blob: 1c31e45336975ac72e955c7aa308b2c6f2169458 [file] [log] [blame]
Misha Brukmanc501f552004-03-01 17:47:27 +00001<!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 +00003<html>
4<head>
5 <title>LLVM Assembly Language Reference Manual</title>
Reid Spencercb84e432004-08-26 20:44:00 +00006 <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 +00008 <meta name="description"
Reid Spencercb84e432004-08-26 20:44:00 +00009 content="LLVM Assembly Language Reference Manual.">
Misha Brukman76307852003-11-08 01:05:38 +000010 <link rel="stylesheet" href="llvm.css" type="text/css">
11</head>
Chris Lattner757528b0b2004-05-23 21:06:01 +000012
Misha Brukman76307852003-11-08 01:05:38 +000013<body>
Chris Lattner757528b0b2004-05-23 21:06:01 +000014
Chris Lattner48b383b02003-11-25 01:02:51 +000015<div class="doc_title"> LLVM Language Reference Manual </div>
Chris Lattner2f7c9632001-06-06 20:29:01 +000016<ol>
Misha Brukman76307852003-11-08 01:05:38 +000017 <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 +000020 <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 +000023 <li><a href="#linkage">Linkage Types</a>
24 <ol>
Bill Wendling8693ef82009-07-20 02:41:50 +000025 <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 +000027 <li><a href="#linkage_linker_private_weak">'<tt>linker_private_weak</tt>' Linkage</a></li>
Bill Wendling578ee402010-08-20 22:05:50 +000028 <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 +000029 <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 +000036 <li><a href="#linkage_linkonce_odr">'<tt>linkonce_odr</tt>' Linkage</a></li>
Bill Wendling8693ef82009-07-20 02:41:50 +000037 <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 +000041 </ol>
42 </li>
Chris Lattner0132aff2005-05-06 22:57:40 +000043 <li><a href="#callingconv">Calling Conventions</a></li>
Chris Lattnerbc088212009-01-11 20:53:49 +000044 <li><a href="#namedtypes">Named Types</a></li>
Chris Lattner6af02f32004-12-09 16:11:40 +000045 <li><a href="#globalvars">Global Variables</a></li>
Chris Lattner91c15c42006-01-23 23:23:47 +000046 <li><a href="#functionstructure">Functions</a></li>
Dan Gohmanef9462f2008-10-14 16:51:45 +000047 <li><a href="#aliasstructure">Aliases</a></li>
Devang Pateld1a89692010-01-11 19:35:55 +000048 <li><a href="#namedmetadatastructure">Named Metadata</a></li>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +000049 <li><a href="#paramattrs">Parameter Attributes</a></li>
Devang Patel9eb525d2008-09-26 23:51:19 +000050 <li><a href="#fnattrs">Function Attributes</a></li>
Gordon Henriksen71183b62007-12-10 03:18:06 +000051 <li><a href="#gc">Garbage Collector Names</a></li>
Chris Lattner91c15c42006-01-23 23:23:47 +000052 <li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
Reid Spencer50c723a2007-02-19 23:54:10 +000053 <li><a href="#datalayout">Data Layout</a></li>
Dan Gohman6154a012009-07-27 18:07:55 +000054 <li><a href="#pointeraliasing">Pointer Aliasing Rules</a></li>
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +000055 <li><a href="#volatile">Volatile Memory Accesses</a></li>
Chris Lattner6af02f32004-12-09 16:11:40 +000056 </ol>
57 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +000058 <li><a href="#typesystem">Type System</a>
59 <ol>
Chris Lattner7824d182008-01-04 04:32:38 +000060 <li><a href="#t_classifications">Type Classifications</a></li>
Eric Christopher455c5772009-12-05 02:46:03 +000061 <li><a href="#t_primitive">Primitive Types</a>
Chris Lattner48b383b02003-11-25 01:02:51 +000062 <ol>
Nick Lewycky84a1eeb2009-09-27 00:45:11 +000063 <li><a href="#t_integer">Integer Type</a></li>
Chris Lattner7824d182008-01-04 04:32:38 +000064 <li><a href="#t_floating">Floating Point Types</a></li>
Dale Johannesen33e5c352010-10-01 00:48:59 +000065 <li><a href="#t_x86mmx">X86mmx Type</a></li>
Chris Lattner7824d182008-01-04 04:32:38 +000066 <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 +000068 <li><a href="#t_metadata">Metadata Type</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +000069 </ol>
70 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +000071 <li><a href="#t_derived">Derived Types</a>
72 <ol>
Chris Lattner392be582010-02-12 20:49:41 +000073 <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 +000078 <li><a href="#t_vector">Vector Type</a></li>
79 </ol>
80 </li>
Misha Brukman76307852003-11-08 01:05:38 +000081 <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 +000083 <li><a href="#t_opaque">Opaque Type</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +000084 </ol>
85 </li>
Chris Lattnercf7a5842009-02-02 07:32:36 +000086 <li><a href="#t_uprefs">Type Up-references</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +000087 </ol>
88 </li>
Chris Lattner6af02f32004-12-09 16:11:40 +000089 <li><a href="#constants">Constants</a>
Chris Lattner74d3f822004-12-09 17:30:23 +000090 <ol>
Dan Gohmanef9462f2008-10-14 16:51:45 +000091 <li><a href="#simpleconstants">Simple Constants</a></li>
Chris Lattner361bfcd2009-02-28 18:32:25 +000092 <li><a href="#complexconstants">Complex Constants</a></li>
Dan Gohmanef9462f2008-10-14 16:51:45 +000093 <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 +000095 <li><a href="#trapvalues">Trap Values</a></li>
Chris Lattner2bfd3202009-10-27 21:19:13 +000096 <li><a href="#blockaddress">Addresses of Basic Blocks</a></li>
Dan Gohmanef9462f2008-10-14 16:51:45 +000097 <li><a href="#constantexprs">Constant Expressions</a></li>
Chris Lattner74d3f822004-12-09 17:30:23 +000098 </ol>
Chris Lattner48b383b02003-11-25 01:02:51 +000099 </li>
Chris Lattner98f013c2006-01-25 23:47:57 +0000100 <li><a href="#othervalues">Other Values</a>
101 <ol>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000102 <li><a href="#inlineasm">Inline Assembler Expressions</a></li>
Devang Pateld1a89692010-01-11 19:35:55 +0000103 <li><a href="#metadata">Metadata Nodes and Metadata Strings</a></li>
Chris Lattner98f013c2006-01-25 23:47:57 +0000104 </ol>
105 </li>
Chris Lattnerae76db52009-07-20 05:55:19 +0000106 <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 +0000109 <li><a href="#intg_compiler_used">The '<tt>llvm.compiler.used</tt>'
110 Global Variable</a></li>
Chris Lattnerae76db52009-07-20 05:55:19 +0000111 <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 +0000117 <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 +0000121 <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 +0000123 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000124 <li><a href="#i_indirectbr">'<tt>indirectbr</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000125 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000126 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000127 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000128 </ol>
129 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000130 <li><a href="#binaryops">Binary Operations</a>
131 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000132 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
Dan Gohmana5b96452009-06-04 22:49:04 +0000133 <li><a href="#i_fadd">'<tt>fadd</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000134 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
Dan Gohmana5b96452009-06-04 22:49:04 +0000135 <li><a href="#i_fsub">'<tt>fsub</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000136 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
Dan Gohmana5b96452009-06-04 22:49:04 +0000137 <li><a href="#i_fmul">'<tt>fmul</tt>' Instruction</a></li>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000138 <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 +0000141 <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 +0000144 </ol>
145 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000146 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
147 <ol>
Reid Spencer2ab01932007-02-02 13:57:07 +0000148 <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 +0000151 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000152 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000153 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000154 </ol>
155 </li>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000156 <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 +0000161 </ol>
162 </li>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000163 <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 +0000169 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000170 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000171 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
Robert Bocchino820bc75b2006-02-17 21:18:08 +0000172 <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 +0000175 </ol>
176 </li>
Reid Spencer97c5fa42006-11-08 01:18:52 +0000177 <li><a href="#convertops">Conversion Operations</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000178 <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 +0000184 <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 +0000188 <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 +0000190 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000191 </ol>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000192 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000193 <li><a href="#otherops">Other Operations</a>
194 <ol>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000195 <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 +0000197 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000198 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000199 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Chris Lattner33337472006-01-13 23:26:01 +0000200 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000201 </ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000202 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000203 </ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000204 </li>
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000205 <li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000206 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000207 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
208 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000209 <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 +0000212 </ol>
213 </li>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000214 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
215 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000216 <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 +0000219 </ol>
220 </li>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000221 <li><a href="#int_codegen">Code Generator Intrinsics</a>
222 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000223 <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 +0000229 <li><a href="#int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswellaa1c3c12004-04-09 16:43:20 +0000230 </ol>
231 </li>
Chris Lattnerfee11462004-02-12 17:01:32 +0000232 <li><a href="#int_libc">Standard C Library Intrinsics</a>
233 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000234 <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 +0000239 <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 +0000242 </ol>
243 </li>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000244 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000245 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000246 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattnerb748c672006-01-16 22:34:14 +0000247 <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 +0000250 </ol>
251 </li>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000252 <li><a href="#int_overflow">Arithmetic with Overflow Intrinsics</a>
253 <ol>
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000254 <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 +0000259 <li><a href="#int_umul_overflow">'<tt>llvm.umul.with.overflow.*</tt> Intrinsics</a></li>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000260 </ol>
261 </li>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000262 <li><a href="#int_fp16">Half Precision Floating Point Intrinsics</a>
263 <ol>
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000264 <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 +0000266 </ol>
267 </li>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000268 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Jim Laskey2211f492007-03-14 19:31:19 +0000269 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sands86e01192007-09-11 14:10:23 +0000270 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
Duncan Sands644f9172007-07-27 12:58:54 +0000271 <ol>
272 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands644f9172007-07-27 12:58:54 +0000273 </ol>
274 </li>
Bill Wendlingf85850f2008-11-18 22:10:53 +0000275 <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 +0000292 <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 +0000300 <li><a href="#int_general">General intrinsics</a>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000301 <ol>
Reid Spencer5b2cb0f2007-07-20 19:59:11 +0000302 <li><a href="#int_var_annotation">
Bill Wendling14313312008-11-19 05:56:17 +0000303 '<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattner293c0372007-09-21 22:59:12 +0000304 <li><a href="#int_annotation">
Bill Wendling14313312008-11-19 05:56:17 +0000305 '<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000306 <li><a href="#int_trap">
Bill Wendling14313312008-11-19 05:56:17 +0000307 '<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 +0000310 <li><a href="#int_objectsize">
311 '<tt>llvm.objectsize</tt>' Intrinsic</a></li>
Tanya Lattner293c0372007-09-21 22:59:12 +0000312 </ol>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000313 </li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000314 </ol>
315 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000316</ol>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000317
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 +0000321</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000322
Chris Lattner2f7c9632001-06-06 20:29:01 +0000323<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000324<div class="doc_section"> <a name="abstract">Abstract </a></div>
325<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000326
Misha Brukman76307852003-11-08 01:05:38 +0000327<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000328
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 +0000335</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000336
Chris Lattner2f7c9632001-06-06 20:29:01 +0000337<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000338<div class="doc_section"> <a name="introduction">Introduction</a> </div>
339<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000340
Misha Brukman76307852003-11-08 01:05:38 +0000341<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000342
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000343<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 +0000351
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000352<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 +0000359 variable is never accessed outside of the current function, allowing it to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000360 be promoted to a simple SSA value instead of a memory location.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000361
Misha Brukman76307852003-11-08 01:05:38 +0000362</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000363
Chris Lattner2f7c9632001-06-06 20:29:01 +0000364<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000365<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000366
Misha Brukman76307852003-11-08 01:05:38 +0000367<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000368
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000369<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 +0000373
Benjamin Kramer79698be2010-07-13 12:26:09 +0000374<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000375%x = <a href="#i_add">add</a> i32 1, %x
Chris Lattner757528b0b2004-05-23 21:06:01 +0000376</pre>
377
Bill Wendling7f4a3362009-11-02 00:24:16 +0000378<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 +0000384
Bill Wendling3716c5d2007-05-29 09:04:49 +0000385</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000386
Chris Lattner87a3dbe2007-10-03 17:34:29 +0000387<!-- Describe the typesetting conventions here. -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000388
Chris Lattner2f7c9632001-06-06 20:29:01 +0000389<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000390<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000391<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000392
Misha Brukman76307852003-11-08 01:05:38 +0000393<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000394
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000395<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 +0000400
Chris Lattner2f7c9632001-06-06 20:29:01 +0000401<ol>
Reid Spencerb23b65f2007-08-07 14:34:28 +0000402 <li>Named values are represented as a string of characters with their prefix.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000403 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 +0000410
Reid Spencerb23b65f2007-08-07 14:34:28 +0000411 <li>Unnamed values are represented as an unsigned numeric value with their
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000412 prefix. For example, <tt>%12</tt>, <tt>@2</tt>, <tt>%44</tt>.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000413
Reid Spencer8f08d802004-12-09 18:02:53 +0000414 <li>Constants, which are described in a <a href="#constants">section about
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000415 constants</a>, below.</li>
Misha Brukman76307852003-11-08 01:05:38 +0000416</ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000417
Reid Spencerb23b65f2007-08-07 14:34:28 +0000418<p>LLVM requires that values start with a prefix for two reasons: Compilers
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000419 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 +0000423
Chris Lattner48b383b02003-11-25 01:02:51 +0000424<p>Reserved words in LLVM are very similar to reserved words in other
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000425 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 +0000433
434<p>Here is an example of LLVM code to multiply the integer variable
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000435 '<tt>%X</tt>' by 8:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000436
Misha Brukman76307852003-11-08 01:05:38 +0000437<p>The easy way:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000438
Benjamin Kramer79698be2010-07-13 12:26:09 +0000439<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000440%result = <a href="#i_mul">mul</a> i32 %X, 8
Chris Lattnerd79749a2004-12-09 16:36:40 +0000441</pre>
442
Misha Brukman76307852003-11-08 01:05:38 +0000443<p>After strength reduction:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000444
Benjamin Kramer79698be2010-07-13 12:26:09 +0000445<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000446%result = <a href="#i_shl">shl</a> i32 %X, i8 3
Chris Lattnerd79749a2004-12-09 16:36:40 +0000447</pre>
448
Misha Brukman76307852003-11-08 01:05:38 +0000449<p>And the hard way:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000450
Benjamin Kramer79698be2010-07-13 12:26:09 +0000451<pre class="doc_code">
Gabor Greifbd0328f2009-10-28 13:05:07 +0000452%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 +0000454%result = <a href="#i_add">add</a> i32 %1, %1
Chris Lattnerd79749a2004-12-09 16:36:40 +0000455</pre>
456
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000457<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 +0000459
Chris Lattner2f7c9632001-06-06 20:29:01 +0000460<ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000461 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000462 line.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000463
464 <li>Unnamed temporaries are created when the result of a computation is not
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000465 assigned to a named value.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000466
Misha Brukman76307852003-11-08 01:05:38 +0000467 <li>Unnamed temporaries are numbered sequentially</li>
468</ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000469
Bill Wendling7f4a3362009-11-02 00:24:16 +0000470<p>It also shows a convention that we follow in this document. When
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000471 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 +0000474
Misha Brukman76307852003-11-08 01:05:38 +0000475</div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000476
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 +0000487<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 +0000493
Benjamin Kramer79698be2010-07-13 12:26:09 +0000494<pre class="doc_code">
Chris Lattner54a7be72010-08-17 17:13:42 +0000495<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 +0000497
Chris Lattner54a7be72010-08-17 17:13:42 +0000498<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 +0000500
501<i>; Definition of main function</i>
Chris Lattner54a7be72010-08-17 17:13:42 +0000502define 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 +0000505
Chris Lattner54a7be72010-08-17 17:13:42 +0000506 <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 +0000510
511<i>; Named metadata</i>
512!1 = metadata !{i32 41}
513!foo = !{!1, null}
Bill Wendling3716c5d2007-05-29 09:04:49 +0000514</pre>
Chris Lattner6af02f32004-12-09 16:11:40 +0000515
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000516<p>This example is made up of a <a href="#globalvars">global variable</a> named
Devang Pateld1a89692010-01-11 19:35:55 +0000517 "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>" function,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000518 a <a href="#functionstructure">function definition</a> for
Devang Pateld1a89692010-01-11 19:35:55 +0000519 "<tt>main</tt>" and <a href="#namedmetadatastructure">named metadata</a>
520 "<tt>foo"</tt>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000521
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000522<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 +0000527
Chris Lattnerd79749a2004-12-09 16:36:40 +0000528</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 +0000537<p>All Global Variables and Functions have one of the following types of
538 linkage:</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000539
540<dl>
Bill Wendling7f4a3362009-11-02 00:24:16 +0000541 <dt><tt><b><a name="linkage_private">private</a></b></tt></dt>
Bill Wendling03bcd6e2010-07-01 21:55:59 +0000542 <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 +0000548
Bill Wendling7f4a3362009-11-02 00:24:16 +0000549 <dt><tt><b><a name="linkage_linker_private">linker_private</a></b></tt></dt>
Bill Wendling03bcd6e2010-07-01 21:55:59 +0000550 <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 +0000560
Bill Wendling578ee402010-08-20 22:05:50 +0000561 <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 +0000570 <dt><tt><b><a name="linkage_internal">internal</a></b></tt></dt>
Bill Wendling36321712010-06-29 22:34:52 +0000571 <dd>Similar to private, but the value shows as a local symbol
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000572 (<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 +0000574
Bill Wendling7f4a3362009-11-02 00:24:16 +0000575 <dt><tt><b><a name="linkage_available_externally">available_externally</a></b></tt></dt>
Chris Lattner184f1be2009-04-13 05:44:34 +0000576 <dd>Globals with "<tt>available_externally</tt>" linkage are never emitted
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000577 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 +0000583
Bill Wendling7f4a3362009-11-02 00:24:16 +0000584 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt></dt>
Chris Lattnere20b4702007-01-14 06:51:48 +0000585 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
Chris Lattner0de4caa2010-01-09 19:15:14 +0000586 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 +0000597
Bill Wendling7f4a3362009-11-02 00:24:16 +0000598 <dt><tt><b><a name="linkage_weak">weak</a></b></tt></dt>
Chris Lattnerd0554882009-08-05 05:21:07 +0000599 <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 +0000604 <dt><tt><b><a name="linkage_common">common</a></b></tt></dt>
Chris Lattnerd0554882009-08-05 05:21:07 +0000605 <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 +0000610 <tt>common</tt> symbols may not have an explicit section,
Eric Christopher455c5772009-12-05 02:46:03 +0000611 must have a zero initializer, and may not be marked '<a
Chris Lattner0aff0b22009-08-05 05:41:44 +0000612 href="#globalvars"><tt>constant</tt></a>'. Functions and aliases may not
613 have common linkage.</dd>
Chris Lattnerd0554882009-08-05 05:21:07 +0000614
Chris Lattnerd79749a2004-12-09 16:36:40 +0000615
Bill Wendling7f4a3362009-11-02 00:24:16 +0000616 <dt><tt><b><a name="linkage_appending">appending</a></b></tt></dt>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000617 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000618 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 +0000622
Bill Wendling7f4a3362009-11-02 00:24:16 +0000623 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000624 <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 +0000627
Bill Wendling7f4a3362009-11-02 00:24:16 +0000628 <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 +0000630 <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 +0000632 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 +0000634 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 +0000637
Chris Lattner6af02f32004-12-09 16:11:40 +0000638 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000639 <dd>If none of the above identifiers are used, the global is externally
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000640 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 +0000642</dl>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000643
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000644<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 +0000647
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000648<dl>
Bill Wendling7f4a3362009-11-02 00:24:16 +0000649 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt></dt>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000650 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000651 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 +0000655
Bill Wendling7f4a3362009-11-02 00:24:16 +0000656 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt></dt>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000657 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000658 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 +0000662</dl>
663
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000664<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 +0000674<p>Aliases can have only <tt>external</tt>, <tt>internal</tt>, <tt>weak</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000675 or <tt>weak_odr</tt> linkages.</p>
676
Chris Lattner6af02f32004-12-09 16:11:40 +0000677</div>
678
679<!-- ======================================================================= -->
680<div class="doc_subsection">
Chris Lattner0132aff2005-05-06 22:57:40 +0000681 <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 +0000687 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 +0000692
693<dl>
694 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000695 <dd>This calling convention (the default if no other calling convention is
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000696 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 +0000700
701 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000702 <dd>This calling convention attempts to make calls as fast as possible
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000703 (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 +0000706 (Application Binary Interface).
707 <a href="CodeGenerator.html#tailcallopt">Tail calls can only be optimized
Chris Lattnera179e4d2010-03-11 00:22:57 +0000708 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 +0000711
712 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000713 <dd>This calling convention attempts to make code in the caller as efficient
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000714 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 +0000719
Chris Lattnera179e4d2010-03-11 00:22:57 +0000720 <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 +0000740 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000741 <dd>Any calling convention may be specified by number, allowing
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000742 target-specific calling conventions to be used. Target specific calling
743 conventions start at 64.</dd>
Chris Lattner573f64e2005-05-07 01:46:40 +0000744</dl>
Chris Lattner0132aff2005-05-06 22:57:40 +0000745
746<p>More calling conventions can be added/defined on an as-needed basis, to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000747 support Pascal conventions or any other well-known target-independent
748 convention.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000749
750</div>
751
752<!-- ======================================================================= -->
753<div class="doc_subsection">
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000754 <a name="visibility">Visibility Styles</a>
755</div>
756
757<div class="doc_text">
758
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000759<p>All Global Variables and Functions have one of the following visibility
760 styles:</p>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000761
762<dl>
763 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
Chris Lattner67c37d12008-08-05 18:29:16 +0000764 <dd>On targets that use the ELF object file format, default visibility means
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000765 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 +0000769
770 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000771 <dd>Two declarations of an object with hidden visibility refer to the same
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000772 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 +0000776
Anton Korobeynikov39f3cff2007-04-29 18:35:00 +0000777 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
Anton Korobeynikov39f3cff2007-04-29 18:35:00 +0000778 <dd>On ELF, protected visibility indicates that the symbol will be placed in
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000779 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 +0000782</dl>
783
784</div>
785
786<!-- ======================================================================= -->
787<div class="doc_subsection">
Chris Lattnerbc088212009-01-11 20:53:49 +0000788 <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 +0000794 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 +0000796
Benjamin Kramer79698be2010-07-13 12:26:09 +0000797<pre class="doc_code">
Chris Lattnerbc088212009-01-11 20:53:49 +0000798%mytype = type { %mytype*, i32 }
799</pre>
Chris Lattnerbc088212009-01-11 20:53:49 +0000800
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000801<p>You may give a name to any <a href="#typesystem">type</a> except
Chris Lattner249b9762010-08-17 23:26:04 +0000802 "<a href="#t_void">void</a>". Type name aliases may be used anywhere a type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000803 is expected with the syntax "%mytype".</p>
Chris Lattnerbc088212009-01-11 20:53:49 +0000804
805<p>Note that type names are aliases for the structural type that they indicate,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000806 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 +0000814
815</div>
816
Chris Lattnerbc088212009-01-11 20:53:49 +0000817<!-- ======================================================================= -->
818<div class="doc_subsection">
Chris Lattner6af02f32004-12-09 16:11:40 +0000819 <a name="globalvars">Global Variables</a>
820</div>
821
822<div class="doc_text">
823
Chris Lattner5d5aede2005-02-12 19:30:21 +0000824<p>Global variables define regions of memory allocated at compilation time
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000825 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 +0000835
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000836<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 +0000842
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000843<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 +0000848
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000849<p>A global variable may be declared to reside in a target-specific numbered
850 address space. For targets that support them, address spaces may affect how
851 optimizations are performed and/or what target instructions are used to
852 access the variable. The default address space is zero. The address space
853 qualifier must precede any other attributes.</p>
Christopher Lamb308121c2007-12-11 09:31:00 +0000854
Chris Lattner662c8722005-11-12 00:45:07 +0000855<p>LLVM allows an explicit section to be specified for globals. If the target
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000856 supports it, it will emit globals to the section specified.</p>
Chris Lattner662c8722005-11-12 00:45:07 +0000857
Chris Lattner78e00bc2010-04-28 00:13:42 +0000858<p>An explicit alignment may be specified for a global, which must be a power
859 of 2. If not present, or if the alignment is set to zero, the alignment of
860 the global is set by the target to whatever it feels convenient. If an
861 explicit alignment is specified, the global is forced to have exactly that
Chris Lattner4bd85e42010-04-28 00:31:12 +0000862 alignment. Targets and optimizers are not allowed to over-align the global
863 if the global has an assigned section. In this case, the extra alignment
864 could be observable: for example, code could assume that the globals are
865 densely packed in their section and try to iterate over them as an array,
866 alignment padding would break this iteration.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000867
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000868<p>For example, the following defines a global in a numbered address space with
869 an initializer, section, and alignment:</p>
Chris Lattner5760c502007-01-14 00:27:09 +0000870
Benjamin Kramer79698be2010-07-13 12:26:09 +0000871<pre class="doc_code">
Dan Gohmanaaa679b2009-01-11 00:40:00 +0000872@G = addrspace(5) constant float 1.0, section "foo", align 4
Chris Lattner5760c502007-01-14 00:27:09 +0000873</pre>
874
Chris Lattner6af02f32004-12-09 16:11:40 +0000875</div>
876
877
878<!-- ======================================================================= -->
879<div class="doc_subsection">
880 <a name="functionstructure">Functions</a>
881</div>
882
883<div class="doc_text">
884
Dan Gohmana269a0a2010-03-01 17:41:39 +0000885<p>LLVM function definitions consist of the "<tt>define</tt>" keyword, an
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000886 optional <a href="#linkage">linkage type</a>, an optional
887 <a href="#visibility">visibility style</a>, an optional
888 <a href="#callingconv">calling convention</a>, a return type, an optional
889 <a href="#paramattrs">parameter attribute</a> for the return type, a function
890 name, a (possibly empty) argument list (each with optional
891 <a href="#paramattrs">parameter attributes</a>), optional
892 <a href="#fnattrs">function attributes</a>, an optional section, an optional
893 alignment, an optional <a href="#gc">garbage collector name</a>, an opening
894 curly brace, a list of basic blocks, and a closing curly brace.</p>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000895
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000896<p>LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
897 optional <a href="#linkage">linkage type</a>, an optional
Eric Christopher455c5772009-12-05 02:46:03 +0000898 <a href="#visibility">visibility style</a>, an optional
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000899 <a href="#callingconv">calling convention</a>, a return type, an optional
900 <a href="#paramattrs">parameter attribute</a> for the return type, a function
901 name, a possibly empty list of arguments, an optional alignment, and an
902 optional <a href="#gc">garbage collector name</a>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000903
Chris Lattner67c37d12008-08-05 18:29:16 +0000904<p>A function definition contains a list of basic blocks, forming the CFG
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000905 (Control Flow Graph) for the function. Each basic block may optionally start
906 with a label (giving the basic block a symbol table entry), contains a list
907 of instructions, and ends with a <a href="#terminators">terminator</a>
908 instruction (such as a branch or function return).</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000909
Chris Lattnera59fb102007-06-08 16:52:14 +0000910<p>The first basic block in a function is special in two ways: it is immediately
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000911 executed on entrance to the function, and it is not allowed to have
912 predecessor basic blocks (i.e. there can not be any branches to the entry
913 block of a function). Because the block can have no predecessors, it also
914 cannot have any <a href="#i_phi">PHI nodes</a>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000915
Chris Lattner662c8722005-11-12 00:45:07 +0000916<p>LLVM allows an explicit section to be specified for functions. If the target
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000917 supports it, it will emit functions to the section specified.</p>
Chris Lattner662c8722005-11-12 00:45:07 +0000918
Chris Lattner54611b42005-11-06 08:02:57 +0000919<p>An explicit alignment may be specified for a function. If not present, or if
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000920 the alignment is set to zero, the alignment of the function is set by the
921 target to whatever it feels convenient. If an explicit alignment is
922 specified, the function is forced to have at least that much alignment. All
923 alignments must be a power of 2.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000924
Bill Wendling30235112009-07-20 02:39:26 +0000925<h5>Syntax:</h5>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000926<pre class="doc_code">
Chris Lattner0ae02092008-10-13 16:55:18 +0000927define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000928 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
929 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
930 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
931 [<a href="#gc">gc</a>] { ... }
932</pre>
Devang Patel02256232008-10-07 17:48:33 +0000933
Chris Lattner6af02f32004-12-09 16:11:40 +0000934</div>
935
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000936<!-- ======================================================================= -->
937<div class="doc_subsection">
938 <a name="aliasstructure">Aliases</a>
939</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000940
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000941<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000942
943<p>Aliases act as "second name" for the aliasee value (which can be either
944 function, global variable, another alias or bitcast of global value). Aliases
945 may have an optional <a href="#linkage">linkage type</a>, and an
946 optional <a href="#visibility">visibility style</a>.</p>
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000947
Bill Wendling30235112009-07-20 02:39:26 +0000948<h5>Syntax:</h5>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000949<pre class="doc_code">
Duncan Sands7e99a942008-09-12 20:48:21 +0000950@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Bill Wendling2d8b9a82007-05-29 09:42:13 +0000951</pre>
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000952
953</div>
954
Chris Lattner91c15c42006-01-23 23:23:47 +0000955<!-- ======================================================================= -->
Devang Pateld1a89692010-01-11 19:35:55 +0000956<div class="doc_subsection">
957 <a name="namedmetadatastructure">Named Metadata</a>
958</div>
959
960<div class="doc_text">
961
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000962<p>Named metadata is a collection of metadata. <a href="#metadata">Metadata
Dan Gohman093cb792010-07-21 18:54:18 +0000963 nodes</a> (but not metadata strings) are the only valid operands for
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000964 a named metadata.</p>
Devang Pateld1a89692010-01-11 19:35:55 +0000965
966<h5>Syntax:</h5>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000967<pre class="doc_code">
Dan Gohman093cb792010-07-21 18:54:18 +0000968; Some unnamed metadata nodes, which are referenced by the named metadata.
969!0 = metadata !{metadata !"zero"}
Devang Pateld1a89692010-01-11 19:35:55 +0000970!1 = metadata !{metadata !"one"}
Dan Gohman093cb792010-07-21 18:54:18 +0000971!2 = metadata !{metadata !"two"}
Dan Gohman58cd65f2010-07-13 19:48:13 +0000972; A named metadata.
Dan Gohman093cb792010-07-21 18:54:18 +0000973!name = !{!0, !1, !2}
Devang Pateld1a89692010-01-11 19:35:55 +0000974</pre>
Devang Pateld1a89692010-01-11 19:35:55 +0000975
976</div>
977
978<!-- ======================================================================= -->
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000979<div class="doc_subsection"><a name="paramattrs">Parameter Attributes</a></div>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000980
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000981<div class="doc_text">
982
983<p>The return type and each parameter of a function type may have a set of
984 <i>parameter attributes</i> associated with them. Parameter attributes are
985 used to communicate additional information about the result or parameters of
986 a function. Parameter attributes are considered to be part of the function,
987 not of the function type, so functions with different parameter attributes
988 can have the same function type.</p>
989
990<p>Parameter attributes are simple keywords that follow the type specified. If
991 multiple parameter attributes are needed, they are space separated. For
992 example:</p>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000993
Benjamin Kramer79698be2010-07-13 12:26:09 +0000994<pre class="doc_code">
Nick Lewyckydac78d82009-02-15 23:06:14 +0000995declare i32 @printf(i8* noalias nocapture, ...)
Chris Lattnerd2597d72008-10-04 18:33:34 +0000996declare i32 @atoi(i8 zeroext)
997declare signext i8 @returns_signed_char()
Bill Wendling3716c5d2007-05-29 09:04:49 +0000998</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000999
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001000<p>Note that any attributes for the function result (<tt>nounwind</tt>,
1001 <tt>readonly</tt>) come immediately after the argument list.</p>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00001002
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001003<p>Currently, only the following parameter attributes are defined:</p>
Chris Lattner5cee13f2008-01-11 06:20:47 +00001004
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001005<dl>
Bill Wendling7f4a3362009-11-02 00:24:16 +00001006 <dt><tt><b>zeroext</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001007 <dd>This indicates to the code generator that the parameter or return value
1008 should be zero-extended to a 32-bit value by the caller (for a parameter)
1009 or the callee (for a return value).</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +00001010
Bill Wendling7f4a3362009-11-02 00:24:16 +00001011 <dt><tt><b>signext</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001012 <dd>This indicates to the code generator that the parameter or return value
1013 should be sign-extended to a 32-bit value by the caller (for a parameter)
1014 or the callee (for a return value).</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +00001015
Bill Wendling7f4a3362009-11-02 00:24:16 +00001016 <dt><tt><b>inreg</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001017 <dd>This indicates that this parameter or return value should be treated in a
1018 special target-dependent fashion during while emitting code for a function
1019 call or return (usually, by putting it in a register as opposed to memory,
1020 though some targets use it to distinguish between two different kinds of
1021 registers). Use of this attribute is target-specific.</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +00001022
Bill Wendling7f4a3362009-11-02 00:24:16 +00001023 <dt><tt><b><a name="byval">byval</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001024 <dd>This indicates that the pointer parameter should really be passed by value
1025 to the function. The attribute implies that a hidden copy of the pointee
1026 is made between the caller and the callee, so the callee is unable to
1027 modify the value in the callee. This attribute is only valid on LLVM
1028 pointer arguments. It is generally used to pass structs and arrays by
1029 value, but is also valid on pointers to scalars. The copy is considered
1030 to belong to the caller not the callee (for example,
1031 <tt><a href="#readonly">readonly</a></tt> functions should not write to
1032 <tt>byval</tt> parameters). This is not a valid attribute for return
1033 values. The byval attribute also supports specifying an alignment with
1034 the align attribute. This has a target-specific effect on the code
1035 generator that usually indicates a desired alignment for the synthesized
1036 stack slot.</dd>
1037
Dan Gohman3770af52010-07-02 23:18:08 +00001038 <dt><tt><b><a name="sret">sret</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001039 <dd>This indicates that the pointer parameter specifies the address of a
1040 structure that is the return value of the function in the source program.
1041 This pointer must be guaranteed by the caller to be valid: loads and
1042 stores to the structure may be assumed by the callee to not to trap. This
1043 may only be applied to the first parameter. This is not a valid attribute
1044 for return values. </dd>
1045
Dan Gohman3770af52010-07-02 23:18:08 +00001046 <dt><tt><b><a name="noalias">noalias</a></b></tt></dt>
Dan Gohmandf12d082010-07-02 18:41:32 +00001047 <dd>This indicates that pointer values
1048 <a href="#pointeraliasing"><i>based</i></a> on the argument or return
Dan Gohmande256292010-07-02 23:46:54 +00001049 value do not alias pointer values which are not <i>based</i> on it,
1050 ignoring certain "irrelevant" dependencies.
1051 For a call to the parent function, dependencies between memory
1052 references from before or after the call and from those during the call
1053 are "irrelevant" to the <tt>noalias</tt> keyword for the arguments and
1054 return value used in that call.
Dan Gohmandf12d082010-07-02 18:41:32 +00001055 The caller shares the responsibility with the callee for ensuring that
1056 these requirements are met.
1057 For further details, please see the discussion of the NoAlias response in
Dan Gohman6c858db2010-07-06 15:26:33 +00001058 <a href="AliasAnalysis.html#MustMayNo">alias analysis</a>.<br>
1059<br>
John McCall72ed8902010-07-06 21:07:14 +00001060 Note that this definition of <tt>noalias</tt> is intentionally
1061 similar to the definition of <tt>restrict</tt> in C99 for function
Chris Lattner5eff9ca2010-07-06 20:51:35 +00001062 arguments, though it is slightly weaker.
Dan Gohman6c858db2010-07-06 15:26:33 +00001063<br>
1064 For function return values, C99's <tt>restrict</tt> is not meaningful,
1065 while LLVM's <tt>noalias</tt> is.
1066 </dd>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001067
Dan Gohman3770af52010-07-02 23:18:08 +00001068 <dt><tt><b><a name="nocapture">nocapture</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001069 <dd>This indicates that the callee does not make any copies of the pointer
1070 that outlive the callee itself. This is not a valid attribute for return
1071 values.</dd>
1072
Dan Gohman3770af52010-07-02 23:18:08 +00001073 <dt><tt><b><a name="nest">nest</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001074 <dd>This indicates that the pointer parameter can be excised using the
1075 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
1076 attribute for return values.</dd>
1077</dl>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00001078
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00001079</div>
1080
1081<!-- ======================================================================= -->
Chris Lattner91c15c42006-01-23 23:23:47 +00001082<div class="doc_subsection">
Gordon Henriksen71183b62007-12-10 03:18:06 +00001083 <a name="gc">Garbage Collector Names</a>
1084</div>
1085
1086<div class="doc_text">
Gordon Henriksen71183b62007-12-10 03:18:06 +00001087
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001088<p>Each function may specify a garbage collector name, which is simply a
1089 string:</p>
1090
Benjamin Kramer79698be2010-07-13 12:26:09 +00001091<pre class="doc_code">
Bill Wendling7f4a3362009-11-02 00:24:16 +00001092define void @f() gc "name" { ... }
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001093</pre>
Gordon Henriksen71183b62007-12-10 03:18:06 +00001094
1095<p>The compiler declares the supported values of <i>name</i>. Specifying a
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001096 collector which will cause the compiler to alter its output in order to
1097 support the named garbage collection algorithm.</p>
1098
Gordon Henriksen71183b62007-12-10 03:18:06 +00001099</div>
1100
1101<!-- ======================================================================= -->
1102<div class="doc_subsection">
Devang Patel9eb525d2008-09-26 23:51:19 +00001103 <a name="fnattrs">Function Attributes</a>
Devang Patelcaacdba2008-09-04 23:05:13 +00001104</div>
1105
1106<div class="doc_text">
Devang Patel9eb525d2008-09-26 23:51:19 +00001107
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001108<p>Function attributes are set to communicate additional information about a
1109 function. Function attributes are considered to be part of the function, not
1110 of the function type, so functions with different parameter attributes can
1111 have the same function type.</p>
Devang Patel9eb525d2008-09-26 23:51:19 +00001112
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001113<p>Function attributes are simple keywords that follow the type specified. If
1114 multiple attributes are needed, they are space separated. For example:</p>
Devang Patelcaacdba2008-09-04 23:05:13 +00001115
Benjamin Kramer79698be2010-07-13 12:26:09 +00001116<pre class="doc_code">
Devang Patel9eb525d2008-09-26 23:51:19 +00001117define void @f() noinline { ... }
1118define void @f() alwaysinline { ... }
1119define void @f() alwaysinline optsize { ... }
Bill Wendling7f4a3362009-11-02 00:24:16 +00001120define void @f() optsize { ... }
Bill Wendlingb175fa42008-09-07 10:26:33 +00001121</pre>
Devang Patelcaacdba2008-09-04 23:05:13 +00001122
Bill Wendlingb175fa42008-09-07 10:26:33 +00001123<dl>
Charles Davisbe5557e2010-02-12 00:31:15 +00001124 <dt><tt><b>alignstack(&lt;<em>n</em>&gt;)</b></tt></dt>
1125 <dd>This attribute indicates that, when emitting the prologue and epilogue,
1126 the backend should forcibly align the stack pointer. Specify the
1127 desired alignment, which must be a power of two, in parentheses.
1128
Bill Wendling7f4a3362009-11-02 00:24:16 +00001129 <dt><tt><b>alwaysinline</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001130 <dd>This attribute indicates that the inliner should attempt to inline this
1131 function into callers whenever possible, ignoring any active inlining size
1132 threshold for this caller.</dd>
Bill Wendlingb175fa42008-09-07 10:26:33 +00001133
Charles Davis22fe1862010-10-25 15:37:09 +00001134 <dt><tt><b>hotpatch</b></tt></dt>
Charles Davis1b2d3722010-10-25 16:29:03 +00001135 <dd>This attribute indicates that the function should be 'hotpatchable',
Charles Davis74205252010-10-25 19:07:39 +00001136 meaning the function can be patched and/or hooked even while it is
1137 loaded into memory. On x86, the function prologue will be preceded
1138 by six bytes of padding and will begin with a two-byte instruction.
1139 Most of the functions in the Windows system DLLs in Windows XP SP2 or
1140 higher were compiled in this fashion.</dd>
Charles Davis22fe1862010-10-25 15:37:09 +00001141
Jakob Stoklund Olesen74bb06c2010-02-06 01:16:28 +00001142 <dt><tt><b>inlinehint</b></tt></dt>
1143 <dd>This attribute indicates that the source code contained a hint that inlining
1144 this function is desirable (such as the "inline" keyword in C/C++). It
1145 is just a hint; it imposes no requirements on the inliner.</dd>
1146
Nick Lewycky14b58da2010-07-06 18:24:09 +00001147 <dt><tt><b>naked</b></tt></dt>
1148 <dd>This attribute disables prologue / epilogue emission for the function.
1149 This can have very system-specific consequences.</dd>
1150
1151 <dt><tt><b>noimplicitfloat</b></tt></dt>
1152 <dd>This attributes disables implicit floating point instructions.</dd>
1153
Bill Wendling7f4a3362009-11-02 00:24:16 +00001154 <dt><tt><b>noinline</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001155 <dd>This attribute indicates that the inliner should never inline this
1156 function in any situation. This attribute may not be used together with
1157 the <tt>alwaysinline</tt> attribute.</dd>
Devang Patel9eb525d2008-09-26 23:51:19 +00001158
Nick Lewycky14b58da2010-07-06 18:24:09 +00001159 <dt><tt><b>noredzone</b></tt></dt>
1160 <dd>This attribute indicates that the code generator should not use a red
1161 zone, even if the target-specific ABI normally permits it.</dd>
Devang Patel9eb525d2008-09-26 23:51:19 +00001162
Bill Wendling7f4a3362009-11-02 00:24:16 +00001163 <dt><tt><b>noreturn</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001164 <dd>This function attribute indicates that the function never returns
1165 normally. This produces undefined behavior at runtime if the function
1166 ever does dynamically return.</dd>
Bill Wendlinga8130172008-11-13 01:02:51 +00001167
Bill Wendling7f4a3362009-11-02 00:24:16 +00001168 <dt><tt><b>nounwind</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001169 <dd>This function attribute indicates that the function never returns with an
1170 unwind or exceptional control flow. If the function does unwind, its
1171 runtime behavior is undefined.</dd>
Bill Wendling0f5541e2008-11-26 19:07:40 +00001172
Nick Lewycky14b58da2010-07-06 18:24:09 +00001173 <dt><tt><b>optsize</b></tt></dt>
1174 <dd>This attribute suggests that optimization passes and code generator passes
1175 make choices that keep the code size of this function low, and otherwise
1176 do optimizations specifically to reduce code size.</dd>
1177
Bill Wendling7f4a3362009-11-02 00:24:16 +00001178 <dt><tt><b>readnone</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001179 <dd>This attribute indicates that the function computes its result (or decides
1180 to unwind an exception) based strictly on its arguments, without
1181 dereferencing any pointer arguments or otherwise accessing any mutable
1182 state (e.g. memory, control registers, etc) visible to caller functions.
1183 It does not write through any pointer arguments
1184 (including <tt><a href="#byval">byval</a></tt> arguments) and never
1185 changes any state visible to callers. This means that it cannot unwind
1186 exceptions by calling the <tt>C++</tt> exception throwing methods, but
1187 could use the <tt>unwind</tt> instruction.</dd>
Devang Patel310fd4a2009-06-12 19:45:19 +00001188
Bill Wendling7f4a3362009-11-02 00:24:16 +00001189 <dt><tt><b><a name="readonly">readonly</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001190 <dd>This attribute indicates that the function does not write through any
1191 pointer arguments (including <tt><a href="#byval">byval</a></tt>
1192 arguments) or otherwise modify any state (e.g. memory, control registers,
1193 etc) visible to caller functions. It may dereference pointer arguments
1194 and read state that may be set in the caller. A readonly function always
1195 returns the same value (or unwinds an exception identically) when called
1196 with the same set of arguments and global state. It cannot unwind an
1197 exception by calling the <tt>C++</tt> exception throwing methods, but may
1198 use the <tt>unwind</tt> instruction.</dd>
Anton Korobeynikovc8ce7b082009-07-17 18:07:26 +00001199
Bill Wendling7f4a3362009-11-02 00:24:16 +00001200 <dt><tt><b><a name="ssp">ssp</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001201 <dd>This attribute indicates that the function should emit a stack smashing
1202 protector. It is in the form of a "canary"&mdash;a random value placed on
1203 the stack before the local variables that's checked upon return from the
1204 function to see if it has been overwritten. A heuristic is used to
1205 determine if a function needs stack protectors or not.<br>
1206<br>
1207 If a function that has an <tt>ssp</tt> attribute is inlined into a
1208 function that doesn't have an <tt>ssp</tt> attribute, then the resulting
1209 function will have an <tt>ssp</tt> attribute.</dd>
1210
Bill Wendling7f4a3362009-11-02 00:24:16 +00001211 <dt><tt><b>sspreq</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001212 <dd>This attribute indicates that the function should <em>always</em> emit a
1213 stack smashing protector. This overrides
Bill Wendling30235112009-07-20 02:39:26 +00001214 the <tt><a href="#ssp">ssp</a></tt> function attribute.<br>
1215<br>
1216 If a function that has an <tt>sspreq</tt> attribute is inlined into a
1217 function that doesn't have an <tt>sspreq</tt> attribute or which has
1218 an <tt>ssp</tt> attribute, then the resulting function will have
1219 an <tt>sspreq</tt> attribute.</dd>
Bill Wendlingb175fa42008-09-07 10:26:33 +00001220</dl>
1221
Devang Patelcaacdba2008-09-04 23:05:13 +00001222</div>
1223
1224<!-- ======================================================================= -->
1225<div class="doc_subsection">
Chris Lattner93564892006-04-08 04:40:53 +00001226 <a name="moduleasm">Module-Level Inline Assembly</a>
Chris Lattner91c15c42006-01-23 23:23:47 +00001227</div>
1228
1229<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001230
1231<p>Modules may contain "module-level inline asm" blocks, which corresponds to
1232 the GCC "file scope inline asm" blocks. These blocks are internally
1233 concatenated by LLVM and treated as a single unit, but may be separated in
1234 the <tt>.ll</tt> file if desired. The syntax is very simple:</p>
Chris Lattner91c15c42006-01-23 23:23:47 +00001235
Benjamin Kramer79698be2010-07-13 12:26:09 +00001236<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +00001237module asm "inline asm code goes here"
1238module asm "more can go here"
1239</pre>
Chris Lattner91c15c42006-01-23 23:23:47 +00001240
1241<p>The strings can contain any character by escaping non-printable characters.
1242 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001243 for the number.</p>
Chris Lattner91c15c42006-01-23 23:23:47 +00001244
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001245<p>The inline asm code is simply printed to the machine code .s file when
1246 assembly code is generated.</p>
1247
Chris Lattner91c15c42006-01-23 23:23:47 +00001248</div>
Chris Lattner6af02f32004-12-09 16:11:40 +00001249
Reid Spencer50c723a2007-02-19 23:54:10 +00001250<!-- ======================================================================= -->
1251<div class="doc_subsection">
1252 <a name="datalayout">Data Layout</a>
1253</div>
1254
1255<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001256
Reid Spencer50c723a2007-02-19 23:54:10 +00001257<p>A module may specify a target specific data layout string that specifies how
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001258 data is to be laid out in memory. The syntax for the data layout is
1259 simply:</p>
1260
Benjamin Kramer79698be2010-07-13 12:26:09 +00001261<pre class="doc_code">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001262target datalayout = "<i>layout specification</i>"
1263</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001264
1265<p>The <i>layout specification</i> consists of a list of specifications
1266 separated by the minus sign character ('-'). Each specification starts with
1267 a letter and may include other information after the letter to define some
1268 aspect of the data layout. The specifications accepted are as follows:</p>
1269
Reid Spencer50c723a2007-02-19 23:54:10 +00001270<dl>
1271 <dt><tt>E</tt></dt>
1272 <dd>Specifies that the target lays out data in big-endian form. That is, the
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001273 bits with the most significance have the lowest address location.</dd>
1274
Reid Spencer50c723a2007-02-19 23:54:10 +00001275 <dt><tt>e</tt></dt>
Chris Lattner67c37d12008-08-05 18:29:16 +00001276 <dd>Specifies that the target lays out data in little-endian form. That is,
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001277 the bits with the least significance have the lowest address
1278 location.</dd>
1279
Reid Spencer50c723a2007-02-19 23:54:10 +00001280 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher455c5772009-12-05 02:46:03 +00001281 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001282 <i>preferred</i> alignments. All sizes are in bits. Specifying
1283 the <i>pref</i> alignment is optional. If omitted, the
1284 preceding <tt>:</tt> should be omitted too.</dd>
1285
Reid Spencer50c723a2007-02-19 23:54:10 +00001286 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1287 <dd>This specifies the alignment for an integer type of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001288 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1289
Reid Spencer50c723a2007-02-19 23:54:10 +00001290 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher455c5772009-12-05 02:46:03 +00001291 <dd>This specifies the alignment for a vector type of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001292 <i>size</i>.</dd>
1293
Reid Spencer50c723a2007-02-19 23:54:10 +00001294 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher455c5772009-12-05 02:46:03 +00001295 <dd>This specifies the alignment for a floating point type of a given bit
Dale Johannesence522852010-05-28 18:54:47 +00001296 <i>size</i>. Only values of <i>size</i> that are supported by the target
1297 will work. 32 (float) and 64 (double) are supported on all targets;
1298 80 or 128 (different flavors of long double) are also supported on some
1299 targets.
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001300
Reid Spencer50c723a2007-02-19 23:54:10 +00001301 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1302 <dd>This specifies the alignment for an aggregate type of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001303 <i>size</i>.</dd>
1304
Daniel Dunbar7921a592009-06-08 22:17:53 +00001305 <dt><tt>s<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1306 <dd>This specifies the alignment for a stack object of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001307 <i>size</i>.</dd>
Chris Lattnera381eff2009-11-07 09:35:34 +00001308
1309 <dt><tt>n<i>size1</i>:<i>size2</i>:<i>size3</i>...</tt></dt>
1310 <dd>This specifies a set of native integer widths for the target CPU
1311 in bits. For example, it might contain "n32" for 32-bit PowerPC,
1312 "n32:64" for PowerPC 64, or "n8:16:32:64" for X86-64. Elements of
Eric Christopher455c5772009-12-05 02:46:03 +00001313 this set are considered to support most general arithmetic
Chris Lattnera381eff2009-11-07 09:35:34 +00001314 operations efficiently.</dd>
Reid Spencer50c723a2007-02-19 23:54:10 +00001315</dl>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001316
Reid Spencer50c723a2007-02-19 23:54:10 +00001317<p>When constructing the data layout for a given target, LLVM starts with a
Dan Gohman61110ae2010-04-28 00:36:01 +00001318 default set of specifications which are then (possibly) overridden by the
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001319 specifications in the <tt>datalayout</tt> keyword. The default specifications
1320 are given in this list:</p>
1321
Reid Spencer50c723a2007-02-19 23:54:10 +00001322<ul>
1323 <li><tt>E</tt> - big endian</li>
Dan Gohman8ad777d2010-02-23 02:44:03 +00001324 <li><tt>p:64:64:64</tt> - 64-bit pointers with 64-bit alignment</li>
Reid Spencer50c723a2007-02-19 23:54:10 +00001325 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1326 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1327 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1328 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattner67c37d12008-08-05 18:29:16 +00001329 <li><tt>i64:32:64</tt> - i64 has ABI alignment of 32-bits but preferred
Reid Spencer50c723a2007-02-19 23:54:10 +00001330 alignment of 64-bits</li>
1331 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1332 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1333 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1334 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1335 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
Daniel Dunbar7921a592009-06-08 22:17:53 +00001336 <li><tt>s0:64:64</tt> - stack objects are 64-bit aligned</li>
Reid Spencer50c723a2007-02-19 23:54:10 +00001337</ul>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001338
1339<p>When LLVM is determining the alignment for a given type, it uses the
1340 following rules:</p>
1341
Reid Spencer50c723a2007-02-19 23:54:10 +00001342<ol>
1343 <li>If the type sought is an exact match for one of the specifications, that
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001344 specification is used.</li>
1345
Reid Spencer50c723a2007-02-19 23:54:10 +00001346 <li>If no match is found, and the type sought is an integer type, then the
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001347 smallest integer type that is larger than the bitwidth of the sought type
1348 is used. If none of the specifications are larger than the bitwidth then
1349 the the largest integer type is used. For example, given the default
1350 specifications above, the i7 type will use the alignment of i8 (next
1351 largest) while both i65 and i256 will use the alignment of i64 (largest
1352 specified).</li>
1353
Reid Spencer50c723a2007-02-19 23:54:10 +00001354 <li>If no match is found, and the type sought is a vector type, then the
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001355 largest vector type that is smaller than the sought vector type will be
1356 used as a fall back. This happens because &lt;128 x double&gt; can be
1357 implemented in terms of 64 &lt;2 x double&gt;, for example.</li>
Reid Spencer50c723a2007-02-19 23:54:10 +00001358</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001359
Reid Spencer50c723a2007-02-19 23:54:10 +00001360</div>
Chris Lattner6af02f32004-12-09 16:11:40 +00001361
Dan Gohman6154a012009-07-27 18:07:55 +00001362<!-- ======================================================================= -->
1363<div class="doc_subsection">
1364 <a name="pointeraliasing">Pointer Aliasing Rules</a>
1365</div>
1366
1367<div class="doc_text">
1368
Andreas Bolka8ae4e242009-07-29 00:02:05 +00001369<p>Any memory access must be done through a pointer value associated
Andreas Bolkae39f0332009-07-27 20:37:10 +00001370with an address range of the memory access, otherwise the behavior
Dan Gohman6154a012009-07-27 18:07:55 +00001371is undefined. Pointer values are associated with address ranges
1372according to the following rules:</p>
1373
1374<ul>
Dan Gohmandf12d082010-07-02 18:41:32 +00001375 <li>A pointer value is associated with the addresses associated with
1376 any value it is <i>based</i> on.
Andreas Bolka8ae4e242009-07-29 00:02:05 +00001377 <li>An address of a global variable is associated with the address
Dan Gohman6154a012009-07-27 18:07:55 +00001378 range of the variable's storage.</li>
1379 <li>The result value of an allocation instruction is associated with
1380 the address range of the allocated storage.</li>
1381 <li>A null pointer in the default address-space is associated with
Andreas Bolka8ae4e242009-07-29 00:02:05 +00001382 no address.</li>
Dan Gohman6154a012009-07-27 18:07:55 +00001383 <li>An integer constant other than zero or a pointer value returned
1384 from a function not defined within LLVM may be associated with address
1385 ranges allocated through mechanisms other than those provided by
Andreas Bolka8ae4e242009-07-29 00:02:05 +00001386 LLVM. Such ranges shall not overlap with any ranges of addresses
Dan Gohman6154a012009-07-27 18:07:55 +00001387 allocated by mechanisms provided by LLVM.</li>
Dan Gohmandf12d082010-07-02 18:41:32 +00001388</ul>
1389
1390<p>A pointer value is <i>based</i> on another pointer value according
1391 to the following rules:</p>
1392
1393<ul>
1394 <li>A pointer value formed from a
1395 <tt><a href="#i_getelementptr">getelementptr</a></tt> operation
1396 is <i>based</i> on the first operand of the <tt>getelementptr</tt>.</li>
1397 <li>The result value of a
1398 <tt><a href="#i_bitcast">bitcast</a></tt> is <i>based</i> on the operand
1399 of the <tt>bitcast</tt>.</li>
1400 <li>A pointer value formed by an
1401 <tt><a href="#i_inttoptr">inttoptr</a></tt> is <i>based</i> on all
1402 pointer values that contribute (directly or indirectly) to the
1403 computation of the pointer's value.</li>
1404 <li>The "<i>based</i> on" relationship is transitive.</li>
1405</ul>
1406
1407<p>Note that this definition of <i>"based"</i> is intentionally
1408 similar to the definition of <i>"based"</i> in C99, though it is
1409 slightly weaker.</p>
Dan Gohman6154a012009-07-27 18:07:55 +00001410
1411<p>LLVM IR does not associate types with memory. The result type of a
Andreas Bolka8ae4e242009-07-29 00:02:05 +00001412<tt><a href="#i_load">load</a></tt> merely indicates the size and
1413alignment of the memory from which to load, as well as the
Dan Gohman4eb47192010-06-17 19:23:50 +00001414interpretation of the value. The first operand type of a
Andreas Bolka8ae4e242009-07-29 00:02:05 +00001415<tt><a href="#i_store">store</a></tt> similarly only indicates the size
1416and alignment of the store.</p>
Dan Gohman6154a012009-07-27 18:07:55 +00001417
1418<p>Consequently, type-based alias analysis, aka TBAA, aka
1419<tt>-fstrict-aliasing</tt>, is not applicable to general unadorned
1420LLVM IR. <a href="#metadata">Metadata</a> may be used to encode
1421additional information which specialized optimization passes may use
1422to implement type-based alias analysis.</p>
1423
1424</div>
1425
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +00001426<!-- ======================================================================= -->
1427<div class="doc_subsection">
1428 <a name="volatile">Volatile Memory Accesses</a>
1429</div>
1430
1431<div class="doc_text">
1432
1433<p>Certain memory accesses, such as <a href="#i_load"><tt>load</tt></a>s, <a
1434href="#i_store"><tt>store</tt></a>s, and <a
1435href="#int_memcpy"><tt>llvm.memcpy</tt></a>s may be marked <tt>volatile</tt>.
1436The optimizers must not change the number of volatile operations or change their
1437order of execution relative to other volatile operations. The optimizers
1438<i>may</i> change the order of volatile operations relative to non-volatile
1439operations. This is not Java's "volatile" and has no cross-thread
1440synchronization behavior.</p>
1441
1442</div>
1443
Chris Lattner2f7c9632001-06-06 20:29:01 +00001444<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +00001445<div class="doc_section"> <a name="typesystem">Type System</a> </div>
1446<!-- *********************************************************************** -->
Chris Lattner6af02f32004-12-09 16:11:40 +00001447
Misha Brukman76307852003-11-08 01:05:38 +00001448<div class="doc_text">
Chris Lattner6af02f32004-12-09 16:11:40 +00001449
Misha Brukman76307852003-11-08 01:05:38 +00001450<p>The LLVM type system is one of the most important features of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001451 intermediate representation. Being typed enables a number of optimizations
1452 to be performed on the intermediate representation directly, without having
1453 to do extra analyses on the side before the transformation. A strong type
1454 system makes it easier to read the generated code and enables novel analyses
1455 and transformations that are not feasible to perform on normal three address
1456 code representations.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +00001457
1458</div>
1459
Chris Lattner2f7c9632001-06-06 20:29:01 +00001460<!-- ======================================================================= -->
Chris Lattner7824d182008-01-04 04:32:38 +00001461<div class="doc_subsection"> <a name="t_classifications">Type
Chris Lattner48b383b02003-11-25 01:02:51 +00001462Classifications</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001463
Misha Brukman76307852003-11-08 01:05:38 +00001464<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001465
1466<p>The types fall into a few useful classifications:</p>
Misha Brukmanc501f552004-03-01 17:47:27 +00001467
1468<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner48b383b02003-11-25 01:02:51 +00001469 <tbody>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001470 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner48b383b02003-11-25 01:02:51 +00001471 <tr>
Chris Lattner7824d182008-01-04 04:32:38 +00001472 <td><a href="#t_integer">integer</a></td>
Reid Spencer138249b2007-05-16 18:44:01 +00001473 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
Chris Lattner48b383b02003-11-25 01:02:51 +00001474 </tr>
1475 <tr>
Chris Lattner7824d182008-01-04 04:32:38 +00001476 <td><a href="#t_floating">floating point</a></td>
1477 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Chris Lattner48b383b02003-11-25 01:02:51 +00001478 </tr>
1479 <tr>
1480 <td><a name="t_firstclass">first class</a></td>
Chris Lattner7824d182008-01-04 04:32:38 +00001481 <td><a href="#t_integer">integer</a>,
1482 <a href="#t_floating">floating point</a>,
1483 <a href="#t_pointer">pointer</a>,
Dan Gohman08783a882008-06-18 18:42:13 +00001484 <a href="#t_vector">vector</a>,
Dan Gohmanb9d66602008-05-12 23:51:09 +00001485 <a href="#t_struct">structure</a>,
1486 <a href="#t_array">array</a>,
Nick Lewyckyadbc2842009-05-30 05:06:04 +00001487 <a href="#t_label">label</a>,
1488 <a href="#t_metadata">metadata</a>.
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00001489 </td>
Chris Lattner48b383b02003-11-25 01:02:51 +00001490 </tr>
Chris Lattner7824d182008-01-04 04:32:38 +00001491 <tr>
1492 <td><a href="#t_primitive">primitive</a></td>
1493 <td><a href="#t_label">label</a>,
1494 <a href="#t_void">void</a>,
Nick Lewyckyadbc2842009-05-30 05:06:04 +00001495 <a href="#t_floating">floating point</a>,
Dale Johannesen33e5c352010-10-01 00:48:59 +00001496 <a href="#t_x86mmx">x86mmx</a>,
Nick Lewyckyadbc2842009-05-30 05:06:04 +00001497 <a href="#t_metadata">metadata</a>.</td>
Chris Lattner7824d182008-01-04 04:32:38 +00001498 </tr>
1499 <tr>
1500 <td><a href="#t_derived">derived</a></td>
Chris Lattner392be582010-02-12 20:49:41 +00001501 <td><a href="#t_array">array</a>,
Chris Lattner7824d182008-01-04 04:32:38 +00001502 <a href="#t_function">function</a>,
1503 <a href="#t_pointer">pointer</a>,
1504 <a href="#t_struct">structure</a>,
1505 <a href="#t_pstruct">packed structure</a>,
1506 <a href="#t_vector">vector</a>,
1507 <a href="#t_opaque">opaque</a>.
Dan Gohman93bf60d2008-10-14 16:32:04 +00001508 </td>
Chris Lattner7824d182008-01-04 04:32:38 +00001509 </tr>
Chris Lattner48b383b02003-11-25 01:02:51 +00001510 </tbody>
Misha Brukman76307852003-11-08 01:05:38 +00001511</table>
Misha Brukmanc501f552004-03-01 17:47:27 +00001512
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001513<p>The <a href="#t_firstclass">first class</a> types are perhaps the most
1514 important. Values of these types are the only ones which can be produced by
Nick Lewycky84a1eeb2009-09-27 00:45:11 +00001515 instructions.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001516
Misha Brukman76307852003-11-08 01:05:38 +00001517</div>
Chris Lattner74d3f822004-12-09 17:30:23 +00001518
Chris Lattner2f7c9632001-06-06 20:29:01 +00001519<!-- ======================================================================= -->
Chris Lattner7824d182008-01-04 04:32:38 +00001520<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Chris Lattner43542b32008-01-04 04:34:14 +00001521
Chris Lattner7824d182008-01-04 04:32:38 +00001522<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001523
Chris Lattner7824d182008-01-04 04:32:38 +00001524<p>The primitive types are the fundamental building blocks of the LLVM
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001525 system.</p>
Chris Lattner7824d182008-01-04 04:32:38 +00001526
Chris Lattner43542b32008-01-04 04:34:14 +00001527</div>
1528
Chris Lattner7824d182008-01-04 04:32:38 +00001529<!-- _______________________________________________________________________ -->
Nick Lewycky84a1eeb2009-09-27 00:45:11 +00001530<div class="doc_subsubsection"> <a name="t_integer">Integer Type</a> </div>
1531
1532<div class="doc_text">
1533
1534<h5>Overview:</h5>
1535<p>The integer type is a very simple type that simply specifies an arbitrary
1536 bit width for the integer type desired. Any bit width from 1 bit to
1537 2<sup>23</sup>-1 (about 8 million) can be specified.</p>
1538
1539<h5>Syntax:</h5>
1540<pre>
1541 iN
1542</pre>
1543
1544<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1545 value.</p>
1546
1547<h5>Examples:</h5>
1548<table class="layout">
1549 <tr class="layout">
1550 <td class="left"><tt>i1</tt></td>
1551 <td class="left">a single-bit integer.</td>
1552 </tr>
1553 <tr class="layout">
1554 <td class="left"><tt>i32</tt></td>
1555 <td class="left">a 32-bit integer.</td>
1556 </tr>
1557 <tr class="layout">
1558 <td class="left"><tt>i1942652</tt></td>
1559 <td class="left">a really big integer of over 1 million bits.</td>
1560 </tr>
1561</table>
1562
Nick Lewycky84a1eeb2009-09-27 00:45:11 +00001563</div>
1564
1565<!-- _______________________________________________________________________ -->
Chris Lattner7824d182008-01-04 04:32:38 +00001566<div class="doc_subsubsection"> <a name="t_floating">Floating Point Types</a> </div>
1567
1568<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001569
1570<table>
1571 <tbody>
1572 <tr><th>Type</th><th>Description</th></tr>
1573 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1574 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1575 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1576 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1577 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1578 </tbody>
1579</table>
1580
Chris Lattner7824d182008-01-04 04:32:38 +00001581</div>
1582
1583<!-- _______________________________________________________________________ -->
Dale Johannesen33e5c352010-10-01 00:48:59 +00001584<div class="doc_subsubsection"> <a name="t_x86mmx">X86mmx Type</a> </div>
1585
1586<div class="doc_text">
1587
1588<h5>Overview:</h5>
1589<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>
1590
1591<h5>Syntax:</h5>
1592<pre>
Dale Johannesenb1f0ff12010-10-01 01:07:02 +00001593 x86mmx
Dale Johannesen33e5c352010-10-01 00:48:59 +00001594</pre>
1595
1596</div>
1597
1598<!-- _______________________________________________________________________ -->
Chris Lattner7824d182008-01-04 04:32:38 +00001599<div class="doc_subsubsection"> <a name="t_void">Void Type</a> </div>
1600
1601<div class="doc_text">
Bill Wendling30235112009-07-20 02:39:26 +00001602
Chris Lattner7824d182008-01-04 04:32:38 +00001603<h5>Overview:</h5>
1604<p>The void type does not represent any value and has no size.</p>
1605
1606<h5>Syntax:</h5>
Chris Lattner7824d182008-01-04 04:32:38 +00001607<pre>
1608 void
1609</pre>
Bill Wendling30235112009-07-20 02:39:26 +00001610
Chris Lattner7824d182008-01-04 04:32:38 +00001611</div>
1612
1613<!-- _______________________________________________________________________ -->
1614<div class="doc_subsubsection"> <a name="t_label">Label Type</a> </div>
1615
1616<div class="doc_text">
Bill Wendling30235112009-07-20 02:39:26 +00001617
Chris Lattner7824d182008-01-04 04:32:38 +00001618<h5>Overview:</h5>
1619<p>The label type represents code labels.</p>
1620
1621<h5>Syntax:</h5>
Chris Lattner7824d182008-01-04 04:32:38 +00001622<pre>
1623 label
1624</pre>
Bill Wendling30235112009-07-20 02:39:26 +00001625
Chris Lattner7824d182008-01-04 04:32:38 +00001626</div>
1627
Nick Lewyckyadbc2842009-05-30 05:06:04 +00001628<!-- _______________________________________________________________________ -->
1629<div class="doc_subsubsection"> <a name="t_metadata">Metadata Type</a> </div>
1630
1631<div class="doc_text">
Bill Wendling30235112009-07-20 02:39:26 +00001632
Nick Lewyckyadbc2842009-05-30 05:06:04 +00001633<h5>Overview:</h5>
Nick Lewycky93e06a52009-09-27 23:27:42 +00001634<p>The metadata type represents embedded metadata. No derived types may be
1635 created from metadata except for <a href="#t_function">function</a>
1636 arguments.
Nick Lewyckyadbc2842009-05-30 05:06:04 +00001637
1638<h5>Syntax:</h5>
Nick Lewyckyadbc2842009-05-30 05:06:04 +00001639<pre>
1640 metadata
1641</pre>
Bill Wendling30235112009-07-20 02:39:26 +00001642
Nick Lewyckyadbc2842009-05-30 05:06:04 +00001643</div>
1644
Chris Lattner7824d182008-01-04 04:32:38 +00001645
1646<!-- ======================================================================= -->
Chris Lattner48b383b02003-11-25 01:02:51 +00001647<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
Chris Lattner74d3f822004-12-09 17:30:23 +00001648
Misha Brukman76307852003-11-08 01:05:38 +00001649<div class="doc_text">
Chris Lattner74d3f822004-12-09 17:30:23 +00001650
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001651<p>The real power in LLVM comes from the derived types in the system. This is
1652 what allows a programmer to represent arrays, functions, pointers, and other
Nick Lewycky84a1eeb2009-09-27 00:45:11 +00001653 useful types. Each of these types contain one or more element types which
1654 may be a primitive type, or another derived type. For example, it is
1655 possible to have a two dimensional array, using an array as the element type
1656 of another array.</p>
Dan Gohman142ccc02009-01-24 15:58:40 +00001657
Chris Lattner392be582010-02-12 20:49:41 +00001658
1659</div>
1660
1661<!-- _______________________________________________________________________ -->
1662<div class="doc_subsubsection"> <a name="t_aggregate">Aggregate Types</a> </div>
1663
1664<div class="doc_text">
1665
1666<p>Aggregate Types are a subset of derived types that can contain multiple
1667 member types. <a href="#t_array">Arrays</a>,
Chris Lattner13ee7952010-08-28 04:09:24 +00001668 <a href="#t_struct">structs</a>, and <a href="#t_vector">vectors</a> are
1669 aggregate types.</p>
Chris Lattner392be582010-02-12 20:49:41 +00001670
1671</div>
1672
Reid Spencer138249b2007-05-16 18:44:01 +00001673<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +00001674<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
Chris Lattner74d3f822004-12-09 17:30:23 +00001675
Misha Brukman76307852003-11-08 01:05:38 +00001676<div class="doc_text">
Chris Lattner74d3f822004-12-09 17:30:23 +00001677
Chris Lattner2f7c9632001-06-06 20:29:01 +00001678<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +00001679<p>The array type is a very simple derived type that arranges elements
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001680 sequentially in memory. The array type requires a size (number of elements)
1681 and an underlying data type.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +00001682
Chris Lattner590645f2002-04-14 06:13:44 +00001683<h5>Syntax:</h5>
Chris Lattner74d3f822004-12-09 17:30:23 +00001684<pre>
1685 [&lt;# elements&gt; x &lt;elementtype&gt;]
1686</pre>
1687
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001688<p>The number of elements is a constant integer value; <tt>elementtype</tt> may
1689 be any type with a size.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +00001690
Chris Lattner590645f2002-04-14 06:13:44 +00001691<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001692<table class="layout">
1693 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +00001694 <td class="left"><tt>[40 x i32]</tt></td>
1695 <td class="left">Array of 40 32-bit integer values.</td>
1696 </tr>
1697 <tr class="layout">
1698 <td class="left"><tt>[41 x i32]</tt></td>
1699 <td class="left">Array of 41 32-bit integer values.</td>
1700 </tr>
1701 <tr class="layout">
1702 <td class="left"><tt>[4 x i8]</tt></td>
1703 <td class="left">Array of 4 8-bit integer values.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001704 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +00001705</table>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001706<p>Here are some examples of multidimensional arrays:</p>
1707<table class="layout">
1708 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +00001709 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1710 <td class="left">3x4 array of 32-bit integer values.</td>
1711 </tr>
1712 <tr class="layout">
1713 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1714 <td class="left">12x10 array of single precision floating point values.</td>
1715 </tr>
1716 <tr class="layout">
1717 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1718 <td class="left">2x3x4 array of 16-bit integer values.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001719 </tr>
1720</table>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +00001721
Dan Gohmanc74bc282009-11-09 19:01:53 +00001722<p>There is no restriction on indexing beyond the end of the array implied by
1723 a static type (though there are restrictions on indexing beyond the bounds
1724 of an allocated object in some cases). This means that single-dimension
1725 'variable sized array' addressing can be implemented in LLVM with a zero
1726 length array type. An implementation of 'pascal style arrays' in LLVM could
1727 use the type "<tt>{ i32, [0 x float]}</tt>", for example.</p>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +00001728
Misha Brukman76307852003-11-08 01:05:38 +00001729</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001730
Chris Lattner2f7c9632001-06-06 20:29:01 +00001731<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +00001732<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001733
Misha Brukman76307852003-11-08 01:05:38 +00001734<div class="doc_text">
Chris Lattnerda508ac2008-04-23 04:59:35 +00001735
Chris Lattner2f7c9632001-06-06 20:29:01 +00001736<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001737<p>The function type can be thought of as a function signature. It consists of
1738 a return type and a list of formal parameter types. The return type of a
Chris Lattner13ee7952010-08-28 04:09:24 +00001739 function type is a first class type or a void type.</p>
Devang Pateld6cff512008-03-10 20:49:15 +00001740
Chris Lattner2f7c9632001-06-06 20:29:01 +00001741<h5>Syntax:</h5>
Chris Lattnerda508ac2008-04-23 04:59:35 +00001742<pre>
Nick Lewycky14d1ccc2009-09-27 07:55:32 +00001743 &lt;returntype&gt; (&lt;parameter list&gt;)
Chris Lattnerda508ac2008-04-23 04:59:35 +00001744</pre>
1745
John Criswell4c0cf7f2005-10-24 16:17:18 +00001746<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001747 specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1748 which indicates that the function takes a variable number of arguments.
1749 Variable argument functions can access their arguments with
1750 the <a href="#int_varargs">variable argument handling intrinsic</a>
Chris Lattner47f2a832010-03-02 06:36:51 +00001751 functions. '<tt>&lt;returntype&gt;</tt>' is any type except
Nick Lewycky93e06a52009-09-27 23:27:42 +00001752 <a href="#t_label">label</a>.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +00001753
Chris Lattner2f7c9632001-06-06 20:29:01 +00001754<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001755<table class="layout">
1756 <tr class="layout">
Reid Spencer58c08712006-12-31 07:18:34 +00001757 <td class="left"><tt>i32 (i32)</tt></td>
1758 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001759 </td>
Reid Spencer58c08712006-12-31 07:18:34 +00001760 </tr><tr class="layout">
Chris Lattner47f2a832010-03-02 06:36:51 +00001761 <td class="left"><tt>float&nbsp;(i16,&nbsp;i32&nbsp;*)&nbsp;*
Reid Spencer655dcc62006-12-31 07:20:23 +00001762 </tt></td>
Eric Christopher455c5772009-12-05 02:46:03 +00001763 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
Chris Lattner47f2a832010-03-02 06:36:51 +00001764 an <tt>i16</tt> and a <a href="#t_pointer">pointer</a> to <tt>i32</tt>,
1765 returning <tt>float</tt>.
Reid Spencer58c08712006-12-31 07:18:34 +00001766 </td>
1767 </tr><tr class="layout">
1768 <td class="left"><tt>i32 (i8*, ...)</tt></td>
Eric Christopher455c5772009-12-05 02:46:03 +00001769 <td class="left">A vararg function that takes at least one
1770 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1771 which returns an integer. This is the signature for <tt>printf</tt> in
Reid Spencer58c08712006-12-31 07:18:34 +00001772 LLVM.
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001773 </td>
Devang Patele3dfc1c2008-03-24 05:35:41 +00001774 </tr><tr class="layout">
1775 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Nick Lewycky14d1ccc2009-09-27 07:55:32 +00001776 <td class="left">A function taking an <tt>i32</tt>, returning a
1777 <a href="#t_struct">structure</a> containing two <tt>i32</tt> values
Devang Patele3dfc1c2008-03-24 05:35:41 +00001778 </td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001779 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +00001780</table>
Misha Brukmanc501f552004-03-01 17:47:27 +00001781
Misha Brukman76307852003-11-08 01:05:38 +00001782</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001783
Chris Lattner2f7c9632001-06-06 20:29:01 +00001784<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +00001785<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001786
Misha Brukman76307852003-11-08 01:05:38 +00001787<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001788
Chris Lattner2f7c9632001-06-06 20:29:01 +00001789<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001790<p>The structure type is used to represent a collection of data members together
1791 in memory. The packing of the field types is defined to match the ABI of the
1792 underlying processor. The elements of a structure may be any type that has a
1793 size.</p>
1794
Jeffrey Yasskinf991bbb2010-01-11 19:19:26 +00001795<p>Structures in memory are accessed using '<tt><a href="#i_load">load</a></tt>'
1796 and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field
1797 with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
1798 Structures in registers are accessed using the
1799 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' and
1800 '<tt><a href="#i_insertvalue">insertvalue</a></tt>' instructions.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +00001801<h5>Syntax:</h5>
Bill Wendling30235112009-07-20 02:39:26 +00001802<pre>
1803 { &lt;type list&gt; }
1804</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001805
Chris Lattner2f7c9632001-06-06 20:29:01 +00001806<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001807<table class="layout">
1808 <tr class="layout">
Jeff Cohen5819f182007-04-22 01:17:39 +00001809 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1810 <td class="left">A triple of three <tt>i32</tt> values</td>
1811 </tr><tr class="layout">
1812 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1813 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1814 second element is a <a href="#t_pointer">pointer</a> to a
1815 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1816 an <tt>i32</tt>.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001817 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +00001818</table>
Dan Gohman142ccc02009-01-24 15:58:40 +00001819
Misha Brukman76307852003-11-08 01:05:38 +00001820</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001821
Chris Lattner2f7c9632001-06-06 20:29:01 +00001822<!-- _______________________________________________________________________ -->
Andrew Lenharth8df88e22006-12-08 17:13:00 +00001823<div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
1824</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001825
Andrew Lenharth8df88e22006-12-08 17:13:00 +00001826<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001827
Andrew Lenharth8df88e22006-12-08 17:13:00 +00001828<h5>Overview:</h5>
1829<p>The packed structure type is used to represent a collection of data members
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001830 together in memory. There is no padding between fields. Further, the
1831 alignment of a packed structure is 1 byte. The elements of a packed
1832 structure may be any type that has a size.</p>
1833
1834<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt> and
1835 '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field with
1836 the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
1837
Andrew Lenharth8df88e22006-12-08 17:13:00 +00001838<h5>Syntax:</h5>
Bill Wendling30235112009-07-20 02:39:26 +00001839<pre>
1840 &lt; { &lt;type list&gt; } &gt;
1841</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001842
Andrew Lenharth8df88e22006-12-08 17:13:00 +00001843<h5>Examples:</h5>
1844<table class="layout">
1845 <tr class="layout">
Jeff Cohen5819f182007-04-22 01:17:39 +00001846 <td class="left"><tt>&lt; { i32, i32, i32 } &gt;</tt></td>
1847 <td class="left">A triple of three <tt>i32</tt> values</td>
1848 </tr><tr class="layout">
Bill Wendlingb175fa42008-09-07 10:26:33 +00001849 <td class="left">
1850<tt>&lt;&nbsp;{&nbsp;float,&nbsp;i32&nbsp;(i32)*&nbsp;}&nbsp;&gt;</tt></td>
Jeff Cohen5819f182007-04-22 01:17:39 +00001851 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1852 second element is a <a href="#t_pointer">pointer</a> to a
1853 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1854 an <tt>i32</tt>.</td>
Andrew Lenharth8df88e22006-12-08 17:13:00 +00001855 </tr>
1856</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001857
Andrew Lenharth8df88e22006-12-08 17:13:00 +00001858</div>
1859
1860<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +00001861<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
Chris Lattner4a67c912009-02-08 19:53:29 +00001862
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001863<div class="doc_text">
1864
1865<h5>Overview:</h5>
Dan Gohman88481112010-02-25 16:50:07 +00001866<p>The pointer type is used to specify memory locations.
1867 Pointers are commonly used to reference objects in memory.</p>
1868
1869<p>Pointer types may have an optional address space attribute defining the
1870 numbered address space where the pointed-to object resides. The default
1871 address space is number zero. The semantics of non-zero address
1872 spaces are target-specific.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001873
1874<p>Note that LLVM does not permit pointers to void (<tt>void*</tt>) nor does it
1875 permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</tt> instead.</p>
Chris Lattner4a67c912009-02-08 19:53:29 +00001876
Chris Lattner590645f2002-04-14 06:13:44 +00001877<h5>Syntax:</h5>
Bill Wendling30235112009-07-20 02:39:26 +00001878<pre>
1879 &lt;type&gt; *
1880</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001881
Chris Lattner590645f2002-04-14 06:13:44 +00001882<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001883<table class="layout">
1884 <tr class="layout">
Dan Gohman623806e2009-01-04 23:44:43 +00001885 <td class="left"><tt>[4 x i32]*</tt></td>
Chris Lattner747359f2007-12-19 05:04:11 +00001886 <td class="left">A <a href="#t_pointer">pointer</a> to <a
1887 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
1888 </tr>
1889 <tr class="layout">
Dan Gohmanaabfdb32010-05-28 17:13:49 +00001890 <td class="left"><tt>i32 (i32*) *</tt></td>
Chris Lattner747359f2007-12-19 05:04:11 +00001891 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00001892 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner747359f2007-12-19 05:04:11 +00001893 <tt>i32</tt>.</td>
1894 </tr>
1895 <tr class="layout">
1896 <td class="left"><tt>i32 addrspace(5)*</tt></td>
1897 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
1898 that resides in address space #5.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001899 </tr>
Misha Brukman76307852003-11-08 01:05:38 +00001900</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001901
Misha Brukman76307852003-11-08 01:05:38 +00001902</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001903
Chris Lattnerc8cb6952004-08-12 19:12:28 +00001904<!-- _______________________________________________________________________ -->
Reid Spencer404a3252007-02-15 03:07:05 +00001905<div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001906
Misha Brukman76307852003-11-08 01:05:38 +00001907<div class="doc_text">
Chris Lattner37b6b092005-04-25 17:34:15 +00001908
Chris Lattnerc8cb6952004-08-12 19:12:28 +00001909<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001910<p>A vector type is a simple derived type that represents a vector of elements.
1911 Vector types are used when multiple primitive data are operated in parallel
1912 using a single instruction (SIMD). A vector type requires a size (number of
Duncan Sands31c0e0e2009-11-27 13:38:03 +00001913 elements) and an underlying primitive data type. Vector types are considered
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001914 <a href="#t_firstclass">first class</a>.</p>
Chris Lattner37b6b092005-04-25 17:34:15 +00001915
Chris Lattnerc8cb6952004-08-12 19:12:28 +00001916<h5>Syntax:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +00001917<pre>
1918 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1919</pre>
1920
Chris Lattnerf11031a2010-10-10 18:20:35 +00001921<p>The number of elements is a constant integer value larger than 0; elementtype
1922 may be any integer or floating point type. Vectors of size zero are not
1923 allowed, and pointers are not allowed as the element type.</p>
Chris Lattner37b6b092005-04-25 17:34:15 +00001924
Chris Lattnerc8cb6952004-08-12 19:12:28 +00001925<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001926<table class="layout">
1927 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +00001928 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
1929 <td class="left">Vector of 4 32-bit integer values.</td>
1930 </tr>
1931 <tr class="layout">
1932 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
1933 <td class="left">Vector of 8 32-bit floating-point values.</td>
1934 </tr>
1935 <tr class="layout">
1936 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
1937 <td class="left">Vector of 2 64-bit integer values.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001938 </tr>
1939</table>
Dan Gohman142ccc02009-01-24 15:58:40 +00001940
Misha Brukman76307852003-11-08 01:05:38 +00001941</div>
1942
Chris Lattner37b6b092005-04-25 17:34:15 +00001943<!-- _______________________________________________________________________ -->
1944<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1945<div class="doc_text">
1946
1947<h5>Overview:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +00001948<p>Opaque types are used to represent unknown types in the system. This
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001949 corresponds (for example) to the C notion of a forward declared structure
1950 type. In LLVM, opaque types can eventually be resolved to any type (not just
1951 a structure type).</p>
Chris Lattner37b6b092005-04-25 17:34:15 +00001952
1953<h5>Syntax:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +00001954<pre>
1955 opaque
1956</pre>
1957
1958<h5>Examples:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +00001959<table class="layout">
1960 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +00001961 <td class="left"><tt>opaque</tt></td>
1962 <td class="left">An opaque type.</td>
Chris Lattner37b6b092005-04-25 17:34:15 +00001963 </tr>
1964</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001965
Chris Lattner37b6b092005-04-25 17:34:15 +00001966</div>
1967
Chris Lattnercf7a5842009-02-02 07:32:36 +00001968<!-- ======================================================================= -->
1969<div class="doc_subsection">
1970 <a name="t_uprefs">Type Up-references</a>
1971</div>
1972
1973<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001974
Chris Lattnercf7a5842009-02-02 07:32:36 +00001975<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001976<p>An "up reference" allows you to refer to a lexically enclosing type without
1977 requiring it to have a name. For instance, a structure declaration may
1978 contain a pointer to any of the types it is lexically a member of. Example
1979 of up references (with their equivalent as named type declarations)
1980 include:</p>
Chris Lattnercf7a5842009-02-02 07:32:36 +00001981
1982<pre>
Chris Lattnerbf1d5452009-02-09 10:00:56 +00001983 { \2 * } %x = type { %x* }
Chris Lattnercf7a5842009-02-02 07:32:36 +00001984 { \2 }* %y = type { %y }*
1985 \1* %z = type %z*
1986</pre>
1987
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001988<p>An up reference is needed by the asmprinter for printing out cyclic types
1989 when there is no declared name for a type in the cycle. Because the
1990 asmprinter does not want to print out an infinite type string, it needs a
1991 syntax to handle recursive types that have no names (all names are optional
1992 in llvm IR).</p>
Chris Lattnercf7a5842009-02-02 07:32:36 +00001993
1994<h5>Syntax:</h5>
1995<pre>
1996 \&lt;level&gt;
1997</pre>
1998
Bill Wendlingd9a66f72009-07-20 02:29:24 +00001999<p>The level is the count of the lexical type that is being referred to.</p>
Chris Lattnercf7a5842009-02-02 07:32:36 +00002000
2001<h5>Examples:</h5>
Chris Lattnercf7a5842009-02-02 07:32:36 +00002002<table class="layout">
2003 <tr class="layout">
2004 <td class="left"><tt>\1*</tt></td>
2005 <td class="left">Self-referential pointer.</td>
2006 </tr>
2007 <tr class="layout">
2008 <td class="left"><tt>{ { \3*, i8 }, i32 }</tt></td>
2009 <td class="left">Recursive structure where the upref refers to the out-most
2010 structure.</td>
2011 </tr>
2012</table>
Chris Lattnercf7a5842009-02-02 07:32:36 +00002013
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002014</div>
Chris Lattner37b6b092005-04-25 17:34:15 +00002015
Chris Lattner74d3f822004-12-09 17:30:23 +00002016<!-- *********************************************************************** -->
2017<div class="doc_section"> <a name="constants">Constants</a> </div>
2018<!-- *********************************************************************** -->
2019
2020<div class="doc_text">
2021
2022<p>LLVM has several different basic types of constants. This section describes
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002023 them all and their syntax.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +00002024
2025</div>
2026
2027<!-- ======================================================================= -->
Reid Spencer8f08d802004-12-09 18:02:53 +00002028<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
Chris Lattner74d3f822004-12-09 17:30:23 +00002029
2030<div class="doc_text">
2031
2032<dl>
2033 <dt><b>Boolean constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +00002034 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
Nick Lewycky84a1eeb2009-09-27 00:45:11 +00002035 constants of the <tt><a href="#t_integer">i1</a></tt> type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +00002036
2037 <dt><b>Integer constants</b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002038 <dd>Standard integers (such as '4') are constants of
2039 the <a href="#t_integer">integer</a> type. Negative numbers may be used
2040 with integer types.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +00002041
2042 <dt><b>Floating point constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +00002043 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002044 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
2045 notation (see below). The assembler requires the exact decimal value of a
2046 floating-point constant. For example, the assembler accepts 1.25 but
2047 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
2048 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Chris Lattner74d3f822004-12-09 17:30:23 +00002049
2050 <dt><b>Null pointer constants</b></dt>
John Criswelldfe6a862004-12-10 15:51:16 +00002051 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002052 and must be of <a href="#t_pointer">pointer type</a>.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +00002053</dl>
2054
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002055<p>The one non-intuitive notation for constants is the hexadecimal form of
2056 floating point constants. For example, the form '<tt>double
2057 0x432ff973cafa8000</tt>' is equivalent to (but harder to read than)
2058 '<tt>double 4.5e+15</tt>'. The only time hexadecimal floating point
2059 constants are required (and the only time that they are generated by the
2060 disassembler) is when a floating point constant must be emitted but it cannot
2061 be represented as a decimal floating point number in a reasonable number of
2062 digits. For example, NaN's, infinities, and other special values are
2063 represented in their IEEE hexadecimal format so that assembly and disassembly
2064 do not cause any bits to change in the constants.</p>
2065
Dale Johannesencd4a3012009-02-11 22:14:51 +00002066<p>When using the hexadecimal form, constants of types float and double are
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002067 represented using the 16-digit form shown above (which matches the IEEE754
2068 representation for double); float values must, however, be exactly
2069 representable as IEE754 single precision. Hexadecimal format is always used
2070 for long double, and there are three forms of long double. The 80-bit format
2071 used by x86 is represented as <tt>0xK</tt> followed by 20 hexadecimal digits.
2072 The 128-bit format used by PowerPC (two adjacent doubles) is represented
2073 by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit format
2074 is represented by <tt>0xL</tt> followed by 32 hexadecimal digits; no
2075 currently supported target uses this format. Long doubles will only work if
2076 they match the long double format on your target. All hexadecimal formats
2077 are big-endian (sign bit at the left).</p>
2078
Dale Johannesen33e5c352010-10-01 00:48:59 +00002079<p>There are no constants of type x86mmx.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +00002080</div>
2081
2082<!-- ======================================================================= -->
Chris Lattner361bfcd2009-02-28 18:32:25 +00002083<div class="doc_subsection">
Bill Wendling972b7202009-07-20 02:32:41 +00002084<a name="aggregateconstants"></a> <!-- old anchor -->
2085<a name="complexconstants">Complex Constants</a>
Chris Lattner74d3f822004-12-09 17:30:23 +00002086</div>
2087
2088<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002089
Chris Lattner361bfcd2009-02-28 18:32:25 +00002090<p>Complex constants are a (potentially recursive) combination of simple
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002091 constants and smaller complex constants.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +00002092
2093<dl>
2094 <dt><b>Structure constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +00002095 <dd>Structure constants are represented with notation similar to structure
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002096 type definitions (a comma separated list of elements, surrounded by braces
2097 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
2098 where "<tt>@G</tt>" is declared as "<tt>@G = external global i32</tt>".
2099 Structure constants must have <a href="#t_struct">structure type</a>, and
2100 the number and types of elements must match those specified by the
2101 type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +00002102
2103 <dt><b>Array constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +00002104 <dd>Array constants are represented with notation similar to array type
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002105 definitions (a comma separated list of elements, surrounded by square
2106 brackets (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74
2107 ]</tt>". Array constants must have <a href="#t_array">array type</a>, and
2108 the number and types of elements must match those specified by the
2109 type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +00002110
Reid Spencer404a3252007-02-15 03:07:05 +00002111 <dt><b>Vector constants</b></dt>
Reid Spencer404a3252007-02-15 03:07:05 +00002112 <dd>Vector constants are represented with notation similar to vector type
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002113 definitions (a comma separated list of elements, surrounded by
2114 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32
2115 42, i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must
2116 have <a href="#t_vector">vector type</a>, and the number and types of
2117 elements must match those specified by the type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +00002118
2119 <dt><b>Zero initialization</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +00002120 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
Chris Lattner392be582010-02-12 20:49:41 +00002121 value to zero of <em>any</em> type, including scalar and
2122 <a href="#t_aggregate">aggregate</a> types.
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002123 This is often used to avoid having to print large zero initializers
2124 (e.g. for large arrays) and is always exactly equivalent to using explicit
2125 zero initializers.</dd>
Nick Lewycky49f89192009-04-04 07:22:01 +00002126
2127 <dt><b>Metadata node</b></dt>
Nick Lewycky8e2c4f42009-05-30 16:08:30 +00002128 <dd>A metadata node is a structure-like constant with
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002129 <a href="#t_metadata">metadata type</a>. For example: "<tt>metadata !{
2130 i32 0, metadata !"test" }</tt>". Unlike other constants that are meant to
2131 be interpreted as part of the instruction stream, metadata is a place to
2132 attach additional information such as debug info.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +00002133</dl>
2134
2135</div>
2136
2137<!-- ======================================================================= -->
2138<div class="doc_subsection">
2139 <a name="globalconstants">Global Variable and Function Addresses</a>
2140</div>
2141
2142<div class="doc_text">
2143
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002144<p>The addresses of <a href="#globalvars">global variables</a>
2145 and <a href="#functionstructure">functions</a> are always implicitly valid
2146 (link-time) constants. These constants are explicitly referenced when
2147 the <a href="#identifiers">identifier for the global</a> is used and always
2148 have <a href="#t_pointer">pointer</a> type. For example, the following is a
2149 legal LLVM file:</p>
Chris Lattner74d3f822004-12-09 17:30:23 +00002150
Benjamin Kramer79698be2010-07-13 12:26:09 +00002151<pre class="doc_code">
Chris Lattner00538a12007-06-06 18:28:13 +00002152@X = global i32 17
2153@Y = global i32 42
2154@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
Chris Lattner74d3f822004-12-09 17:30:23 +00002155</pre>
2156
2157</div>
2158
2159<!-- ======================================================================= -->
Reid Spencer641f5c92004-12-09 18:13:12 +00002160<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
Chris Lattner74d3f822004-12-09 17:30:23 +00002161<div class="doc_text">
2162
Chris Lattnerec72b9b2009-09-07 22:52:39 +00002163<p>The string '<tt>undef</tt>' can be used anywhere a constant is expected, and
Benjamin Kramer0f420382009-10-12 14:46:08 +00002164 indicates that the user of the value may receive an unspecified bit-pattern.
Bill Wendling6bbe0912010-10-27 01:07:41 +00002165 Undefined values may be of any type (other than '<tt>label</tt>'
2166 or '<tt>void</tt>') and be used anywhere a constant is permitted.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002167
Chris Lattner92ada5d2009-09-11 01:49:31 +00002168<p>Undefined values are useful because they indicate to the compiler that the
Chris Lattnerec72b9b2009-09-07 22:52:39 +00002169 program is well defined no matter what value is used. This gives the
2170 compiler more freedom to optimize. Here are some examples of (potentially
2171 surprising) transformations that are valid (in pseudo IR):</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002172
Chris Lattnerec72b9b2009-09-07 22:52:39 +00002173
Benjamin Kramer79698be2010-07-13 12:26:09 +00002174<pre class="doc_code">
Chris Lattnerec72b9b2009-09-07 22:52:39 +00002175 %A = add %X, undef
2176 %B = sub %X, undef
2177 %C = xor %X, undef
2178Safe:
2179 %A = undef
2180 %B = undef
2181 %C = undef
2182</pre>
Chris Lattnerec72b9b2009-09-07 22:52:39 +00002183
2184<p>This is safe because all of the output bits are affected by the undef bits.
Bill Wendling6bbe0912010-10-27 01:07:41 +00002185 Any output bit can have a zero or one depending on the input bits.</p>
Chris Lattnerec72b9b2009-09-07 22:52:39 +00002186
Benjamin Kramer79698be2010-07-13 12:26:09 +00002187<pre class="doc_code">
Chris Lattnerec72b9b2009-09-07 22:52:39 +00002188 %A = or %X, undef
2189 %B = and %X, undef
2190Safe:
2191 %A = -1
2192 %B = 0
2193Unsafe:
2194 %A = undef
2195 %B = undef
2196</pre>
Chris Lattnerec72b9b2009-09-07 22:52:39 +00002197
2198<p>These logical operations have bits that are not always affected by the input.
Bill Wendling6bbe0912010-10-27 01:07:41 +00002199 For example, if <tt>%X</tt> has a zero bit, then the output of the
2200 '<tt>and</tt>' operation will always be a zero for that bit, no matter what
2201 the corresponding bit from the '<tt>undef</tt>' is. As such, it is unsafe to
2202 optimize or assume that the result of the '<tt>and</tt>' is '<tt>undef</tt>'.
2203 However, it is safe to assume that all bits of the '<tt>undef</tt>' could be
2204 0, and optimize the '<tt>and</tt>' to 0. Likewise, it is safe to assume that
2205 all the bits of the '<tt>undef</tt>' operand to the '<tt>or</tt>' could be
2206 set, allowing the '<tt>or</tt>' to be folded to -1.</p>
Chris Lattnerec72b9b2009-09-07 22:52:39 +00002207
Benjamin Kramer79698be2010-07-13 12:26:09 +00002208<pre class="doc_code">
Chris Lattnerec72b9b2009-09-07 22:52:39 +00002209 %A = select undef, %X, %Y
2210 %B = select undef, 42, %Y
2211 %C = select %X, %Y, undef
2212Safe:
2213 %A = %X (or %Y)
2214 %B = 42 (or %Y)
2215 %C = %Y
2216Unsafe:
2217 %A = undef
2218 %B = undef
2219 %C = undef
2220</pre>
Chris Lattnerec72b9b2009-09-07 22:52:39 +00002221
Bill Wendling6bbe0912010-10-27 01:07:41 +00002222<p>This set of examples shows that undefined '<tt>select</tt>' (and conditional
2223 branch) conditions can go <em>either way</em>, but they have to come from one
2224 of the two operands. In the <tt>%A</tt> example, if <tt>%X</tt> and
2225 <tt>%Y</tt> were both known to have a clear low bit, then <tt>%A</tt> would
2226 have to have a cleared low bit. However, in the <tt>%C</tt> example, the
2227 optimizer is allowed to assume that the '<tt>undef</tt>' operand could be the
2228 same as <tt>%Y</tt>, allowing the whole '<tt>select</tt>' to be
2229 eliminated.</p>
Chris Lattnerec72b9b2009-09-07 22:52:39 +00002230
Benjamin Kramer79698be2010-07-13 12:26:09 +00002231<pre class="doc_code">
Chris Lattnerec72b9b2009-09-07 22:52:39 +00002232 %A = xor undef, undef
Eric Christopher455c5772009-12-05 02:46:03 +00002233
Chris Lattnerec72b9b2009-09-07 22:52:39 +00002234 %B = undef
2235 %C = xor %B, %B
2236
2237 %D = undef
2238 %E = icmp lt %D, 4
2239 %F = icmp gte %D, 4
2240
2241Safe:
2242 %A = undef
2243 %B = undef
2244 %C = undef
2245 %D = undef
2246 %E = undef
2247 %F = undef
2248</pre>
Chris Lattnerec72b9b2009-09-07 22:52:39 +00002249
Bill Wendling6bbe0912010-10-27 01:07:41 +00002250<p>This example points out that two '<tt>undef</tt>' operands are not
2251 necessarily the same. This can be surprising to people (and also matches C
2252 semantics) where they assume that "<tt>X^X</tt>" is always zero, even
2253 if <tt>X</tt> is undefined. This isn't true for a number of reasons, but the
2254 short answer is that an '<tt>undef</tt>' "variable" can arbitrarily change
2255 its value over its "live range". This is true because the variable doesn't
2256 actually <em>have a live range</em>. Instead, the value is logically read
2257 from arbitrary registers that happen to be around when needed, so the value
2258 is not necessarily consistent over time. In fact, <tt>%A</tt> and <tt>%C</tt>
2259 need to have the same semantics or the core LLVM "replace all uses with"
2260 concept would not hold.</p>
Chris Lattnera34a7182009-09-07 23:33:52 +00002261
Benjamin Kramer79698be2010-07-13 12:26:09 +00002262<pre class="doc_code">
Chris Lattnera34a7182009-09-07 23:33:52 +00002263 %A = fdiv undef, %X
2264 %B = fdiv %X, undef
2265Safe:
2266 %A = undef
2267b: unreachable
2268</pre>
Chris Lattnera34a7182009-09-07 23:33:52 +00002269
2270<p>These examples show the crucial difference between an <em>undefined
Bill Wendling6bbe0912010-10-27 01:07:41 +00002271 value</em> and <em>undefined behavior</em>. An undefined value (like
2272 '<tt>undef</tt>') is allowed to have an arbitrary bit-pattern. This means that
2273 the <tt>%A</tt> operation can be constant folded to '<tt>undef</tt>', because
2274 the '<tt>undef</tt>' could be an SNaN, and <tt>fdiv</tt> is not (currently)
2275 defined on SNaN's. However, in the second example, we can make a more
2276 aggressive assumption: because the <tt>undef</tt> is allowed to be an
2277 arbitrary value, we are allowed to assume that it could be zero. Since a
2278 divide by zero has <em>undefined behavior</em>, we are allowed to assume that
2279 the operation does not execute at all. This allows us to delete the divide and
2280 all code after it. Because the undefined operation "can't happen", the
2281 optimizer can assume that it occurs in dead code.</p>
Eric Christopher455c5772009-12-05 02:46:03 +00002282
Benjamin Kramer79698be2010-07-13 12:26:09 +00002283<pre class="doc_code">
Chris Lattnera34a7182009-09-07 23:33:52 +00002284a: store undef -> %X
2285b: store %X -> undef
2286Safe:
2287a: &lt;deleted&gt;
2288b: unreachable
2289</pre>
Chris Lattnera34a7182009-09-07 23:33:52 +00002290
Bill Wendling6bbe0912010-10-27 01:07:41 +00002291<p>These examples reiterate the <tt>fdiv</tt> example: a store <em>of</em> an
2292 undefined value can be assumed to not have any effect; we can assume that the
2293 value is overwritten with bits that happen to match what was already there.
2294 However, a store <em>to</em> an undefined location could clobber arbitrary
2295 memory, therefore, it has undefined behavior.</p>
Chris Lattnera34a7182009-09-07 23:33:52 +00002296
Chris Lattner74d3f822004-12-09 17:30:23 +00002297</div>
2298
2299<!-- ======================================================================= -->
Dan Gohmanffc9a6b2010-04-22 23:14:21 +00002300<div class="doc_subsection"><a name="trapvalues">Trap Values</a></div>
2301<div class="doc_text">
2302
Dan Gohmanb2a709b2010-04-26 20:21:21 +00002303<p>Trap values are similar to <a href="#undefvalues">undef values</a>, however
Dan Gohmanffc9a6b2010-04-22 23:14:21 +00002304 instead of representing an unspecified bit pattern, they represent the
2305 fact that an instruction or constant expression which cannot evoke side
2306 effects has nevertheless detected a condition which results in undefined
Dan Gohmanb2a709b2010-04-26 20:21:21 +00002307 behavior.</p>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +00002308
Dan Gohman2f1ae062010-04-28 00:49:41 +00002309<p>There is currently no way of representing a trap value in the IR; they
Dan Gohmanac355aa2010-05-03 14:51:43 +00002310 only exist when produced by operations such as
Dan Gohman2f1ae062010-04-28 00:49:41 +00002311 <a href="#i_add"><tt>add</tt></a> with the <tt>nsw</tt> flag.</p>
Dan Gohmanb8b85c12010-04-26 23:36:52 +00002312
Dan Gohman2f1ae062010-04-28 00:49:41 +00002313<p>Trap value behavior is defined in terms of value <i>dependence</i>:</p>
Dan Gohmanb8b85c12010-04-26 23:36:52 +00002314
Dan Gohman2f1ae062010-04-28 00:49:41 +00002315<ul>
2316<li>Values other than <a href="#i_phi"><tt>phi</tt></a> nodes depend on
2317 their operands.</li>
2318
2319<li><a href="#i_phi"><tt>Phi</tt></a> nodes depend on the operand corresponding
2320 to their dynamic predecessor basic block.</li>
2321
2322<li>Function arguments depend on the corresponding actual argument values in
2323 the dynamic callers of their functions.</li>
2324
2325<li><a href="#i_call"><tt>Call</tt></a> instructions depend on the
2326 <a href="#i_ret"><tt>ret</tt></a> instructions that dynamically transfer
2327 control back to them.</li>
2328
Dan Gohman7292a752010-05-03 14:55:22 +00002329<li><a href="#i_invoke"><tt>Invoke</tt></a> instructions depend on the
2330 <a href="#i_ret"><tt>ret</tt></a>, <a href="#i_unwind"><tt>unwind</tt></a>,
2331 or exception-throwing call instructions that dynamically transfer control
2332 back to them.</li>
2333
Dan Gohman2f1ae062010-04-28 00:49:41 +00002334<li>Non-volatile loads and stores depend on the most recent stores to all of the
2335 referenced memory addresses, following the order in the IR
2336 (including loads and stores implied by intrinsics such as
2337 <a href="#int_memcpy"><tt>@llvm.memcpy</tt></a>.)</li>
2338
Dan Gohman3513ea52010-05-03 14:59:34 +00002339<!-- TODO: In the case of multiple threads, this only applies if the store
2340 "happens-before" the load or store. -->
Dan Gohmanb8b85c12010-04-26 23:36:52 +00002341
Dan Gohman2f1ae062010-04-28 00:49:41 +00002342<!-- TODO: floating-point exception state -->
Dan Gohmanb8b85c12010-04-26 23:36:52 +00002343
Dan Gohman2f1ae062010-04-28 00:49:41 +00002344<li>An instruction with externally visible side effects depends on the most
2345 recent preceding instruction with externally visible side effects, following
Dan Gohman6c858db2010-07-06 15:26:33 +00002346 the order in the IR. (This includes
2347 <a href="#volatile">volatile operations</a>.)</li>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +00002348
Dan Gohman7292a752010-05-03 14:55:22 +00002349<li>An instruction <i>control-depends</i> on a
2350 <a href="#terminators">terminator instruction</a>
2351 if the terminator instruction has multiple successors and the instruction
2352 is always executed when control transfers to one of the successors, and
2353 may not be executed when control is transfered to another.</li>
Dan Gohman2f1ae062010-04-28 00:49:41 +00002354
2355<li>Dependence is transitive.</li>
2356
2357</ul>
Dan Gohman2f1ae062010-04-28 00:49:41 +00002358
2359<p>Whenever a trap value is generated, all values which depend on it evaluate
2360 to trap. If they have side effects, the evoke their side effects as if each
2361 operand with a trap value were undef. If they have externally-visible side
2362 effects, the behavior is undefined.</p>
2363
2364<p>Here are some examples:</p>
Dan Gohman48a25882010-04-26 20:54:53 +00002365
Benjamin Kramer79698be2010-07-13 12:26:09 +00002366<pre class="doc_code">
Dan Gohmanb8b85c12010-04-26 23:36:52 +00002367entry:
2368 %trap = sub nuw i32 0, 1 ; Results in a trap value.
Dan Gohman2f1ae062010-04-28 00:49:41 +00002369 %still_trap = and i32 %trap, 0 ; Whereas (and i32 undef, 0) would return 0.
2370 %trap_yet_again = getelementptr i32* @h, i32 %still_trap
2371 store i32 0, i32* %trap_yet_again ; undefined behavior
2372
2373 store i32 %trap, i32* @g ; Trap value conceptually stored to memory.
2374 %trap2 = load i32* @g ; Returns a trap value, not just undef.
2375
2376 volatile store i32 %trap, i32* @g ; External observation; undefined behavior.
2377
2378 %narrowaddr = bitcast i32* @g to i16*
2379 %wideaddr = bitcast i32* @g to i64*
2380 %trap3 = load 16* %narrowaddr ; Returns a trap value.
2381 %trap4 = load i64* %widaddr ; Returns a trap value.
2382
2383 %cmp = icmp i32 slt %trap, 0 ; Returns a trap value.
Dan Gohmanb8b85c12010-04-26 23:36:52 +00002384 %br i1 %cmp, %true, %end ; Branch to either destination.
2385
2386true:
Dan Gohman2f1ae062010-04-28 00:49:41 +00002387 volatile store i32 0, i32* @g ; This is control-dependent on %cmp, so
2388 ; it has undefined behavior.
Dan Gohmanb8b85c12010-04-26 23:36:52 +00002389 br label %end
2390
2391end:
2392 %p = phi i32 [ 0, %entry ], [ 1, %true ]
2393 ; Both edges into this PHI are
2394 ; control-dependent on %cmp, so this
Dan Gohman2f1ae062010-04-28 00:49:41 +00002395 ; always results in a trap value.
Dan Gohmanb8b85c12010-04-26 23:36:52 +00002396
2397 volatile store i32 0, i32* @g ; %end is control-equivalent to %entry
2398 ; so this is defined (ignoring earlier
2399 ; undefined behavior in this example).
Dan Gohmanb8b85c12010-04-26 23:36:52 +00002400</pre>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +00002401
Dan Gohmanffc9a6b2010-04-22 23:14:21 +00002402</div>
2403
2404<!-- ======================================================================= -->
Chris Lattner2bfd3202009-10-27 21:19:13 +00002405<div class="doc_subsection"><a name="blockaddress">Addresses of Basic
2406 Blocks</a></div>
Chris Lattnere4801f72009-10-27 21:01:34 +00002407<div class="doc_text">
2408
Chris Lattneraa99c942009-11-01 01:27:45 +00002409<p><b><tt>blockaddress(@function, %block)</tt></b></p>
Chris Lattnere4801f72009-10-27 21:01:34 +00002410
2411<p>The '<tt>blockaddress</tt>' constant computes the address of the specified
Chris Lattner5c5f0ac2009-10-27 21:49:40 +00002412 basic block in the specified function, and always has an i8* type. Taking
Chris Lattneraa99c942009-11-01 01:27:45 +00002413 the address of the entry block is illegal.</p>
Eric Christopher455c5772009-12-05 02:46:03 +00002414
Chris Lattnere4801f72009-10-27 21:01:34 +00002415<p>This value only has defined behavior when used as an operand to the
Bill Wendling6bbe0912010-10-27 01:07:41 +00002416 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>' instruction, or for
2417 comparisons against null. Pointer equality tests between labels addresses
2418 results in undefined behavior &mdash; though, again, comparison against null
2419 is ok, and no label is equal to the null pointer. This may be passed around
2420 as an opaque pointer sized value as long as the bits are not inspected. This
2421 allows <tt>ptrtoint</tt> and arithmetic to be performed on these values so
2422 long as the original value is reconstituted before the <tt>indirectbr</tt>
2423 instruction.</p>
Eric Christopher455c5772009-12-05 02:46:03 +00002424
Bill Wendling6bbe0912010-10-27 01:07:41 +00002425<p>Finally, some targets may provide defined semantics when using the value as
2426 the operand to an inline assembly, but that is target specific.</p>
Chris Lattnere4801f72009-10-27 21:01:34 +00002427
2428</div>
2429
2430
2431<!-- ======================================================================= -->
Chris Lattner74d3f822004-12-09 17:30:23 +00002432<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
2433</div>
2434
2435<div class="doc_text">
2436
2437<p>Constant expressions are used to allow expressions involving other constants
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002438 to be used as constants. Constant expressions may be of
2439 any <a href="#t_firstclass">first class</a> type and may involve any LLVM
2440 operation that does not have side effects (e.g. load and call are not
Bill Wendling6bbe0912010-10-27 01:07:41 +00002441 supported). The following is the syntax for constant expressions:</p>
Chris Lattner74d3f822004-12-09 17:30:23 +00002442
2443<dl>
Dan Gohmand6a6f612010-05-28 17:07:41 +00002444 <dt><b><tt>trunc (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002445 <dd>Truncate a constant to another type. The bit size of CST must be larger
2446 than the bit size of TYPE. Both types must be integers.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +00002447
Dan Gohmand6a6f612010-05-28 17:07:41 +00002448 <dt><b><tt>zext (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002449 <dd>Zero extend a constant to another type. The bit size of CST must be
Duncan Sandsa522e562010-07-13 12:06:14 +00002450 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00002451
Dan Gohmand6a6f612010-05-28 17:07:41 +00002452 <dt><b><tt>sext (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002453 <dd>Sign extend a constant to another type. The bit size of CST must be
Duncan Sandsa522e562010-07-13 12:06:14 +00002454 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00002455
Dan Gohmand6a6f612010-05-28 17:07:41 +00002456 <dt><b><tt>fptrunc (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002457 <dd>Truncate a floating point constant to another floating point type. The
2458 size of CST must be larger than the size of TYPE. Both types must be
2459 floating point.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00002460
Dan Gohmand6a6f612010-05-28 17:07:41 +00002461 <dt><b><tt>fpext (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002462 <dd>Floating point extend a constant to another type. The size of CST must be
2463 smaller or equal to the size of TYPE. Both types must be floating
2464 point.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00002465
Dan Gohmand6a6f612010-05-28 17:07:41 +00002466 <dt><b><tt>fptoui (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00002467 <dd>Convert a floating point constant to the corresponding unsigned integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002468 constant. TYPE must be a scalar or vector integer type. CST must be of
2469 scalar or vector floating point type. Both CST and TYPE must be scalars,
2470 or vectors of the same number of elements. If the value won't fit in the
2471 integer type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00002472
Dan Gohmand6a6f612010-05-28 17:07:41 +00002473 <dt><b><tt>fptosi (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00002474 <dd>Convert a floating point constant to the corresponding signed integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002475 constant. TYPE must be a scalar or vector integer type. CST must be of
2476 scalar or vector floating point type. Both CST and TYPE must be scalars,
2477 or vectors of the same number of elements. If the value won't fit in the
2478 integer type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00002479
Dan Gohmand6a6f612010-05-28 17:07:41 +00002480 <dt><b><tt>uitofp (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00002481 <dd>Convert an unsigned integer constant to the corresponding floating point
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002482 constant. TYPE must be a scalar or vector floating point type. CST must be
2483 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2484 vectors of the same number of elements. If the value won't fit in the
2485 floating point type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00002486
Dan Gohmand6a6f612010-05-28 17:07:41 +00002487 <dt><b><tt>sitofp (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00002488 <dd>Convert a signed integer constant to the corresponding floating point
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002489 constant. TYPE must be a scalar or vector floating point type. CST must be
2490 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2491 vectors of the same number of elements. If the value won't fit in the
2492 floating point type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00002493
Dan Gohmand6a6f612010-05-28 17:07:41 +00002494 <dt><b><tt>ptrtoint (CST to TYPE)</tt></b></dt>
Reid Spencer5b950642006-11-11 23:08:07 +00002495 <dd>Convert a pointer typed constant to the corresponding integer constant
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002496 <tt>TYPE</tt> must be an integer type. <tt>CST</tt> must be of pointer
2497 type. The <tt>CST</tt> value is zero extended, truncated, or unchanged to
2498 make it fit in <tt>TYPE</tt>.</dd>
Reid Spencer5b950642006-11-11 23:08:07 +00002499
Dan Gohmand6a6f612010-05-28 17:07:41 +00002500 <dt><b><tt>inttoptr (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002501 <dd>Convert a integer constant to a pointer constant. TYPE must be a pointer
2502 type. CST must be of integer type. The CST value is zero extended,
2503 truncated, or unchanged to make it fit in a pointer size. This one is
2504 <i>really</i> dangerous!</dd>
Reid Spencer5b950642006-11-11 23:08:07 +00002505
Dan Gohmand6a6f612010-05-28 17:07:41 +00002506 <dt><b><tt>bitcast (CST to TYPE)</tt></b></dt>
Chris Lattner789dee32009-02-28 18:27:03 +00002507 <dd>Convert a constant, CST, to another TYPE. The constraints of the operands
2508 are the same as those for the <a href="#i_bitcast">bitcast
2509 instruction</a>.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +00002510
Dan Gohmand6a6f612010-05-28 17:07:41 +00002511 <dt><b><tt>getelementptr (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
2512 <dt><b><tt>getelementptr inbounds (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +00002513 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002514 constants. As with the <a href="#i_getelementptr">getelementptr</a>
2515 instruction, the index list may have zero or more indexes, which are
2516 required to make sense for the type of "CSTPTR".</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +00002517
Dan Gohmand6a6f612010-05-28 17:07:41 +00002518 <dt><b><tt>select (COND, VAL1, VAL2)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002519 <dd>Perform the <a href="#i_select">select operation</a> on constants.</dd>
Reid Spencer9965ee72006-12-04 19:23:19 +00002520
Dan Gohmand6a6f612010-05-28 17:07:41 +00002521 <dt><b><tt>icmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer9965ee72006-12-04 19:23:19 +00002522 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
2523
Dan Gohmand6a6f612010-05-28 17:07:41 +00002524 <dt><b><tt>fcmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer9965ee72006-12-04 19:23:19 +00002525 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
Robert Bocchino7e97a6d2006-01-10 19:31:34 +00002526
Dan Gohmand6a6f612010-05-28 17:07:41 +00002527 <dt><b><tt>extractelement (VAL, IDX)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002528 <dd>Perform the <a href="#i_extractelement">extractelement operation</a> on
2529 constants.</dd>
Robert Bocchino7e97a6d2006-01-10 19:31:34 +00002530
Dan Gohmand6a6f612010-05-28 17:07:41 +00002531 <dt><b><tt>insertelement (VAL, ELT, IDX)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002532 <dd>Perform the <a href="#i_insertelement">insertelement operation</a> on
2533 constants.</dd>
Chris Lattner016a0e52006-04-08 00:13:41 +00002534
Dan Gohmand6a6f612010-05-28 17:07:41 +00002535 <dt><b><tt>shufflevector (VEC1, VEC2, IDXMASK)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002536 <dd>Perform the <a href="#i_shufflevector">shufflevector operation</a> on
2537 constants.</dd>
Chris Lattner016a0e52006-04-08 00:13:41 +00002538
Nick Lewycky9ab9a7f2010-05-29 06:44:15 +00002539 <dt><b><tt>extractvalue (VAL, IDX0, IDX1, ...)</tt></b></dt>
2540 <dd>Perform the <a href="#i_extractvalue">extractvalue operation</a> on
2541 constants. The index list is interpreted in a similar manner as indices in
2542 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2543 index value must be specified.</dd>
2544
2545 <dt><b><tt>insertvalue (VAL, ELT, IDX0, IDX1, ...)</tt></b></dt>
2546 <dd>Perform the <a href="#i_insertvalue">insertvalue operation</a> on
2547 constants. The index list is interpreted in a similar manner as indices in
2548 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2549 index value must be specified.</dd>
2550
Dan Gohmand6a6f612010-05-28 17:07:41 +00002551 <dt><b><tt>OPCODE (LHS, RHS)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002552 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
2553 be any of the <a href="#binaryops">binary</a>
2554 or <a href="#bitwiseops">bitwise binary</a> operations. The constraints
2555 on operands are the same as those for the corresponding instruction
2556 (e.g. no bitwise operations on floating point values are allowed).</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +00002557</dl>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002558
Chris Lattner74d3f822004-12-09 17:30:23 +00002559</div>
Chris Lattnerb1652612004-03-08 16:49:10 +00002560
Chris Lattner2f7c9632001-06-06 20:29:01 +00002561<!-- *********************************************************************** -->
Chris Lattner98f013c2006-01-25 23:47:57 +00002562<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
2563<!-- *********************************************************************** -->
2564
2565<!-- ======================================================================= -->
2566<div class="doc_subsection">
2567<a name="inlineasm">Inline Assembler Expressions</a>
2568</div>
2569
2570<div class="doc_text">
2571
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002572<p>LLVM supports inline assembler expressions (as opposed
2573 to <a href="#moduleasm"> Module-Level Inline Assembly</a>) through the use of
2574 a special value. This value represents the inline assembler as a string
2575 (containing the instructions to emit), a list of operand constraints (stored
Dale Johannesen63c94fe2009-10-13 21:56:55 +00002576 as a string), a flag that indicates whether or not the inline asm
Dale Johannesen1cfb9582009-10-21 23:28:00 +00002577 expression has side effects, and a flag indicating whether the function
2578 containing the asm needs to align its stack conservatively. An example
2579 inline assembler expression is:</p>
Chris Lattner98f013c2006-01-25 23:47:57 +00002580
Benjamin Kramer79698be2010-07-13 12:26:09 +00002581<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +00002582i32 (i32) asm "bswap $0", "=r,r"
Chris Lattner98f013c2006-01-25 23:47:57 +00002583</pre>
2584
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002585<p>Inline assembler expressions may <b>only</b> be used as the callee operand of
2586 a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we
2587 have:</p>
Chris Lattner98f013c2006-01-25 23:47:57 +00002588
Benjamin Kramer79698be2010-07-13 12:26:09 +00002589<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +00002590%X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
Chris Lattner98f013c2006-01-25 23:47:57 +00002591</pre>
2592
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002593<p>Inline asms with side effects not visible in the constraint list must be
2594 marked as having side effects. This is done through the use of the
2595 '<tt>sideeffect</tt>' keyword, like so:</p>
Chris Lattner98f013c2006-01-25 23:47:57 +00002596
Benjamin Kramer79698be2010-07-13 12:26:09 +00002597<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +00002598call void asm sideeffect "eieio", ""()
Chris Lattner98f013c2006-01-25 23:47:57 +00002599</pre>
2600
Dale Johannesen1cfb9582009-10-21 23:28:00 +00002601<p>In some cases inline asms will contain code that will not work unless the
2602 stack is aligned in some way, such as calls or SSE instructions on x86,
2603 yet will not contain code that does that alignment within the asm.
2604 The compiler should make conservative assumptions about what the asm might
2605 contain and should generate its usual stack alignment code in the prologue
2606 if the '<tt>alignstack</tt>' keyword is present:</p>
Dale Johannesen63c94fe2009-10-13 21:56:55 +00002607
Benjamin Kramer79698be2010-07-13 12:26:09 +00002608<pre class="doc_code">
Dale Johannesen1cfb9582009-10-21 23:28:00 +00002609call void asm alignstack "eieio", ""()
Dale Johannesen63c94fe2009-10-13 21:56:55 +00002610</pre>
Dale Johannesen63c94fe2009-10-13 21:56:55 +00002611
2612<p>If both keywords appear the '<tt>sideeffect</tt>' keyword must come
2613 first.</p>
2614
Chris Lattner98f013c2006-01-25 23:47:57 +00002615<p>TODO: The format of the asm and constraints string still need to be
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002616 documented here. Constraints on what can be done (e.g. duplication, moving,
2617 etc need to be documented). This is probably best done by reference to
2618 another document that covers inline asm from a holistic perspective.</p>
Chris Lattner51065562010-04-07 05:38:05 +00002619</div>
2620
2621<div class="doc_subsubsection">
2622<a name="inlineasm_md">Inline Asm Metadata</a>
2623</div>
2624
2625<div class="doc_text">
2626
2627<p>The call instructions that wrap inline asm nodes may have a "!srcloc" MDNode
Chris Lattner79ffdc72010-11-17 08:20:42 +00002628 attached to it that contains a list of constant integers. If present, the
2629 code generator will use the integer as the location cookie value when report
Chris Lattner51065562010-04-07 05:38:05 +00002630 errors through the LLVMContext error reporting mechanisms. This allows a
Dan Gohman61110ae2010-04-28 00:36:01 +00002631 front-end to correlate backend errors that occur with inline asm back to the
Chris Lattner51065562010-04-07 05:38:05 +00002632 source code that produced it. For example:</p>
2633
Benjamin Kramer79698be2010-07-13 12:26:09 +00002634<pre class="doc_code">
Chris Lattner51065562010-04-07 05:38:05 +00002635call void asm sideeffect "something bad", ""()<b>, !srcloc !42</b>
2636...
2637!42 = !{ i32 1234567 }
2638</pre>
Chris Lattner51065562010-04-07 05:38:05 +00002639
2640<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 +00002641 IR. If the MDNode contains multiple constants, the code generator will use
2642 the one that corresponds to the line of the asm that the error occurs on.</p>
Chris Lattner98f013c2006-01-25 23:47:57 +00002643
2644</div>
2645
Chris Lattnerc2f8f162010-01-15 21:50:19 +00002646<!-- ======================================================================= -->
2647<div class="doc_subsection"><a name="metadata">Metadata Nodes and Metadata
2648 Strings</a>
2649</div>
2650
2651<div class="doc_text">
2652
2653<p>LLVM IR allows metadata to be attached to instructions in the program that
2654 can convey extra information about the code to the optimizers and code
2655 generator. One example application of metadata is source-level debug
2656 information. There are two metadata primitives: strings and nodes. All
2657 metadata has the <tt>metadata</tt> type and is identified in syntax by a
2658 preceding exclamation point ('<tt>!</tt>').</p>
2659
2660<p>A metadata string is a string surrounded by double quotes. It can contain
2661 any character by escaping non-printable characters with "\xx" where "xx" is
2662 the two digit hex code. For example: "<tt>!"test\00"</tt>".</p>
2663
2664<p>Metadata nodes are represented with notation similar to structure constants
2665 (a comma separated list of elements, surrounded by braces and preceded by an
2666 exclamation point). For example: "<tt>!{ metadata !"test\00", i32
2667 10}</tt>". Metadata nodes can have any values as their operand.</p>
2668
2669<p>A <a href="#namedmetadatastructure">named metadata</a> is a collection of
2670 metadata nodes, which can be looked up in the module symbol table. For
2671 example: "<tt>!foo = metadata !{!4, !3}</tt>".
2672
Devang Patel9984bd62010-03-04 23:44:48 +00002673<p>Metadata can be used as function arguments. Here <tt>llvm.dbg.value</tt>
Benjamin Kramer79698be2010-07-13 12:26:09 +00002674 function is using two metadata arguments.</p>
Devang Patel9984bd62010-03-04 23:44:48 +00002675
Benjamin Kramer79698be2010-07-13 12:26:09 +00002676 <pre class="doc_code">
Devang Patel9984bd62010-03-04 23:44:48 +00002677 call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2678 </pre>
Devang Patel9984bd62010-03-04 23:44:48 +00002679
2680<p>Metadata can be attached with an instruction. Here metadata <tt>!21</tt> is
Benjamin Kramer79698be2010-07-13 12:26:09 +00002681 attached with <tt>add</tt> instruction using <tt>!dbg</tt> identifier.</p>
Devang Patel9984bd62010-03-04 23:44:48 +00002682
Benjamin Kramer79698be2010-07-13 12:26:09 +00002683 <pre class="doc_code">
Devang Patel9984bd62010-03-04 23:44:48 +00002684 %indvar.next = add i64 %indvar, 1, !dbg !21
2685 </pre>
Chris Lattnerc2f8f162010-01-15 21:50:19 +00002686</div>
2687
Chris Lattnerae76db52009-07-20 05:55:19 +00002688
2689<!-- *********************************************************************** -->
2690<div class="doc_section">
2691 <a name="intrinsic_globals">Intrinsic Global Variables</a>
2692</div>
2693<!-- *********************************************************************** -->
2694
2695<p>LLVM has a number of "magic" global variables that contain data that affect
2696code generation or other IR semantics. These are documented here. All globals
Chris Lattner58f9bb22009-07-20 06:14:25 +00002697of this sort should have a section specified as "<tt>llvm.metadata</tt>". This
2698section and all globals that start with "<tt>llvm.</tt>" are reserved for use
2699by LLVM.</p>
Chris Lattnerae76db52009-07-20 05:55:19 +00002700
2701<!-- ======================================================================= -->
2702<div class="doc_subsection">
2703<a name="intg_used">The '<tt>llvm.used</tt>' Global Variable</a>
2704</div>
2705
2706<div class="doc_text">
2707
2708<p>The <tt>@llvm.used</tt> global is an array with i8* element type which has <a
2709href="#linkage_appending">appending linkage</a>. This array contains a list of
2710pointers to global variables and functions which may optionally have a pointer
2711cast formed of bitcast or getelementptr. For example, a legal use of it is:</p>
2712
2713<pre>
2714 @X = global i8 4
2715 @Y = global i32 123
2716
2717 @llvm.used = appending global [2 x i8*] [
2718 i8* @X,
2719 i8* bitcast (i32* @Y to i8*)
2720 ], section "llvm.metadata"
2721</pre>
2722
2723<p>If a global variable appears in the <tt>@llvm.used</tt> list, then the
2724compiler, assembler, and linker are required to treat the symbol as if there is
2725a reference to the global that it cannot see. For example, if a variable has
2726internal linkage and no references other than that from the <tt>@llvm.used</tt>
2727list, it cannot be deleted. This is commonly used to represent references from
2728inline asms and other things the compiler cannot "see", and corresponds to
2729"attribute((used))" in GNU C.</p>
2730
2731<p>On some targets, the code generator must emit a directive to the assembler or
2732object file to prevent the assembler and linker from molesting the symbol.</p>
2733
2734</div>
2735
2736<!-- ======================================================================= -->
2737<div class="doc_subsection">
Chris Lattner58f9bb22009-07-20 06:14:25 +00002738<a name="intg_compiler_used">The '<tt>llvm.compiler.used</tt>' Global Variable</a>
2739</div>
2740
2741<div class="doc_text">
2742
2743<p>The <tt>@llvm.compiler.used</tt> directive is the same as the
2744<tt>@llvm.used</tt> directive, except that it only prevents the compiler from
2745touching the symbol. On targets that support it, this allows an intelligent
2746linker to optimize references to the symbol without being impeded as it would be
2747by <tt>@llvm.used</tt>.</p>
2748
2749<p>This is a rare construct that should only be used in rare circumstances, and
2750should not be exposed to source languages.</p>
2751
2752</div>
2753
2754<!-- ======================================================================= -->
2755<div class="doc_subsection">
Chris Lattnerae76db52009-07-20 05:55:19 +00002756<a name="intg_global_ctors">The '<tt>llvm.global_ctors</tt>' Global Variable</a>
2757</div>
2758
2759<div class="doc_text">
David Chisnalla9d4a6f2010-04-30 19:23:49 +00002760<pre>
2761%0 = type { i32, void ()* }
David Chisnallb492b812010-04-30 19:27:35 +00002762@llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor }]
David Chisnalla9d4a6f2010-04-30 19:23:49 +00002763</pre>
2764<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.
2765</p>
Chris Lattnerae76db52009-07-20 05:55:19 +00002766
2767</div>
2768
2769<!-- ======================================================================= -->
2770<div class="doc_subsection">
2771<a name="intg_global_dtors">The '<tt>llvm.global_dtors</tt>' Global Variable</a>
2772</div>
2773
2774<div class="doc_text">
David Chisnalla9d4a6f2010-04-30 19:23:49 +00002775<pre>
2776%0 = type { i32, void ()* }
David Chisnallb492b812010-04-30 19:27:35 +00002777@llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor }]
David Chisnalla9d4a6f2010-04-30 19:23:49 +00002778</pre>
Chris Lattnerae76db52009-07-20 05:55:19 +00002779
David Chisnalla9d4a6f2010-04-30 19:23:49 +00002780<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.
2781</p>
Chris Lattnerae76db52009-07-20 05:55:19 +00002782
2783</div>
2784
2785
Chris Lattner98f013c2006-01-25 23:47:57 +00002786<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +00002787<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
2788<!-- *********************************************************************** -->
Chris Lattner74d3f822004-12-09 17:30:23 +00002789
Misha Brukman76307852003-11-08 01:05:38 +00002790<div class="doc_text">
Chris Lattner74d3f822004-12-09 17:30:23 +00002791
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002792<p>The LLVM instruction set consists of several different classifications of
2793 instructions: <a href="#terminators">terminator
2794 instructions</a>, <a href="#binaryops">binary instructions</a>,
2795 <a href="#bitwiseops">bitwise binary instructions</a>,
2796 <a href="#memoryops">memory instructions</a>, and
2797 <a href="#otherops">other instructions</a>.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +00002798
Misha Brukman76307852003-11-08 01:05:38 +00002799</div>
Chris Lattner74d3f822004-12-09 17:30:23 +00002800
Chris Lattner2f7c9632001-06-06 20:29:01 +00002801<!-- ======================================================================= -->
Chris Lattner48b383b02003-11-25 01:02:51 +00002802<div class="doc_subsection"> <a name="terminators">Terminator
2803Instructions</a> </div>
Chris Lattner74d3f822004-12-09 17:30:23 +00002804
Misha Brukman76307852003-11-08 01:05:38 +00002805<div class="doc_text">
Chris Lattner74d3f822004-12-09 17:30:23 +00002806
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002807<p>As mentioned <a href="#functionstructure">previously</a>, every basic block
2808 in a program ends with a "Terminator" instruction, which indicates which
2809 block should be executed after the current block is finished. These
2810 terminator instructions typically yield a '<tt>void</tt>' value: they produce
2811 control flow, not values (the one exception being the
2812 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
2813
Duncan Sands626b0242010-04-15 20:35:54 +00002814<p>There are seven different terminator instructions: the
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002815 '<a href="#i_ret"><tt>ret</tt></a>' instruction, the
2816 '<a href="#i_br"><tt>br</tt></a>' instruction, the
2817 '<a href="#i_switch"><tt>switch</tt></a>' instruction, the
Bill Wendling33fef7e2009-11-02 00:25:26 +00002818 '<a href="#i_indirectbr">'<tt>indirectbr</tt></a>' Instruction, the
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002819 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the
2820 '<a href="#i_unwind"><tt>unwind</tt></a>' instruction, and the
2821 '<a href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +00002822
Misha Brukman76307852003-11-08 01:05:38 +00002823</div>
Chris Lattner74d3f822004-12-09 17:30:23 +00002824
Chris Lattner2f7c9632001-06-06 20:29:01 +00002825<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +00002826<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
2827Instruction</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002828
Misha Brukman76307852003-11-08 01:05:38 +00002829<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002830
Chris Lattner2f7c9632001-06-06 20:29:01 +00002831<h5>Syntax:</h5>
Dan Gohmancc3132e2008-10-04 19:00:07 +00002832<pre>
2833 ret &lt;type&gt; &lt;value&gt; <i>; Return a value from a non-void function</i>
Chris Lattner590645f2002-04-14 06:13:44 +00002834 ret void <i>; Return from void function</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002835</pre>
Chris Lattnerda508ac2008-04-23 04:59:35 +00002836
Chris Lattner2f7c9632001-06-06 20:29:01 +00002837<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002838<p>The '<tt>ret</tt>' instruction is used to return control flow (and optionally
2839 a value) from a function back to the caller.</p>
2840
2841<p>There are two forms of the '<tt>ret</tt>' instruction: one that returns a
2842 value and then causes control flow, and one that just causes control flow to
2843 occur.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +00002844
Chris Lattner2f7c9632001-06-06 20:29:01 +00002845<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002846<p>The '<tt>ret</tt>' instruction optionally accepts a single argument, the
2847 return value. The type of the return value must be a
2848 '<a href="#t_firstclass">first class</a>' type.</p>
Dan Gohmancc3132e2008-10-04 19:00:07 +00002849
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002850<p>A function is not <a href="#wellformed">well formed</a> if it it has a
2851 non-void return type and contains a '<tt>ret</tt>' instruction with no return
2852 value or a return value with a type that does not match its type, or if it
2853 has a void return type and contains a '<tt>ret</tt>' instruction with a
2854 return value.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +00002855
Chris Lattner2f7c9632001-06-06 20:29:01 +00002856<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002857<p>When the '<tt>ret</tt>' instruction is executed, control flow returns back to
2858 the calling function's context. If the caller is a
2859 "<a href="#i_call"><tt>call</tt></a>" instruction, execution continues at the
2860 instruction after the call. If the caller was an
2861 "<a href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues at
2862 the beginning of the "normal" destination block. If the instruction returns
2863 a value, that value shall set the call or invoke instruction's return
2864 value.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +00002865
Chris Lattner2f7c9632001-06-06 20:29:01 +00002866<h5>Example:</h5>
Chris Lattnerda508ac2008-04-23 04:59:35 +00002867<pre>
2868 ret i32 5 <i>; Return an integer value of 5</i>
Chris Lattner590645f2002-04-14 06:13:44 +00002869 ret void <i>; Return from a void function</i>
Bill Wendling050ee8f2009-02-28 22:12:54 +00002870 ret { i32, i8 } { i32 4, i8 2 } <i>; Return a struct of values 4 and 2</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002871</pre>
Dan Gohman3065b612009-01-12 23:12:39 +00002872
Misha Brukman76307852003-11-08 01:05:38 +00002873</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002874<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +00002875<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002876
Misha Brukman76307852003-11-08 01:05:38 +00002877<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002878
Chris Lattner2f7c9632001-06-06 20:29:01 +00002879<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002880<pre>
2881 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 +00002882</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002883
Chris Lattner2f7c9632001-06-06 20:29:01 +00002884<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002885<p>The '<tt>br</tt>' instruction is used to cause control flow to transfer to a
2886 different basic block in the current function. There are two forms of this
2887 instruction, corresponding to a conditional branch and an unconditional
2888 branch.</p>
2889
Chris Lattner2f7c9632001-06-06 20:29:01 +00002890<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002891<p>The conditional branch form of the '<tt>br</tt>' instruction takes a single
2892 '<tt>i1</tt>' value and two '<tt>label</tt>' values. The unconditional form
2893 of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>' value as a
2894 target.</p>
2895
Chris Lattner2f7c9632001-06-06 20:29:01 +00002896<h5>Semantics:</h5>
Reid Spencer36a15422007-01-12 03:35:51 +00002897<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002898 argument is evaluated. If the value is <tt>true</tt>, control flows to the
2899 '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
2900 control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
2901
Chris Lattner2f7c9632001-06-06 20:29:01 +00002902<h5>Example:</h5>
Bill Wendling30235112009-07-20 02:39:26 +00002903<pre>
2904Test:
2905 %cond = <a href="#i_icmp">icmp</a> eq i32 %a, %b
2906 br i1 %cond, label %IfEqual, label %IfUnequal
2907IfEqual:
2908 <a href="#i_ret">ret</a> i32 1
2909IfUnequal:
2910 <a href="#i_ret">ret</a> i32 0
2911</pre>
2912
Misha Brukman76307852003-11-08 01:05:38 +00002913</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002914
Chris Lattner2f7c9632001-06-06 20:29:01 +00002915<!-- _______________________________________________________________________ -->
Chris Lattnercf96c6c2004-02-24 04:54:45 +00002916<div class="doc_subsubsection">
2917 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
2918</div>
2919
Misha Brukman76307852003-11-08 01:05:38 +00002920<div class="doc_text">
Chris Lattnercf96c6c2004-02-24 04:54:45 +00002921
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002922<h5>Syntax:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +00002923<pre>
2924 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
2925</pre>
2926
Chris Lattner2f7c9632001-06-06 20:29:01 +00002927<h5>Overview:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +00002928<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002929 several different places. It is a generalization of the '<tt>br</tt>'
2930 instruction, allowing a branch to occur to one of many possible
2931 destinations.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +00002932
Chris Lattner2f7c9632001-06-06 20:29:01 +00002933<h5>Arguments:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +00002934<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002935 comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination,
2936 and an array of pairs of comparison value constants and '<tt>label</tt>'s.
2937 The table is not allowed to contain duplicate constant entries.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +00002938
Chris Lattner2f7c9632001-06-06 20:29:01 +00002939<h5>Semantics:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +00002940<p>The <tt>switch</tt> instruction specifies a table of values and
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002941 destinations. When the '<tt>switch</tt>' instruction is executed, this table
2942 is searched for the given value. If the value is found, control flow is
Benjamin Kramer0f420382009-10-12 14:46:08 +00002943 transferred to the corresponding destination; otherwise, control flow is
2944 transferred to the default destination.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002945
Chris Lattnercf96c6c2004-02-24 04:54:45 +00002946<h5>Implementation:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +00002947<p>Depending on properties of the target machine and the particular
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002948 <tt>switch</tt> instruction, this instruction may be code generated in
2949 different ways. For example, it could be generated as a series of chained
2950 conditional branches or with a lookup table.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +00002951
2952<h5>Example:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +00002953<pre>
2954 <i>; Emulate a conditional br instruction</i>
Reid Spencer36a15422007-01-12 03:35:51 +00002955 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Dan Gohman623806e2009-01-04 23:44:43 +00002956 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
Chris Lattnercf96c6c2004-02-24 04:54:45 +00002957
2958 <i>; Emulate an unconditional br instruction</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00002959 switch i32 0, label %dest [ ]
Chris Lattnercf96c6c2004-02-24 04:54:45 +00002960
2961 <i>; Implement a jump table:</i>
Dan Gohman623806e2009-01-04 23:44:43 +00002962 switch i32 %val, label %otherwise [ i32 0, label %onzero
2963 i32 1, label %onone
2964 i32 2, label %ontwo ]
Chris Lattner2f7c9632001-06-06 20:29:01 +00002965</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00002966
Misha Brukman76307852003-11-08 01:05:38 +00002967</div>
Chris Lattner0132aff2005-05-06 22:57:40 +00002968
Chris Lattner3ed871f2009-10-27 19:13:16 +00002969
2970<!-- _______________________________________________________________________ -->
2971<div class="doc_subsubsection">
Chris Lattnerd04cb6d2009-10-28 00:19:10 +00002972 <a name="i_indirectbr">'<tt>indirectbr</tt>' Instruction</a>
Chris Lattner3ed871f2009-10-27 19:13:16 +00002973</div>
2974
2975<div class="doc_text">
2976
2977<h5>Syntax:</h5>
2978<pre>
Chris Lattnerd04cb6d2009-10-28 00:19:10 +00002979 indirectbr &lt;somety&gt;* &lt;address&gt;, [ label &lt;dest1&gt;, label &lt;dest2&gt;, ... ]
Chris Lattner3ed871f2009-10-27 19:13:16 +00002980</pre>
2981
2982<h5>Overview:</h5>
2983
Chris Lattnerd04cb6d2009-10-28 00:19:10 +00002984<p>The '<tt>indirectbr</tt>' instruction implements an indirect branch to a label
Chris Lattner3ed871f2009-10-27 19:13:16 +00002985 within the current function, whose address is specified by
Chris Lattnere4801f72009-10-27 21:01:34 +00002986 "<tt>address</tt>". Address must be derived from a <a
2987 href="#blockaddress">blockaddress</a> constant.</p>
Chris Lattner3ed871f2009-10-27 19:13:16 +00002988
2989<h5>Arguments:</h5>
2990
2991<p>The '<tt>address</tt>' argument is the address of the label to jump to. The
2992 rest of the arguments indicate the full set of possible destinations that the
2993 address may point to. Blocks are allowed to occur multiple times in the
2994 destination list, though this isn't particularly useful.</p>
Eric Christopher455c5772009-12-05 02:46:03 +00002995
Chris Lattner3ed871f2009-10-27 19:13:16 +00002996<p>This destination list is required so that dataflow analysis has an accurate
2997 understanding of the CFG.</p>
2998
2999<h5>Semantics:</h5>
3000
3001<p>Control transfers to the block specified in the address argument. All
3002 possible destination blocks must be listed in the label list, otherwise this
3003 instruction has undefined behavior. This implies that jumps to labels
3004 defined in other functions have undefined behavior as well.</p>
3005
3006<h5>Implementation:</h5>
3007
3008<p>This is typically implemented with a jump through a register.</p>
3009
3010<h5>Example:</h5>
3011<pre>
Chris Lattnerd04cb6d2009-10-28 00:19:10 +00003012 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
Chris Lattner3ed871f2009-10-27 19:13:16 +00003013</pre>
3014
3015</div>
3016
3017
Chris Lattner2f7c9632001-06-06 20:29:01 +00003018<!-- _______________________________________________________________________ -->
Chris Lattner0132aff2005-05-06 22:57:40 +00003019<div class="doc_subsubsection">
3020 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
3021</div>
3022
Misha Brukman76307852003-11-08 01:05:38 +00003023<div class="doc_text">
Chris Lattner0132aff2005-05-06 22:57:40 +00003024
Chris Lattner2f7c9632001-06-06 20:29:01 +00003025<h5>Syntax:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +00003026<pre>
Devang Patel02256232008-10-07 17:48:33 +00003027 &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 +00003028 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
Chris Lattner0132aff2005-05-06 22:57:40 +00003029</pre>
3030
Chris Lattnera8292f32002-05-06 22:08:29 +00003031<h5>Overview:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +00003032<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003033 function, with the possibility of control flow transfer to either the
3034 '<tt>normal</tt>' label or the '<tt>exception</tt>' label. If the callee
3035 function returns with the "<tt><a href="#i_ret">ret</a></tt>" instruction,
3036 control flow will return to the "normal" label. If the callee (or any
3037 indirect callees) returns with the "<a href="#i_unwind"><tt>unwind</tt></a>"
3038 instruction, control is interrupted and continued at the dynamically nearest
3039 "exception" label.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +00003040
Chris Lattner2f7c9632001-06-06 20:29:01 +00003041<h5>Arguments:</h5>
Misha Brukman76307852003-11-08 01:05:38 +00003042<p>This instruction requires several arguments:</p>
Chris Lattner0132aff2005-05-06 22:57:40 +00003043
Chris Lattner2f7c9632001-06-06 20:29:01 +00003044<ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003045 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
3046 convention</a> the call should use. If none is specified, the call
3047 defaults to using C calling conventions.</li>
Devang Patel7e9b05e2008-10-06 18:50:38 +00003048
3049 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003050 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
3051 '<tt>inreg</tt>' attributes are valid here.</li>
Devang Patel7e9b05e2008-10-06 18:50:38 +00003052
Chris Lattner0132aff2005-05-06 22:57:40 +00003053 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003054 function value being invoked. In most cases, this is a direct function
3055 invocation, but indirect <tt>invoke</tt>s are just as possible, branching
3056 off an arbitrary pointer to function value.</li>
Chris Lattner0132aff2005-05-06 22:57:40 +00003057
3058 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003059 function to be invoked. </li>
Chris Lattner0132aff2005-05-06 22:57:40 +00003060
3061 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner47f2a832010-03-02 06:36:51 +00003062 signature argument types and parameter attributes. All arguments must be
3063 of <a href="#t_firstclass">first class</a> type. If the function
3064 signature indicates the function accepts a variable number of arguments,
3065 the extra arguments can be specified.</li>
Chris Lattner0132aff2005-05-06 22:57:40 +00003066
3067 <li>'<tt>normal label</tt>': the label reached when the called function
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003068 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
Chris Lattner0132aff2005-05-06 22:57:40 +00003069
3070 <li>'<tt>exception label</tt>': the label reached when a callee returns with
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003071 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
Chris Lattner0132aff2005-05-06 22:57:40 +00003072
Devang Patel02256232008-10-07 17:48:33 +00003073 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003074 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
3075 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner2f7c9632001-06-06 20:29:01 +00003076</ol>
Chris Lattner0132aff2005-05-06 22:57:40 +00003077
Chris Lattner2f7c9632001-06-06 20:29:01 +00003078<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003079<p>This instruction is designed to operate as a standard
3080 '<tt><a href="#i_call">call</a></tt>' instruction in most regards. The
3081 primary difference is that it establishes an association with a label, which
3082 is used by the runtime library to unwind the stack.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +00003083
3084<p>This instruction is used in languages with destructors to ensure that proper
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003085 cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
3086 exception. Additionally, this is important for implementation of
3087 '<tt>catch</tt>' clauses in high-level languages that support them.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +00003088
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003089<p>For the purposes of the SSA form, the definition of the value returned by the
3090 '<tt>invoke</tt>' instruction is deemed to occur on the edge from the current
3091 block to the "normal" label. If the callee unwinds then no return value is
3092 available.</p>
Dan Gohman9069d892009-05-22 21:47:08 +00003093
Chris Lattner97257f82010-01-15 18:08:37 +00003094<p>Note that the code generator does not yet completely support unwind, and
3095that the invoke/unwind semantics are likely to change in future versions.</p>
3096
Chris Lattner2f7c9632001-06-06 20:29:01 +00003097<h5>Example:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +00003098<pre>
Nick Lewycky084ab472008-03-16 07:18:12 +00003099 %retval = invoke i32 @Test(i32 15) to label %Continue
Jeff Cohen5819f182007-04-22 01:17:39 +00003100 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewycky084ab472008-03-16 07:18:12 +00003101 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Jeff Cohen5819f182007-04-22 01:17:39 +00003102 unwind label %TestCleanup <i>; {i32}:retval set</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00003103</pre>
Chris Lattner08b7d5b2004-10-16 18:04:13 +00003104
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003105</div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +00003106
Chris Lattner5ed60612003-09-03 00:41:47 +00003107<!-- _______________________________________________________________________ -->
Chris Lattner08b7d5b2004-10-16 18:04:13 +00003108
Chris Lattner48b383b02003-11-25 01:02:51 +00003109<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
3110Instruction</a> </div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +00003111
Misha Brukman76307852003-11-08 01:05:38 +00003112<div class="doc_text">
Chris Lattner08b7d5b2004-10-16 18:04:13 +00003113
Chris Lattner5ed60612003-09-03 00:41:47 +00003114<h5>Syntax:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +00003115<pre>
3116 unwind
3117</pre>
3118
Chris Lattner5ed60612003-09-03 00:41:47 +00003119<h5>Overview:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +00003120<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003121 at the first callee in the dynamic call stack which used
3122 an <a href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call.
3123 This is primarily used to implement exception handling.</p>
Chris Lattner08b7d5b2004-10-16 18:04:13 +00003124
Chris Lattner5ed60612003-09-03 00:41:47 +00003125<h5>Semantics:</h5>
Chris Lattnerfe8519c2008-04-19 21:01:16 +00003126<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003127 immediately halt. The dynamic call stack is then searched for the
3128 first <a href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack.
3129 Once found, execution continues at the "exceptional" destination block
3130 specified by the <tt>invoke</tt> instruction. If there is no <tt>invoke</tt>
3131 instruction in the dynamic call chain, undefined behavior results.</p>
3132
Chris Lattner97257f82010-01-15 18:08:37 +00003133<p>Note that the code generator does not yet completely support unwind, and
3134that the invoke/unwind semantics are likely to change in future versions.</p>
3135
Misha Brukman76307852003-11-08 01:05:38 +00003136</div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +00003137
3138<!-- _______________________________________________________________________ -->
3139
3140<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
3141Instruction</a> </div>
3142
3143<div class="doc_text">
3144
3145<h5>Syntax:</h5>
3146<pre>
3147 unreachable
3148</pre>
3149
3150<h5>Overview:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +00003151<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003152 instruction is used to inform the optimizer that a particular portion of the
3153 code is not reachable. This can be used to indicate that the code after a
3154 no-return function cannot be reached, and other facts.</p>
Chris Lattner08b7d5b2004-10-16 18:04:13 +00003155
3156<h5>Semantics:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +00003157<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003158
Chris Lattner08b7d5b2004-10-16 18:04:13 +00003159</div>
3160
Chris Lattner2f7c9632001-06-06 20:29:01 +00003161<!-- ======================================================================= -->
Chris Lattner48b383b02003-11-25 01:02:51 +00003162<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003163
Misha Brukman76307852003-11-08 01:05:38 +00003164<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003165
3166<p>Binary operators are used to do most of the computation in a program. They
3167 require two operands of the same type, execute an operation on them, and
3168 produce a single value. The operands might represent multiple data, as is
3169 the case with the <a href="#t_vector">vector</a> data type. The result value
3170 has the same type as its operands.</p>
3171
Misha Brukman76307852003-11-08 01:05:38 +00003172<p>There are several different binary operators:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003173
Misha Brukman76307852003-11-08 01:05:38 +00003174</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003175
Chris Lattner2f7c9632001-06-06 20:29:01 +00003176<!-- _______________________________________________________________________ -->
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003177<div class="doc_subsubsection">
3178 <a name="i_add">'<tt>add</tt>' Instruction</a>
3179</div>
3180
Misha Brukman76307852003-11-08 01:05:38 +00003181<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003182
Chris Lattner2f7c9632001-06-06 20:29:01 +00003183<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003184<pre>
Dan Gohmanb07de442009-07-20 22:41:19 +00003185 &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 +00003186 &lt;result&gt; = add nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3187 &lt;result&gt; = add nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3188 &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 +00003189</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003190
Chris Lattner2f7c9632001-06-06 20:29:01 +00003191<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +00003192<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003193
Chris Lattner2f7c9632001-06-06 20:29:01 +00003194<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003195<p>The two arguments to the '<tt>add</tt>' instruction must
3196 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3197 integer values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003198
Chris Lattner2f7c9632001-06-06 20:29:01 +00003199<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +00003200<p>The value produced is the integer sum of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003201
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003202<p>If the sum has unsigned overflow, the result returned is the mathematical
3203 result modulo 2<sup>n</sup>, where n is the bit width of the result.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003204
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003205<p>Because LLVM integers use a two's complement representation, this instruction
3206 is appropriate for both signed and unsigned integers.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003207
Dan Gohman902dfff2009-07-22 22:44:56 +00003208<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3209 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3210 <tt>nsw</tt> keywords are present, the result value of the <tt>add</tt>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +00003211 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3212 respectively, occurs.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +00003213
Chris Lattner2f7c9632001-06-06 20:29:01 +00003214<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003215<pre>
3216 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00003217</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003218
Misha Brukman76307852003-11-08 01:05:38 +00003219</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003220
Chris Lattner2f7c9632001-06-06 20:29:01 +00003221<!-- _______________________________________________________________________ -->
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003222<div class="doc_subsubsection">
Dan Gohmana5b96452009-06-04 22:49:04 +00003223 <a name="i_fadd">'<tt>fadd</tt>' Instruction</a>
3224</div>
3225
3226<div class="doc_text">
3227
3228<h5>Syntax:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +00003229<pre>
3230 &lt;result&gt; = fadd &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3231</pre>
3232
3233<h5>Overview:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +00003234<p>The '<tt>fadd</tt>' instruction returns the sum of its two operands.</p>
3235
3236<h5>Arguments:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +00003237<p>The two arguments to the '<tt>fadd</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003238 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3239 floating point values. Both arguments must have identical types.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +00003240
3241<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +00003242<p>The value produced is the floating point sum of the two operands.</p>
3243
3244<h5>Example:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +00003245<pre>
3246 &lt;result&gt; = fadd float 4.0, %var <i>; yields {float}:result = 4.0 + %var</i>
3247</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003248
Dan Gohmana5b96452009-06-04 22:49:04 +00003249</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003250
Dan Gohmana5b96452009-06-04 22:49:04 +00003251<!-- _______________________________________________________________________ -->
3252<div class="doc_subsubsection">
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003253 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
3254</div>
3255
Misha Brukman76307852003-11-08 01:05:38 +00003256<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003257
Chris Lattner2f7c9632001-06-06 20:29:01 +00003258<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003259<pre>
Dan Gohman902dfff2009-07-22 22:44:56 +00003260 &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 +00003261 &lt;result&gt; = sub nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3262 &lt;result&gt; = sub nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3263 &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 +00003264</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003265
Chris Lattner2f7c9632001-06-06 20:29:01 +00003266<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +00003267<p>The '<tt>sub</tt>' instruction returns the difference of its two
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003268 operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003269
3270<p>Note that the '<tt>sub</tt>' instruction is used to represent the
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003271 '<tt>neg</tt>' instruction present in most other intermediate
3272 representations.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003273
Chris Lattner2f7c9632001-06-06 20:29:01 +00003274<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003275<p>The two arguments to the '<tt>sub</tt>' instruction must
3276 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3277 integer values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003278
Chris Lattner2f7c9632001-06-06 20:29:01 +00003279<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +00003280<p>The value produced is the integer difference of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003281
Dan Gohmana5b96452009-06-04 22:49:04 +00003282<p>If the difference has unsigned overflow, the result returned is the
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003283 mathematical result modulo 2<sup>n</sup>, where n is the bit width of the
3284 result.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003285
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003286<p>Because LLVM integers use a two's complement representation, this instruction
3287 is appropriate for both signed and unsigned integers.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003288
Dan Gohman902dfff2009-07-22 22:44:56 +00003289<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3290 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3291 <tt>nsw</tt> keywords are present, the result value of the <tt>sub</tt>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +00003292 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3293 respectively, occurs.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +00003294
Chris Lattner2f7c9632001-06-06 20:29:01 +00003295<h5>Example:</h5>
Bill Wendling2d8b9a82007-05-29 09:42:13 +00003296<pre>
3297 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003298 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00003299</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003300
Misha Brukman76307852003-11-08 01:05:38 +00003301</div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003302
Chris Lattner2f7c9632001-06-06 20:29:01 +00003303<!-- _______________________________________________________________________ -->
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003304<div class="doc_subsubsection">
Dan Gohmana5b96452009-06-04 22:49:04 +00003305 <a name="i_fsub">'<tt>fsub</tt>' Instruction</a>
3306</div>
3307
3308<div class="doc_text">
3309
3310<h5>Syntax:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +00003311<pre>
3312 &lt;result&gt; = fsub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3313</pre>
3314
3315<h5>Overview:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +00003316<p>The '<tt>fsub</tt>' instruction returns the difference of its two
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003317 operands.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +00003318
3319<p>Note that the '<tt>fsub</tt>' instruction is used to represent the
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003320 '<tt>fneg</tt>' instruction present in most other intermediate
3321 representations.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +00003322
3323<h5>Arguments:</h5>
Bill Wendling972b7202009-07-20 02:32:41 +00003324<p>The two arguments to the '<tt>fsub</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003325 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3326 floating point values. Both arguments must have identical types.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +00003327
3328<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +00003329<p>The value produced is the floating point difference of the two operands.</p>
3330
3331<h5>Example:</h5>
3332<pre>
3333 &lt;result&gt; = fsub float 4.0, %var <i>; yields {float}:result = 4.0 - %var</i>
3334 &lt;result&gt; = fsub float -0.0, %val <i>; yields {float}:result = -%var</i>
3335</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003336
Dan Gohmana5b96452009-06-04 22:49:04 +00003337</div>
3338
3339<!-- _______________________________________________________________________ -->
3340<div class="doc_subsubsection">
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003341 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
3342</div>
3343
Misha Brukman76307852003-11-08 01:05:38 +00003344<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003345
Chris Lattner2f7c9632001-06-06 20:29:01 +00003346<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003347<pre>
Dan Gohman902dfff2009-07-22 22:44:56 +00003348 &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 +00003349 &lt;result&gt; = mul nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3350 &lt;result&gt; = mul nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3351 &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 +00003352</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003353
Chris Lattner2f7c9632001-06-06 20:29:01 +00003354<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003355<p>The '<tt>mul</tt>' instruction returns the product of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003356
Chris Lattner2f7c9632001-06-06 20:29:01 +00003357<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003358<p>The two arguments to the '<tt>mul</tt>' instruction must
3359 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3360 integer values. Both arguments must have identical types.</p>
Eric Christopher455c5772009-12-05 02:46:03 +00003361
Chris Lattner2f7c9632001-06-06 20:29:01 +00003362<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +00003363<p>The value produced is the integer product of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003364
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003365<p>If the result of the multiplication has unsigned overflow, the result
3366 returned is the mathematical result modulo 2<sup>n</sup>, where n is the bit
3367 width of the result.</p>
3368
3369<p>Because LLVM integers use a two's complement representation, and the result
3370 is the same width as the operands, this instruction returns the correct
3371 result for both signed and unsigned integers. If a full product
3372 (e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands should
3373 be sign-extended or zero-extended as appropriate to the width of the full
3374 product.</p>
3375
Dan Gohman902dfff2009-07-22 22:44:56 +00003376<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3377 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3378 <tt>nsw</tt> keywords are present, the result value of the <tt>mul</tt>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +00003379 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3380 respectively, occurs.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +00003381
Chris Lattner2f7c9632001-06-06 20:29:01 +00003382<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003383<pre>
3384 &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00003385</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003386
Misha Brukman76307852003-11-08 01:05:38 +00003387</div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003388
Chris Lattner2f7c9632001-06-06 20:29:01 +00003389<!-- _______________________________________________________________________ -->
Dan Gohmana5b96452009-06-04 22:49:04 +00003390<div class="doc_subsubsection">
3391 <a name="i_fmul">'<tt>fmul</tt>' Instruction</a>
3392</div>
3393
3394<div class="doc_text">
3395
3396<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003397<pre>
3398 &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 +00003399</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003400
Dan Gohmana5b96452009-06-04 22:49:04 +00003401<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003402<p>The '<tt>fmul</tt>' instruction returns the product of its two operands.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +00003403
3404<h5>Arguments:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +00003405<p>The two arguments to the '<tt>fmul</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003406 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3407 floating point values. Both arguments must have identical types.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +00003408
3409<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +00003410<p>The value produced is the floating point product of the two operands.</p>
3411
3412<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003413<pre>
3414 &lt;result&gt; = fmul float 4.0, %var <i>; yields {float}:result = 4.0 * %var</i>
Dan Gohmana5b96452009-06-04 22:49:04 +00003415</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003416
Dan Gohmana5b96452009-06-04 22:49:04 +00003417</div>
3418
3419<!-- _______________________________________________________________________ -->
Reid Spencer7e80b0b2006-10-26 06:15:43 +00003420<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
3421</a></div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003422
Reid Spencer7e80b0b2006-10-26 06:15:43 +00003423<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003424
Reid Spencer7e80b0b2006-10-26 06:15:43 +00003425<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003426<pre>
3427 &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 +00003428</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003429
Reid Spencer7e80b0b2006-10-26 06:15:43 +00003430<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003431<p>The '<tt>udiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003432
Reid Spencer7e80b0b2006-10-26 06:15:43 +00003433<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +00003434<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003435 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3436 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003437
Reid Spencer7e80b0b2006-10-26 06:15:43 +00003438<h5>Semantics:</h5>
Chris Lattner2f2427e2008-01-28 00:36:27 +00003439<p>The value produced is the unsigned integer quotient of the two operands.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003440
Chris Lattner2f2427e2008-01-28 00:36:27 +00003441<p>Note that unsigned integer division and signed integer division are distinct
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003442 operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
3443
Chris Lattner2f2427e2008-01-28 00:36:27 +00003444<p>Division by zero leads to undefined behavior.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003445
Reid Spencer7e80b0b2006-10-26 06:15:43 +00003446<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003447<pre>
3448 &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +00003449</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003450
Reid Spencer7e80b0b2006-10-26 06:15:43 +00003451</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003452
Reid Spencer7e80b0b2006-10-26 06:15:43 +00003453<!-- _______________________________________________________________________ -->
3454<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
3455</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003456
Reid Spencer7e80b0b2006-10-26 06:15:43 +00003457<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003458
Reid Spencer7e80b0b2006-10-26 06:15:43 +00003459<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003460<pre>
Dan Gohmanb07de442009-07-20 22:41:19 +00003461 &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 +00003462 &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 +00003463</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003464
Reid Spencer7e80b0b2006-10-26 06:15:43 +00003465<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003466<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003467
Reid Spencer7e80b0b2006-10-26 06:15:43 +00003468<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +00003469<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003470 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3471 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003472
Reid Spencer7e80b0b2006-10-26 06:15:43 +00003473<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003474<p>The value produced is the signed integer quotient of the two operands rounded
3475 towards zero.</p>
3476
Chris Lattner2f2427e2008-01-28 00:36:27 +00003477<p>Note that signed integer division and unsigned integer division are distinct
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003478 operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
3479
Chris Lattner2f2427e2008-01-28 00:36:27 +00003480<p>Division by zero leads to undefined behavior. Overflow also leads to
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003481 undefined behavior; this is a rare case, but can occur, for example, by doing
3482 a 32-bit division of -2147483648 by -1.</p>
3483
Dan Gohman71dfd782009-07-22 00:04:19 +00003484<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohman57255802010-04-23 15:23:32 +00003485 <tt>sdiv</tt> is a <a href="#trapvalues">trap value</a> if the result would
Dan Gohmane501ff72010-07-11 00:08:34 +00003486 be rounded.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +00003487
Reid Spencer7e80b0b2006-10-26 06:15:43 +00003488<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003489<pre>
3490 &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +00003491</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003492
Reid Spencer7e80b0b2006-10-26 06:15:43 +00003493</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003494
Reid Spencer7e80b0b2006-10-26 06:15:43 +00003495<!-- _______________________________________________________________________ -->
3496<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
Chris Lattner48b383b02003-11-25 01:02:51 +00003497Instruction</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003498
Misha Brukman76307852003-11-08 01:05:38 +00003499<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003500
Chris Lattner2f7c9632001-06-06 20:29:01 +00003501<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003502<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +00003503 &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 +00003504</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003505
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003506<h5>Overview:</h5>
3507<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003508
Chris Lattner48b383b02003-11-25 01:02:51 +00003509<h5>Arguments:</h5>
Jeff Cohen5819f182007-04-22 01:17:39 +00003510<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003511 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3512 floating point values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003513
Chris Lattner48b383b02003-11-25 01:02:51 +00003514<h5>Semantics:</h5>
Reid Spencer7e80b0b2006-10-26 06:15:43 +00003515<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003516
Chris Lattner48b383b02003-11-25 01:02:51 +00003517<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003518<pre>
3519 &lt;result&gt; = fdiv float 4.0, %var <i>; yields {float}:result = 4.0 / %var</i>
Chris Lattner48b383b02003-11-25 01:02:51 +00003520</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003521
Chris Lattner48b383b02003-11-25 01:02:51 +00003522</div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003523
Chris Lattner48b383b02003-11-25 01:02:51 +00003524<!-- _______________________________________________________________________ -->
Reid Spencer7eb55b32006-11-02 01:53:59 +00003525<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
3526</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003527
Reid Spencer7eb55b32006-11-02 01:53:59 +00003528<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003529
Reid Spencer7eb55b32006-11-02 01:53:59 +00003530<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003531<pre>
3532 &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 +00003533</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003534
Reid Spencer7eb55b32006-11-02 01:53:59 +00003535<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003536<p>The '<tt>urem</tt>' instruction returns the remainder from the unsigned
3537 division of its two arguments.</p>
3538
Reid Spencer7eb55b32006-11-02 01:53:59 +00003539<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +00003540<p>The two arguments to the '<tt>urem</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003541 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3542 values. Both arguments must have identical types.</p>
3543
Reid Spencer7eb55b32006-11-02 01:53:59 +00003544<h5>Semantics:</h5>
3545<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003546 This instruction always performs an unsigned division to get the
3547 remainder.</p>
3548
Chris Lattner2f2427e2008-01-28 00:36:27 +00003549<p>Note that unsigned integer remainder and signed integer remainder are
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003550 distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
3551
Chris Lattner2f2427e2008-01-28 00:36:27 +00003552<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003553
Reid Spencer7eb55b32006-11-02 01:53:59 +00003554<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003555<pre>
3556 &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer7eb55b32006-11-02 01:53:59 +00003557</pre>
3558
3559</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003560
Reid Spencer7eb55b32006-11-02 01:53:59 +00003561<!-- _______________________________________________________________________ -->
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003562<div class="doc_subsubsection">
3563 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
3564</div>
3565
Chris Lattner48b383b02003-11-25 01:02:51 +00003566<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003567
Chris Lattner48b383b02003-11-25 01:02:51 +00003568<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003569<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +00003570 &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 +00003571</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003572
Chris Lattner48b383b02003-11-25 01:02:51 +00003573<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003574<p>The '<tt>srem</tt>' instruction returns the remainder from the signed
3575 division of its two operands. This instruction can also take
3576 <a href="#t_vector">vector</a> versions of the values in which case the
3577 elements must be integers.</p>
Chris Lattnerb8f816e2008-01-04 04:33:49 +00003578
Chris Lattner48b383b02003-11-25 01:02:51 +00003579<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +00003580<p>The two arguments to the '<tt>srem</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003581 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3582 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003583
Chris Lattner48b383b02003-11-25 01:02:51 +00003584<h5>Semantics:</h5>
Reid Spencer7eb55b32006-11-02 01:53:59 +00003585<p>This instruction returns the <i>remainder</i> of a division (where the result
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003586 has the same sign as the dividend, <tt>op1</tt>), not the <i>modulo</i>
3587 operator (where the result has the same sign as the divisor, <tt>op2</tt>) of
3588 a value. For more information about the difference,
3589 see <a href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
3590 Math Forum</a>. For a table of how this is implemented in various languages,
3591 please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
3592 Wikipedia: modulo operation</a>.</p>
3593
Chris Lattner2f2427e2008-01-28 00:36:27 +00003594<p>Note that signed integer remainder and unsigned integer remainder are
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003595 distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
3596
Chris Lattner2f2427e2008-01-28 00:36:27 +00003597<p>Taking the remainder of a division by zero leads to undefined behavior.
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003598 Overflow also leads to undefined behavior; this is a rare case, but can
3599 occur, for example, by taking the remainder of a 32-bit division of
3600 -2147483648 by -1. (The remainder doesn't actually overflow, but this rule
3601 lets srem be implemented using instructions that return both the result of
3602 the division and the remainder.)</p>
3603
Chris Lattner48b383b02003-11-25 01:02:51 +00003604<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003605<pre>
3606 &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer7eb55b32006-11-02 01:53:59 +00003607</pre>
3608
3609</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003610
Reid Spencer7eb55b32006-11-02 01:53:59 +00003611<!-- _______________________________________________________________________ -->
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003612<div class="doc_subsubsection">
3613 <a name="i_frem">'<tt>frem</tt>' Instruction</a> </div>
3614
Reid Spencer7eb55b32006-11-02 01:53:59 +00003615<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003616
Reid Spencer7eb55b32006-11-02 01:53:59 +00003617<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003618<pre>
3619 &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 +00003620</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003621
Reid Spencer7eb55b32006-11-02 01:53:59 +00003622<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003623<p>The '<tt>frem</tt>' instruction returns the remainder from the division of
3624 its two operands.</p>
3625
Reid Spencer7eb55b32006-11-02 01:53:59 +00003626<h5>Arguments:</h5>
3627<p>The two arguments to the '<tt>frem</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003628 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3629 floating point values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003630
Reid Spencer7eb55b32006-11-02 01:53:59 +00003631<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003632<p>This instruction returns the <i>remainder</i> of a division. The remainder
3633 has the same sign as the dividend.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003634
Reid Spencer7eb55b32006-11-02 01:53:59 +00003635<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003636<pre>
3637 &lt;result&gt; = frem float 4.0, %var <i>; yields {float}:result = 4.0 % %var</i>
Chris Lattner48b383b02003-11-25 01:02:51 +00003638</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003639
Misha Brukman76307852003-11-08 01:05:38 +00003640</div>
Robert Bocchino820bc75b2006-02-17 21:18:08 +00003641
Reid Spencer2ab01932007-02-02 13:57:07 +00003642<!-- ======================================================================= -->
3643<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
3644Operations</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003645
Reid Spencer2ab01932007-02-02 13:57:07 +00003646<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003647
3648<p>Bitwise binary operators are used to do various forms of bit-twiddling in a
3649 program. They are generally very efficient instructions and can commonly be
3650 strength reduced from other instructions. They require two operands of the
3651 same type, execute an operation on them, and produce a single value. The
3652 resulting value is the same type as its operands.</p>
3653
Reid Spencer2ab01932007-02-02 13:57:07 +00003654</div>
3655
Reid Spencer04e259b2007-01-31 21:39:12 +00003656<!-- _______________________________________________________________________ -->
3657<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
3658Instruction</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003659
Reid Spencer04e259b2007-01-31 21:39:12 +00003660<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003661
Reid Spencer04e259b2007-01-31 21:39:12 +00003662<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003663<pre>
3664 &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 +00003665</pre>
Chris Lattnerf0e50112007-10-03 21:01:14 +00003666
Reid Spencer04e259b2007-01-31 21:39:12 +00003667<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003668<p>The '<tt>shl</tt>' instruction returns the first operand shifted to the left
3669 a specified number of bits.</p>
Chris Lattnerf0e50112007-10-03 21:01:14 +00003670
Reid Spencer04e259b2007-01-31 21:39:12 +00003671<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003672<p>Both arguments to the '<tt>shl</tt>' instruction must be the
3673 same <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3674 integer type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Eric Christopher455c5772009-12-05 02:46:03 +00003675
Reid Spencer04e259b2007-01-31 21:39:12 +00003676<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003677<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod
3678 2<sup>n</sup>, where <tt>n</tt> is the width of the result. If <tt>op2</tt>
3679 is (statically or dynamically) negative or equal to or larger than the number
3680 of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3681 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3682 shift amount in <tt>op2</tt>.</p>
Chris Lattnerf0e50112007-10-03 21:01:14 +00003683
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003684<h5>Example:</h5>
3685<pre>
Reid Spencer04e259b2007-01-31 21:39:12 +00003686 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
3687 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
3688 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattnerf0e50112007-10-03 21:01:14 +00003689 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Mon P Wang4dd832d2008-12-09 05:46:39 +00003690 &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 +00003691</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003692
Reid Spencer04e259b2007-01-31 21:39:12 +00003693</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003694
Reid Spencer04e259b2007-01-31 21:39:12 +00003695<!-- _______________________________________________________________________ -->
3696<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
3697Instruction</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003698
Reid Spencer04e259b2007-01-31 21:39:12 +00003699<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003700
Reid Spencer04e259b2007-01-31 21:39:12 +00003701<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003702<pre>
3703 &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 +00003704</pre>
3705
3706<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003707<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
3708 operand shifted to the right a specified number of bits with zero fill.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +00003709
3710<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +00003711<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003712 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3713 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +00003714
3715<h5>Semantics:</h5>
3716<p>This instruction always performs a logical shift right operation. The most
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003717 significant bits of the result will be filled with zero bits after the shift.
3718 If <tt>op2</tt> is (statically or dynamically) equal to or larger than the
3719 number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3720 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3721 shift amount in <tt>op2</tt>.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +00003722
3723<h5>Example:</h5>
3724<pre>
3725 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
3726 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
3727 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
3728 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattnerf0e50112007-10-03 21:01:14 +00003729 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Mon P Wang4dd832d2008-12-09 05:46:39 +00003730 &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 +00003731</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003732
Reid Spencer04e259b2007-01-31 21:39:12 +00003733</div>
3734
Reid Spencer2ab01932007-02-02 13:57:07 +00003735<!-- _______________________________________________________________________ -->
Reid Spencer04e259b2007-01-31 21:39:12 +00003736<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
3737Instruction</a> </div>
3738<div class="doc_text">
3739
3740<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003741<pre>
3742 &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 +00003743</pre>
3744
3745<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003746<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
3747 operand shifted to the right a specified number of bits with sign
3748 extension.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +00003749
3750<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +00003751<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003752 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3753 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +00003754
3755<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003756<p>This instruction always performs an arithmetic shift right operation, The
3757 most significant bits of the result will be filled with the sign bit
3758 of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
3759 larger than the number of bits in <tt>op1</tt>, the result is undefined. If
3760 the arguments are vectors, each vector element of <tt>op1</tt> is shifted by
3761 the corresponding shift amount in <tt>op2</tt>.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +00003762
3763<h5>Example:</h5>
3764<pre>
3765 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
3766 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
3767 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
3768 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattnerf0e50112007-10-03 21:01:14 +00003769 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Mon P Wang4dd832d2008-12-09 05:46:39 +00003770 &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 +00003771</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003772
Reid Spencer04e259b2007-01-31 21:39:12 +00003773</div>
3774
Chris Lattner2f7c9632001-06-06 20:29:01 +00003775<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +00003776<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
3777Instruction</a> </div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003778
Misha Brukman76307852003-11-08 01:05:38 +00003779<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003780
Chris Lattner2f7c9632001-06-06 20:29:01 +00003781<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003782<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +00003783 &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 +00003784</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003785
Chris Lattner2f7c9632001-06-06 20:29:01 +00003786<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003787<p>The '<tt>and</tt>' instruction returns the bitwise logical and of its two
3788 operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003789
Chris Lattner2f7c9632001-06-06 20:29:01 +00003790<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +00003791<p>The two arguments to the '<tt>and</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003792 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3793 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003794
Chris Lattner2f7c9632001-06-06 20:29:01 +00003795<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +00003796<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003797
Misha Brukman76307852003-11-08 01:05:38 +00003798<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner48b383b02003-11-25 01:02:51 +00003799 <tbody>
3800 <tr>
3801 <td>In0</td>
3802 <td>In1</td>
3803 <td>Out</td>
3804 </tr>
3805 <tr>
3806 <td>0</td>
3807 <td>0</td>
3808 <td>0</td>
3809 </tr>
3810 <tr>
3811 <td>0</td>
3812 <td>1</td>
3813 <td>0</td>
3814 </tr>
3815 <tr>
3816 <td>1</td>
3817 <td>0</td>
3818 <td>0</td>
3819 </tr>
3820 <tr>
3821 <td>1</td>
3822 <td>1</td>
3823 <td>1</td>
3824 </tr>
3825 </tbody>
3826</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003827
Chris Lattner2f7c9632001-06-06 20:29:01 +00003828<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003829<pre>
3830 &lt;result&gt; = and i32 4, %var <i>; yields {i32}:result = 4 &amp; %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003831 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
3832 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00003833</pre>
Misha Brukman76307852003-11-08 01:05:38 +00003834</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +00003835<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +00003836<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003837
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003838<div class="doc_text">
3839
3840<h5>Syntax:</h5>
3841<pre>
3842 &lt;result&gt; = or &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3843</pre>
3844
3845<h5>Overview:</h5>
3846<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive or of its
3847 two operands.</p>
3848
3849<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +00003850<p>The two arguments to the '<tt>or</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003851 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3852 values. Both arguments must have identical types.</p>
3853
Chris Lattner2f7c9632001-06-06 20:29:01 +00003854<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +00003855<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003856
Chris Lattner48b383b02003-11-25 01:02:51 +00003857<table border="1" cellspacing="0" cellpadding="4">
3858 <tbody>
3859 <tr>
3860 <td>In0</td>
3861 <td>In1</td>
3862 <td>Out</td>
3863 </tr>
3864 <tr>
3865 <td>0</td>
3866 <td>0</td>
3867 <td>0</td>
3868 </tr>
3869 <tr>
3870 <td>0</td>
3871 <td>1</td>
3872 <td>1</td>
3873 </tr>
3874 <tr>
3875 <td>1</td>
3876 <td>0</td>
3877 <td>1</td>
3878 </tr>
3879 <tr>
3880 <td>1</td>
3881 <td>1</td>
3882 <td>1</td>
3883 </tr>
3884 </tbody>
3885</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003886
Chris Lattner2f7c9632001-06-06 20:29:01 +00003887<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003888<pre>
3889 &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003890 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
3891 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00003892</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003893
Misha Brukman76307852003-11-08 01:05:38 +00003894</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003895
Chris Lattner2f7c9632001-06-06 20:29:01 +00003896<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +00003897<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
3898Instruction</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003899
Misha Brukman76307852003-11-08 01:05:38 +00003900<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003901
Chris Lattner2f7c9632001-06-06 20:29:01 +00003902<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003903<pre>
3904 &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 +00003905</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003906
Chris Lattner2f7c9632001-06-06 20:29:01 +00003907<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003908<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive or of
3909 its two operands. The <tt>xor</tt> is used to implement the "one's
3910 complement" operation, which is the "~" operator in C.</p>
3911
Chris Lattner2f7c9632001-06-06 20:29:01 +00003912<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +00003913<p>The two arguments to the '<tt>xor</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003914 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3915 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003916
Chris Lattner2f7c9632001-06-06 20:29:01 +00003917<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +00003918<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003919
Chris Lattner48b383b02003-11-25 01:02:51 +00003920<table border="1" cellspacing="0" cellpadding="4">
3921 <tbody>
3922 <tr>
3923 <td>In0</td>
3924 <td>In1</td>
3925 <td>Out</td>
3926 </tr>
3927 <tr>
3928 <td>0</td>
3929 <td>0</td>
3930 <td>0</td>
3931 </tr>
3932 <tr>
3933 <td>0</td>
3934 <td>1</td>
3935 <td>1</td>
3936 </tr>
3937 <tr>
3938 <td>1</td>
3939 <td>0</td>
3940 <td>1</td>
3941 </tr>
3942 <tr>
3943 <td>1</td>
3944 <td>1</td>
3945 <td>0</td>
3946 </tr>
3947 </tbody>
3948</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003949
Chris Lattner2f7c9632001-06-06 20:29:01 +00003950<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003951<pre>
3952 &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003953 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
3954 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
3955 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00003956</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003957
Misha Brukman76307852003-11-08 01:05:38 +00003958</div>
Chris Lattner54611b42005-11-06 08:02:57 +00003959
Chris Lattner2f7c9632001-06-06 20:29:01 +00003960<!-- ======================================================================= -->
Eric Christopher455c5772009-12-05 02:46:03 +00003961<div class="doc_subsection">
Chris Lattnerce83bff2006-04-08 23:07:04 +00003962 <a name="vectorops">Vector Operations</a>
3963</div>
3964
3965<div class="doc_text">
3966
3967<p>LLVM supports several instructions to represent vector operations in a
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003968 target-independent manner. These instructions cover the element-access and
3969 vector-specific operations needed to process vectors effectively. While LLVM
3970 does directly support these vector operations, many sophisticated algorithms
3971 will want to use target-specific intrinsics to take full advantage of a
3972 specific target.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +00003973
3974</div>
3975
3976<!-- _______________________________________________________________________ -->
3977<div class="doc_subsubsection">
3978 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
3979</div>
3980
3981<div class="doc_text">
3982
3983<h5>Syntax:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +00003984<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003985 &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 +00003986</pre>
3987
3988<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003989<p>The '<tt>extractelement</tt>' instruction extracts a single scalar element
3990 from a vector at a specified index.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +00003991
3992
3993<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00003994<p>The first operand of an '<tt>extractelement</tt>' instruction is a value
3995 of <a href="#t_vector">vector</a> type. The second operand is an index
3996 indicating the position from which to extract the element. The index may be
3997 a variable.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +00003998
3999<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004000<p>The result is a scalar of the same type as the element type of
4001 <tt>val</tt>. Its value is the value at position <tt>idx</tt> of
4002 <tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4003 results are undefined.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +00004004
4005<h5>Example:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +00004006<pre>
Gabor Greif03ab4dc2009-10-28 13:14:50 +00004007 &lt;result&gt; = extractelement &lt;4 x i32&gt; %vec, i32 0 <i>; yields i32</i>
Chris Lattnerce83bff2006-04-08 23:07:04 +00004008</pre>
Chris Lattnerce83bff2006-04-08 23:07:04 +00004009
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004010</div>
Chris Lattnerce83bff2006-04-08 23:07:04 +00004011
4012<!-- _______________________________________________________________________ -->
4013<div class="doc_subsubsection">
4014 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
4015</div>
4016
4017<div class="doc_text">
4018
4019<h5>Syntax:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +00004020<pre>
Dan Gohman43ba0672008-05-12 23:38:42 +00004021 &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 +00004022</pre>
4023
4024<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004025<p>The '<tt>insertelement</tt>' instruction inserts a scalar element into a
4026 vector at a specified index.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +00004027
4028<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004029<p>The first operand of an '<tt>insertelement</tt>' instruction is a value
4030 of <a href="#t_vector">vector</a> type. The second operand is a scalar value
4031 whose type must equal the element type of the first operand. The third
4032 operand is an index indicating the position at which to insert the value.
4033 The index may be a variable.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +00004034
4035<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004036<p>The result is a vector of the same type as <tt>val</tt>. Its element values
4037 are those of <tt>val</tt> except at position <tt>idx</tt>, where it gets the
4038 value <tt>elt</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4039 results are undefined.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +00004040
4041<h5>Example:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +00004042<pre>
Gabor Greif03ab4dc2009-10-28 13:14:50 +00004043 &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 +00004044</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004045
Chris Lattnerce83bff2006-04-08 23:07:04 +00004046</div>
4047
4048<!-- _______________________________________________________________________ -->
4049<div class="doc_subsubsection">
4050 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
4051</div>
4052
4053<div class="doc_text">
4054
4055<h5>Syntax:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +00004056<pre>
Mon P Wang25f01062008-11-10 04:46:22 +00004057 &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 +00004058</pre>
4059
4060<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004061<p>The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
4062 from two input vectors, returning a vector with the same element type as the
4063 input and length that is the same as the shuffle mask.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +00004064
4065<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004066<p>The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
4067 with types that match each other. The third argument is a shuffle mask whose
4068 element type is always 'i32'. The result of the instruction is a vector
4069 whose length is the same as the shuffle mask and whose element type is the
4070 same as the element type of the first two operands.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +00004071
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004072<p>The shuffle mask operand is required to be a constant vector with either
4073 constant integer or undef values.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +00004074
4075<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004076<p>The elements of the two input vectors are numbered from left to right across
4077 both of the vectors. The shuffle mask operand specifies, for each element of
4078 the result vector, which element of the two input vectors the result element
4079 gets. The element selector may be undef (meaning "don't care") and the
4080 second operand may be undef if performing a shuffle from only one vector.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +00004081
4082<h5>Example:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +00004083<pre>
Eric Christopher455c5772009-12-05 02:46:03 +00004084 &lt;result&gt; = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
Jeff Cohen5819f182007-04-22 01:17:39 +00004085 &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 +00004086 &lt;result&gt; = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; undef,
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00004087 &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 +00004088 &lt;result&gt; = shufflevector &lt;8 x i32&gt; %v1, &lt;8 x i32&gt; undef,
Mon P Wang25f01062008-11-10 04:46:22 +00004089 &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 +00004090 &lt;result&gt; = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
Mon P Wang25f01062008-11-10 04:46:22 +00004091 &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 +00004092</pre>
Chris Lattnerce83bff2006-04-08 23:07:04 +00004093
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004094</div>
Tanya Lattnerb138bbe2006-04-14 19:24:33 +00004095
Chris Lattnerce83bff2006-04-08 23:07:04 +00004096<!-- ======================================================================= -->
Eric Christopher455c5772009-12-05 02:46:03 +00004097<div class="doc_subsection">
Dan Gohmanb9d66602008-05-12 23:51:09 +00004098 <a name="aggregateops">Aggregate Operations</a>
4099</div>
4100
4101<div class="doc_text">
4102
Chris Lattner392be582010-02-12 20:49:41 +00004103<p>LLVM supports several instructions for working with
4104 <a href="#t_aggregate">aggregate</a> values.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +00004105
4106</div>
4107
4108<!-- _______________________________________________________________________ -->
4109<div class="doc_subsubsection">
4110 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
4111</div>
4112
4113<div class="doc_text">
4114
4115<h5>Syntax:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +00004116<pre>
4117 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
4118</pre>
4119
4120<h5>Overview:</h5>
Chris Lattner392be582010-02-12 20:49:41 +00004121<p>The '<tt>extractvalue</tt>' instruction extracts the value of a member field
4122 from an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +00004123
4124<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004125<p>The first operand of an '<tt>extractvalue</tt>' instruction is a value
Chris Lattner13ee7952010-08-28 04:09:24 +00004126 of <a href="#t_struct">struct</a> or
Chris Lattner392be582010-02-12 20:49:41 +00004127 <a href="#t_array">array</a> type. The operands are constant indices to
4128 specify which value to extract in a similar manner as indices in a
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004129 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +00004130
4131<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004132<p>The result is the value at the position in the aggregate specified by the
4133 index operands.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +00004134
4135<h5>Example:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +00004136<pre>
Gabor Greif03ab4dc2009-10-28 13:14:50 +00004137 &lt;result&gt; = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohmanb9d66602008-05-12 23:51:09 +00004138</pre>
Dan Gohmanb9d66602008-05-12 23:51:09 +00004139
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004140</div>
Dan Gohmanb9d66602008-05-12 23:51:09 +00004141
4142<!-- _______________________________________________________________________ -->
4143<div class="doc_subsubsection">
4144 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
4145</div>
4146
4147<div class="doc_text">
4148
4149<h5>Syntax:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +00004150<pre>
Jeffrey Yasskinf991bbb2010-01-11 19:19:26 +00004151 &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 +00004152</pre>
4153
4154<h5>Overview:</h5>
Chris Lattner392be582010-02-12 20:49:41 +00004155<p>The '<tt>insertvalue</tt>' instruction inserts a value into a member field
4156 in an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +00004157
4158<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004159<p>The first operand of an '<tt>insertvalue</tt>' instruction is a value
Chris Lattner13ee7952010-08-28 04:09:24 +00004160 of <a href="#t_struct">struct</a> or
Chris Lattner392be582010-02-12 20:49:41 +00004161 <a href="#t_array">array</a> type. The second operand is a first-class
4162 value to insert. The following operands are constant indices indicating
4163 the position at which to insert the value in a similar manner as indices in a
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004164 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction. The
4165 value to insert must have the same type as the value identified by the
4166 indices.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +00004167
4168<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004169<p>The result is an aggregate of the same type as <tt>val</tt>. Its value is
4170 that of <tt>val</tt> except that the value at the position specified by the
4171 indices is that of <tt>elt</tt>.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +00004172
4173<h5>Example:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +00004174<pre>
Jeffrey Yasskinf991bbb2010-01-11 19:19:26 +00004175 %agg1 = insertvalue {i32, float} undef, i32 1, 0 <i>; yields {i32 1, float undef}</i>
4176 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 <i>; yields {i32 1, float %val}</i>
Dan Gohmanb9d66602008-05-12 23:51:09 +00004177</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004178
Dan Gohmanb9d66602008-05-12 23:51:09 +00004179</div>
4180
4181
4182<!-- ======================================================================= -->
Eric Christopher455c5772009-12-05 02:46:03 +00004183<div class="doc_subsection">
Chris Lattner6ab66722006-08-15 00:45:58 +00004184 <a name="memoryops">Memory Access and Addressing Operations</a>
Chris Lattner54611b42005-11-06 08:02:57 +00004185</div>
4186
Misha Brukman76307852003-11-08 01:05:38 +00004187<div class="doc_text">
Chris Lattner54611b42005-11-06 08:02:57 +00004188
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004189<p>A key design point of an SSA-based representation is how it represents
4190 memory. In LLVM, no memory locations are in SSA form, which makes things
Victor Hernandeza70c6df2009-10-26 23:44:29 +00004191 very simple. This section describes how to read, write, and allocate
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004192 memory in LLVM.</p>
Chris Lattner54611b42005-11-06 08:02:57 +00004193
Misha Brukman76307852003-11-08 01:05:38 +00004194</div>
Chris Lattner54611b42005-11-06 08:02:57 +00004195
Chris Lattner2f7c9632001-06-06 20:29:01 +00004196<!-- _______________________________________________________________________ -->
Chris Lattner54611b42005-11-06 08:02:57 +00004197<div class="doc_subsubsection">
Chris Lattner54611b42005-11-06 08:02:57 +00004198 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
4199</div>
4200
Misha Brukman76307852003-11-08 01:05:38 +00004201<div class="doc_text">
Chris Lattner54611b42005-11-06 08:02:57 +00004202
Chris Lattner2f7c9632001-06-06 20:29:01 +00004203<h5>Syntax:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +00004204<pre>
Dan Gohman2140a742010-05-28 01:14:11 +00004205 &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 +00004206</pre>
Chris Lattner54611b42005-11-06 08:02:57 +00004207
Chris Lattner2f7c9632001-06-06 20:29:01 +00004208<h5>Overview:</h5>
Jeff Cohen5819f182007-04-22 01:17:39 +00004209<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004210 currently executing function, to be automatically released when this function
4211 returns to its caller. The object is always allocated in the generic address
4212 space (address space zero).</p>
Chris Lattner54611b42005-11-06 08:02:57 +00004213
Chris Lattner2f7c9632001-06-06 20:29:01 +00004214<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004215<p>The '<tt>alloca</tt>' instruction
4216 allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt> bytes of memory on the
4217 runtime stack, returning a pointer of the appropriate type to the program.
4218 If "NumElements" is specified, it is the number of elements allocated,
4219 otherwise "NumElements" is defaulted to be one. If a constant alignment is
4220 specified, the value result of the allocation is guaranteed to be aligned to
4221 at least that boundary. If not specified, or if zero, the target can choose
4222 to align the allocation on any convenient boundary compatible with the
4223 type.</p>
Chris Lattner54611b42005-11-06 08:02:57 +00004224
Misha Brukman76307852003-11-08 01:05:38 +00004225<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner54611b42005-11-06 08:02:57 +00004226
Chris Lattner2f7c9632001-06-06 20:29:01 +00004227<h5>Semantics:</h5>
Bill Wendling9ee6a312009-05-08 20:49:29 +00004228<p>Memory is allocated; a pointer is returned. The operation is undefined if
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004229 there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
4230 memory is automatically released when the function returns. The
4231 '<tt>alloca</tt>' instruction is commonly used to represent automatic
4232 variables that must have an address available. When the function returns
4233 (either with the <tt><a href="#i_ret">ret</a></tt>
4234 or <tt><a href="#i_unwind">unwind</a></tt> instructions), the memory is
4235 reclaimed. Allocating zero bytes is legal, but the result is undefined.</p>
Chris Lattner54611b42005-11-06 08:02:57 +00004236
Chris Lattner2f7c9632001-06-06 20:29:01 +00004237<h5>Example:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +00004238<pre>
Dan Gohman7a5acb52009-01-04 23:49:44 +00004239 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
4240 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
4241 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
4242 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00004243</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004244
Misha Brukman76307852003-11-08 01:05:38 +00004245</div>
Chris Lattner54611b42005-11-06 08:02:57 +00004246
Chris Lattner2f7c9632001-06-06 20:29:01 +00004247<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +00004248<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
4249Instruction</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004250
Misha Brukman76307852003-11-08 01:05:38 +00004251<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004252
Chris Lattner095735d2002-05-06 03:03:22 +00004253<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004254<pre>
Bill Wendlingbc4024f2010-02-25 21:23:24 +00004255 &lt;result&gt; = load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;]
4256 &lt;result&gt; = volatile load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;]
4257 !&lt;index&gt; = !{ i32 1 }
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004258</pre>
4259
Chris Lattner095735d2002-05-06 03:03:22 +00004260<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +00004261<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004262
Chris Lattner095735d2002-05-06 03:03:22 +00004263<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004264<p>The argument to the '<tt>load</tt>' instruction specifies the memory address
4265 from which to load. The pointer must point to
4266 a <a href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
4267 marked as <tt>volatile</tt>, then the optimizer is not allowed to modify the
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +00004268 number or order of execution of this <tt>load</tt> with other <a
4269 href="#volatile">volatile operations</a>.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004270
Bill Wendlingbc4024f2010-02-25 21:23:24 +00004271<p>The optional constant <tt>align</tt> argument specifies the alignment of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004272 operation (that is, the alignment of the memory address). A value of 0 or an
Bill Wendlingbc4024f2010-02-25 21:23:24 +00004273 omitted <tt>align</tt> argument means that the operation has the preferential
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004274 alignment for the target. It is the responsibility of the code emitter to
4275 ensure that the alignment information is correct. Overestimating the
Bill Wendlingbc4024f2010-02-25 21:23:24 +00004276 alignment results in undefined behavior. Underestimating the alignment may
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004277 produce less efficient code. An alignment of 1 is always safe.</p>
4278
Bill Wendlingbc4024f2010-02-25 21:23:24 +00004279<p>The optional <tt>!nontemporal</tt> metadata must reference a single
4280 metatadata name &lt;index&gt; corresponding to a metadata node with
Dan Gohmana269a0a2010-03-01 17:41:39 +00004281 one <tt>i32</tt> entry of value 1. The existence of
Bill Wendlingbc4024f2010-02-25 21:23:24 +00004282 the <tt>!nontemporal</tt> metatadata on the instruction tells the optimizer
4283 and code generator that this load is not expected to be reused in the cache.
4284 The code generator may select special instructions to save cache bandwidth,
Dan Gohmana269a0a2010-03-01 17:41:39 +00004285 such as the <tt>MOVNT</tt> instruction on x86.</p>
David Greene9641d062010-02-16 20:50:18 +00004286
Chris Lattner095735d2002-05-06 03:03:22 +00004287<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004288<p>The location of memory pointed to is loaded. If the value being loaded is of
4289 scalar type then the number of bytes read does not exceed the minimum number
4290 of bytes needed to hold all bits of the type. For example, loading an
4291 <tt>i24</tt> reads at most three bytes. When loading a value of a type like
4292 <tt>i20</tt> with a size that is not an integral number of bytes, the result
4293 is undefined if the value was not originally written using a store of the
4294 same type.</p>
4295
Chris Lattner095735d2002-05-06 03:03:22 +00004296<h5>Examples:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004297<pre>
4298 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
4299 <a href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00004300 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner095735d2002-05-06 03:03:22 +00004301</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004302
Misha Brukman76307852003-11-08 01:05:38 +00004303</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004304
Chris Lattner095735d2002-05-06 03:03:22 +00004305<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +00004306<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
4307Instruction</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004308
Reid Spencera89fb182006-11-09 21:18:01 +00004309<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004310
Chris Lattner095735d2002-05-06 03:03:22 +00004311<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004312<pre>
Benjamin Kramer79698be2010-07-13 12:26:09 +00004313 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>
4314 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 +00004315</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004316
Chris Lattner095735d2002-05-06 03:03:22 +00004317<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +00004318<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004319
Chris Lattner095735d2002-05-06 03:03:22 +00004320<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004321<p>There are two arguments to the '<tt>store</tt>' instruction: a value to store
4322 and an address at which to store it. The type of the
4323 '<tt>&lt;pointer&gt;</tt>' operand must be a pointer to
4324 the <a href="#t_firstclass">first class</a> type of the
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +00004325 '<tt>&lt;value&gt;</tt>' operand. If the <tt>store</tt> is marked as
4326 <tt>volatile</tt>, then the optimizer is not allowed to modify the number or
4327 order of execution of this <tt>store</tt> with other <a
4328 href="#volatile">volatile operations</a>.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004329
4330<p>The optional constant "align" argument specifies the alignment of the
4331 operation (that is, the alignment of the memory address). A value of 0 or an
4332 omitted "align" argument means that the operation has the preferential
4333 alignment for the target. It is the responsibility of the code emitter to
4334 ensure that the alignment information is correct. Overestimating the
4335 alignment results in an undefined behavior. Underestimating the alignment may
4336 produce less efficient code. An alignment of 1 is always safe.</p>
4337
David Greene9641d062010-02-16 20:50:18 +00004338<p>The optional !nontemporal metadata must reference a single metatadata
Benjamin Kramer79698be2010-07-13 12:26:09 +00004339 name &lt;index&gt; corresponding to a metadata node with one i32 entry of
Dan Gohmana269a0a2010-03-01 17:41:39 +00004340 value 1. The existence of the !nontemporal metatadata on the
David Greene9641d062010-02-16 20:50:18 +00004341 instruction tells the optimizer and code generator that this load is
4342 not expected to be reused in the cache. The code generator may
4343 select special instructions to save cache bandwidth, such as the
Dan Gohmana269a0a2010-03-01 17:41:39 +00004344 MOVNT instruction on x86.</p>
David Greene9641d062010-02-16 20:50:18 +00004345
4346
Chris Lattner48b383b02003-11-25 01:02:51 +00004347<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004348<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>' at the
4349 location specified by the '<tt>&lt;pointer&gt;</tt>' operand. If
4350 '<tt>&lt;value&gt;</tt>' is of scalar type then the number of bytes written
4351 does not exceed the minimum number of bytes needed to hold all bits of the
4352 type. For example, storing an <tt>i24</tt> writes at most three bytes. When
4353 writing a value of a type like <tt>i20</tt> with a size that is not an
4354 integral number of bytes, it is unspecified what happens to the extra bits
4355 that do not belong to the type, but they will typically be overwritten.</p>
4356
Chris Lattner095735d2002-05-06 03:03:22 +00004357<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004358<pre>
4359 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling8830ffe2007-10-22 05:10:05 +00004360 store i32 3, i32* %ptr <i>; yields {void}</i>
4361 %val = <a href="#i_load">load</a> i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner095735d2002-05-06 03:03:22 +00004362</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004363
Reid Spencer443460a2006-11-09 21:15:49 +00004364</div>
4365
Chris Lattner095735d2002-05-06 03:03:22 +00004366<!-- _______________________________________________________________________ -->
Chris Lattner33fd7022004-04-05 01:30:49 +00004367<div class="doc_subsubsection">
4368 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
4369</div>
4370
Misha Brukman76307852003-11-08 01:05:38 +00004371<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004372
Chris Lattner590645f2002-04-14 06:13:44 +00004373<h5>Syntax:</h5>
Chris Lattner33fd7022004-04-05 01:30:49 +00004374<pre>
Matthijs Kooijman0e268272008-10-13 13:44:15 +00004375 &lt;result&gt; = getelementptr &lt;pty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
Dan Gohman1639c392009-07-27 21:53:46 +00004376 &lt;result&gt; = getelementptr inbounds &lt;pty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
Chris Lattner33fd7022004-04-05 01:30:49 +00004377</pre>
4378
Chris Lattner590645f2002-04-14 06:13:44 +00004379<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004380<p>The '<tt>getelementptr</tt>' instruction is used to get the address of a
Chris Lattner392be582010-02-12 20:49:41 +00004381 subelement of an <a href="#t_aggregate">aggregate</a> data structure.
4382 It performs address calculation only and does not access memory.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +00004383
Chris Lattner590645f2002-04-14 06:13:44 +00004384<h5>Arguments:</h5>
Matthijs Kooijman0e268272008-10-13 13:44:15 +00004385<p>The first argument is always a pointer, and forms the basis of the
Chris Lattnera40b9122009-07-29 06:44:13 +00004386 calculation. The remaining arguments are indices that indicate which of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004387 elements of the aggregate object are indexed. The interpretation of each
4388 index is dependent on the type being indexed into. The first index always
4389 indexes the pointer value given as the first argument, the second index
4390 indexes a value of the type pointed to (not necessarily the value directly
4391 pointed to, since the first index can be non-zero), etc. The first type
Chris Lattner392be582010-02-12 20:49:41 +00004392 indexed into must be a pointer value, subsequent types can be arrays,
Chris Lattner13ee7952010-08-28 04:09:24 +00004393 vectors, and structs. Note that subsequent types being indexed into
Chris Lattner392be582010-02-12 20:49:41 +00004394 can never be pointers, since that would require loading the pointer before
4395 continuing calculation.</p>
Matthijs Kooijman0e268272008-10-13 13:44:15 +00004396
4397<p>The type of each index argument depends on the type it is indexing into.
Chris Lattner13ee7952010-08-28 04:09:24 +00004398 When indexing into a (optionally packed) structure, only <tt>i32</tt>
Chris Lattner392be582010-02-12 20:49:41 +00004399 integer <b>constants</b> are allowed. When indexing into an array, pointer
4400 or vector, integers of any width are allowed, and they are not required to be
Chris Lattnera40b9122009-07-29 06:44:13 +00004401 constant.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +00004402
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004403<p>For example, let's consider a C code fragment and how it gets compiled to
4404 LLVM:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +00004405
Benjamin Kramer79698be2010-07-13 12:26:09 +00004406<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +00004407struct RT {
4408 char A;
Chris Lattnera446f1b2007-05-29 15:43:56 +00004409 int B[10][20];
Bill Wendling3716c5d2007-05-29 09:04:49 +00004410 char C;
4411};
4412struct ST {
Chris Lattnera446f1b2007-05-29 15:43:56 +00004413 int X;
Bill Wendling3716c5d2007-05-29 09:04:49 +00004414 double Y;
4415 struct RT Z;
4416};
Chris Lattner33fd7022004-04-05 01:30:49 +00004417
Chris Lattnera446f1b2007-05-29 15:43:56 +00004418int *foo(struct ST *s) {
Bill Wendling3716c5d2007-05-29 09:04:49 +00004419 return &amp;s[1].Z.B[5][13];
4420}
Chris Lattner33fd7022004-04-05 01:30:49 +00004421</pre>
4422
Misha Brukman76307852003-11-08 01:05:38 +00004423<p>The LLVM code generated by the GCC frontend is:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +00004424
Benjamin Kramer79698be2010-07-13 12:26:09 +00004425<pre class="doc_code">
Chris Lattnerbc088212009-01-11 20:53:49 +00004426%RT = <a href="#namedtypes">type</a> { i8 , [10 x [20 x i32]], i8 }
4427%ST = <a href="#namedtypes">type</a> { i32, double, %RT }
Chris Lattner33fd7022004-04-05 01:30:49 +00004428
Dan Gohman6b867702009-07-25 02:23:48 +00004429define i32* @foo(%ST* %s) {
Bill Wendling3716c5d2007-05-29 09:04:49 +00004430entry:
4431 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
4432 ret i32* %reg
4433}
Chris Lattner33fd7022004-04-05 01:30:49 +00004434</pre>
4435
Chris Lattner590645f2002-04-14 06:13:44 +00004436<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +00004437<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004438 type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
4439 }</tt>' type, a structure. The second index indexes into the third element
4440 of the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
4441 i8 }</tt>' type, another structure. The third index indexes into the second
4442 element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
4443 array. The two dimensions of the array are subscripted into, yielding an
4444 '<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a
4445 pointer to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +00004446
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004447<p>Note that it is perfectly legal to index partially through a structure,
4448 returning a pointer to an inner element. Because of this, the LLVM code for
4449 the given testcase is equivalent to:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +00004450
4451<pre>
Dan Gohman6b867702009-07-25 02:23:48 +00004452 define i32* @foo(%ST* %s) {
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00004453 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
Jeff Cohen5819f182007-04-22 01:17:39 +00004454 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
4455 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00004456 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
4457 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
4458 ret i32* %t5
Chris Lattner33fd7022004-04-05 01:30:49 +00004459 }
Chris Lattnera8292f32002-05-06 22:08:29 +00004460</pre>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +00004461
Dan Gohman1639c392009-07-27 21:53:46 +00004462<p>If the <tt>inbounds</tt> keyword is present, the result value of the
Dan Gohman57255802010-04-23 15:23:32 +00004463 <tt>getelementptr</tt> is a <a href="#trapvalues">trap value</a> if the
4464 base pointer is not an <i>in bounds</i> address of an allocated object,
4465 or if any of the addresses that would be formed by successive addition of
4466 the offsets implied by the indices to the base address with infinitely
4467 precise arithmetic are not an <i>in bounds</i> address of that allocated
4468 object. The <i>in bounds</i> addresses for an allocated object are all
4469 the addresses that point into the object, plus the address one byte past
4470 the end.</p>
Dan Gohman1639c392009-07-27 21:53:46 +00004471
4472<p>If the <tt>inbounds</tt> keyword is not present, the offsets are added to
4473 the base address with silently-wrapping two's complement arithmetic, and
4474 the result value of the <tt>getelementptr</tt> may be outside the object
4475 pointed to by the base pointer. The result value may not necessarily be
4476 used to access memory though, even if it happens to point into allocated
4477 storage. See the <a href="#pointeraliasing">Pointer Aliasing Rules</a>
4478 section for more information.</p>
4479
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004480<p>The getelementptr instruction is often confusing. For some more insight into
4481 how it works, see <a href="GetElementPtr.html">the getelementptr FAQ</a>.</p>
Chris Lattner6ab66722006-08-15 00:45:58 +00004482
Chris Lattner590645f2002-04-14 06:13:44 +00004483<h5>Example:</h5>
Chris Lattner33fd7022004-04-05 01:30:49 +00004484<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00004485 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijman0e268272008-10-13 13:44:15 +00004486 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
4487 <i>; yields i8*:vptr</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +00004488 %vptr = getelementptr {i32, &lt;2 x i8&gt;}* %svptr, i64 0, i32 1, i32 1
Matthijs Kooijman0e268272008-10-13 13:44:15 +00004489 <i>; yields i8*:eptr</i>
4490 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Sanjiv Gupta0c155e62009-04-25 07:27:44 +00004491 <i>; yields i32*:iptr</i>
Sanjiv Gupta77abea02009-04-24 16:38:13 +00004492 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
Chris Lattner33fd7022004-04-05 01:30:49 +00004493</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004494
Chris Lattner33fd7022004-04-05 01:30:49 +00004495</div>
Reid Spencer443460a2006-11-09 21:15:49 +00004496
Chris Lattner2f7c9632001-06-06 20:29:01 +00004497<!-- ======================================================================= -->
Reid Spencer97c5fa42006-11-08 01:18:52 +00004498<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
Misha Brukman76307852003-11-08 01:05:38 +00004499</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004500
Misha Brukman76307852003-11-08 01:05:38 +00004501<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004502
Reid Spencer97c5fa42006-11-08 01:18:52 +00004503<p>The instructions in this category are the conversion instructions (casting)
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004504 which all take a single operand and a type. They perform various bit
4505 conversions on the operand.</p>
4506
Misha Brukman76307852003-11-08 01:05:38 +00004507</div>
Chris Lattnerb53c28d2004-03-12 05:50:16 +00004508
Chris Lattnera8292f32002-05-06 22:08:29 +00004509<!-- _______________________________________________________________________ -->
Chris Lattnerb53c28d2004-03-12 05:50:16 +00004510<div class="doc_subsubsection">
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004511 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
4512</div>
4513<div class="doc_text">
4514
4515<h5>Syntax:</h5>
4516<pre>
4517 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4518</pre>
4519
4520<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004521<p>The '<tt>trunc</tt>' instruction truncates its operand to the
4522 type <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004523
4524<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004525<p>The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
4526 be an <a href="#t_integer">integer</a> type, and a type that specifies the
4527 size and type of the result, which must be
4528 an <a href="#t_integer">integer</a> type. The bit size of <tt>value</tt> must
4529 be larger than the bit size of <tt>ty2</tt>. Equal sized types are not
4530 allowed.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004531
4532<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004533<p>The '<tt>trunc</tt>' instruction truncates the high order bits
4534 in <tt>value</tt> and converts the remaining bits to <tt>ty2</tt>. Since the
4535 source size must be larger than the destination size, <tt>trunc</tt> cannot
4536 be a <i>no-op cast</i>. It will always truncate bits.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004537
4538<h5>Example:</h5>
4539<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00004540 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
Reid Spencer36a15422007-01-12 03:35:51 +00004541 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
Gabor Greiff50fd572009-10-28 09:21:30 +00004542 %Z = trunc i32 122 to i1 <i>; yields i1:false</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004543</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004544
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004545</div>
4546
4547<!-- _______________________________________________________________________ -->
4548<div class="doc_subsubsection">
4549 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
4550</div>
4551<div class="doc_text">
4552
4553<h5>Syntax:</h5>
4554<pre>
4555 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4556</pre>
4557
4558<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +00004559<p>The '<tt>zext</tt>' instruction zero extends its operand to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004560 <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004561
4562
4563<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +00004564<p>The '<tt>zext</tt>' instruction takes a value to cast, which must be of
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004565 <a href="#t_integer">integer</a> type, and a type to cast it to, which must
4566 also be of <a href="#t_integer">integer</a> type. The bit size of the
Eric Christopher455c5772009-12-05 02:46:03 +00004567 <tt>value</tt> must be smaller than the bit size of the destination type,
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004568 <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004569
4570<h5>Semantics:</h5>
4571<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004572 bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004573
Reid Spencer07c9c682007-01-12 15:46:11 +00004574<p>When zero extending from i1, the result will always be either 0 or 1.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004575
4576<h5>Example:</h5>
4577<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00004578 %X = zext i32 257 to i64 <i>; yields i64:257</i>
Reid Spencer36a15422007-01-12 03:35:51 +00004579 %Y = zext i1 true to i32 <i>; yields i32:1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004580</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004581
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004582</div>
4583
4584<!-- _______________________________________________________________________ -->
4585<div class="doc_subsubsection">
4586 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
4587</div>
4588<div class="doc_text">
4589
4590<h5>Syntax:</h5>
4591<pre>
4592 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4593</pre>
4594
4595<h5>Overview:</h5>
4596<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
4597
4598<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +00004599<p>The '<tt>sext</tt>' instruction takes a value to cast, which must be of
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004600 <a href="#t_integer">integer</a> type, and a type to cast it to, which must
4601 also be of <a href="#t_integer">integer</a> type. The bit size of the
Eric Christopher455c5772009-12-05 02:46:03 +00004602 <tt>value</tt> must be smaller than the bit size of the destination type,
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004603 <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004604
4605<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004606<p>The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
4607 bit (highest order bit) of the <tt>value</tt> until it reaches the bit size
4608 of the type <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004609
Reid Spencer36a15422007-01-12 03:35:51 +00004610<p>When sign extending from i1, the extension always results in -1 or 0.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004611
4612<h5>Example:</h5>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004613<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00004614 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
Reid Spencer36a15422007-01-12 03:35:51 +00004615 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004616</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004617
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004618</div>
4619
4620<!-- _______________________________________________________________________ -->
4621<div class="doc_subsubsection">
Reid Spencer2e2740d2006-11-09 21:48:10 +00004622 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
4623</div>
4624
4625<div class="doc_text">
4626
4627<h5>Syntax:</h5>
Reid Spencer2e2740d2006-11-09 21:48:10 +00004628<pre>
4629 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4630</pre>
4631
4632<h5>Overview:</h5>
4633<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004634 <tt>ty2</tt>.</p>
Reid Spencer2e2740d2006-11-09 21:48:10 +00004635
4636<h5>Arguments:</h5>
4637<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004638 point</a> value to cast and a <a href="#t_floating">floating point</a> type
4639 to cast it to. The size of <tt>value</tt> must be larger than the size of
Eric Christopher455c5772009-12-05 02:46:03 +00004640 <tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004641 <i>no-op cast</i>.</p>
Reid Spencer2e2740d2006-11-09 21:48:10 +00004642
4643<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004644<p>The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
Eric Christopher455c5772009-12-05 02:46:03 +00004645 <a href="#t_floating">floating point</a> type to a smaller
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004646 <a href="#t_floating">floating point</a> type. If the value cannot fit
4647 within the destination type, <tt>ty2</tt>, then the results are
4648 undefined.</p>
Reid Spencer2e2740d2006-11-09 21:48:10 +00004649
4650<h5>Example:</h5>
4651<pre>
4652 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
4653 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
4654</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004655
Reid Spencer2e2740d2006-11-09 21:48:10 +00004656</div>
4657
4658<!-- _______________________________________________________________________ -->
4659<div class="doc_subsubsection">
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004660 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
4661</div>
4662<div class="doc_text">
4663
4664<h5>Syntax:</h5>
4665<pre>
4666 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4667</pre>
4668
4669<h5>Overview:</h5>
4670<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004671 floating point value.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004672
4673<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +00004674<p>The '<tt>fpext</tt>' instruction takes a
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004675 <a href="#t_floating">floating point</a> <tt>value</tt> to cast, and
4676 a <a href="#t_floating">floating point</a> type to cast it to. The source
4677 type must be smaller than the destination type.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004678
4679<h5>Semantics:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +00004680<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004681 <a href="#t_floating">floating point</a> type to a larger
4682 <a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
4683 used to make a <i>no-op cast</i> because it always changes bits. Use
4684 <tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004685
4686<h5>Example:</h5>
4687<pre>
4688 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
4689 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
4690</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004691
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004692</div>
4693
4694<!-- _______________________________________________________________________ -->
4695<div class="doc_subsubsection">
Reid Spencer2eadb532007-01-21 00:29:26 +00004696 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004697</div>
4698<div class="doc_text">
4699
4700<h5>Syntax:</h5>
4701<pre>
Reid Spencer753163d2007-07-31 14:40:14 +00004702 &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 +00004703</pre>
4704
4705<h5>Overview:</h5>
Reid Spencer753163d2007-07-31 14:40:14 +00004706<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004707 unsigned integer equivalent of type <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004708
4709<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004710<p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
4711 scalar or vector <a href="#t_floating">floating point</a> value, and a type
4712 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4713 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4714 vector integer type with the same number of elements as <tt>ty</tt></p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004715
4716<h5>Semantics:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +00004717<p>The '<tt>fptoui</tt>' instruction converts its
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004718 <a href="#t_floating">floating point</a> operand into the nearest (rounding
4719 towards zero) unsigned integer value. If the value cannot fit
4720 in <tt>ty2</tt>, the results are undefined.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004721
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004722<h5>Example:</h5>
4723<pre>
Reid Spencer753163d2007-07-31 14:40:14 +00004724 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner5b95a172007-09-22 03:17:52 +00004725 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Gabor Greiff50fd572009-10-28 09:21:30 +00004726 %Z = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004727</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004728
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004729</div>
4730
4731<!-- _______________________________________________________________________ -->
4732<div class="doc_subsubsection">
Reid Spencer51b07252006-11-09 23:03:26 +00004733 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004734</div>
4735<div class="doc_text">
4736
4737<h5>Syntax:</h5>
4738<pre>
Reid Spencer51b07252006-11-09 23:03:26 +00004739 &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 +00004740</pre>
4741
4742<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +00004743<p>The '<tt>fptosi</tt>' instruction converts
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004744 <a href="#t_floating">floating point</a> <tt>value</tt> to
4745 type <tt>ty2</tt>.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +00004746
Chris Lattnera8292f32002-05-06 22:08:29 +00004747<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004748<p>The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
4749 scalar or vector <a href="#t_floating">floating point</a> value, and a type
4750 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4751 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4752 vector integer type with the same number of elements as <tt>ty</tt></p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +00004753
Chris Lattnera8292f32002-05-06 22:08:29 +00004754<h5>Semantics:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +00004755<p>The '<tt>fptosi</tt>' instruction converts its
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004756 <a href="#t_floating">floating point</a> operand into the nearest (rounding
4757 towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
4758 the results are undefined.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +00004759
Chris Lattner70de6632001-07-09 00:26:23 +00004760<h5>Example:</h5>
Chris Lattnerb53c28d2004-03-12 05:50:16 +00004761<pre>
Reid Spencer36a15422007-01-12 03:35:51 +00004762 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner5b95a172007-09-22 03:17:52 +00004763 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Gabor Greiff50fd572009-10-28 09:21:30 +00004764 %Z = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004765</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004766
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004767</div>
4768
4769<!-- _______________________________________________________________________ -->
4770<div class="doc_subsubsection">
Reid Spencer51b07252006-11-09 23:03:26 +00004771 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004772</div>
4773<div class="doc_text">
4774
4775<h5>Syntax:</h5>
4776<pre>
Reid Spencer51b07252006-11-09 23:03:26 +00004777 &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 +00004778</pre>
4779
4780<h5>Overview:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +00004781<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004782 integer and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004783
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004784<h5>Arguments:</h5>
Nate Begemand4d45c22007-11-17 03:58:34 +00004785<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be a
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004786 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
4787 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4788 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4789 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004790
4791<h5>Semantics:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +00004792<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004793 integer quantity and converts it to the corresponding floating point
4794 value. If the value cannot fit in the floating point value, the results are
4795 undefined.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004796
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004797<h5>Example:</h5>
4798<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00004799 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +00004800 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004801</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004802
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004803</div>
4804
4805<!-- _______________________________________________________________________ -->
4806<div class="doc_subsubsection">
Reid Spencer51b07252006-11-09 23:03:26 +00004807 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004808</div>
4809<div class="doc_text">
4810
4811<h5>Syntax:</h5>
4812<pre>
Reid Spencer51b07252006-11-09 23:03:26 +00004813 &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 +00004814</pre>
4815
4816<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004817<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed integer
4818 and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004819
4820<h5>Arguments:</h5>
Nate Begemand4d45c22007-11-17 03:58:34 +00004821<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be a
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004822 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
4823 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4824 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4825 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004826
4827<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004828<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed integer
4829 quantity and converts it to the corresponding floating point value. If the
4830 value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004831
4832<h5>Example:</h5>
4833<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00004834 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +00004835 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004836</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004837
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004838</div>
4839
4840<!-- _______________________________________________________________________ -->
4841<div class="doc_subsubsection">
Reid Spencerb7344ff2006-11-11 21:00:47 +00004842 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
4843</div>
4844<div class="doc_text">
4845
4846<h5>Syntax:</h5>
4847<pre>
4848 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4849</pre>
4850
4851<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004852<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
4853 the integer type <tt>ty2</tt>.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +00004854
4855<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004856<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
4857 must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
4858 <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> type.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +00004859
4860<h5>Semantics:</h5>
4861<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004862 <tt>ty2</tt> by interpreting the pointer value as an integer and either
4863 truncating or zero extending that value to the size of the integer type. If
4864 <tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
4865 <tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
4866 are the same size, then nothing is done (<i>no-op cast</i>) other than a type
4867 change.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +00004868
4869<h5>Example:</h5>
4870<pre>
Jeff Cohen222a8a42007-04-29 01:07:00 +00004871 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
4872 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
Reid Spencerb7344ff2006-11-11 21:00:47 +00004873</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004874
Reid Spencerb7344ff2006-11-11 21:00:47 +00004875</div>
4876
4877<!-- _______________________________________________________________________ -->
4878<div class="doc_subsubsection">
4879 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
4880</div>
4881<div class="doc_text">
4882
4883<h5>Syntax:</h5>
4884<pre>
4885 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4886</pre>
4887
4888<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004889<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to a
4890 pointer type, <tt>ty2</tt>.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +00004891
4892<h5>Arguments:</h5>
Duncan Sands16f122e2007-03-30 12:22:09 +00004893<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004894 value to cast, and a type to cast it to, which must be a
4895 <a href="#t_pointer">pointer</a> type.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +00004896
4897<h5>Semantics:</h5>
4898<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004899 <tt>ty2</tt> by applying either a zero extension or a truncation depending on
4900 the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
4901 size of a pointer then a truncation is done. If <tt>value</tt> is smaller
4902 than the size of a pointer then a zero extension is done. If they are the
4903 same size, nothing is done (<i>no-op cast</i>).</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +00004904
4905<h5>Example:</h5>
4906<pre>
Jeff Cohen222a8a42007-04-29 01:07:00 +00004907 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
Gabor Greiff50fd572009-10-28 09:21:30 +00004908 %Y = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
4909 %Z = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
Reid Spencerb7344ff2006-11-11 21:00:47 +00004910</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004911
Reid Spencerb7344ff2006-11-11 21:00:47 +00004912</div>
4913
4914<!-- _______________________________________________________________________ -->
4915<div class="doc_subsubsection">
Reid Spencer5b950642006-11-11 23:08:07 +00004916 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004917</div>
4918<div class="doc_text">
4919
4920<h5>Syntax:</h5>
4921<pre>
Reid Spencer5b950642006-11-11 23:08:07 +00004922 &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 +00004923</pre>
4924
4925<h5>Overview:</h5>
Reid Spencer5b950642006-11-11 23:08:07 +00004926<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004927 <tt>ty2</tt> without changing any bits.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004928
4929<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004930<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be a
4931 non-aggregate first class value, and a type to cast it to, which must also be
4932 a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes
4933 of <tt>value</tt> and the destination type, <tt>ty2</tt>, must be
4934 identical. If the source type is a pointer, the destination type must also be
4935 a pointer. This instruction supports bitwise conversion of vectors to
4936 integers and to vectors of other types (as long as they have the same
4937 size).</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004938
4939<h5>Semantics:</h5>
Reid Spencer5b950642006-11-11 23:08:07 +00004940<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004941 <tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
4942 this conversion. The conversion is done as if the <tt>value</tt> had been
4943 stored to memory and read back as type <tt>ty2</tt>. Pointer types may only
4944 be converted to other pointer types with this instruction. To convert
4945 pointers to other types, use the <a href="#i_inttoptr">inttoptr</a> or
4946 <a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00004947
4948<h5>Example:</h5>
4949<pre>
Jeff Cohen222a8a42007-04-29 01:07:00 +00004950 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00004951 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Eric Christopher455c5772009-12-05 02:46:03 +00004952 %Z = bitcast &lt;2 x int&gt; %V to i64; <i>; yields i64: %V</i>
Chris Lattner70de6632001-07-09 00:26:23 +00004953</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004954
Misha Brukman76307852003-11-08 01:05:38 +00004955</div>
Chris Lattnerb53c28d2004-03-12 05:50:16 +00004956
Reid Spencer97c5fa42006-11-08 01:18:52 +00004957<!-- ======================================================================= -->
4958<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004959
Reid Spencer97c5fa42006-11-08 01:18:52 +00004960<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004961
4962<p>The instructions in this category are the "miscellaneous" instructions, which
4963 defy better classification.</p>
4964
Reid Spencer97c5fa42006-11-08 01:18:52 +00004965</div>
Reid Spencerc828a0e2006-11-18 21:50:54 +00004966
4967<!-- _______________________________________________________________________ -->
4968<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
4969</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004970
Reid Spencerc828a0e2006-11-18 21:50:54 +00004971<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004972
Reid Spencerc828a0e2006-11-18 21:50:54 +00004973<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004974<pre>
4975 &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 +00004976</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004977
Reid Spencerc828a0e2006-11-18 21:50:54 +00004978<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004979<p>The '<tt>icmp</tt>' instruction returns a boolean value or a vector of
4980 boolean values based on comparison of its two integer, integer vector, or
4981 pointer operands.</p>
4982
Reid Spencerc828a0e2006-11-18 21:50:54 +00004983<h5>Arguments:</h5>
4984<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
Bill Wendlingd9a66f72009-07-20 02:29:24 +00004985 the condition code indicating the kind of comparison to perform. It is not a
4986 value, just a keyword. The possible condition code are:</p>
4987
Reid Spencerc828a0e2006-11-18 21:50:54 +00004988<ol>
4989 <li><tt>eq</tt>: equal</li>
4990 <li><tt>ne</tt>: not equal </li>
4991 <li><tt>ugt</tt>: unsigned greater than</li>
4992 <li><tt>uge</tt>: unsigned greater or equal</li>
4993 <li><tt>ult</tt>: unsigned less than</li>
4994 <li><tt>ule</tt>: unsigned less or equal</li>
4995 <li><tt>sgt</tt>: signed greater than</li>
4996 <li><tt>sge</tt>: signed greater or equal</li>
4997 <li><tt>slt</tt>: signed less than</li>
4998 <li><tt>sle</tt>: signed less or equal</li>
4999</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005000
Chris Lattnerc0f423a2007-01-15 01:54:13 +00005001<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005002 <a href="#t_pointer">pointer</a> or integer <a href="#t_vector">vector</a>
5003 typed. They must also be identical types.</p>
5004
Reid Spencerc828a0e2006-11-18 21:50:54 +00005005<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005006<p>The '<tt>icmp</tt>' compares <tt>op1</tt> and <tt>op2</tt> according to the
5007 condition code given as <tt>cond</tt>. The comparison performed always yields
Nick Lewycky84a1eeb2009-09-27 00:45:11 +00005008 either an <a href="#t_integer"><tt>i1</tt></a> or vector of <tt>i1</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005009 result, as follows:</p>
5010
Reid Spencerc828a0e2006-11-18 21:50:54 +00005011<ol>
Eric Christopher455c5772009-12-05 02:46:03 +00005012 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005013 <tt>false</tt> otherwise. No sign interpretation is necessary or
5014 performed.</li>
5015
Eric Christopher455c5772009-12-05 02:46:03 +00005016 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005017 <tt>false</tt> otherwise. No sign interpretation is necessary or
5018 performed.</li>
5019
Reid Spencerc828a0e2006-11-18 21:50:54 +00005020 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005021 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5022
Reid Spencerc828a0e2006-11-18 21:50:54 +00005023 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005024 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5025 to <tt>op2</tt>.</li>
5026
Reid Spencerc828a0e2006-11-18 21:50:54 +00005027 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005028 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5029
Reid Spencerc828a0e2006-11-18 21:50:54 +00005030 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005031 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5032
Reid Spencerc828a0e2006-11-18 21:50:54 +00005033 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005034 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5035
Reid Spencerc828a0e2006-11-18 21:50:54 +00005036 <li><tt>sge</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005037 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5038 to <tt>op2</tt>.</li>
5039
Reid Spencerc828a0e2006-11-18 21:50:54 +00005040 <li><tt>slt</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005041 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5042
Reid Spencerc828a0e2006-11-18 21:50:54 +00005043 <li><tt>sle</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005044 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +00005045</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005046
Reid Spencerc828a0e2006-11-18 21:50:54 +00005047<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005048 values are compared as if they were integers.</p>
5049
5050<p>If the operands are integer vectors, then they are compared element by
5051 element. The result is an <tt>i1</tt> vector with the same number of elements
5052 as the values being compared. Otherwise, the result is an <tt>i1</tt>.</p>
Reid Spencerc828a0e2006-11-18 21:50:54 +00005053
5054<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005055<pre>
5056 &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00005057 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
5058 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
5059 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
5060 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
5061 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
Reid Spencerc828a0e2006-11-18 21:50:54 +00005062</pre>
Dan Gohmana5127ab2009-01-22 01:39:38 +00005063
5064<p>Note that the code generator does not yet support vector types with
5065 the <tt>icmp</tt> instruction.</p>
5066
Reid Spencerc828a0e2006-11-18 21:50:54 +00005067</div>
5068
5069<!-- _______________________________________________________________________ -->
5070<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
5071</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005072
Reid Spencerc828a0e2006-11-18 21:50:54 +00005073<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005074
Reid Spencerc828a0e2006-11-18 21:50:54 +00005075<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005076<pre>
5077 &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 +00005078</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005079
Reid Spencerc828a0e2006-11-18 21:50:54 +00005080<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005081<p>The '<tt>fcmp</tt>' instruction returns a boolean value or vector of boolean
5082 values based on comparison of its operands.</p>
5083
5084<p>If the operands are floating point scalars, then the result type is a boolean
Nick Lewycky84a1eeb2009-09-27 00:45:11 +00005085(<a href="#t_integer"><tt>i1</tt></a>).</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005086
5087<p>If the operands are floating point vectors, then the result type is a vector
5088 of boolean with the same number of elements as the operands being
5089 compared.</p>
5090
Reid Spencerc828a0e2006-11-18 21:50:54 +00005091<h5>Arguments:</h5>
5092<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005093 the condition code indicating the kind of comparison to perform. It is not a
5094 value, just a keyword. The possible condition code are:</p>
5095
Reid Spencerc828a0e2006-11-18 21:50:54 +00005096<ol>
Reid Spencerf69acf32006-11-19 03:00:14 +00005097 <li><tt>false</tt>: no comparison, always returns false</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +00005098 <li><tt>oeq</tt>: ordered and equal</li>
5099 <li><tt>ogt</tt>: ordered and greater than </li>
5100 <li><tt>oge</tt>: ordered and greater than or equal</li>
5101 <li><tt>olt</tt>: ordered and less than </li>
5102 <li><tt>ole</tt>: ordered and less than or equal</li>
5103 <li><tt>one</tt>: ordered and not equal</li>
5104 <li><tt>ord</tt>: ordered (no nans)</li>
5105 <li><tt>ueq</tt>: unordered or equal</li>
5106 <li><tt>ugt</tt>: unordered or greater than </li>
5107 <li><tt>uge</tt>: unordered or greater than or equal</li>
5108 <li><tt>ult</tt>: unordered or less than </li>
5109 <li><tt>ule</tt>: unordered or less than or equal</li>
5110 <li><tt>une</tt>: unordered or not equal</li>
5111 <li><tt>uno</tt>: unordered (either nans)</li>
Reid Spencerf69acf32006-11-19 03:00:14 +00005112 <li><tt>true</tt>: no comparison, always returns true</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +00005113</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005114
Jeff Cohen222a8a42007-04-29 01:07:00 +00005115<p><i>Ordered</i> means that neither operand is a QNAN while
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005116 <i>unordered</i> means that either operand may be a QNAN.</p>
5117
5118<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be either
5119 a <a href="#t_floating">floating point</a> type or
5120 a <a href="#t_vector">vector</a> of floating point type. They must have
5121 identical types.</p>
5122
Reid Spencerc828a0e2006-11-18 21:50:54 +00005123<h5>Semantics:</h5>
Gabor Greif0f75ad02008-08-07 21:46:00 +00005124<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005125 according to the condition code given as <tt>cond</tt>. If the operands are
5126 vectors, then the vectors are compared element by element. Each comparison
Nick Lewycky84a1eeb2009-09-27 00:45:11 +00005127 performed always yields an <a href="#t_integer">i1</a> result, as
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005128 follows:</p>
5129
Reid Spencerc828a0e2006-11-18 21:50:54 +00005130<ol>
5131 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005132
Eric Christopher455c5772009-12-05 02:46:03 +00005133 <li><tt>oeq</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005134 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5135
Reid Spencerf69acf32006-11-19 03:00:14 +00005136 <li><tt>ogt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Dan Gohmana269a0a2010-03-01 17:41:39 +00005137 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005138
Eric Christopher455c5772009-12-05 02:46:03 +00005139 <li><tt>oge</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005140 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5141
Eric Christopher455c5772009-12-05 02:46:03 +00005142 <li><tt>olt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005143 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5144
Eric Christopher455c5772009-12-05 02:46:03 +00005145 <li><tt>ole</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005146 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5147
Eric Christopher455c5772009-12-05 02:46:03 +00005148 <li><tt>one</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005149 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5150
Reid Spencerf69acf32006-11-19 03:00:14 +00005151 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005152
Eric Christopher455c5772009-12-05 02:46:03 +00005153 <li><tt>ueq</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005154 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5155
Eric Christopher455c5772009-12-05 02:46:03 +00005156 <li><tt>ugt</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005157 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5158
Eric Christopher455c5772009-12-05 02:46:03 +00005159 <li><tt>uge</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005160 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5161
Eric Christopher455c5772009-12-05 02:46:03 +00005162 <li><tt>ult</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005163 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5164
Eric Christopher455c5772009-12-05 02:46:03 +00005165 <li><tt>ule</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005166 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5167
Eric Christopher455c5772009-12-05 02:46:03 +00005168 <li><tt>une</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005169 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5170
Reid Spencerf69acf32006-11-19 03:00:14 +00005171 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005172
Reid Spencerc828a0e2006-11-18 21:50:54 +00005173 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
5174</ol>
Reid Spencerc828a0e2006-11-18 21:50:54 +00005175
5176<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005177<pre>
5178 &lt;result&gt; = fcmp oeq float 4.0, 5.0 <i>; yields: result=false</i>
Dan Gohmanc579d972008-09-09 01:02:47 +00005179 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
5180 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
5181 &lt;result&gt; = fcmp ueq double 1.0, 2.0 <i>; yields: result=false</i>
Reid Spencerc828a0e2006-11-18 21:50:54 +00005182</pre>
Dan Gohmana5127ab2009-01-22 01:39:38 +00005183
5184<p>Note that the code generator does not yet support vector types with
5185 the <tt>fcmp</tt> instruction.</p>
5186
Reid Spencerc828a0e2006-11-18 21:50:54 +00005187</div>
5188
Reid Spencer97c5fa42006-11-08 01:18:52 +00005189<!-- _______________________________________________________________________ -->
Nate Begemand2195702008-05-12 19:01:56 +00005190<div class="doc_subsubsection">
Chris Lattnerb54c30f2008-05-20 20:48:21 +00005191 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
5192</div>
5193
Reid Spencer97c5fa42006-11-08 01:18:52 +00005194<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +00005195
Reid Spencer97c5fa42006-11-08 01:18:52 +00005196<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005197<pre>
5198 &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...
5199</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00005200
Reid Spencer97c5fa42006-11-08 01:18:52 +00005201<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005202<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in the
5203 SSA graph representing the function.</p>
5204
Reid Spencer97c5fa42006-11-08 01:18:52 +00005205<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005206<p>The type of the incoming values is specified with the first type field. After
5207 this, the '<tt>phi</tt>' instruction takes a list of pairs as arguments, with
5208 one pair for each predecessor basic block of the current block. Only values
5209 of <a href="#t_firstclass">first class</a> type may be used as the value
5210 arguments to the PHI node. Only labels may be used as the label
5211 arguments.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00005212
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005213<p>There must be no non-phi instructions between the start of a basic block and
5214 the PHI instructions: i.e. PHI instructions must be first in a basic
5215 block.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00005216
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005217<p>For the purposes of the SSA form, the use of each incoming value is deemed to
5218 occur on the edge from the corresponding predecessor block to the current
5219 block (but after any definition of an '<tt>invoke</tt>' instruction's return
5220 value on the same edge).</p>
Jay Foad1a4eea52009-06-03 10:20:10 +00005221
Reid Spencer97c5fa42006-11-08 01:18:52 +00005222<h5>Semantics:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +00005223<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005224 specified by the pair corresponding to the predecessor basic block that
5225 executed just prior to the current block.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00005226
Reid Spencer97c5fa42006-11-08 01:18:52 +00005227<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00005228<pre>
5229Loop: ; Infinite loop that counts from 0 on up...
5230 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
5231 %nextindvar = add i32 %indvar, 1
5232 br label %Loop
5233</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005234
Reid Spencer97c5fa42006-11-08 01:18:52 +00005235</div>
5236
Chris Lattnerb53c28d2004-03-12 05:50:16 +00005237<!-- _______________________________________________________________________ -->
5238<div class="doc_subsubsection">
5239 <a name="i_select">'<tt>select</tt>' Instruction</a>
5240</div>
5241
5242<div class="doc_text">
5243
5244<h5>Syntax:</h5>
Chris Lattnerb53c28d2004-03-12 05:50:16 +00005245<pre>
Dan Gohmanc579d972008-09-09 01:02:47 +00005246 &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>
5247
Dan Gohmanef9462f2008-10-14 16:51:45 +00005248 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Chris Lattnerb53c28d2004-03-12 05:50:16 +00005249</pre>
5250
5251<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005252<p>The '<tt>select</tt>' instruction is used to choose one value based on a
5253 condition, without branching.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +00005254
5255
5256<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005257<p>The '<tt>select</tt>' instruction requires an 'i1' value or a vector of 'i1'
5258 values indicating the condition, and two values of the
5259 same <a href="#t_firstclass">first class</a> type. If the val1/val2 are
5260 vectors and the condition is a scalar, then entire vectors are selected, not
5261 individual elements.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +00005262
5263<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005264<p>If the condition is an i1 and it evaluates to 1, the instruction returns the
5265 first value argument; otherwise, it returns the second value argument.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +00005266
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005267<p>If the condition is a vector of i1, then the value arguments must be vectors
5268 of the same size, and the selection is done element by element.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +00005269
5270<h5>Example:</h5>
Chris Lattnerb53c28d2004-03-12 05:50:16 +00005271<pre>
Reid Spencer36a15422007-01-12 03:35:51 +00005272 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
Chris Lattnerb53c28d2004-03-12 05:50:16 +00005273</pre>
Dan Gohmana5127ab2009-01-22 01:39:38 +00005274
5275<p>Note that the code generator does not yet support conditions
5276 with vector type.</p>
5277
Chris Lattnerb53c28d2004-03-12 05:50:16 +00005278</div>
5279
Robert Bocchinof72fdfe2006-01-15 20:48:27 +00005280<!-- _______________________________________________________________________ -->
5281<div class="doc_subsubsection">
Chris Lattnere23c1392005-05-06 05:47:36 +00005282 <a name="i_call">'<tt>call</tt>' Instruction</a>
5283</div>
5284
Misha Brukman76307852003-11-08 01:05:38 +00005285<div class="doc_text">
Chris Lattnere23c1392005-05-06 05:47:36 +00005286
Chris Lattner2f7c9632001-06-06 20:29:01 +00005287<h5>Syntax:</h5>
Chris Lattnere23c1392005-05-06 05:47:36 +00005288<pre>
Devang Patel02256232008-10-07 17:48:33 +00005289 &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 +00005290</pre>
5291
Chris Lattner2f7c9632001-06-06 20:29:01 +00005292<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +00005293<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattnere23c1392005-05-06 05:47:36 +00005294
Chris Lattner2f7c9632001-06-06 20:29:01 +00005295<h5>Arguments:</h5>
Misha Brukman76307852003-11-08 01:05:38 +00005296<p>This instruction requires several arguments:</p>
Chris Lattnere23c1392005-05-06 05:47:36 +00005297
Chris Lattnera8292f32002-05-06 22:08:29 +00005298<ol>
Jeffrey Yasskinb8677462010-01-09 19:44:16 +00005299 <li>The optional "tail" marker indicates that the callee function does not
5300 access any allocas or varargs in the caller. Note that calls may be
5301 marked "tail" even if they do not occur before
5302 a <a href="#i_ret"><tt>ret</tt></a> instruction. If the "tail" marker is
5303 present, the function call is eligible for tail call optimization,
5304 but <a href="CodeGenerator.html#tailcallopt">might not in fact be
Evan Cheng59676492010-03-08 21:05:02 +00005305 optimized into a jump</a>. The code generator may optimize calls marked
5306 "tail" with either 1) automatic <a href="CodeGenerator.html#sibcallopt">
5307 sibling call optimization</a> when the caller and callee have
5308 matching signatures, or 2) forced tail call optimization when the
5309 following extra requirements are met:
Jeffrey Yasskinb8677462010-01-09 19:44:16 +00005310 <ul>
5311 <li>Caller and callee both have the calling
5312 convention <tt>fastcc</tt>.</li>
5313 <li>The call is in tail position (ret immediately follows call and ret
5314 uses value of call or is void).</li>
5315 <li>Option <tt>-tailcallopt</tt> is enabled,
Dan Gohman6232f732010-03-02 01:08:11 +00005316 or <code>llvm::GuaranteedTailCallOpt</code> is <code>true</code>.</li>
Jeffrey Yasskinb8677462010-01-09 19:44:16 +00005317 <li><a href="CodeGenerator.html#tailcallopt">Platform specific
5318 constraints are met.</a></li>
5319 </ul>
5320 </li>
Devang Patel7e9b05e2008-10-06 18:50:38 +00005321
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005322 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
5323 convention</a> the call should use. If none is specified, the call
Jeffrey Yasskinb8677462010-01-09 19:44:16 +00005324 defaults to using C calling conventions. The calling convention of the
5325 call must match the calling convention of the target function, or else the
5326 behavior is undefined.</li>
Devang Patel7e9b05e2008-10-06 18:50:38 +00005327
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005328 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
5329 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
5330 '<tt>inreg</tt>' attributes are valid here.</li>
5331
5332 <li>'<tt>ty</tt>': the type of the call instruction itself which is also the
5333 type of the return value. Functions that return no value are marked
5334 <tt><a href="#t_void">void</a></tt>.</li>
5335
5336 <li>'<tt>fnty</tt>': shall be the signature of the pointer to function value
5337 being invoked. The argument types must match the types implied by this
5338 signature. This type can be omitted if the function is not varargs and if
5339 the function type does not return a pointer to a function.</li>
5340
5341 <li>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
5342 be invoked. In most cases, this is a direct function invocation, but
5343 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
5344 to function value.</li>
5345
5346 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner47f2a832010-03-02 06:36:51 +00005347 signature argument types and parameter attributes. All arguments must be
5348 of <a href="#t_firstclass">first class</a> type. If the function
5349 signature indicates the function accepts a variable number of arguments,
5350 the extra arguments can be specified.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005351
5352 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
5353 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
5354 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattnera8292f32002-05-06 22:08:29 +00005355</ol>
Chris Lattnere23c1392005-05-06 05:47:36 +00005356
Chris Lattner2f7c9632001-06-06 20:29:01 +00005357<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005358<p>The '<tt>call</tt>' instruction is used to cause control flow to transfer to
5359 a specified function, with its incoming arguments bound to the specified
5360 values. Upon a '<tt><a href="#i_ret">ret</a></tt>' instruction in the called
5361 function, control flow continues with the instruction after the function
5362 call, and the return value of the function is bound to the result
5363 argument.</p>
Chris Lattnere23c1392005-05-06 05:47:36 +00005364
Chris Lattner2f7c9632001-06-06 20:29:01 +00005365<h5>Example:</h5>
Chris Lattnere23c1392005-05-06 05:47:36 +00005366<pre>
Nick Lewyckya9b13d52007-09-08 13:57:50 +00005367 %retval = call i32 @test(i32 %argc)
Dan Gohmanaabfdb32010-05-28 17:13:49 +00005368 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) <i>; yields i32</i>
Chris Lattnerfb7c88d2008-03-21 17:24:17 +00005369 %X = tail call i32 @foo() <i>; yields i32</i>
5370 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
5371 call void %foo(i8 97 signext)
Devang Pateld6cff512008-03-10 20:49:15 +00005372
5373 %struct.A = type { i32, i8 }
Devang Patel7e9b05e2008-10-06 18:50:38 +00005374 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohmancc3132e2008-10-04 19:00:07 +00005375 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
5376 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattner6cbe8e92008-10-08 06:26:11 +00005377 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijmaneefa7df2008-10-07 10:03:45 +00005378 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Chris Lattnere23c1392005-05-06 05:47:36 +00005379</pre>
5380
Dale Johannesen68f971b2009-09-24 18:38:21 +00005381<p>llvm treats calls to some functions with names and arguments that match the
Dale Johannesen722212d2009-09-25 17:04:42 +00005382standard C99 library as being the C99 library functions, and may perform
5383optimizations or generate code for them under that assumption. This is
5384something we'd like to change in the future to provide better support for
Dan Gohmana269a0a2010-03-01 17:41:39 +00005385freestanding environments and non-C-based languages.</p>
Dale Johannesen68f971b2009-09-24 18:38:21 +00005386
Misha Brukman76307852003-11-08 01:05:38 +00005387</div>
Chris Lattner6a4a0492004-09-27 21:51:25 +00005388
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00005389<!-- _______________________________________________________________________ -->
Chris Lattner6a4a0492004-09-27 21:51:25 +00005390<div class="doc_subsubsection">
Chris Lattner33337472006-01-13 23:26:01 +00005391 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
Chris Lattner6a4a0492004-09-27 21:51:25 +00005392</div>
5393
Misha Brukman76307852003-11-08 01:05:38 +00005394<div class="doc_text">
Chris Lattner6a4a0492004-09-27 21:51:25 +00005395
Chris Lattner26ca62e2003-10-18 05:51:36 +00005396<h5>Syntax:</h5>
Chris Lattner6a4a0492004-09-27 21:51:25 +00005397<pre>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +00005398 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
Chris Lattner6a4a0492004-09-27 21:51:25 +00005399</pre>
5400
Chris Lattner26ca62e2003-10-18 05:51:36 +00005401<h5>Overview:</h5>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +00005402<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005403 the "variable argument" area of a function call. It is used to implement the
5404 <tt>va_arg</tt> macro in C.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +00005405
Chris Lattner26ca62e2003-10-18 05:51:36 +00005406<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005407<p>This instruction takes a <tt>va_list*</tt> value and the type of the
5408 argument. It returns a value of the specified argument type and increments
5409 the <tt>va_list</tt> to point to the next argument. The actual type
5410 of <tt>va_list</tt> is target specific.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +00005411
Chris Lattner26ca62e2003-10-18 05:51:36 +00005412<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005413<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified type
5414 from the specified <tt>va_list</tt> and causes the <tt>va_list</tt> to point
5415 to the next argument. For more information, see the variable argument
5416 handling <a href="#int_varargs">Intrinsic Functions</a>.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +00005417
5418<p>It is legal for this instruction to be called in a function which does not
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005419 take a variable number of arguments, for example, the <tt>vfprintf</tt>
5420 function.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +00005421
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005422<p><tt>va_arg</tt> is an LLVM instruction instead of
5423 an <a href="#intrinsics">intrinsic function</a> because it takes a type as an
5424 argument.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +00005425
Chris Lattner26ca62e2003-10-18 05:51:36 +00005426<h5>Example:</h5>
Chris Lattner6a4a0492004-09-27 21:51:25 +00005427<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
5428
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005429<p>Note that the code generator does not yet fully support va_arg on many
5430 targets. Also, it does not currently support va_arg with aggregate types on
5431 any target.</p>
Dan Gohman3065b612009-01-12 23:12:39 +00005432
Misha Brukman76307852003-11-08 01:05:38 +00005433</div>
Chris Lattner941515c2004-01-06 05:31:32 +00005434
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00005435<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +00005436<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
5437<!-- *********************************************************************** -->
Chris Lattner941515c2004-01-06 05:31:32 +00005438
Misha Brukman76307852003-11-08 01:05:38 +00005439<div class="doc_text">
Chris Lattnerfee11462004-02-12 17:01:32 +00005440
5441<p>LLVM supports the notion of an "intrinsic function". These functions have
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005442 well known names and semantics and are required to follow certain
5443 restrictions. Overall, these intrinsics represent an extension mechanism for
5444 the LLVM language that does not require changing all of the transformations
5445 in LLVM when adding to the language (or the bitcode reader/writer, the
5446 parser, etc...).</p>
Chris Lattnerfee11462004-02-12 17:01:32 +00005447
John Criswell88190562005-05-16 16:17:45 +00005448<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005449 prefix is reserved in LLVM for intrinsic names; thus, function names may not
5450 begin with this prefix. Intrinsic functions must always be external
5451 functions: you cannot define the body of intrinsic functions. Intrinsic
5452 functions may only be used in call or invoke instructions: it is illegal to
5453 take the address of an intrinsic function. Additionally, because intrinsic
5454 functions are part of the LLVM language, it is required if any are added that
5455 they be documented here.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +00005456
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005457<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents a
5458 family of functions that perform the same operation but on different data
5459 types. Because LLVM can represent over 8 million different integer types,
5460 overloading is used commonly to allow an intrinsic function to operate on any
5461 integer type. One or more of the argument types or the result type can be
5462 overloaded to accept any integer type. Argument types may also be defined as
5463 exactly matching a previous argument's type or the result type. This allows
5464 an intrinsic function which accepts multiple arguments, but needs all of them
5465 to be of the same type, to only be overloaded with respect to a single
5466 argument or the result.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +00005467
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005468<p>Overloaded intrinsics will have the names of its overloaded argument types
5469 encoded into its function name, each preceded by a period. Only those types
5470 which are overloaded result in a name suffix. Arguments whose type is matched
5471 against another type do not. For example, the <tt>llvm.ctpop</tt> function
5472 can take an integer of any width and returns an integer of exactly the same
5473 integer width. This leads to a family of functions such as
5474 <tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29
5475 %val)</tt>. Only one type, the return type, is overloaded, and only one type
5476 suffix is required. Because the argument's type is matched against the return
5477 type, it does not require its own name suffix.</p>
Reid Spencer4eefaab2007-04-01 08:04:23 +00005478
Eric Christopher455c5772009-12-05 02:46:03 +00005479<p>To learn how to add an intrinsic function, please see the
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005480 <a href="ExtendingLLVM.html">Extending LLVM Guide</a>.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +00005481
Misha Brukman76307852003-11-08 01:05:38 +00005482</div>
Chris Lattner941515c2004-01-06 05:31:32 +00005483
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00005484<!-- ======================================================================= -->
Chris Lattner941515c2004-01-06 05:31:32 +00005485<div class="doc_subsection">
5486 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
5487</div>
5488
Misha Brukman76307852003-11-08 01:05:38 +00005489<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +00005490
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005491<p>Variable argument support is defined in LLVM with
5492 the <a href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
5493 intrinsic functions. These functions are related to the similarly named
5494 macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005495
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005496<p>All of these functions operate on arguments that use a target-specific value
5497 type "<tt>va_list</tt>". The LLVM assembly language reference manual does
5498 not define what this type is, so all transformations should be prepared to
5499 handle these functions regardless of the type used.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005500
Chris Lattner30b868d2006-05-15 17:26:46 +00005501<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005502 instruction and the variable argument handling intrinsic functions are
5503 used.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005504
Benjamin Kramer79698be2010-07-13 12:26:09 +00005505<pre class="doc_code">
Anton Korobeynikov640bbe02007-03-21 23:58:04 +00005506define i32 @test(i32 %X, ...) {
Chris Lattnerfee11462004-02-12 17:01:32 +00005507 ; Initialize variable argument processing
Jeff Cohen222a8a42007-04-29 01:07:00 +00005508 %ap = alloca i8*
Chris Lattnerdb0790c2007-01-08 07:55:15 +00005509 %ap2 = bitcast i8** %ap to i8*
Anton Korobeynikov640bbe02007-03-21 23:58:04 +00005510 call void @llvm.va_start(i8* %ap2)
Chris Lattnerfee11462004-02-12 17:01:32 +00005511
5512 ; Read a single integer argument
Jeff Cohen222a8a42007-04-29 01:07:00 +00005513 %tmp = va_arg i8** %ap, i32
Chris Lattnerfee11462004-02-12 17:01:32 +00005514
5515 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Jeff Cohen222a8a42007-04-29 01:07:00 +00005516 %aq = alloca i8*
Chris Lattnerdb0790c2007-01-08 07:55:15 +00005517 %aq2 = bitcast i8** %aq to i8*
Jeff Cohen222a8a42007-04-29 01:07:00 +00005518 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
Anton Korobeynikov640bbe02007-03-21 23:58:04 +00005519 call void @llvm.va_end(i8* %aq2)
Chris Lattnerfee11462004-02-12 17:01:32 +00005520
5521 ; Stop processing of arguments.
Anton Korobeynikov640bbe02007-03-21 23:58:04 +00005522 call void @llvm.va_end(i8* %ap2)
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00005523 ret i32 %tmp
Chris Lattnerfee11462004-02-12 17:01:32 +00005524}
Anton Korobeynikov640bbe02007-03-21 23:58:04 +00005525
5526declare void @llvm.va_start(i8*)
5527declare void @llvm.va_copy(i8*, i8*)
5528declare void @llvm.va_end(i8*)
Chris Lattnerfee11462004-02-12 17:01:32 +00005529</pre>
Chris Lattner941515c2004-01-06 05:31:32 +00005530
Bill Wendling3716c5d2007-05-29 09:04:49 +00005531</div>
5532
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00005533<!-- _______________________________________________________________________ -->
Chris Lattner941515c2004-01-06 05:31:32 +00005534<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00005535 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
Chris Lattner941515c2004-01-06 05:31:32 +00005536</div>
5537
5538
Misha Brukman76307852003-11-08 01:05:38 +00005539<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005540
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00005541<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005542<pre>
5543 declare void %llvm.va_start(i8* &lt;arglist&gt;)
5544</pre>
5545
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00005546<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005547<p>The '<tt>llvm.va_start</tt>' intrinsic initializes <tt>*&lt;arglist&gt;</tt>
5548 for subsequent use by <tt><a href="#i_va_arg">va_arg</a></tt>.</p>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +00005549
5550<h5>Arguments:</h5>
Dan Gohmanef9462f2008-10-14 16:51:45 +00005551<p>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +00005552
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00005553<h5>Semantics:</h5>
Dan Gohmanef9462f2008-10-14 16:51:45 +00005554<p>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005555 macro available in C. In a target-dependent way, it initializes
5556 the <tt>va_list</tt> element to which the argument points, so that the next
5557 call to <tt>va_arg</tt> will produce the first variable argument passed to
5558 the function. Unlike the C <tt>va_start</tt> macro, this intrinsic does not
5559 need to know the last argument of the function as the compiler can figure
5560 that out.</p>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +00005561
Misha Brukman76307852003-11-08 01:05:38 +00005562</div>
Chris Lattner941515c2004-01-06 05:31:32 +00005563
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00005564<!-- _______________________________________________________________________ -->
Chris Lattner941515c2004-01-06 05:31:32 +00005565<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00005566 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
Chris Lattner941515c2004-01-06 05:31:32 +00005567</div>
5568
Misha Brukman76307852003-11-08 01:05:38 +00005569<div class="doc_text">
Chris Lattnerdb0790c2007-01-08 07:55:15 +00005570
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005571<h5>Syntax:</h5>
5572<pre>
5573 declare void @llvm.va_end(i8* &lt;arglist&gt;)
5574</pre>
5575
5576<h5>Overview:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +00005577<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005578 which has been initialized previously
5579 with <tt><a href="#int_va_start">llvm.va_start</a></tt>
5580 or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerdb0790c2007-01-08 07:55:15 +00005581
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00005582<h5>Arguments:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +00005583<p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
Chris Lattnerdb0790c2007-01-08 07:55:15 +00005584
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00005585<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +00005586<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005587 macro available in C. In a target-dependent way, it destroys
5588 the <tt>va_list</tt> element to which the argument points. Calls
5589 to <a href="#int_va_start"><tt>llvm.va_start</tt></a>
5590 and <a href="#int_va_copy"> <tt>llvm.va_copy</tt></a> must be matched exactly
5591 with calls to <tt>llvm.va_end</tt>.</p>
Chris Lattnerdb0790c2007-01-08 07:55:15 +00005592
Misha Brukman76307852003-11-08 01:05:38 +00005593</div>
Chris Lattner941515c2004-01-06 05:31:32 +00005594
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00005595<!-- _______________________________________________________________________ -->
Chris Lattner941515c2004-01-06 05:31:32 +00005596<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00005597 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
Chris Lattner941515c2004-01-06 05:31:32 +00005598</div>
5599
Misha Brukman76307852003-11-08 01:05:38 +00005600<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +00005601
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00005602<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005603<pre>
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +00005604 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
Chris Lattner757528b0b2004-05-23 21:06:01 +00005605</pre>
5606
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00005607<h5>Overview:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +00005608<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005609 from the source argument list to the destination argument list.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005610
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00005611<h5>Arguments:</h5>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +00005612<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005613 The second argument is a pointer to a <tt>va_list</tt> element to copy
5614 from.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005615
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00005616<h5>Semantics:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +00005617<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005618 macro available in C. In a target-dependent way, it copies the
5619 source <tt>va_list</tt> element into the destination <tt>va_list</tt>
5620 element. This intrinsic is necessary because
5621 the <tt><a href="#int_va_start"> llvm.va_start</a></tt> intrinsic may be
5622 arbitrarily complex and require, for example, memory allocation.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005623
Misha Brukman76307852003-11-08 01:05:38 +00005624</div>
Chris Lattner941515c2004-01-06 05:31:32 +00005625
Chris Lattnerfee11462004-02-12 17:01:32 +00005626<!-- ======================================================================= -->
5627<div class="doc_subsection">
Chris Lattner757528b0b2004-05-23 21:06:01 +00005628 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
5629</div>
5630
5631<div class="doc_text">
5632
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005633<p>LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattner67c37d12008-08-05 18:29:16 +00005634Collection</a> (GC) requires the implementation and generation of these
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005635intrinsics. These intrinsics allow identification of <a href="#int_gcroot">GC
5636roots on the stack</a>, as well as garbage collector implementations that
5637require <a href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a>
5638barriers. Front-ends for type-safe garbage collected languages should generate
5639these intrinsics to make use of the LLVM garbage collectors. For more details,
5640see <a href="GarbageCollection.html">Accurate Garbage Collection with
5641LLVM</a>.</p>
Christopher Lamb55c6d4f2007-12-17 01:00:21 +00005642
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005643<p>The garbage collection intrinsics only operate on objects in the generic
5644 address space (address space zero).</p>
Christopher Lamb55c6d4f2007-12-17 01:00:21 +00005645
Chris Lattner757528b0b2004-05-23 21:06:01 +00005646</div>
5647
5648<!-- _______________________________________________________________________ -->
5649<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00005650 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005651</div>
5652
5653<div class="doc_text">
5654
5655<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005656<pre>
Chris Lattner12477732007-09-21 17:30:40 +00005657 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Chris Lattner757528b0b2004-05-23 21:06:01 +00005658</pre>
5659
5660<h5>Overview:</h5>
John Criswelldfe6a862004-12-10 15:51:16 +00005661<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005662 the code generator, and allows some metadata to be associated with it.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005663
5664<h5>Arguments:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005665<p>The first argument specifies the address of a stack object that contains the
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005666 root pointer. The second pointer (which must be either a constant or a
5667 global value address) contains the meta-data to be associated with the
5668 root.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005669
5670<h5>Semantics:</h5>
Chris Lattner851b7712008-04-24 05:59:56 +00005671<p>At runtime, a call to this intrinsic stores a null pointer into the "ptrloc"
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005672 location. At compile-time, the code generator generates information to allow
5673 the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
5674 intrinsic may only be used in a function which <a href="#gc">specifies a GC
5675 algorithm</a>.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005676
5677</div>
5678
Chris Lattner757528b0b2004-05-23 21:06:01 +00005679<!-- _______________________________________________________________________ -->
5680<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00005681 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005682</div>
5683
5684<div class="doc_text">
5685
5686<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005687<pre>
Chris Lattner12477732007-09-21 17:30:40 +00005688 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Chris Lattner757528b0b2004-05-23 21:06:01 +00005689</pre>
5690
5691<h5>Overview:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005692<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005693 locations, allowing garbage collector implementations that require read
5694 barriers.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005695
5696<h5>Arguments:</h5>
Chris Lattnerf9228072006-03-14 20:02:51 +00005697<p>The second argument is the address to read from, which should be an address
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005698 allocated from the garbage collector. The first object is a pointer to the
5699 start of the referenced object, if needed by the language runtime (otherwise
5700 null).</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005701
5702<h5>Semantics:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005703<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005704 instruction, but may be replaced with substantially more complex code by the
5705 garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
5706 may only be used in a function which <a href="#gc">specifies a GC
5707 algorithm</a>.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005708
5709</div>
5710
Chris Lattner757528b0b2004-05-23 21:06:01 +00005711<!-- _______________________________________________________________________ -->
5712<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00005713 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005714</div>
5715
5716<div class="doc_text">
5717
5718<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005719<pre>
Chris Lattner12477732007-09-21 17:30:40 +00005720 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Chris Lattner757528b0b2004-05-23 21:06:01 +00005721</pre>
5722
5723<h5>Overview:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005724<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005725 locations, allowing garbage collector implementations that require write
5726 barriers (such as generational or reference counting collectors).</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005727
5728<h5>Arguments:</h5>
Chris Lattnerf9228072006-03-14 20:02:51 +00005729<p>The first argument is the reference to store, the second is the start of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005730 object to store it to, and the third is the address of the field of Obj to
5731 store to. If the runtime does not require a pointer to the object, Obj may
5732 be null.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005733
5734<h5>Semantics:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005735<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005736 instruction, but may be replaced with substantially more complex code by the
5737 garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
5738 may only be used in a function which <a href="#gc">specifies a GC
5739 algorithm</a>.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00005740
5741</div>
5742
Chris Lattner757528b0b2004-05-23 21:06:01 +00005743<!-- ======================================================================= -->
5744<div class="doc_subsection">
Chris Lattner3649c3a2004-02-14 04:08:35 +00005745 <a name="int_codegen">Code Generator Intrinsics</a>
5746</div>
5747
5748<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005749
5750<p>These intrinsics are provided by LLVM to expose special features that may
5751 only be implemented with code generator support.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +00005752
5753</div>
5754
5755<!-- _______________________________________________________________________ -->
5756<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00005757 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
Chris Lattner3649c3a2004-02-14 04:08:35 +00005758</div>
5759
5760<div class="doc_text">
5761
5762<h5>Syntax:</h5>
5763<pre>
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +00005764 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +00005765</pre>
5766
5767<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005768<p>The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
5769 target-specific value indicating the return address of the current function
5770 or one of its callers.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +00005771
5772<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005773<p>The argument to this intrinsic indicates which function to return the address
5774 for. Zero indicates the calling function, one indicates its caller, etc.
5775 The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +00005776
5777<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005778<p>The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer
5779 indicating the return address of the specified call frame, or zero if it
5780 cannot be identified. The value returned by this intrinsic is likely to be
5781 incorrect or 0 for arguments other than zero, so it should only be used for
5782 debugging purposes.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +00005783
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005784<p>Note that calling this intrinsic does not prevent function inlining or other
5785 aggressive transformations, so the value returned may not be that of the
5786 obvious source-language caller.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +00005787
Chris Lattner3649c3a2004-02-14 04:08:35 +00005788</div>
5789
Chris Lattner3649c3a2004-02-14 04:08:35 +00005790<!-- _______________________________________________________________________ -->
5791<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00005792 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
Chris Lattner3649c3a2004-02-14 04:08:35 +00005793</div>
5794
5795<div class="doc_text">
5796
5797<h5>Syntax:</h5>
5798<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +00005799 declare i8* @llvm.frameaddress(i32 &lt;level&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +00005800</pre>
5801
5802<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005803<p>The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
5804 target-specific frame pointer value for the specified stack frame.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +00005805
5806<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005807<p>The argument to this intrinsic indicates which function to return the frame
5808 pointer for. Zero indicates the calling function, one indicates its caller,
5809 etc. The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +00005810
5811<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005812<p>The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer
5813 indicating the frame address of the specified call frame, or zero if it
5814 cannot be identified. The value returned by this intrinsic is likely to be
5815 incorrect or 0 for arguments other than zero, so it should only be used for
5816 debugging purposes.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +00005817
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005818<p>Note that calling this intrinsic does not prevent function inlining or other
5819 aggressive transformations, so the value returned may not be that of the
5820 obvious source-language caller.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +00005821
Chris Lattner3649c3a2004-02-14 04:08:35 +00005822</div>
5823
Chris Lattnerc8a2c222005-02-28 19:24:19 +00005824<!-- _______________________________________________________________________ -->
5825<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00005826 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
Chris Lattner2f0f0012006-01-13 02:03:13 +00005827</div>
5828
5829<div class="doc_text">
5830
5831<h5>Syntax:</h5>
5832<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +00005833 declare i8* @llvm.stacksave()
Chris Lattner2f0f0012006-01-13 02:03:13 +00005834</pre>
5835
5836<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005837<p>The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state
5838 of the function stack, for use
5839 with <a href="#int_stackrestore"> <tt>llvm.stackrestore</tt></a>. This is
5840 useful for implementing language features like scoped automatic variable
5841 sized arrays in C99.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +00005842
5843<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005844<p>This intrinsic returns a opaque pointer value that can be passed
5845 to <a href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When
5846 an <tt>llvm.stackrestore</tt> intrinsic is executed with a value saved
5847 from <tt>llvm.stacksave</tt>, it effectively restores the state of the stack
5848 to the state it was in when the <tt>llvm.stacksave</tt> intrinsic executed.
5849 In practice, this pops any <a href="#i_alloca">alloca</a> blocks from the
5850 stack that were allocated after the <tt>llvm.stacksave</tt> was executed.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +00005851
5852</div>
5853
5854<!-- _______________________________________________________________________ -->
5855<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00005856 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
Chris Lattner2f0f0012006-01-13 02:03:13 +00005857</div>
5858
5859<div class="doc_text">
5860
5861<h5>Syntax:</h5>
5862<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +00005863 declare void @llvm.stackrestore(i8* %ptr)
Chris Lattner2f0f0012006-01-13 02:03:13 +00005864</pre>
5865
5866<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005867<p>The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
5868 the function stack to the state it was in when the
5869 corresponding <a href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic
5870 executed. This is useful for implementing language features like scoped
5871 automatic variable sized arrays in C99.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +00005872
5873<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005874<p>See the description
5875 for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +00005876
5877</div>
5878
Chris Lattner2f0f0012006-01-13 02:03:13 +00005879<!-- _______________________________________________________________________ -->
5880<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00005881 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
Chris Lattnerc8a2c222005-02-28 19:24:19 +00005882</div>
5883
5884<div class="doc_text">
5885
5886<h5>Syntax:</h5>
5887<pre>
Chris Lattner12477732007-09-21 17:30:40 +00005888 declare void @llvm.prefetch(i8* &lt;address&gt;, i32 &lt;rw&gt;, i32 &lt;locality&gt;)
Chris Lattnerc8a2c222005-02-28 19:24:19 +00005889</pre>
5890
5891<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005892<p>The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to
5893 insert a prefetch instruction if supported; otherwise, it is a noop.
5894 Prefetches have no effect on the behavior of the program but can change its
5895 performance characteristics.</p>
Chris Lattnerc8a2c222005-02-28 19:24:19 +00005896
5897<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005898<p><tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the
5899 specifier determining if the fetch should be for a read (0) or write (1),
5900 and <tt>locality</tt> is a temporal locality specifier ranging from (0) - no
5901 locality, to (3) - extremely local keep in cache. The <tt>rw</tt>
5902 and <tt>locality</tt> arguments must be constant integers.</p>
Chris Lattnerc8a2c222005-02-28 19:24:19 +00005903
5904<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005905<p>This intrinsic does not modify the behavior of the program. In particular,
5906 prefetches cannot trap and do not produce a value. On targets that support
5907 this intrinsic, the prefetch can provide hints to the processor cache for
5908 better performance.</p>
Chris Lattnerc8a2c222005-02-28 19:24:19 +00005909
5910</div>
5911
Andrew Lenharthb4427912005-03-28 20:05:49 +00005912<!-- _______________________________________________________________________ -->
5913<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00005914 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
Andrew Lenharthb4427912005-03-28 20:05:49 +00005915</div>
5916
5917<div class="doc_text">
5918
5919<h5>Syntax:</h5>
5920<pre>
Chris Lattner12477732007-09-21 17:30:40 +00005921 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Andrew Lenharthb4427912005-03-28 20:05:49 +00005922</pre>
5923
5924<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005925<p>The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program
5926 Counter (PC) in a region of code to simulators and other tools. The method
5927 is target specific, but it is expected that the marker will use exported
5928 symbols to transmit the PC of the marker. The marker makes no guarantees
5929 that it will remain with any specific instruction after optimizations. It is
5930 possible that the presence of a marker will inhibit optimizations. The
5931 intended use is to be inserted after optimizations to allow correlations of
5932 simulation runs.</p>
Andrew Lenharthb4427912005-03-28 20:05:49 +00005933
5934<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005935<p><tt>id</tt> is a numerical id identifying the marker.</p>
Andrew Lenharthb4427912005-03-28 20:05:49 +00005936
5937<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005938<p>This intrinsic does not modify the behavior of the program. Backends that do
Dan Gohmana269a0a2010-03-01 17:41:39 +00005939 not support this intrinsic may ignore it.</p>
Andrew Lenharthb4427912005-03-28 20:05:49 +00005940
5941</div>
5942
Andrew Lenharth01aa5632005-11-11 16:47:30 +00005943<!-- _______________________________________________________________________ -->
5944<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00005945 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
Andrew Lenharth01aa5632005-11-11 16:47:30 +00005946</div>
5947
5948<div class="doc_text">
5949
5950<h5>Syntax:</h5>
5951<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +00005952 declare i64 @llvm.readcyclecounter()
Andrew Lenharth01aa5632005-11-11 16:47:30 +00005953</pre>
5954
5955<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005956<p>The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
5957 counter register (or similar low latency, high accuracy clocks) on those
5958 targets that support it. On X86, it should map to RDTSC. On Alpha, it
5959 should map to RPCC. As the backing counters overflow quickly (on the order
5960 of 9 seconds on alpha), this should only be used for small timings.</p>
Andrew Lenharth01aa5632005-11-11 16:47:30 +00005961
5962<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005963<p>When directly supported, reading the cycle counter should not modify any
5964 memory. Implementations are allowed to either return a application specific
5965 value or a system wide value. On backends without support, this is lowered
5966 to a constant 0.</p>
Andrew Lenharth01aa5632005-11-11 16:47:30 +00005967
5968</div>
5969
Chris Lattner3649c3a2004-02-14 04:08:35 +00005970<!-- ======================================================================= -->
5971<div class="doc_subsection">
Chris Lattnerfee11462004-02-12 17:01:32 +00005972 <a name="int_libc">Standard C Library Intrinsics</a>
5973</div>
5974
5975<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005976
5977<p>LLVM provides intrinsics for a few important standard C library functions.
5978 These intrinsics allow source-language front-ends to pass information about
5979 the alignment of the pointer arguments to the code generator, providing
5980 opportunity for more efficient code generation.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +00005981
5982</div>
5983
5984<!-- _______________________________________________________________________ -->
5985<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00005986 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
Chris Lattnerfee11462004-02-12 17:01:32 +00005987</div>
5988
5989<div class="doc_text">
5990
5991<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005992<p>This is an overloaded intrinsic. You can use <tt>llvm.memcpy</tt> on any
Mon P Wang508127b2010-04-07 06:35:53 +00005993 integer bit width and for different address spaces. Not all targets support
5994 all bit widths however.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00005995
Chris Lattnerfee11462004-02-12 17:01:32 +00005996<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +00005997 declare void @llvm.memcpy.p0i8.p0i8.i32(i8* &lt;dest&gt;, i8* &lt;src&gt;,
Chris Lattnerbd4ca622010-04-08 00:53:57 +00005998 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanaabfdb32010-05-28 17:13:49 +00005999 declare void @llvm.memcpy.p0i8.p0i8.i64(i8* &lt;dest&gt;, i8* &lt;src&gt;,
Chris Lattnerbd4ca622010-04-08 00:53:57 +00006000 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattnerfee11462004-02-12 17:01:32 +00006001</pre>
6002
6003<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006004<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6005 source location to the destination location.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +00006006
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006007<p>Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
Chris Lattnerbd4ca622010-04-08 00:53:57 +00006008 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6009 and the pointers can be in specified address spaces.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +00006010
6011<h5>Arguments:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +00006012
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006013<p>The first argument is a pointer to the destination, the second is a pointer
6014 to the source. The third argument is an integer argument specifying the
Chris Lattnerbd4ca622010-04-08 00:53:57 +00006015 number of bytes to copy, the fourth argument is the alignment of the
6016 source and destination locations, and the fifth is a boolean indicating a
6017 volatile access.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +00006018
Dan Gohmana269a0a2010-03-01 17:41:39 +00006019<p>If the call to this intrinsic has an alignment value that is not 0 or 1,
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006020 then the caller guarantees that both the source and destination pointers are
6021 aligned to that boundary.</p>
Chris Lattner4c67c482004-02-12 21:18:15 +00006022
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +00006023<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6024 <tt>llvm.memcpy</tt> call is a <a href="#volatile">volatile operation</a>.
6025 The detailed access behavior is not very cleanly specified and it is unwise
6026 to depend on it.</p>
Chris Lattnerbd4ca622010-04-08 00:53:57 +00006027
Chris Lattnerfee11462004-02-12 17:01:32 +00006028<h5>Semantics:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +00006029
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006030<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6031 source location to the destination location, which are not allowed to
6032 overlap. It copies "len" bytes of memory over. If the argument is known to
6033 be aligned to some boundary, this can be specified as the fourth argument,
6034 otherwise it should be set to 0 or 1.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +00006035
Chris Lattnerfee11462004-02-12 17:01:32 +00006036</div>
6037
Chris Lattnerf30152e2004-02-12 18:10:10 +00006038<!-- _______________________________________________________________________ -->
6039<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00006040 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
Chris Lattnerf30152e2004-02-12 18:10:10 +00006041</div>
6042
6043<div class="doc_text">
6044
6045<h5>Syntax:</h5>
Chris Lattnerdd708342008-11-21 16:42:48 +00006046<p>This is an overloaded intrinsic. You can use llvm.memmove on any integer bit
Mon P Wang508127b2010-04-07 06:35:53 +00006047 width and for different address space. Not all targets support all bit
6048 widths however.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006049
Chris Lattnerf30152e2004-02-12 18:10:10 +00006050<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +00006051 declare void @llvm.memmove.p0i8.p0i8.i32(i8* &lt;dest&gt;, i8* &lt;src&gt;,
Chris Lattnerbd4ca622010-04-08 00:53:57 +00006052 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanaabfdb32010-05-28 17:13:49 +00006053 declare void @llvm.memmove.p0i8.p0i8.i64(i8* &lt;dest&gt;, i8* &lt;src&gt;,
Chris Lattnerbd4ca622010-04-08 00:53:57 +00006054 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattnerf30152e2004-02-12 18:10:10 +00006055</pre>
6056
6057<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006058<p>The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the
6059 source location to the destination location. It is similar to the
6060 '<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to
6061 overlap.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +00006062
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006063<p>Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
Chris Lattnerbd4ca622010-04-08 00:53:57 +00006064 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6065 and the pointers can be in specified address spaces.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +00006066
6067<h5>Arguments:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +00006068
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006069<p>The first argument is a pointer to the destination, the second is a pointer
6070 to the source. The third argument is an integer argument specifying the
Chris Lattnerbd4ca622010-04-08 00:53:57 +00006071 number of bytes to copy, the fourth argument is the alignment of the
6072 source and destination locations, and the fifth is a boolean indicating a
6073 volatile access.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +00006074
Dan Gohmana269a0a2010-03-01 17:41:39 +00006075<p>If the call to this intrinsic has an alignment value that is not 0 or 1,
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006076 then the caller guarantees that the source and destination pointers are
6077 aligned to that boundary.</p>
Chris Lattner4c67c482004-02-12 21:18:15 +00006078
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +00006079<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6080 <tt>llvm.memmove</tt> call is a <a href="#volatile">volatile operation</a>.
6081 The detailed access behavior is not very cleanly specified and it is unwise
6082 to depend on it.</p>
Chris Lattnerbd4ca622010-04-08 00:53:57 +00006083
Chris Lattnerf30152e2004-02-12 18:10:10 +00006084<h5>Semantics:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +00006085
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006086<p>The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the
6087 source location to the destination location, which may overlap. It copies
6088 "len" bytes of memory over. If the argument is known to be aligned to some
6089 boundary, this can be specified as the fourth argument, otherwise it should
6090 be set to 0 or 1.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +00006091
Chris Lattnerf30152e2004-02-12 18:10:10 +00006092</div>
6093
Chris Lattner3649c3a2004-02-14 04:08:35 +00006094<!-- _______________________________________________________________________ -->
6095<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00006096 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
Chris Lattner3649c3a2004-02-14 04:08:35 +00006097</div>
6098
6099<div class="doc_text">
6100
6101<h5>Syntax:</h5>
Chris Lattnerdd708342008-11-21 16:42:48 +00006102<p>This is an overloaded intrinsic. You can use llvm.memset on any integer bit
John Criswellad05ae42010-07-30 16:30:28 +00006103 width and for different address spaces. However, not all targets support all
6104 bit widths.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006105
Chris Lattner3649c3a2004-02-14 04:08:35 +00006106<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +00006107 declare void @llvm.memset.p0i8.i32(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattner685db9d2010-04-08 00:54:34 +00006108 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanaabfdb32010-05-28 17:13:49 +00006109 declare void @llvm.memset.p0i8.i64(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattner685db9d2010-04-08 00:54:34 +00006110 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +00006111</pre>
6112
6113<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006114<p>The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a
6115 particular byte value.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +00006116
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006117<p>Note that, unlike the standard libc function, the <tt>llvm.memset</tt>
John Criswellad05ae42010-07-30 16:30:28 +00006118 intrinsic does not return a value and takes extra alignment/volatile
6119 arguments. Also, the destination can be in an arbitrary address space.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +00006120
6121<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006122<p>The first argument is a pointer to the destination to fill, the second is the
John Criswellad05ae42010-07-30 16:30:28 +00006123 byte value with which to fill it, the third argument is an integer argument
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006124 specifying the number of bytes to fill, and the fourth argument is the known
John Criswellad05ae42010-07-30 16:30:28 +00006125 alignment of the destination location.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +00006126
Dan Gohmana269a0a2010-03-01 17:41:39 +00006127<p>If the call to this intrinsic has an alignment value that is not 0 or 1,
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006128 then the caller guarantees that the destination pointer is aligned to that
6129 boundary.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +00006130
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +00006131<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6132 <tt>llvm.memset</tt> call is a <a href="#volatile">volatile operation</a>.
6133 The detailed access behavior is not very cleanly specified and it is unwise
6134 to depend on it.</p>
Chris Lattnerbd4ca622010-04-08 00:53:57 +00006135
Chris Lattner3649c3a2004-02-14 04:08:35 +00006136<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006137<p>The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting
6138 at the destination location. If the argument is known to be aligned to some
6139 boundary, this can be specified as the fourth argument, otherwise it should
6140 be set to 0 or 1.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +00006141
Chris Lattner3649c3a2004-02-14 04:08:35 +00006142</div>
6143
Chris Lattner3b4f4372004-06-11 02:28:03 +00006144<!-- _______________________________________________________________________ -->
6145<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00006146 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
Chris Lattner8a8f2e52005-07-21 01:29:16 +00006147</div>
6148
6149<div class="doc_text">
6150
6151<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006152<p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
6153 floating point or vector of floating point type. Not all targets support all
6154 types however.</p>
6155
Chris Lattner8a8f2e52005-07-21 01:29:16 +00006156<pre>
Dale Johannesendd89d272007-10-02 17:47:38 +00006157 declare float @llvm.sqrt.f32(float %Val)
6158 declare double @llvm.sqrt.f64(double %Val)
6159 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
6160 declare fp128 @llvm.sqrt.f128(fp128 %Val)
6161 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Chris Lattner8a8f2e52005-07-21 01:29:16 +00006162</pre>
6163
6164<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006165<p>The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
6166 returning the same value as the libm '<tt>sqrt</tt>' functions would.
6167 Unlike <tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined
6168 behavior for negative numbers other than -0.0 (which allows for better
6169 optimization, because there is no need to worry about errno being
6170 set). <tt>llvm.sqrt(-0.0)</tt> is defined to return -0.0 like IEEE sqrt.</p>
Chris Lattner8a8f2e52005-07-21 01:29:16 +00006171
6172<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006173<p>The argument and return value are floating point numbers of the same
6174 type.</p>
Chris Lattner8a8f2e52005-07-21 01:29:16 +00006175
6176<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006177<p>This function returns the sqrt of the specified operand if it is a
6178 nonnegative floating point number.</p>
Chris Lattner8a8f2e52005-07-21 01:29:16 +00006179
Chris Lattner8a8f2e52005-07-21 01:29:16 +00006180</div>
6181
Chris Lattner33b73f92006-09-08 06:34:02 +00006182<!-- _______________________________________________________________________ -->
6183<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00006184 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
Chris Lattner33b73f92006-09-08 06:34:02 +00006185</div>
6186
6187<div class="doc_text">
6188
6189<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006190<p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
6191 floating point or vector of floating point type. Not all targets support all
6192 types however.</p>
6193
Chris Lattner33b73f92006-09-08 06:34:02 +00006194<pre>
Dale Johannesendd89d272007-10-02 17:47:38 +00006195 declare float @llvm.powi.f32(float %Val, i32 %power)
6196 declare double @llvm.powi.f64(double %Val, i32 %power)
6197 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
6198 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
6199 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Chris Lattner33b73f92006-09-08 06:34:02 +00006200</pre>
6201
6202<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006203<p>The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
6204 specified (positive or negative) power. The order of evaluation of
6205 multiplications is not defined. When a vector of floating point type is
6206 used, the second argument remains a scalar integer value.</p>
Chris Lattner33b73f92006-09-08 06:34:02 +00006207
6208<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006209<p>The second argument is an integer power, and the first is a value to raise to
6210 that power.</p>
Chris Lattner33b73f92006-09-08 06:34:02 +00006211
6212<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006213<p>This function returns the first value raised to the second power with an
6214 unspecified sequence of rounding operations.</p>
Chris Lattner33b73f92006-09-08 06:34:02 +00006215
Chris Lattner33b73f92006-09-08 06:34:02 +00006216</div>
6217
Dan Gohmanb6324c12007-10-15 20:30:11 +00006218<!-- _______________________________________________________________________ -->
6219<div class="doc_subsubsection">
6220 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
6221</div>
6222
6223<div class="doc_text">
6224
6225<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006226<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
6227 floating point or vector of floating point type. Not all targets support all
6228 types however.</p>
6229
Dan Gohmanb6324c12007-10-15 20:30:11 +00006230<pre>
6231 declare float @llvm.sin.f32(float %Val)
6232 declare double @llvm.sin.f64(double %Val)
6233 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
6234 declare fp128 @llvm.sin.f128(fp128 %Val)
6235 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
6236</pre>
6237
6238<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006239<p>The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +00006240
6241<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006242<p>The argument and return value are floating point numbers of the same
6243 type.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +00006244
6245<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006246<p>This function returns the sine of the specified operand, returning the same
6247 values as the libm <tt>sin</tt> functions would, and handles error conditions
6248 in the same way.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +00006249
Dan Gohmanb6324c12007-10-15 20:30:11 +00006250</div>
6251
6252<!-- _______________________________________________________________________ -->
6253<div class="doc_subsubsection">
6254 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
6255</div>
6256
6257<div class="doc_text">
6258
6259<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006260<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
6261 floating point or vector of floating point type. Not all targets support all
6262 types however.</p>
6263
Dan Gohmanb6324c12007-10-15 20:30:11 +00006264<pre>
6265 declare float @llvm.cos.f32(float %Val)
6266 declare double @llvm.cos.f64(double %Val)
6267 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
6268 declare fp128 @llvm.cos.f128(fp128 %Val)
6269 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
6270</pre>
6271
6272<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006273<p>The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +00006274
6275<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006276<p>The argument and return value are floating point numbers of the same
6277 type.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +00006278
6279<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006280<p>This function returns the cosine of the specified operand, returning the same
6281 values as the libm <tt>cos</tt> functions would, and handles error conditions
6282 in the same way.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +00006283
Dan Gohmanb6324c12007-10-15 20:30:11 +00006284</div>
6285
6286<!-- _______________________________________________________________________ -->
6287<div class="doc_subsubsection">
6288 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
6289</div>
6290
6291<div class="doc_text">
6292
6293<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006294<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
6295 floating point or vector of floating point type. Not all targets support all
6296 types however.</p>
6297
Dan Gohmanb6324c12007-10-15 20:30:11 +00006298<pre>
6299 declare float @llvm.pow.f32(float %Val, float %Power)
6300 declare double @llvm.pow.f64(double %Val, double %Power)
6301 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
6302 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
6303 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
6304</pre>
6305
6306<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006307<p>The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
6308 specified (positive or negative) power.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +00006309
6310<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006311<p>The second argument is a floating point power, and the first is a value to
6312 raise to that power.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +00006313
6314<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006315<p>This function returns the first value raised to the second power, returning
6316 the same values as the libm <tt>pow</tt> functions would, and handles error
6317 conditions in the same way.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +00006318
Dan Gohmanb6324c12007-10-15 20:30:11 +00006319</div>
6320
Andrew Lenharth1d463522005-05-03 18:01:48 +00006321<!-- ======================================================================= -->
6322<div class="doc_subsection">
Nate Begeman0f223bb2006-01-13 23:26:38 +00006323 <a name="int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharth1d463522005-05-03 18:01:48 +00006324</div>
6325
6326<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006327
6328<p>LLVM provides intrinsics for a few important bit manipulation operations.
6329 These allow efficient code generation for some algorithms.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +00006330
6331</div>
6332
6333<!-- _______________________________________________________________________ -->
6334<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00006335 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
Nate Begeman0f223bb2006-01-13 23:26:38 +00006336</div>
6337
6338<div class="doc_text">
6339
6340<h5>Syntax:</h5>
Reid Spencer4eefaab2007-04-01 08:04:23 +00006341<p>This is an overloaded intrinsic function. You can use bswap on any integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006342 type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
6343
Nate Begeman0f223bb2006-01-13 23:26:38 +00006344<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +00006345 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
6346 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
6347 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Nate Begeman0f223bb2006-01-13 23:26:38 +00006348</pre>
6349
6350<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006351<p>The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
6352 values with an even number of bytes (positive multiple of 16 bits). These
6353 are useful for performing operations on data that is not in the target's
6354 native byte order.</p>
Nate Begeman0f223bb2006-01-13 23:26:38 +00006355
6356<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006357<p>The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
6358 and low byte of the input i16 swapped. Similarly,
6359 the <tt>llvm.bswap.i32</tt> intrinsic returns an i32 value that has the four
6360 bytes of the input i32 swapped, so that if the input bytes are numbered 0, 1,
6361 2, 3 then the returned i32 will have its bytes in 3, 2, 1, 0 order.
6362 The <tt>llvm.bswap.i48</tt>, <tt>llvm.bswap.i64</tt> and other intrinsics
6363 extend this concept to additional even-byte lengths (6 bytes, 8 bytes and
6364 more, respectively).</p>
Nate Begeman0f223bb2006-01-13 23:26:38 +00006365
6366</div>
6367
6368<!-- _______________________________________________________________________ -->
6369<div class="doc_subsubsection">
Reid Spencerb4f9a6f2006-01-16 21:12:35 +00006370 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
Andrew Lenharth1d463522005-05-03 18:01:48 +00006371</div>
6372
6373<div class="doc_text">
6374
6375<h5>Syntax:</h5>
Reid Spencer4eefaab2007-04-01 08:04:23 +00006376<p>This is an overloaded intrinsic. You can use llvm.ctpop on any integer bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006377 width. Not all targets support all bit widths however.</p>
6378
Andrew Lenharth1d463522005-05-03 18:01:48 +00006379<pre>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006380 declare i8 @llvm.ctpop.i8(i8 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +00006381 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +00006382 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +00006383 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
6384 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Andrew Lenharth1d463522005-05-03 18:01:48 +00006385</pre>
6386
6387<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006388<p>The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set
6389 in a value.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +00006390
6391<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006392<p>The only argument is the value to be counted. The argument may be of any
6393 integer type. The return type must match the argument type.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +00006394
6395<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006396<p>The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +00006397
Andrew Lenharth1d463522005-05-03 18:01:48 +00006398</div>
6399
6400<!-- _______________________________________________________________________ -->
6401<div class="doc_subsubsection">
Chris Lattnerb748c672006-01-16 22:34:14 +00006402 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
Andrew Lenharth1d463522005-05-03 18:01:48 +00006403</div>
6404
6405<div class="doc_text">
6406
6407<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006408<p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
6409 integer bit width. Not all targets support all bit widths however.</p>
6410
Andrew Lenharth1d463522005-05-03 18:01:48 +00006411<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +00006412 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
6413 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +00006414 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +00006415 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
6416 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Andrew Lenharth1d463522005-05-03 18:01:48 +00006417</pre>
6418
6419<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006420<p>The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
6421 leading zeros in a variable.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +00006422
6423<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006424<p>The only argument is the value to be counted. The argument may be of any
6425 integer type. The return type must match the argument type.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +00006426
6427<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006428<p>The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant)
6429 zeros in a variable. If the src == 0 then the result is the size in bits of
6430 the type of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +00006431
Andrew Lenharth1d463522005-05-03 18:01:48 +00006432</div>
Chris Lattner3b4f4372004-06-11 02:28:03 +00006433
Chris Lattnerefa20fa2005-05-15 19:39:26 +00006434<!-- _______________________________________________________________________ -->
6435<div class="doc_subsubsection">
Chris Lattnerb748c672006-01-16 22:34:14 +00006436 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
Chris Lattnerefa20fa2005-05-15 19:39:26 +00006437</div>
6438
6439<div class="doc_text">
6440
6441<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006442<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
6443 integer bit width. Not all targets support all bit widths however.</p>
6444
Chris Lattnerefa20fa2005-05-15 19:39:26 +00006445<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +00006446 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
6447 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +00006448 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +00006449 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
6450 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Chris Lattnerefa20fa2005-05-15 19:39:26 +00006451</pre>
6452
6453<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006454<p>The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
6455 trailing zeros.</p>
Chris Lattnerefa20fa2005-05-15 19:39:26 +00006456
6457<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006458<p>The only argument is the value to be counted. The argument may be of any
6459 integer type. The return type must match the argument type.</p>
Chris Lattnerefa20fa2005-05-15 19:39:26 +00006460
6461<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006462<p>The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant)
6463 zeros in a variable. If the src == 0 then the result is the size in bits of
6464 the type of src. For example, <tt>llvm.cttz(2) = 1</tt>.</p>
Chris Lattnerefa20fa2005-05-15 19:39:26 +00006465
Chris Lattnerefa20fa2005-05-15 19:39:26 +00006466</div>
6467
Bill Wendlingfd2bd722009-02-08 04:04:40 +00006468<!-- ======================================================================= -->
6469<div class="doc_subsection">
6470 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
6471</div>
6472
6473<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006474
6475<p>LLVM provides intrinsics for some arithmetic with overflow operations.</p>
Bill Wendlingfd2bd722009-02-08 04:04:40 +00006476
6477</div>
6478
Bill Wendlingf4d70622009-02-08 01:40:31 +00006479<!-- _______________________________________________________________________ -->
6480<div class="doc_subsubsection">
Bill Wendlingfd2bd722009-02-08 04:04:40 +00006481 <a name="int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics</a>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006482</div>
6483
6484<div class="doc_text">
6485
6486<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006487<p>This is an overloaded intrinsic. You can use <tt>llvm.sadd.with.overflow</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006488 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006489
6490<pre>
6491 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
6492 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6493 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
6494</pre>
6495
6496<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006497<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006498 a signed addition of the two arguments, and indicate whether an overflow
6499 occurred during the signed summation.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006500
6501<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006502<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006503 be of integer types of any bit width, but they must have the same bit
6504 width. The second element of the result structure must be of
6505 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6506 undergo signed addition.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006507
6508<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006509<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006510 a signed addition of the two variables. They return a structure &mdash; the
6511 first element of which is the signed summation, and the second element of
6512 which is a bit specifying if the signed summation resulted in an
6513 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006514
6515<h5>Examples:</h5>
6516<pre>
6517 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6518 %sum = extractvalue {i32, i1} %res, 0
6519 %obit = extractvalue {i32, i1} %res, 1
6520 br i1 %obit, label %overflow, label %normal
6521</pre>
6522
6523</div>
6524
6525<!-- _______________________________________________________________________ -->
6526<div class="doc_subsubsection">
Bill Wendlingfd2bd722009-02-08 04:04:40 +00006527 <a name="int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics</a>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006528</div>
6529
6530<div class="doc_text">
6531
6532<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006533<p>This is an overloaded intrinsic. You can use <tt>llvm.uadd.with.overflow</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006534 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006535
6536<pre>
6537 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
6538 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6539 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
6540</pre>
6541
6542<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006543<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006544 an unsigned addition of the two arguments, and indicate whether a carry
6545 occurred during the unsigned summation.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006546
6547<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006548<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006549 be of integer types of any bit width, but they must have the same bit
6550 width. The second element of the result structure must be of
6551 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6552 undergo unsigned addition.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006553
6554<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006555<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006556 an unsigned addition of the two arguments. They return a structure &mdash;
6557 the first element of which is the sum, and the second element of which is a
6558 bit specifying if the unsigned summation resulted in a carry.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006559
6560<h5>Examples:</h5>
6561<pre>
6562 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6563 %sum = extractvalue {i32, i1} %res, 0
6564 %obit = extractvalue {i32, i1} %res, 1
6565 br i1 %obit, label %carry, label %normal
6566</pre>
6567
6568</div>
6569
6570<!-- _______________________________________________________________________ -->
6571<div class="doc_subsubsection">
Bill Wendlingfd2bd722009-02-08 04:04:40 +00006572 <a name="int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics</a>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006573</div>
6574
6575<div class="doc_text">
6576
6577<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006578<p>This is an overloaded intrinsic. You can use <tt>llvm.ssub.with.overflow</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006579 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006580
6581<pre>
6582 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
6583 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6584 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
6585</pre>
6586
6587<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006588<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006589 a signed subtraction of the two arguments, and indicate whether an overflow
6590 occurred during the signed subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006591
6592<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006593<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006594 be of integer types of any bit width, but they must have the same bit
6595 width. The second element of the result structure must be of
6596 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6597 undergo signed subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006598
6599<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006600<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006601 a signed subtraction of the two arguments. They return a structure &mdash;
6602 the first element of which is the subtraction, and the second element of
6603 which is a bit specifying if the signed subtraction resulted in an
6604 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006605
6606<h5>Examples:</h5>
6607<pre>
6608 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6609 %sum = extractvalue {i32, i1} %res, 0
6610 %obit = extractvalue {i32, i1} %res, 1
6611 br i1 %obit, label %overflow, label %normal
6612</pre>
6613
6614</div>
6615
6616<!-- _______________________________________________________________________ -->
6617<div class="doc_subsubsection">
Bill Wendlingfd2bd722009-02-08 04:04:40 +00006618 <a name="int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt>' Intrinsics</a>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006619</div>
6620
6621<div class="doc_text">
6622
6623<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006624<p>This is an overloaded intrinsic. You can use <tt>llvm.usub.with.overflow</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006625 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006626
6627<pre>
6628 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
6629 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6630 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
6631</pre>
6632
6633<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006634<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006635 an unsigned subtraction of the two arguments, and indicate whether an
6636 overflow occurred during the unsigned subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006637
6638<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006639<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006640 be of integer types of any bit width, but they must have the same bit
6641 width. The second element of the result structure must be of
6642 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6643 undergo unsigned subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006644
6645<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006646<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006647 an unsigned subtraction of the two arguments. They return a structure &mdash;
6648 the first element of which is the subtraction, and the second element of
6649 which is a bit specifying if the unsigned subtraction resulted in an
6650 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006651
6652<h5>Examples:</h5>
6653<pre>
6654 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6655 %sum = extractvalue {i32, i1} %res, 0
6656 %obit = extractvalue {i32, i1} %res, 1
6657 br i1 %obit, label %overflow, label %normal
6658</pre>
6659
6660</div>
6661
6662<!-- _______________________________________________________________________ -->
6663<div class="doc_subsubsection">
Bill Wendlingfd2bd722009-02-08 04:04:40 +00006664 <a name="int_smul_overflow">'<tt>llvm.smul.with.overflow.*</tt>' Intrinsics</a>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006665</div>
6666
6667<div class="doc_text">
6668
6669<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006670<p>This is an overloaded intrinsic. You can use <tt>llvm.smul.with.overflow</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006671 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006672
6673<pre>
6674 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
6675 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6676 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
6677</pre>
6678
6679<h5>Overview:</h5>
6680
6681<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006682 a signed multiplication of the two arguments, and indicate whether an
6683 overflow occurred during the signed multiplication.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006684
6685<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006686<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006687 be of integer types of any bit width, but they must have the same bit
6688 width. The second element of the result structure must be of
6689 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6690 undergo signed multiplication.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006691
6692<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006693<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006694 a signed multiplication of the two arguments. They return a structure &mdash;
6695 the first element of which is the multiplication, and the second element of
6696 which is a bit specifying if the signed multiplication resulted in an
6697 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +00006698
6699<h5>Examples:</h5>
6700<pre>
6701 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6702 %sum = extractvalue {i32, i1} %res, 0
6703 %obit = extractvalue {i32, i1} %res, 1
6704 br i1 %obit, label %overflow, label %normal
6705</pre>
6706
Reid Spencer5bf54c82007-04-11 23:23:49 +00006707</div>
6708
Bill Wendlingb9a73272009-02-08 23:00:09 +00006709<!-- _______________________________________________________________________ -->
6710<div class="doc_subsubsection">
6711 <a name="int_umul_overflow">'<tt>llvm.umul.with.overflow.*</tt>' Intrinsics</a>
6712</div>
6713
6714<div class="doc_text">
6715
6716<h5>Syntax:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +00006717<p>This is an overloaded intrinsic. You can use <tt>llvm.umul.with.overflow</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006718 on any integer bit width.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +00006719
6720<pre>
6721 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
6722 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6723 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
6724</pre>
6725
6726<h5>Overview:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +00006727<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006728 a unsigned multiplication of the two arguments, and indicate whether an
6729 overflow occurred during the unsigned multiplication.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +00006730
6731<h5>Arguments:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +00006732<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006733 be of integer types of any bit width, but they must have the same bit
6734 width. The second element of the result structure must be of
6735 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6736 undergo unsigned multiplication.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +00006737
6738<h5>Semantics:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +00006739<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006740 an unsigned multiplication of the two arguments. They return a structure
6741 &mdash; the first element of which is the multiplication, and the second
6742 element of which is a bit specifying if the unsigned multiplication resulted
6743 in an overflow.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +00006744
6745<h5>Examples:</h5>
6746<pre>
6747 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6748 %sum = extractvalue {i32, i1} %res, 0
6749 %obit = extractvalue {i32, i1} %res, 1
6750 br i1 %obit, label %overflow, label %normal
6751</pre>
6752
6753</div>
6754
Chris Lattner941515c2004-01-06 05:31:32 +00006755<!-- ======================================================================= -->
6756<div class="doc_subsection">
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +00006757 <a name="int_fp16">Half Precision Floating Point Intrinsics</a>
6758</div>
6759
6760<div class="doc_text">
6761
Chris Lattner022a9fb2010-03-15 04:12:21 +00006762<p>Half precision floating point is a storage-only format. This means that it is
6763 a dense encoding (in memory) but does not support computation in the
6764 format.</p>
Chris Lattnerbbd8bd32010-03-14 23:03:31 +00006765
Chris Lattner022a9fb2010-03-15 04:12:21 +00006766<p>This means that code must first load the half-precision floating point
Chris Lattnerbbd8bd32010-03-14 23:03:31 +00006767 value as an i16, then convert it to float with <a
6768 href="#int_convert_from_fp16"><tt>llvm.convert.from.fp16</tt></a>.
6769 Computation can then be performed on the float value (including extending to
Chris Lattner022a9fb2010-03-15 04:12:21 +00006770 double etc). To store the value back to memory, it is first converted to
6771 float if needed, then converted to i16 with
Chris Lattnerbbd8bd32010-03-14 23:03:31 +00006772 <a href="#int_convert_to_fp16"><tt>llvm.convert.to.fp16</tt></a>, then
6773 storing as an i16 value.</p>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +00006774</div>
6775
6776<!-- _______________________________________________________________________ -->
6777<div class="doc_subsubsection">
Chris Lattnerbbd8bd32010-03-14 23:03:31 +00006778 <a name="int_convert_to_fp16">'<tt>llvm.convert.to.fp16</tt>' Intrinsic</a>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +00006779</div>
6780
6781<div class="doc_text">
6782
6783<h5>Syntax:</h5>
6784<pre>
6785 declare i16 @llvm.convert.to.fp16(f32 %a)
6786</pre>
6787
6788<h5>Overview:</h5>
6789<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
6790 a conversion from single precision floating point format to half precision
6791 floating point format.</p>
6792
6793<h5>Arguments:</h5>
6794<p>The intrinsic function contains single argument - the value to be
6795 converted.</p>
6796
6797<h5>Semantics:</h5>
6798<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
6799 a conversion from single precision floating point format to half precision
Chris Lattner022a9fb2010-03-15 04:12:21 +00006800 floating point format. The return value is an <tt>i16</tt> which
Chris Lattnerbbd8bd32010-03-14 23:03:31 +00006801 contains the converted number.</p>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +00006802
6803<h5>Examples:</h5>
6804<pre>
6805 %res = call i16 @llvm.convert.to.fp16(f32 %a)
6806 store i16 %res, i16* @x, align 2
6807</pre>
6808
6809</div>
6810
6811<!-- _______________________________________________________________________ -->
6812<div class="doc_subsubsection">
Chris Lattnerbbd8bd32010-03-14 23:03:31 +00006813 <a name="int_convert_from_fp16">'<tt>llvm.convert.from.fp16</tt>' Intrinsic</a>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +00006814</div>
6815
6816<div class="doc_text">
6817
6818<h5>Syntax:</h5>
6819<pre>
6820 declare f32 @llvm.convert.from.fp16(i16 %a)
6821</pre>
6822
6823<h5>Overview:</h5>
6824<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs
6825 a conversion from half precision floating point format to single precision
6826 floating point format.</p>
6827
6828<h5>Arguments:</h5>
6829<p>The intrinsic function contains single argument - the value to be
6830 converted.</p>
6831
6832<h5>Semantics:</h5>
6833<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs a
Chris Lattner022a9fb2010-03-15 04:12:21 +00006834 conversion from half single precision floating point format to single
Chris Lattnerbbd8bd32010-03-14 23:03:31 +00006835 precision floating point format. The input half-float value is represented by
6836 an <tt>i16</tt> value.</p>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +00006837
6838<h5>Examples:</h5>
6839<pre>
6840 %a = load i16* @x, align 2
6841 %res = call f32 @llvm.convert.from.fp16(i16 %a)
6842</pre>
6843
6844</div>
6845
6846<!-- ======================================================================= -->
6847<div class="doc_subsection">
Chris Lattner941515c2004-01-06 05:31:32 +00006848 <a name="int_debugger">Debugger Intrinsics</a>
6849</div>
6850
6851<div class="doc_text">
Chris Lattner941515c2004-01-06 05:31:32 +00006852
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006853<p>The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt>
6854 prefix), are described in
6855 the <a href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source
6856 Level Debugging</a> document.</p>
6857
6858</div>
Chris Lattner941515c2004-01-06 05:31:32 +00006859
Jim Laskey2211f492007-03-14 19:31:19 +00006860<!-- ======================================================================= -->
6861<div class="doc_subsection">
6862 <a name="int_eh">Exception Handling Intrinsics</a>
6863</div>
6864
6865<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006866
6867<p>The LLVM exception handling intrinsics (which all start with
6868 <tt>llvm.eh.</tt> prefix), are described in
6869 the <a href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
6870 Handling</a> document.</p>
6871
Jim Laskey2211f492007-03-14 19:31:19 +00006872</div>
6873
Tanya Lattnercb1b9602007-06-15 20:50:54 +00006874<!-- ======================================================================= -->
6875<div class="doc_subsection">
Duncan Sands86e01192007-09-11 14:10:23 +00006876 <a name="int_trampoline">Trampoline Intrinsic</a>
Duncan Sands644f9172007-07-27 12:58:54 +00006877</div>
6878
6879<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006880
6881<p>This intrinsic makes it possible to excise one parameter, marked with
Dan Gohman3770af52010-07-02 23:18:08 +00006882 the <a href="#nest"><tt>nest</tt></a> attribute, from a function.
6883 The result is a callable
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006884 function pointer lacking the nest parameter - the caller does not need to
6885 provide a value for it. Instead, the value to use is stored in advance in a
6886 "trampoline", a block of memory usually allocated on the stack, which also
6887 contains code to splice the nest value into the argument list. This is used
6888 to implement the GCC nested function address extension.</p>
6889
6890<p>For example, if the function is
6891 <tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
6892 pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as
6893 follows:</p>
6894
Benjamin Kramer79698be2010-07-13 12:26:09 +00006895<pre class="doc_code">
Duncan Sands86e01192007-09-11 14:10:23 +00006896 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
6897 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
Dan Gohmand6a6f612010-05-28 17:07:41 +00006898 %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 +00006899 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands644f9172007-07-27 12:58:54 +00006900</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006901
Dan Gohmand6a6f612010-05-28 17:07:41 +00006902<p>The call <tt>%val = call i32 %fp(i32 %x, i32 %y)</tt> is then equivalent
6903 to <tt>%val = call i32 %f(i8* %nval, i32 %x, i32 %y)</tt>.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006904
Duncan Sands644f9172007-07-27 12:58:54 +00006905</div>
6906
6907<!-- _______________________________________________________________________ -->
6908<div class="doc_subsubsection">
6909 <a name="int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a>
6910</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006911
Duncan Sands644f9172007-07-27 12:58:54 +00006912<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006913
Duncan Sands644f9172007-07-27 12:58:54 +00006914<h5>Syntax:</h5>
6915<pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006916 declare i8* @llvm.init.trampoline(i8* &lt;tramp&gt;, i8* &lt;func&gt;, i8* &lt;nval&gt;)
Duncan Sands644f9172007-07-27 12:58:54 +00006917</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006918
Duncan Sands644f9172007-07-27 12:58:54 +00006919<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006920<p>This fills the memory pointed to by <tt>tramp</tt> with code and returns a
6921 function pointer suitable for executing it.</p>
6922
Duncan Sands644f9172007-07-27 12:58:54 +00006923<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006924<p>The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
6925 pointers. The <tt>tramp</tt> argument must point to a sufficiently large and
6926 sufficiently aligned block of memory; this memory is written to by the
6927 intrinsic. Note that the size and the alignment are target-specific - LLVM
6928 currently provides no portable way of determining them, so a front-end that
6929 generates this intrinsic needs to have some target-specific knowledge.
6930 The <tt>func</tt> argument must hold a function bitcast to
6931 an <tt>i8*</tt>.</p>
6932
Duncan Sands644f9172007-07-27 12:58:54 +00006933<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006934<p>The block of memory pointed to by <tt>tramp</tt> is filled with target
6935 dependent code, turning it into a function. A pointer to this function is
6936 returned, but needs to be bitcast to an <a href="#int_trampoline">appropriate
6937 function pointer type</a> before being called. The new function's signature
6938 is the same as that of <tt>func</tt> with any arguments marked with
6939 the <tt>nest</tt> attribute removed. At most one such <tt>nest</tt> argument
6940 is allowed, and it must be of pointer type. Calling the new function is
6941 equivalent to calling <tt>func</tt> with the same argument list, but
6942 with <tt>nval</tt> used for the missing <tt>nest</tt> argument. If, after
6943 calling <tt>llvm.init.trampoline</tt>, the memory pointed to
6944 by <tt>tramp</tt> is modified, then the effect of any later call to the
6945 returned function pointer is undefined.</p>
6946
Duncan Sands644f9172007-07-27 12:58:54 +00006947</div>
6948
6949<!-- ======================================================================= -->
6950<div class="doc_subsection">
Andrew Lenharth9b254ee2008-02-16 01:24:58 +00006951 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
6952</div>
6953
6954<div class="doc_text">
Andrew Lenharth9b254ee2008-02-16 01:24:58 +00006955
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006956<p>These intrinsic functions expand the "universal IR" of LLVM to represent
6957 hardware constructs for atomic operations and memory synchronization. This
6958 provides an interface to the hardware, not an interface to the programmer. It
6959 is aimed at a low enough level to allow any programming models or APIs
6960 (Application Programming Interfaces) which need atomic behaviors to map
6961 cleanly onto it. It is also modeled primarily on hardware behavior. Just as
6962 hardware provides a "universal IR" for source languages, it also provides a
6963 starting point for developing a "universal" atomic operation and
6964 synchronization IR.</p>
6965
6966<p>These do <em>not</em> form an API such as high-level threading libraries,
6967 software transaction memory systems, atomic primitives, and intrinsic
6968 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
6969 application libraries. The hardware interface provided by LLVM should allow
6970 a clean implementation of all of these APIs and parallel programming models.
6971 No one model or paradigm should be selected above others unless the hardware
6972 itself ubiquitously does so.</p>
6973
Andrew Lenharth9b254ee2008-02-16 01:24:58 +00006974</div>
6975
6976<!-- _______________________________________________________________________ -->
6977<div class="doc_subsubsection">
6978 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
6979</div>
6980<div class="doc_text">
6981<h5>Syntax:</h5>
6982<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +00006983 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 +00006984</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006985
Andrew Lenharth9b254ee2008-02-16 01:24:58 +00006986<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006987<p>The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
6988 specific pairs of memory access types.</p>
6989
Andrew Lenharth9b254ee2008-02-16 01:24:58 +00006990<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006991<p>The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
6992 The first four arguments enables a specific barrier as listed below. The
Dan Gohmana269a0a2010-03-01 17:41:39 +00006993 fifth argument specifies that the barrier applies to io or device or uncached
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006994 memory.</p>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +00006995
Bill Wendlingd9a66f72009-07-20 02:29:24 +00006996<ul>
6997 <li><tt>ll</tt>: load-load barrier</li>
6998 <li><tt>ls</tt>: load-store barrier</li>
6999 <li><tt>sl</tt>: store-load barrier</li>
7000 <li><tt>ss</tt>: store-store barrier</li>
7001 <li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
7002</ul>
7003
Andrew Lenharth9b254ee2008-02-16 01:24:58 +00007004<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007005<p>This intrinsic causes the system to enforce some ordering constraints upon
7006 the loads and stores of the program. This barrier does not
7007 indicate <em>when</em> any events will occur, it only enforces
7008 an <em>order</em> in which they occur. For any of the specified pairs of load
7009 and store operations (f.ex. load-load, or store-load), all of the first
7010 operations preceding the barrier will complete before any of the second
7011 operations succeeding the barrier begin. Specifically the semantics for each
7012 pairing is as follows:</p>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +00007013
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007014<ul>
7015 <li><tt>ll</tt>: All loads before the barrier must complete before any load
7016 after the barrier begins.</li>
Eric Christopher455c5772009-12-05 02:46:03 +00007017 <li><tt>ls</tt>: All loads before the barrier must complete before any
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007018 store after the barrier begins.</li>
Eric Christopher455c5772009-12-05 02:46:03 +00007019 <li><tt>ss</tt>: All stores before the barrier must complete before any
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007020 store after the barrier begins.</li>
Eric Christopher455c5772009-12-05 02:46:03 +00007021 <li><tt>sl</tt>: All stores before the barrier must complete before any
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007022 load after the barrier begins.</li>
7023</ul>
7024
7025<p>These semantics are applied with a logical "and" behavior when more than one
7026 is enabled in a single memory barrier intrinsic.</p>
7027
7028<p>Backends may implement stronger barriers than those requested when they do
7029 not support as fine grained a barrier as requested. Some architectures do
7030 not need all types of barriers and on such architectures, these become
7031 noops.</p>
7032
Andrew Lenharth9b254ee2008-02-16 01:24:58 +00007033<h5>Example:</h5>
7034<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +00007035%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7036%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth9b254ee2008-02-16 01:24:58 +00007037 store i32 4, %ptr
7038
7039%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007040 call void @llvm.memory.barrier(i1 false, i1 true, i1 false, i1 false)
Andrew Lenharth9b254ee2008-02-16 01:24:58 +00007041 <i>; guarantee the above finishes</i>
7042 store i32 8, %ptr <i>; before this begins</i>
7043</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007044
Andrew Lenharth9b254ee2008-02-16 01:24:58 +00007045</div>
7046
Andrew Lenharth95528942008-02-21 06:45:13 +00007047<!-- _______________________________________________________________________ -->
7048<div class="doc_subsubsection">
Mon P Wang6a490372008-06-25 08:15:39 +00007049 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
Andrew Lenharth95528942008-02-21 06:45:13 +00007050</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007051
Andrew Lenharth95528942008-02-21 06:45:13 +00007052<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007053
Andrew Lenharth95528942008-02-21 06:45:13 +00007054<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007055<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
7056 any integer bit width and for different address spaces. Not all targets
7057 support all bit widths however.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +00007058
7059<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007060 declare i8 @llvm.atomic.cmp.swap.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt;)
7061 declare i16 @llvm.atomic.cmp.swap.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt;)
7062 declare i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt;)
7063 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 +00007064</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007065
Andrew Lenharth95528942008-02-21 06:45:13 +00007066<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007067<p>This loads a value in memory and compares it to a given value. If they are
7068 equal, it stores a new value into the memory.</p>
7069
Andrew Lenharth95528942008-02-21 06:45:13 +00007070<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007071<p>The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result
7072 as well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
7073 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
7074 this integer type. While any bit width integer may be used, targets may only
7075 lower representations they support in hardware.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +00007076
Andrew Lenharth95528942008-02-21 06:45:13 +00007077<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007078<p>This entire intrinsic must be executed atomically. It first loads the value
7079 in memory pointed to by <tt>ptr</tt> and compares it with the
7080 value <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the
7081 memory. The loaded value is yielded in all cases. This provides the
7082 equivalent of an atomic compare-and-swap operation within the SSA
7083 framework.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +00007084
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007085<h5>Examples:</h5>
Andrew Lenharth95528942008-02-21 06:45:13 +00007086<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +00007087%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7088%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth95528942008-02-21 06:45:13 +00007089 store i32 4, %ptr
7090
7091%val1 = add i32 4, 4
Dan Gohmand6a6f612010-05-28 17:07:41 +00007092%result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %ptr, i32 4, %val1)
Andrew Lenharth95528942008-02-21 06:45:13 +00007093 <i>; yields {i32}:result1 = 4</i>
7094%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
7095%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
7096
7097%val2 = add i32 1, 1
Dan Gohmand6a6f612010-05-28 17:07:41 +00007098%result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %ptr, i32 5, %val2)
Andrew Lenharth95528942008-02-21 06:45:13 +00007099 <i>; yields {i32}:result2 = 8</i>
7100%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
7101
7102%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
7103</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007104
Andrew Lenharth95528942008-02-21 06:45:13 +00007105</div>
7106
7107<!-- _______________________________________________________________________ -->
7108<div class="doc_subsubsection">
7109 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
7110</div>
7111<div class="doc_text">
7112<h5>Syntax:</h5>
7113
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007114<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
7115 integer bit width. Not all targets support all bit widths however.</p>
7116
Andrew Lenharth95528942008-02-21 06:45:13 +00007117<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007118 declare i8 @llvm.atomic.swap.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;val&gt;)
7119 declare i16 @llvm.atomic.swap.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;val&gt;)
7120 declare i32 @llvm.atomic.swap.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;val&gt;)
7121 declare i64 @llvm.atomic.swap.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;val&gt;)
Andrew Lenharth95528942008-02-21 06:45:13 +00007122</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007123
Andrew Lenharth95528942008-02-21 06:45:13 +00007124<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007125<p>This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
7126 the value from memory. It then stores the value in <tt>val</tt> in the memory
7127 at <tt>ptr</tt>.</p>
7128
Andrew Lenharth95528942008-02-21 06:45:13 +00007129<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007130<p>The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both
7131 the <tt>val</tt> argument and the result must be integers of the same bit
7132 width. The first argument, <tt>ptr</tt>, must be a pointer to a value of this
7133 integer type. The targets may only lower integer representations they
7134 support.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +00007135
Andrew Lenharth95528942008-02-21 06:45:13 +00007136<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007137<p>This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
7138 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
7139 equivalent of an atomic swap operation within the SSA framework.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +00007140
Andrew Lenharth95528942008-02-21 06:45:13 +00007141<h5>Examples:</h5>
7142<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +00007143%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7144%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth95528942008-02-21 06:45:13 +00007145 store i32 4, %ptr
7146
7147%val1 = add i32 4, 4
Dan Gohmand6a6f612010-05-28 17:07:41 +00007148%result1 = call i32 @llvm.atomic.swap.i32.p0i32(i32* %ptr, i32 %val1)
Andrew Lenharth95528942008-02-21 06:45:13 +00007149 <i>; yields {i32}:result1 = 4</i>
7150%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
7151%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
7152
7153%val2 = add i32 1, 1
Dan Gohmand6a6f612010-05-28 17:07:41 +00007154%result2 = call i32 @llvm.atomic.swap.i32.p0i32(i32* %ptr, i32 %val2)
Andrew Lenharth95528942008-02-21 06:45:13 +00007155 <i>; yields {i32}:result2 = 8</i>
7156
7157%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
7158%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
7159</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007160
Andrew Lenharth95528942008-02-21 06:45:13 +00007161</div>
7162
7163<!-- _______________________________________________________________________ -->
7164<div class="doc_subsubsection">
Mon P Wang6a490372008-06-25 08:15:39 +00007165 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
Andrew Lenharth95528942008-02-21 06:45:13 +00007166
7167</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007168
Andrew Lenharth95528942008-02-21 06:45:13 +00007169<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007170
Andrew Lenharth95528942008-02-21 06:45:13 +00007171<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007172<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.add</tt> on
7173 any integer bit width. Not all targets support all bit widths however.</p>
7174
Andrew Lenharth95528942008-02-21 06:45:13 +00007175<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007176 declare i8 @llvm.atomic.load.add.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7177 declare i16 @llvm.atomic.load.add.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7178 declare i32 @llvm.atomic.load.add.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7179 declare i64 @llvm.atomic.load.add.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Andrew Lenharth95528942008-02-21 06:45:13 +00007180</pre>
Andrew Lenharth95528942008-02-21 06:45:13 +00007181
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007182<h5>Overview:</h5>
7183<p>This intrinsic adds <tt>delta</tt> to the value stored in memory
7184 at <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
7185
7186<h5>Arguments:</h5>
7187<p>The intrinsic takes two arguments, the first a pointer to an integer value
7188 and the second an integer value. The result is also an integer value. These
7189 integer types can have any bit width, but they must all have the same bit
7190 width. The targets may only lower integer representations they support.</p>
7191
Andrew Lenharth95528942008-02-21 06:45:13 +00007192<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007193<p>This intrinsic does a series of operations atomically. It first loads the
7194 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
7195 to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +00007196
7197<h5>Examples:</h5>
7198<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +00007199%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7200%ptr = bitcast i8* %mallocP to i32*
7201 store i32 4, %ptr
Dan Gohmand6a6f612010-05-28 17:07:41 +00007202%result1 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 4)
Andrew Lenharth95528942008-02-21 06:45:13 +00007203 <i>; yields {i32}:result1 = 4</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007204%result2 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 2)
Andrew Lenharth95528942008-02-21 06:45:13 +00007205 <i>; yields {i32}:result2 = 8</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007206%result3 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 5)
Andrew Lenharth95528942008-02-21 06:45:13 +00007207 <i>; yields {i32}:result3 = 10</i>
Mon P Wang6a490372008-06-25 08:15:39 +00007208%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
Andrew Lenharth95528942008-02-21 06:45:13 +00007209</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007210
Andrew Lenharth95528942008-02-21 06:45:13 +00007211</div>
7212
Mon P Wang6a490372008-06-25 08:15:39 +00007213<!-- _______________________________________________________________________ -->
7214<div class="doc_subsubsection">
7215 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
7216
7217</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007218
Mon P Wang6a490372008-06-25 08:15:39 +00007219<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007220
Mon P Wang6a490372008-06-25 08:15:39 +00007221<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007222<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.sub</tt> on
7223 any integer bit width and for different address spaces. Not all targets
7224 support all bit widths however.</p>
7225
Mon P Wang6a490372008-06-25 08:15:39 +00007226<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007227 declare i8 @llvm.atomic.load.sub.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7228 declare i16 @llvm.atomic.load.sub.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7229 declare i32 @llvm.atomic.load.sub.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7230 declare i64 @llvm.atomic.load.sub.i64.p0i32(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +00007231</pre>
Mon P Wang6a490372008-06-25 08:15:39 +00007232
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007233<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +00007234<p>This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007235 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
7236
7237<h5>Arguments:</h5>
7238<p>The intrinsic takes two arguments, the first a pointer to an integer value
7239 and the second an integer value. The result is also an integer value. These
7240 integer types can have any bit width, but they must all have the same bit
7241 width. The targets may only lower integer representations they support.</p>
7242
Mon P Wang6a490372008-06-25 08:15:39 +00007243<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007244<p>This intrinsic does a series of operations atomically. It first loads the
7245 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
7246 result to <tt>ptr</tt>. It yields the original value stored
7247 at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +00007248
7249<h5>Examples:</h5>
7250<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +00007251%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7252%ptr = bitcast i8* %mallocP to i32*
7253 store i32 8, %ptr
Dan Gohmand6a6f612010-05-28 17:07:41 +00007254%result1 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 4)
Mon P Wang6a490372008-06-25 08:15:39 +00007255 <i>; yields {i32}:result1 = 8</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007256%result2 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 2)
Mon P Wang6a490372008-06-25 08:15:39 +00007257 <i>; yields {i32}:result2 = 4</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007258%result3 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 5)
Mon P Wang6a490372008-06-25 08:15:39 +00007259 <i>; yields {i32}:result3 = 2</i>
7260%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
7261</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007262
Mon P Wang6a490372008-06-25 08:15:39 +00007263</div>
7264
7265<!-- _______________________________________________________________________ -->
7266<div class="doc_subsubsection">
7267 <a name="int_atomic_load_and">'<tt>llvm.atomic.load.and.*</tt>' Intrinsic</a><br>
7268 <a name="int_atomic_load_nand">'<tt>llvm.atomic.load.nand.*</tt>' Intrinsic</a><br>
7269 <a name="int_atomic_load_or">'<tt>llvm.atomic.load.or.*</tt>' Intrinsic</a><br>
7270 <a name="int_atomic_load_xor">'<tt>llvm.atomic.load.xor.*</tt>' Intrinsic</a><br>
Mon P Wang6a490372008-06-25 08:15:39 +00007271</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007272
Mon P Wang6a490372008-06-25 08:15:39 +00007273<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007274
Mon P Wang6a490372008-06-25 08:15:39 +00007275<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007276<p>These are overloaded intrinsics. You can
7277 use <tt>llvm.atomic.load_and</tt>, <tt>llvm.atomic.load_nand</tt>,
7278 <tt>llvm.atomic.load_or</tt>, and <tt>llvm.atomic.load_xor</tt> on any integer
7279 bit width and for different address spaces. Not all targets support all bit
7280 widths however.</p>
Mon P Wang6a490372008-06-25 08:15:39 +00007281
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007282<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007283 declare i8 @llvm.atomic.load.and.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7284 declare i16 @llvm.atomic.load.and.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7285 declare i32 @llvm.atomic.load.and.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7286 declare i64 @llvm.atomic.load.and.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +00007287</pre>
7288
7289<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007290 declare i8 @llvm.atomic.load.or.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7291 declare i16 @llvm.atomic.load.or.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7292 declare i32 @llvm.atomic.load.or.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7293 declare i64 @llvm.atomic.load.or.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +00007294</pre>
7295
7296<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007297 declare i8 @llvm.atomic.load.nand.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7298 declare i16 @llvm.atomic.load.nand.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7299 declare i32 @llvm.atomic.load.nand.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7300 declare i64 @llvm.atomic.load.nand.i64.p0i32(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +00007301</pre>
7302
7303<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007304 declare i8 @llvm.atomic.load.xor.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7305 declare i16 @llvm.atomic.load.xor.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7306 declare i32 @llvm.atomic.load.xor.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7307 declare i64 @llvm.atomic.load.xor.i64.p0i32(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +00007308</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007309
Mon P Wang6a490372008-06-25 08:15:39 +00007310<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007311<p>These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
7312 the value stored in memory at <tt>ptr</tt>. It yields the original value
7313 at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +00007314
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007315<h5>Arguments:</h5>
7316<p>These intrinsics take two arguments, the first a pointer to an integer value
7317 and the second an integer value. The result is also an integer value. These
7318 integer types can have any bit width, but they must all have the same bit
7319 width. The targets may only lower integer representations they support.</p>
7320
Mon P Wang6a490372008-06-25 08:15:39 +00007321<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007322<p>These intrinsics does a series of operations atomically. They first load the
7323 value stored at <tt>ptr</tt>. They then do the bitwise
7324 operation <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the
7325 original value stored at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +00007326
7327<h5>Examples:</h5>
7328<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +00007329%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7330%ptr = bitcast i8* %mallocP to i32*
7331 store i32 0x0F0F, %ptr
Dan Gohmand6a6f612010-05-28 17:07:41 +00007332%result0 = call i32 @llvm.atomic.load.nand.i32.p0i32(i32* %ptr, i32 0xFF)
Mon P Wang6a490372008-06-25 08:15:39 +00007333 <i>; yields {i32}:result0 = 0x0F0F</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007334%result1 = call i32 @llvm.atomic.load.and.i32.p0i32(i32* %ptr, i32 0xFF)
Mon P Wang6a490372008-06-25 08:15:39 +00007335 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007336%result2 = call i32 @llvm.atomic.load.or.i32.p0i32(i32* %ptr, i32 0F)
Mon P Wang6a490372008-06-25 08:15:39 +00007337 <i>; yields {i32}:result2 = 0xF0</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007338%result3 = call i32 @llvm.atomic.load.xor.i32.p0i32(i32* %ptr, i32 0F)
Mon P Wang6a490372008-06-25 08:15:39 +00007339 <i>; yields {i32}:result3 = FF</i>
7340%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
7341</pre>
Mon P Wang6a490372008-06-25 08:15:39 +00007342
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007343</div>
Mon P Wang6a490372008-06-25 08:15:39 +00007344
7345<!-- _______________________________________________________________________ -->
7346<div class="doc_subsubsection">
7347 <a name="int_atomic_load_max">'<tt>llvm.atomic.load.max.*</tt>' Intrinsic</a><br>
7348 <a name="int_atomic_load_min">'<tt>llvm.atomic.load.min.*</tt>' Intrinsic</a><br>
7349 <a name="int_atomic_load_umax">'<tt>llvm.atomic.load.umax.*</tt>' Intrinsic</a><br>
7350 <a name="int_atomic_load_umin">'<tt>llvm.atomic.load.umin.*</tt>' Intrinsic</a><br>
Mon P Wang6a490372008-06-25 08:15:39 +00007351</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007352
Mon P Wang6a490372008-06-25 08:15:39 +00007353<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007354
Mon P Wang6a490372008-06-25 08:15:39 +00007355<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007356<p>These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
7357 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
7358 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
7359 address spaces. Not all targets support all bit widths however.</p>
Mon P Wang6a490372008-06-25 08:15:39 +00007360
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007361<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007362 declare i8 @llvm.atomic.load.max.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7363 declare i16 @llvm.atomic.load.max.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7364 declare i32 @llvm.atomic.load.max.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7365 declare i64 @llvm.atomic.load.max.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +00007366</pre>
7367
7368<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007369 declare i8 @llvm.atomic.load.min.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7370 declare i16 @llvm.atomic.load.min.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7371 declare i32 @llvm.atomic.load.min.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7372 declare i64 @llvm.atomic.load.min.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +00007373</pre>
7374
7375<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007376 declare i8 @llvm.atomic.load.umax.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7377 declare i16 @llvm.atomic.load.umax.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7378 declare i32 @llvm.atomic.load.umax.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7379 declare i64 @llvm.atomic.load.umax.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +00007380</pre>
7381
7382<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007383 declare i8 @llvm.atomic.load.umin.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7384 declare i16 @llvm.atomic.load.umin.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7385 declare i32 @llvm.atomic.load.umin.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7386 declare i64 @llvm.atomic.load.umin.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +00007387</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007388
Mon P Wang6a490372008-06-25 08:15:39 +00007389<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +00007390<p>These intrinsics takes the signed or unsigned minimum or maximum of
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007391 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
7392 original value at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +00007393
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007394<h5>Arguments:</h5>
7395<p>These intrinsics take two arguments, the first a pointer to an integer value
7396 and the second an integer value. The result is also an integer value. These
7397 integer types can have any bit width, but they must all have the same bit
7398 width. The targets may only lower integer representations they support.</p>
7399
Mon P Wang6a490372008-06-25 08:15:39 +00007400<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007401<p>These intrinsics does a series of operations atomically. They first load the
7402 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or
7403 max <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They
7404 yield the original value stored at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +00007405
7406<h5>Examples:</h5>
7407<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +00007408%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7409%ptr = bitcast i8* %mallocP to i32*
7410 store i32 7, %ptr
Dan Gohmand6a6f612010-05-28 17:07:41 +00007411%result0 = call i32 @llvm.atomic.load.min.i32.p0i32(i32* %ptr, i32 -2)
Mon P Wang6a490372008-06-25 08:15:39 +00007412 <i>; yields {i32}:result0 = 7</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007413%result1 = call i32 @llvm.atomic.load.max.i32.p0i32(i32* %ptr, i32 8)
Mon P Wang6a490372008-06-25 08:15:39 +00007414 <i>; yields {i32}:result1 = -2</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007415%result2 = call i32 @llvm.atomic.load.umin.i32.p0i32(i32* %ptr, i32 10)
Mon P Wang6a490372008-06-25 08:15:39 +00007416 <i>; yields {i32}:result2 = 8</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007417%result3 = call i32 @llvm.atomic.load.umax.i32.p0i32(i32* %ptr, i32 30)
Mon P Wang6a490372008-06-25 08:15:39 +00007418 <i>; yields {i32}:result3 = 8</i>
7419%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
7420</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007421
Mon P Wang6a490372008-06-25 08:15:39 +00007422</div>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +00007423
Nick Lewycky6f7d8342009-10-13 07:03:23 +00007424
7425<!-- ======================================================================= -->
7426<div class="doc_subsection">
7427 <a name="int_memorymarkers">Memory Use Markers</a>
7428</div>
7429
7430<div class="doc_text">
7431
7432<p>This class of intrinsics exists to information about the lifetime of memory
7433 objects and ranges where variables are immutable.</p>
7434
7435</div>
7436
7437<!-- _______________________________________________________________________ -->
7438<div class="doc_subsubsection">
7439 <a name="int_lifetime_start">'<tt>llvm.lifetime.start</tt>' Intrinsic</a>
7440</div>
7441
7442<div class="doc_text">
7443
7444<h5>Syntax:</h5>
7445<pre>
7446 declare void @llvm.lifetime.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7447</pre>
7448
7449<h5>Overview:</h5>
7450<p>The '<tt>llvm.lifetime.start</tt>' intrinsic specifies the start of a memory
7451 object's lifetime.</p>
7452
7453<h5>Arguments:</h5>
Nick Lewycky9bc89042009-10-13 07:57:33 +00007454<p>The first argument is a constant integer representing the size of the
7455 object, or -1 if it is variable sized. The second argument is a pointer to
7456 the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +00007457
7458<h5>Semantics:</h5>
7459<p>This intrinsic indicates that before this point in the code, the value of the
7460 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
Nick Lewyckyd20fd592009-10-27 16:56:58 +00007461 never be used and has an undefined value. A load from the pointer that
7462 precedes this intrinsic can be replaced with
Nick Lewycky6f7d8342009-10-13 07:03:23 +00007463 <tt>'<a href="#undefvalues">undef</a>'</tt>.</p>
7464
7465</div>
7466
7467<!-- _______________________________________________________________________ -->
7468<div class="doc_subsubsection">
7469 <a name="int_lifetime_end">'<tt>llvm.lifetime.end</tt>' Intrinsic</a>
7470</div>
7471
7472<div class="doc_text">
7473
7474<h5>Syntax:</h5>
7475<pre>
7476 declare void @llvm.lifetime.end(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7477</pre>
7478
7479<h5>Overview:</h5>
7480<p>The '<tt>llvm.lifetime.end</tt>' intrinsic specifies the end of a memory
7481 object's lifetime.</p>
7482
7483<h5>Arguments:</h5>
Nick Lewycky9bc89042009-10-13 07:57:33 +00007484<p>The first argument is a constant integer representing the size of the
7485 object, or -1 if it is variable sized. The second argument is a pointer to
7486 the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +00007487
7488<h5>Semantics:</h5>
7489<p>This intrinsic indicates that after this point in the code, the value of the
7490 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
7491 never be used and has an undefined value. Any stores into the memory object
7492 following this intrinsic may be removed as dead.
7493
7494</div>
7495
7496<!-- _______________________________________________________________________ -->
7497<div class="doc_subsubsection">
7498 <a name="int_invariant_start">'<tt>llvm.invariant.start</tt>' Intrinsic</a>
7499</div>
7500
7501<div class="doc_text">
7502
7503<h5>Syntax:</h5>
7504<pre>
7505 declare {}* @llvm.invariant.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;) readonly
7506</pre>
7507
7508<h5>Overview:</h5>
7509<p>The '<tt>llvm.invariant.start</tt>' intrinsic specifies that the contents of
7510 a memory object will not change.</p>
7511
7512<h5>Arguments:</h5>
Nick Lewycky9bc89042009-10-13 07:57:33 +00007513<p>The first argument is a constant integer representing the size of the
7514 object, or -1 if it is variable sized. The second argument is a pointer to
7515 the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +00007516
7517<h5>Semantics:</h5>
7518<p>This intrinsic indicates that until an <tt>llvm.invariant.end</tt> that uses
7519 the return value, the referenced memory location is constant and
7520 unchanging.</p>
7521
7522</div>
7523
7524<!-- _______________________________________________________________________ -->
7525<div class="doc_subsubsection">
7526 <a name="int_invariant_end">'<tt>llvm.invariant.end</tt>' Intrinsic</a>
7527</div>
7528
7529<div class="doc_text">
7530
7531<h5>Syntax:</h5>
7532<pre>
7533 declare void @llvm.invariant.end({}* &lt;start&gt;, i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7534</pre>
7535
7536<h5>Overview:</h5>
7537<p>The '<tt>llvm.invariant.end</tt>' intrinsic specifies that the contents of
7538 a memory object are mutable.</p>
7539
7540<h5>Arguments:</h5>
7541<p>The first argument is the matching <tt>llvm.invariant.start</tt> intrinsic.
Nick Lewycky9bc89042009-10-13 07:57:33 +00007542 The second argument is a constant integer representing the size of the
7543 object, or -1 if it is variable sized and the third argument is a pointer
7544 to the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +00007545
7546<h5>Semantics:</h5>
7547<p>This intrinsic indicates that the memory is mutable again.</p>
7548
7549</div>
7550
Andrew Lenharth9b254ee2008-02-16 01:24:58 +00007551<!-- ======================================================================= -->
7552<div class="doc_subsection">
Tanya Lattnercb1b9602007-06-15 20:50:54 +00007553 <a name="int_general">General Intrinsics</a>
7554</div>
7555
7556<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007557
7558<p>This class of intrinsics is designed to be generic and has no specific
7559 purpose.</p>
7560
Tanya Lattnercb1b9602007-06-15 20:50:54 +00007561</div>
7562
7563<!-- _______________________________________________________________________ -->
7564<div class="doc_subsubsection">
7565 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
7566</div>
7567
7568<div class="doc_text">
7569
7570<h5>Syntax:</h5>
7571<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007572 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 +00007573</pre>
7574
7575<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007576<p>The '<tt>llvm.var.annotation</tt>' intrinsic.</p>
Tanya Lattnercb1b9602007-06-15 20:50:54 +00007577
7578<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007579<p>The first argument is a pointer to a value, the second is a pointer to a
7580 global string, the third is a pointer to a global string which is the source
7581 file name, and the last argument is the line number.</p>
Tanya Lattnercb1b9602007-06-15 20:50:54 +00007582
7583<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007584<p>This intrinsic allows annotation of local variables with arbitrary strings.
7585 This can be useful for special purpose optimizations that want to look for
7586 these annotations. These have no other defined use, they are ignored by code
7587 generation and optimization.</p>
Tanya Lattnercb1b9602007-06-15 20:50:54 +00007588
Tanya Lattnercb1b9602007-06-15 20:50:54 +00007589</div>
7590
Tanya Lattner293c0372007-09-21 22:59:12 +00007591<!-- _______________________________________________________________________ -->
7592<div class="doc_subsubsection">
Tanya Lattner0186a652007-09-21 23:57:59 +00007593 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
Tanya Lattner293c0372007-09-21 22:59:12 +00007594</div>
7595
7596<div class="doc_text">
7597
7598<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007599<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
7600 any integer bit width.</p>
7601
Tanya Lattner293c0372007-09-21 22:59:12 +00007602<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007603 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7604 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7605 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7606 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7607 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 +00007608</pre>
7609
7610<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007611<p>The '<tt>llvm.annotation</tt>' intrinsic.</p>
Tanya Lattner293c0372007-09-21 22:59:12 +00007612
7613<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007614<p>The first argument is an integer value (result of some expression), the
7615 second is a pointer to a global string, the third is a pointer to a global
7616 string which is the source file name, and the last argument is the line
7617 number. It returns the value of the first argument.</p>
Tanya Lattner293c0372007-09-21 22:59:12 +00007618
7619<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007620<p>This intrinsic allows annotations to be put on arbitrary expressions with
7621 arbitrary strings. This can be useful for special purpose optimizations that
7622 want to look for these annotations. These have no other defined use, they
7623 are ignored by code generation and optimization.</p>
Tanya Lattner293c0372007-09-21 22:59:12 +00007624
Tanya Lattner293c0372007-09-21 22:59:12 +00007625</div>
Jim Laskey2211f492007-03-14 19:31:19 +00007626
Anton Korobeynikov06cbb652008-01-15 22:31:34 +00007627<!-- _______________________________________________________________________ -->
7628<div class="doc_subsubsection">
7629 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
7630</div>
7631
7632<div class="doc_text">
7633
7634<h5>Syntax:</h5>
7635<pre>
7636 declare void @llvm.trap()
7637</pre>
7638
7639<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007640<p>The '<tt>llvm.trap</tt>' intrinsic.</p>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +00007641
7642<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007643<p>None.</p>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +00007644
7645<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007646<p>This intrinsics is lowered to the target dependent trap instruction. If the
7647 target does not have a trap instruction, this intrinsic will be lowered to
7648 the call of the <tt>abort()</tt> function.</p>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +00007649
Anton Korobeynikov06cbb652008-01-15 22:31:34 +00007650</div>
7651
Bill Wendling14313312008-11-19 05:56:17 +00007652<!-- _______________________________________________________________________ -->
7653<div class="doc_subsubsection">
Misha Brukman50de2b22008-11-22 23:55:29 +00007654 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
Bill Wendling14313312008-11-19 05:56:17 +00007655</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007656
Bill Wendling14313312008-11-19 05:56:17 +00007657<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007658
Bill Wendling14313312008-11-19 05:56:17 +00007659<h5>Syntax:</h5>
7660<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007661 declare void @llvm.stackprotector(i8* &lt;guard&gt;, i8** &lt;slot&gt;)
Bill Wendling14313312008-11-19 05:56:17 +00007662</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007663
Bill Wendling14313312008-11-19 05:56:17 +00007664<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007665<p>The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and
7666 stores it onto the stack at <tt>slot</tt>. The stack slot is adjusted to
7667 ensure that it is placed on the stack before local variables.</p>
7668
Bill Wendling14313312008-11-19 05:56:17 +00007669<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007670<p>The <tt>llvm.stackprotector</tt> intrinsic requires two pointer
7671 arguments. The first argument is the value loaded from the stack
7672 guard <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt>
7673 that has enough space to hold the value of the guard.</p>
7674
Bill Wendling14313312008-11-19 05:56:17 +00007675<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007676<p>This intrinsic causes the prologue/epilogue inserter to force the position of
7677 the <tt>AllocaInst</tt> stack slot to be before local variables on the
7678 stack. This is to ensure that if a local variable on the stack is
7679 overwritten, it will destroy the value of the guard. When the function exits,
Bill Wendling6bbe0912010-10-27 01:07:41 +00007680 the guard on the stack is checked against the original guard. If they are
Bill Wendlingd9a66f72009-07-20 02:29:24 +00007681 different, then the program aborts by calling the <tt>__stack_chk_fail()</tt>
7682 function.</p>
7683
Bill Wendling14313312008-11-19 05:56:17 +00007684</div>
7685
Eric Christopher73484322009-11-30 08:03:53 +00007686<!-- _______________________________________________________________________ -->
7687<div class="doc_subsubsection">
7688 <a name="int_objectsize">'<tt>llvm.objectsize</tt>' Intrinsic</a>
7689</div>
7690
7691<div class="doc_text">
7692
7693<h5>Syntax:</h5>
7694<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +00007695 declare i32 @llvm.objectsize.i32(i8* &lt;object&gt;, i1 &lt;type&gt;)
7696 declare i64 @llvm.objectsize.i64(i8* &lt;object&gt;, i1 &lt;type&gt;)
Eric Christopher73484322009-11-30 08:03:53 +00007697</pre>
7698
7699<h5>Overview:</h5>
Bill Wendling6bbe0912010-10-27 01:07:41 +00007700<p>The <tt>llvm.objectsize</tt> intrinsic is designed to provide information to
7701 the optimizers to determine at compile time whether a) an operation (like
7702 memcpy) will overflow a buffer that corresponds to an object, or b) that a
7703 runtime check for overflow isn't necessary. An object in this context means
7704 an allocation of a specific class, structure, array, or other object.</p>
Eric Christopher73484322009-11-30 08:03:53 +00007705
7706<h5>Arguments:</h5>
Bill Wendling6bbe0912010-10-27 01:07:41 +00007707<p>The <tt>llvm.objectsize</tt> intrinsic takes two arguments. The first
Eric Christopher31e39bd2009-12-23 00:29:49 +00007708 argument is a pointer to or into the <tt>object</tt>. The second argument
Bill Wendling6bbe0912010-10-27 01:07:41 +00007709 is a boolean 0 or 1. This argument determines whether you want the
7710 maximum (0) or minimum (1) bytes remaining. This needs to be a literal 0 or
Eric Christopher31e39bd2009-12-23 00:29:49 +00007711 1, variables are not allowed.</p>
7712
Eric Christopher73484322009-11-30 08:03:53 +00007713<h5>Semantics:</h5>
7714<p>The <tt>llvm.objectsize</tt> intrinsic is lowered to either a constant
Bill Wendling6bbe0912010-10-27 01:07:41 +00007715 representing the size of the object concerned, or <tt>i32/i64 -1 or 0</tt>,
7716 depending on the <tt>type</tt> argument, if the size cannot be determined at
7717 compile time.</p>
Eric Christopher73484322009-11-30 08:03:53 +00007718
7719</div>
7720
Chris Lattner2f7c9632001-06-06 20:29:01 +00007721<!-- *********************************************************************** -->
Chris Lattner2f7c9632001-06-06 20:29:01 +00007722<hr>
Misha Brukmanc501f552004-03-01 17:47:27 +00007723<address>
7724 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman86242e12008-12-11 17:34:48 +00007725 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukmanc501f552004-03-01 17:47:27 +00007726 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman86242e12008-12-11 17:34:48 +00007727 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
Misha Brukmanc501f552004-03-01 17:47:27 +00007728
7729 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
Reid Spencerca058542006-03-14 05:39:39 +00007730 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
Misha Brukmanc501f552004-03-01 17:47:27 +00007731 Last modified: $Date$
7732</address>
Chris Lattnerb8f816e2008-01-04 04:33:49 +00007733
Misha Brukman76307852003-11-08 01:05:38 +00007734</body>
7735</html>