.from.fp16(i16 %a)
6843</pre>
6844
6845<h5>Overview:</h5>
6846<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs
6847 a conversion from half precision floating point format to single precision
6848 floating point format.</p>
6849
6850<h5>Arguments:</h5>
6851<p>The intrinsic function contains single argument - the value to be
6852 converted.</p>
6853
6854<h5>Semantics:</h5>
6855<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs a
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]6856 conversion from half single precision floating point format to single
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]6857 precision floating point format. The input half-float value is represented by
6858 an <tt>i16</tt> value.</p>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]6859
6860<h5>Examples:</h5>
6861<pre>
6862 %a = load i16* @x, align 2
6863 %res = call f32 @llvm.convert.from.fp16(i16 %a)
6864</pre>
6865
6866</div>
6867
6868<!-- ======================================================================= -->
6869<div class="doc_subsection">
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6870 <a name="int_debugger">Debugger Intrinsics</a>
6871</div>
6872
6873<div class="doc_text">
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6874
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6875<p>The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt>
6876 prefix), are described in
6877 the <a href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source
6878 Level Debugging</a> document.</p>
6879
6880</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6881
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]6882<!-- ======================================================================= -->
6883<div class="doc_subsection">
6884 <a name="int_eh">Exception Handling Intrinsics</a>
6885</div>
6886
6887<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6888
6889<p>The LLVM exception handling intrinsics (which all start with
6890 <tt>llvm.eh.</tt> prefix), are described in
6891 the <a href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
6892 Handling</a> document.</p>
6893
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]6894</div>
6895
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]6896<!-- ======================================================================= -->
6897<div class="doc_subsection">
Duncan Sands86e01192007-09-11 14:10:23 +0000[diff] [blame]6898 <a name="int_trampoline">Trampoline Intrinsic</a>
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]6899</div>
6900
6901<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6902
6903<p>This intrinsic makes it possible to excise one parameter, marked with
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]6904 the <a href="#nest"><tt>nest</tt></a> attribute, from a function.
6905 The result is a callable
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6906 function pointer lacking the nest parameter - the caller does not need to
6907 provide a value for it. Instead, the value to use is stored in advance in a
6908 "trampoline", a block of memory usually allocated on the stack, which also
6909 contains code to splice the nest value into the argument list. This is used
6910 to implement the GCC nested function address extension.</p>
6911
6912<p>For example, if the function is
6913 <tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
6914 pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as
6915 follows:</p>
6916
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]6917<pre class="doc_code">
Duncan Sands86e01192007-09-11 14:10:23 +0000[diff] [blame]6918 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
6919 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]6920 %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]6921 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]6922</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6923
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]6924<p>The call <tt>%val = call i32 %fp(i32 %x, i32 %y)</tt> is then equivalent
6925 to <tt>%val = call i32 %f(i8* %nval, i32 %x, i32 %y)</tt>.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6926
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]6927</div>
6928
6929<!-- _______________________________________________________________________ -->
6930<div class="doc_subsubsection">
6931 <a name="int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a>
6932</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6933
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]6934<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6935
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]6936<h5>Syntax:</h5>
6937<pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6938 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]6939</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6940
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]6941<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6942<p>This fills the memory pointed to by <tt>tramp</tt> with code and returns a
6943 function pointer suitable for executing it.</p>
6944
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]6945<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6946<p>The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
6947 pointers. The <tt>tramp</tt> argument must point to a sufficiently large and
6948 sufficiently aligned block of memory; this memory is written to by the
6949 intrinsic. Note that the size and the alignment are target-specific - LLVM
6950 currently provides no portable way of determining them, so a front-end that
6951 generates this intrinsic needs to have some target-specific knowledge.
6952 The <tt>func</tt> argument must hold a function bitcast to
6953 an <tt>i8*</tt>.</p>
6954
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]6955<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6956<p>The block of memory pointed to by <tt>tramp</tt> is filled with target
6957 dependent code, turning it into a function. A pointer to this function is
6958 returned, but needs to be bitcast to an <a href="#int_trampoline">appropriate
6959 function pointer type</a> before being called. The new function's signature
6960 is the same as that of <tt>func</tt> with any arguments marked with
6961 the <tt>nest</tt> attribute removed. At most one such <tt>nest</tt> argument
6962 is allowed, and it must be of pointer type. Calling the new function is
6963 equivalent to calling <tt>func</tt> with the same argument list, but
6964 with <tt>nval</tt> used for the missing <tt>nest</tt> argument. If, after
6965 calling <tt>llvm.init.trampoline</tt>, the memory pointed to
6966 by <tt>tramp</tt> is modified, then the effect of any later call to the
6967 returned function pointer is undefined.</p>
6968
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]6969</div>
6970
6971<!-- ======================================================================= -->
6972<div class="doc_subsection">
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]6973 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
6974</div>
6975
6976<div class="doc_text">
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]6977
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6978<p>These intrinsic functions expand the "universal IR" of LLVM to represent
6979 hardware constructs for atomic operations and memory synchronization. This
6980 provides an interface to the hardware, not an interface to the programmer. It
6981 is aimed at a low enough level to allow any programming models or APIs
6982 (Application Programming Interfaces) which need atomic behaviors to map
6983 cleanly onto it. It is also modeled primarily on hardware behavior. Just as
6984 hardware provides a "universal IR" for source languages, it also provides a
6985 starting point for developing a "universal" atomic operation and
6986 synchronization IR.</p>
6987
6988<p>These do <em>not</em> form an API such as high-level threading libraries,
6989 software transaction memory systems, atomic primitives, and intrinsic
6990 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
6991 application libraries. The hardware interface provided by LLVM should allow
6992 a clean implementation of all of these APIs and parallel programming models.
6993 No one model or paradigm should be selected above others unless the hardware
6994 itself ubiquitously does so.</p>
6995
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]6996</div>
6997
6998<!-- _______________________________________________________________________ -->
6999<div class="doc_subsubsection">
7000 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
7001</div>
7002<div class="doc_text">
7003<h5>Syntax:</h5>
7004<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7005 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]7006</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7007
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7008<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7009<p>The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
7010 specific pairs of memory access types.</p>
7011
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7012<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7013<p>The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
7014 The first four arguments enables a specific barrier as listed below. The
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]7015 fifth argument specifies that the barrier applies to io or device or uncached
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7016 memory.</p>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7017
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7018<ul>
7019 <li><tt>ll</tt>: load-load barrier</li>
7020 <li><tt>ls</tt>: load-store barrier</li>
7021 <li><tt>sl</tt>: store-load barrier</li>
7022 <li><tt>ss</tt>: store-store barrier</li>
7023 <li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
7024</ul>
7025
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7026<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7027<p>This intrinsic causes the system to enforce some ordering constraints upon
7028 the loads and stores of the program. This barrier does not
7029 indicate <em>when</em> any events will occur, it only enforces
7030 an <em>order</em> in which they occur. For any of the specified pairs of load
7031 and store operations (f.ex. load-load, or store-load), all of the first
7032 operations preceding the barrier will complete before any of the second
7033 operations succeeding the barrier begin. Specifically the semantics for each
7034 pairing is as follows:</p>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7035
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7036<ul>
7037 <li><tt>ll</tt>: All loads before the barrier must complete before any load
7038 after the barrier begins.</li>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7039 <li><tt>ls</tt>: All loads before the barrier must complete before any
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7040 store after the barrier begins.</li>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7041 <li><tt>ss</tt>: All stores before the barrier must complete before any
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7042 store after the barrier begins.</li>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7043 <li><tt>sl</tt>: All stores before the barrier must complete before any
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7044 load after the barrier begins.</li>
7045</ul>
7046
7047<p>These semantics are applied with a logical "and" behavior when more than one
7048 is enabled in a single memory barrier intrinsic.</p>
7049
7050<p>Backends may implement stronger barriers than those requested when they do
7051 not support as fine grained a barrier as requested. Some architectures do
7052 not need all types of barriers and on such architectures, these become
7053 noops.</p>
7054
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7055<h5>Example:</h5>
7056<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7057%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7058%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7059 store i32 4, %ptr
7060
7061%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7062 call void @llvm.memory.barrier(i1 false, i1 true, i1 false, i1 false)
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7063 <i>; guarantee the above finishes</i>
7064 store i32 8, %ptr <i>; before this begins</i>
7065</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7066
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7067</div>
7068
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7069<!-- _______________________________________________________________________ -->
7070<div class="doc_subsubsection">
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7071 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7072</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7073
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7074<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7075
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7076<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7077<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
7078 any integer bit width and for different address spaces. Not all targets
7079 support all bit widths however.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7080
7081<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7082 declare i8 @llvm.atomic.cmp.swap.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt;)
7083 declare i16 @llvm.atomic.cmp.swap.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt;)
7084 declare i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt;)
7085 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]7086</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7087
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7088<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7089<p>This loads a value in memory and compares it to a given value. If they are
7090 equal, it stores a new value into the memory.</p>
7091
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7092<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7093<p>The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result
7094 as well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
7095 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
7096 this integer type. While any bit width integer may be used, targets may only
7097 lower representations they support in hardware.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7098
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7099<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7100<p>This entire intrinsic must be executed atomically. It first loads the value
7101 in memory pointed to by <tt>ptr</tt> and compares it with the
7102 value <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the
7103 memory. The loaded value is yielded in all cases. This provides the
7104 equivalent of an atomic compare-and-swap operation within the SSA
7105 framework.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7106
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7107<h5>Examples:</h5>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7108<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7109%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7110%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7111 store i32 4, %ptr
7112
7113%val1 = add i32 4, 4
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7114%result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %ptr, i32 4, %val1)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7115 <i>; yields {i32}:result1 = 4</i>
7116%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
7117%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
7118
7119%val2 = add i32 1, 1
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7120%result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %ptr, i32 5, %val2)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7121 <i>; yields {i32}:result2 = 8</i>
7122%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
7123
7124%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
7125</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7126
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7127</div>
7128
7129<!-- _______________________________________________________________________ -->
7130<div class="doc_subsubsection">
7131 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
7132</div>
7133<div class="doc_text">
7134<h5>Syntax:</h5>
7135
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7136<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
7137 integer bit width. Not all targets support all bit widths however.</p>
7138
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7139<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7140 declare i8 @llvm.atomic.swap.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;val&gt;)
7141 declare i16 @llvm.atomic.swap.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;val&gt;)
7142 declare i32 @llvm.atomic.swap.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;val&gt;)
7143 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]7144</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7145
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7146<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7147<p>This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
7148 the value from memory. It then stores the value in <tt>val</tt> in the memory
7149 at <tt>ptr</tt>.</p>
7150
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7151<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7152<p>The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both
7153 the <tt>val</tt> argument and the result must be integers of the same bit
7154 width. The first argument, <tt>ptr</tt>, must be a pointer to a value of this
7155 integer type. The targets may only lower integer representations they
7156 support.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7157
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7158<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7159<p>This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
7160 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
7161 equivalent of an atomic swap operation within the SSA framework.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7162
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7163<h5>Examples:</h5>
7164<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7165%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7166%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7167 store i32 4, %ptr
7168
7169%val1 = add i32 4, 4
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7170%result1 = call i32 @llvm.atomic.swap.i32.p0i32(i32* %ptr, i32 %val1)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7171 <i>; yields {i32}:result1 = 4</i>
7172%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
7173%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
7174
7175%val2 = add i32 1, 1
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7176%result2 = call i32 @llvm.atomic.swap.i32.p0i32(i32* %ptr, i32 %val2)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7177 <i>; yields {i32}:result2 = 8</i>
7178
7179%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
7180%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
7181</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7182
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7183</div>
7184
7185<!-- _______________________________________________________________________ -->
7186<div class="doc_subsubsection">
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7187 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7188
7189</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7190
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7191<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7192
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7193<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7194<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.add</tt> on
7195 any integer bit width. Not all targets support all bit widths however.</p>
7196
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7197<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7198 declare i8 @llvm.atomic.load.add.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7199 declare i16 @llvm.atomic.load.add.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7200 declare i32 @llvm.atomic.load.add.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7201 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]7202</pre>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7203
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7204<h5>Overview:</h5>
7205<p>This intrinsic adds <tt>delta</tt> to the value stored in memory
7206 at <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
7207
7208<h5>Arguments:</h5>
7209<p>The intrinsic takes two arguments, the first a pointer to an integer value
7210 and the second an integer value. The result is also an integer value. These
7211 integer types can have any bit width, but they must all have the same bit
7212 width. The targets may only lower integer representations they support.</p>
7213
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7214<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7215<p>This intrinsic does a series of operations atomically. It first loads the
7216 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
7217 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]7218
7219<h5>Examples:</h5>
7220<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7221%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7222%ptr = bitcast i8* %mallocP to i32*
7223 store i32 4, %ptr
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7224%result1 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 4)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7225 <i>; yields {i32}:result1 = 4</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7226%result2 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 2)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7227 <i>; yields {i32}:result2 = 8</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7228%result3 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 5)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7229 <i>; yields {i32}:result3 = 10</i>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7230%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7231</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7232
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7233</div>
7234
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7235<!-- _______________________________________________________________________ -->
7236<div class="doc_subsubsection">
7237 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
7238
7239</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7240
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7241<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7242
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7243<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7244<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.sub</tt> on
7245 any integer bit width and for different address spaces. Not all targets
7246 support all bit widths however.</p>
7247
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7248<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7249 declare i8 @llvm.atomic.load.sub.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7250 declare i16 @llvm.atomic.load.sub.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7251 declare i32 @llvm.atomic.load.sub.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7252 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]7253</pre>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7254
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7255<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7256<p>This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7257 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
7258
7259<h5>Arguments:</h5>
7260<p>The intrinsic takes two arguments, the first a pointer to an integer value
7261 and the second an integer value. The result is also an integer value. These
7262 integer types can have any bit width, but they must all have the same bit
7263 width. The targets may only lower integer representations they support.</p>
7264
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7265<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7266<p>This intrinsic does a series of operations atomically. It first loads the
7267 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
7268 result to <tt>ptr</tt>. It yields the original value stored
7269 at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7270
7271<h5>Examples:</h5>
7272<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7273%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7274%ptr = bitcast i8* %mallocP to i32*
7275 store i32 8, %ptr
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7276%result1 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 4)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7277 <i>; yields {i32}:result1 = 8</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7278%result2 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 2)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7279 <i>; yields {i32}:result2 = 4</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7280%result3 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 5)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7281 <i>; yields {i32}:result3 = 2</i>
7282%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
7283</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7284
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7285</div>
7286
7287<!-- _______________________________________________________________________ -->
7288<div class="doc_subsubsection">
7289 <a name="int_atomic_load_and">'<tt>llvm.atomic.load.and.*</tt>' Intrinsic</a><br>
7290 <a name="int_atomic_load_nand">'<tt>llvm.atomic.load.nand.*</tt>' Intrinsic</a><br>
7291 <a name="int_atomic_load_or">'<tt>llvm.atomic.load.or.*</tt>' Intrinsic</a><br>
7292 <a name="int_atomic_load_xor">'<tt>llvm.atomic.load.xor.*</tt>' Intrinsic</a><br>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7293</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7294
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7295<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7296
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7297<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7298<p>These are overloaded intrinsics. You can
7299 use <tt>llvm.atomic.load_and</tt>, <tt>llvm.atomic.load_nand</tt>,
7300 <tt>llvm.atomic.load_or</tt>, and <tt>llvm.atomic.load_xor</tt> on any integer
7301 bit width and for different address spaces. Not all targets support all bit
7302 widths however.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7303
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7304<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7305 declare i8 @llvm.atomic.load.and.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7306 declare i16 @llvm.atomic.load.and.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7307 declare i32 @llvm.atomic.load.and.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7308 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]7309</pre>
7310
7311<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7312 declare i8 @llvm.atomic.load.or.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7313 declare i16 @llvm.atomic.load.or.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7314 declare i32 @llvm.atomic.load.or.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7315 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]7316</pre>
7317
7318<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7319 declare i8 @llvm.atomic.load.nand.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7320 declare i16 @llvm.atomic.load.nand.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7321 declare i32 @llvm.atomic.load.nand.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7322 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]7323</pre>
7324
7325<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7326 declare i8 @llvm.atomic.load.xor.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7327 declare i16 @llvm.atomic.load.xor.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7328 declare i32 @llvm.atomic.load.xor.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7329 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]7330</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7331
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7332<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7333<p>These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
7334 the value stored in memory at <tt>ptr</tt>. It yields the original value
7335 at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7336
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7337<h5>Arguments:</h5>
7338<p>These intrinsics take two arguments, the first a pointer to an integer value
7339 and the second an integer value. The result is also an integer value. These
7340 integer types can have any bit width, but they must all have the same bit
7341 width. The targets may only lower integer representations they support.</p>
7342
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7343<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7344<p>These intrinsics does a series of operations atomically. They first load the
7345 value stored at <tt>ptr</tt>. They then do the bitwise
7346 operation <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the
7347 original value stored at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7348
7349<h5>Examples:</h5>
7350<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7351%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7352%ptr = bitcast i8* %mallocP to i32*
7353 store i32 0x0F0F, %ptr
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7354%result0 = call i32 @llvm.atomic.load.nand.i32.p0i32(i32* %ptr, i32 0xFF)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7355 <i>; yields {i32}:result0 = 0x0F0F</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7356%result1 = call i32 @llvm.atomic.load.and.i32.p0i32(i32* %ptr, i32 0xFF)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7357 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7358%result2 = call i32 @llvm.atomic.load.or.i32.p0i32(i32* %ptr, i32 0F)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7359 <i>; yields {i32}:result2 = 0xF0</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7360%result3 = call i32 @llvm.atomic.load.xor.i32.p0i32(i32* %ptr, i32 0F)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7361 <i>; yields {i32}:result3 = FF</i>
7362%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
7363</pre>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7364
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7365</div>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7366
7367<!-- _______________________________________________________________________ -->
7368<div class="doc_subsubsection">
7369 <a name="int_atomic_load_max">'<tt>llvm.atomic.load.max.*</tt>' Intrinsic</a><br>
7370 <a name="int_atomic_load_min">'<tt>llvm.atomic.load.min.*</tt>' Intrinsic</a><br>
7371 <a name="int_atomic_load_umax">'<tt>llvm.atomic.load.umax.*</tt>' Intrinsic</a><br>
7372 <a name="int_atomic_load_umin">'<tt>llvm.atomic.load.umin.*</tt>' Intrinsic</a><br>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7373</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7374
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7375<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7376
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7377<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7378<p>These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
7379 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
7380 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
7381 address spaces. Not all targets support all bit widths however.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7382
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7383<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7384 declare i8 @llvm.atomic.load.max.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7385 declare i16 @llvm.atomic.load.max.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7386 declare i32 @llvm.atomic.load.max.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7387 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]7388</pre>
7389
7390<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7391 declare i8 @llvm.atomic.load.min.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7392 declare i16 @llvm.atomic.load.min.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7393 declare i32 @llvm.atomic.load.min.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7394 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]7395</pre>
7396
7397<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7398 declare i8 @llvm.atomic.load.umax.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7399 declare i16 @llvm.atomic.load.umax.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7400 declare i32 @llvm.atomic.load.umax.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7401 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]7402</pre>
7403
7404<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7405 declare i8 @llvm.atomic.load.umin.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7406 declare i16 @llvm.atomic.load.umin.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7407 declare i32 @llvm.atomic.load.umin.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7408 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]7409</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7410
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7411<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7412<p>These intrinsics takes the signed or unsigned minimum or maximum of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7413 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
7414 original value at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7415
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7416<h5>Arguments:</h5>
7417<p>These intrinsics take two arguments, the first a pointer to an integer value
7418 and the second an integer value. The result is also an integer value. These
7419 integer types can have any bit width, but they must all have the same bit
7420 width. The targets may only lower integer representations they support.</p>
7421
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7422<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7423<p>These intrinsics does a series of operations atomically. They first load the
7424 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or
7425 max <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They
7426 yield the original value stored at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7427
7428<h5>Examples:</h5>
7429<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7430%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7431%ptr = bitcast i8* %mallocP to i32*
7432 store i32 7, %ptr
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7433%result0 = call i32 @llvm.atomic.load.min.i32.p0i32(i32* %ptr, i32 -2)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7434 <i>; yields {i32}:result0 = 7</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7435%result1 = call i32 @llvm.atomic.load.max.i32.p0i32(i32* %ptr, i32 8)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7436 <i>; yields {i32}:result1 = -2</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7437%result2 = call i32 @llvm.atomic.load.umin.i32.p0i32(i32* %ptr, i32 10)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7438 <i>; yields {i32}:result2 = 8</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7439%result3 = call i32 @llvm.atomic.load.umax.i32.p0i32(i32* %ptr, i32 30)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7440 <i>; yields {i32}:result3 = 8</i>
7441%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
7442</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7443
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7444</div>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7445
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7446
7447<!-- ======================================================================= -->
7448<div class="doc_subsection">
7449 <a name="int_memorymarkers">Memory Use Markers</a>
7450</div>
7451
7452<div class="doc_text">
7453
7454<p>This class of intrinsics exists to information about the lifetime of memory
7455 objects and ranges where variables are immutable.</p>
7456
7457</div>
7458
7459<!-- _______________________________________________________________________ -->
7460<div class="doc_subsubsection">
7461 <a name="int_lifetime_start">'<tt>llvm.lifetime.start</tt>' Intrinsic</a>
7462</div>
7463
7464<div class="doc_text">
7465
7466<h5>Syntax:</h5>
7467<pre>
7468 declare void @llvm.lifetime.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7469</pre>
7470
7471<h5>Overview:</h5>
7472<p>The '<tt>llvm.lifetime.start</tt>' intrinsic specifies the start of a memory
7473 object's lifetime.</p>
7474
7475<h5>Arguments:</h5>
Nick Lewycky9bc89042009-10-13 07:57:33 +0000[diff] [blame]7476<p>The first argument is a constant integer representing the size of the
7477 object, or -1 if it is variable sized. The second argument is a pointer to
7478 the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7479
7480<h5>Semantics:</h5>
7481<p>This intrinsic indicates that before this point in the code, the value of the
7482 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]7483 never be used and has an undefined value. A load from the pointer that
7484 precedes this intrinsic can be replaced with
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7485 <tt>'<a href="#undefvalues">undef</a>'</tt>.</p>
7486
7487</div>
7488
7489<!-- _______________________________________________________________________ -->
7490<div class="doc_subsubsection">
7491 <a name="int_lifetime_end">'<tt>llvm.lifetime.end</tt>' Intrinsic</a>
7492</div>
7493
7494<div class="doc_text">
7495
7496<h5>Syntax:</h5>
7497<pre>
7498 declare void @llvm.lifetime.end(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7499</pre>
7500
7501<h5>Overview:</h5>
7502<p>The '<tt>llvm.lifetime.end</tt>' intrinsic specifies the end of a memory
7503 object's lifetime.</p>
7504
7505<h5>Arguments:</h5>
Nick Lewycky9bc89042009-10-13 07:57:33 +0000[diff] [blame]7506<p>The first argument is a constant integer representing the size of the
7507 object, or -1 if it is variable sized. The second argument is a pointer to
7508 the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7509
7510<h5>Semantics:</h5>
7511<p>This intrinsic indicates that after this point in the code, the value of the
7512 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
7513 never be used and has an undefined value. Any stores into the memory object
7514 following this intrinsic may be removed as dead.
7515
7516</div>
7517
7518<!-- _______________________________________________________________________ -->
7519<div class="doc_subsubsection">
7520 <a name="int_invariant_start">'<tt>llvm.invariant.start</tt>' Intrinsic</a>
7521</div>
7522
7523<div class="doc_text">
7524
7525<h5>Syntax:</h5>
7526<pre>
Nick Lewycky2965d3e2010-11-30 04:13:41 +0000[diff] [blame]7527 declare {}* @llvm.invariant.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7528</pre>
7529
7530<h5>Overview:</h5>
7531<p>The '<tt>llvm.invariant.start</tt>' intrinsic specifies that the contents of
7532 a memory object will not change.</p>
7533
7534<h5>Arguments:</h5>
Nick Lewycky9bc89042009-10-13 07:57:33 +0000[diff] [blame]7535<p>The first argument is a constant integer representing the size of the
7536 object, or -1 if it is variable sized. The second argument is a pointer to
7537 the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7538
7539<h5>Semantics:</h5>
7540<p>This intrinsic indicates that until an <tt>llvm.invariant.end</tt> that uses
7541 the return value, the referenced memory location is constant and
7542 unchanging.</p>
7543
7544</div>
7545
7546<!-- _______________________________________________________________________ -->
7547<div class="doc_subsubsection">
7548 <a name="int_invariant_end">'<tt>llvm.invariant.end</tt>' Intrinsic</a>
7549</div>
7550
7551<div class="doc_text">
7552
7553<h5>Syntax:</h5>
7554<pre>
7555 declare void @llvm.invariant.end({}* &lt;start&gt;, i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7556</pre>
7557
7558<h5>Overview:</h5>
7559<p>The '<tt>llvm.invariant.end</tt>' intrinsic specifies that the contents of
7560 a memory object are mutable.</p>
7561
7562<h5>Arguments:</h5>
7563<p>The first argument is the matching <tt>llvm.invariant.start</tt> intrinsic.
Nick Lewycky9bc89042009-10-13 07:57:33 +0000[diff] [blame]7564 The second argument is a constant integer representing the size of the
7565 object, or -1 if it is variable sized and the third argument is a pointer
7566 to the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7567
7568<h5>Semantics:</h5>
7569<p>This intrinsic indicates that the memory is mutable again.</p>
7570
7571</div>
7572
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7573<!-- ======================================================================= -->
7574<div class="doc_subsection">
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7575 <a name="int_general">General Intrinsics</a>
7576</div>
7577
7578<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7579
7580<p>This class of intrinsics is designed to be generic and has no specific
7581 purpose.</p>
7582
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7583</div>
7584
7585<!-- _______________________________________________________________________ -->
7586<div class="doc_subsubsection">
7587 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
7588</div>
7589
7590<div class="doc_text">
7591
7592<h5>Syntax:</h5>
7593<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7594 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]7595</pre>
7596
7597<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7598<p>The '<tt>llvm.var.annotation</tt>' intrinsic.</p>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7599
7600<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7601<p>The first argument is a pointer to a value, the second is a pointer to a
7602 global string, the third is a pointer to a global string which is the source
7603 file name, and the last argument is the line number.</p>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7604
7605<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7606<p>This intrinsic allows annotation of local variables with arbitrary strings.
7607 This can be useful for special purpose optimizations that want to look for
7608 these annotations. These have no other defined use, they are ignored by code
7609 generation and optimization.</p>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7610
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7611</div>
7612
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7613<!-- _______________________________________________________________________ -->
7614<div class="doc_subsubsection">
Tanya Lattner0186a652007-09-21 23:57:59 +0000[diff] [blame]7615 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7616</div>
7617
7618<div class="doc_text">
7619
7620<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7621<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
7622 any integer bit width.</p>
7623
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7624<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7625 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7626 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7627 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7628 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7629 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]7630</pre>
7631
7632<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7633<p>The '<tt>llvm.annotation</tt>' intrinsic.</p>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7634
7635<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7636<p>The first argument is an integer value (result of some expression), the
7637 second is a pointer to a global string, the third is a pointer to a global
7638 string which is the source file name, and the last argument is the line
7639 number. It returns the value of the first argument.</p>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7640
7641<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7642<p>This intrinsic allows annotations to be put on arbitrary expressions with
7643 arbitrary strings. This can be useful for special purpose optimizations that
7644 want to look for these annotations. These have no other defined use, they
7645 are ignored by code generation and optimization.</p>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7646
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7647</div>
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]7648
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]7649<!-- _______________________________________________________________________ -->
7650<div class="doc_subsubsection">
7651 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
7652</div>
7653
7654<div class="doc_text">
7655
7656<h5>Syntax:</h5>
7657<pre>
7658 declare void @llvm.trap()
7659</pre>
7660
7661<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7662<p>The '<tt>llvm.trap</tt>' intrinsic.</p>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]7663
7664<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7665<p>None.</p>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]7666
7667<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7668<p>This intrinsics is lowered to the target dependent trap instruction. If the
7669 target does not have a trap instruction, this intrinsic will be lowered to
7670 the call of the <tt>abort()</tt> function.</p>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]7671
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]7672</div>
7673
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7674<!-- _______________________________________________________________________ -->
7675<div class="doc_subsubsection">
Misha Brukman50de2b22008-11-22 23:55:29 +0000[diff] [blame]7676 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7677</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7678
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7679<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7680
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7681<h5>Syntax:</h5>
7682<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7683 declare void @llvm.stackprotector(i8* &lt;guard&gt;, i8** &lt;slot&gt;)
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7684</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7685
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7686<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7687<p>The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and
7688 stores it onto the stack at <tt>slot</tt>. The stack slot is adjusted to
7689 ensure that it is placed on the stack before local variables.</p>
7690
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7691<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7692<p>The <tt>llvm.stackprotector</tt> intrinsic requires two pointer
7693 arguments. The first argument is the value loaded from the stack
7694 guard <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt>
7695 that has enough space to hold the value of the guard.</p>
7696
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7697<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7698<p>This intrinsic causes the prologue/epilogue inserter to force the position of
7699 the <tt>AllocaInst</tt> stack slot to be before local variables on the
7700 stack. This is to ensure that if a local variable on the stack is
7701 overwritten, it will destroy the value of the guard. When the function exits,
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]7702 the guard on the stack is checked against the original guard. If they are
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7703 different, then the program aborts by calling the <tt>__stack_chk_fail()</tt>
7704 function.</p>
7705
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7706</div>
7707
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]7708<!-- _______________________________________________________________________ -->
7709<div class="doc_subsubsection">
7710 <a name="int_objectsize">'<tt>llvm.objectsize</tt>' Intrinsic</a>
7711</div>
7712
7713<div class="doc_text">
7714
7715<h5>Syntax:</h5>
7716<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7717 declare i32 @llvm.objectsize.i32(i8* &lt;object&gt;, i1 &lt;type&gt;)
7718 declare i64 @llvm.objectsize.i64(i8* &lt;object&gt;, i1 &lt;type&gt;)
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]7719</pre>
7720
7721<h5>Overview:</h5>
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]7722<p>The <tt>llvm.objectsize</tt> intrinsic is designed to provide information to
7723 the optimizers to determine at compile time whether a) an operation (like
7724 memcpy) will overflow a buffer that corresponds to an object, or b) that a
7725 runtime check for overflow isn't necessary. An object in this context means
7726 an allocation of a specific class, structure, array, or other object.</p>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]7727
7728<h5>Arguments:</h5>
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]7729<p>The <tt>llvm.objectsize</tt> intrinsic takes two arguments. The first
Eric Christopher31e39bd2009-12-23 00:29:49 +0000[diff] [blame]7730 argument is a pointer to or into the <tt>object</tt>. The second argument
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]7731 is a boolean 0 or 1. This argument determines whether you want the
7732 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]7733 1, variables are not allowed.</p>
7734
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]7735<h5>Semantics:</h5>
7736<p>The <tt>llvm.objectsize</tt> intrinsic is lowered to either a constant
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]7737 representing the size of the object concerned, or <tt>i32/i64 -1 or 0</tt>,
7738 depending on the <tt>type</tt> argument, if the size cannot be determined at
7739 compile time.</p>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]7740
7741</div>
7742
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]7743<!-- *********************************************************************** -->
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]7744<hr>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]7745<address>
7746 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman86242e12008-12-11 17:34:48 +0000[diff] [blame]7747 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]7748 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman86242e12008-12-11 17:34:48 +0000[diff] [blame]7749 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]7750
7751 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
Reid Spencerca058542006-03-14 05:39:39 +0000[diff] [blame]7752 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]7753 Last modified: $Date$
7754</address>
Chris Lattnerb8f816e2008-01-04 04:33:49 +0000[diff] [blame]7755
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]7756</body>
7757</html>