Gitiles
Code Review Sign In
gerrit-public.fairphone.software / toolchain / llvm-project / e63da933eba10f5ea5bf39af47a0ded719c2cf5d / . / llvm / docs / LangRef.html
blob: 69e0a75951f4fdff9b78c6965bc363baf79954c7 [file] [log] [blame]
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3<html>
4<head>
5 <title>LLVM Assembly Language Reference Manual</title>
Reid Spencercb84e432004-08-26 20:44:00 +0000[diff] [blame]6 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
7 <meta name="author" content="Chris Lattner">
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]8 <meta name="description"
Reid Spencercb84e432004-08-26 20:44:00 +0000[diff] [blame]9 content="LLVM Assembly Language Reference Manual.">
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]10 <link rel="stylesheet" href="llvm.css" type="text/css">
11</head>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]12
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]13<body>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]14
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]15<h1>LLVM Language Reference Manual</h1>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]16<ol>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]17 <li><a href="#abstract">Abstract</a></li>
18 <li><a href="#introduction">Introduction</a></li>
19 <li><a href="#identifiers">Identifiers</a></li>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]20 <li><a href="#highlevel">High Level Structure</a>
21 <ol>
22 <li><a href="#modulestructure">Module Structure</a></li>
Bill Wendlinga3c6f6b2009-07-20 01:03:30 +0000[diff] [blame]23 <li><a href="#linkage">Linkage Types</a>
24 <ol>
Bill Wendling8693ef82009-07-20 02:41:50 +0000[diff] [blame]25 <li><a href="#linkage_private">'<tt>private</tt>' Linkage</a></li>
26 <li><a href="#linkage_linker_private">'<tt>linker_private</tt>' Linkage</a></li>
Bill Wendling03bcd6e2010-07-01 21:55:59 +0000[diff] [blame]27 <li><a href="#linkage_linker_private_weak">'<tt>linker_private_weak</tt>' Linkage</a></li>
Bill Wendling578ee402010-08-20 22:05:50 +0000[diff] [blame]28 <li><a href="#linkage_linker_private_weak_def_auto">'<tt>linker_private_weak_def_auto</tt>' Linkage</a></li>
Bill Wendling8693ef82009-07-20 02:41:50 +0000[diff] [blame]29 <li><a href="#linkage_internal">'<tt>internal</tt>' Linkage</a></li>
30 <li><a href="#linkage_available_externally">'<tt>available_externally</tt>' Linkage</a></li>
31 <li><a href="#linkage_linkonce">'<tt>linkonce</tt>' Linkage</a></li>
32 <li><a href="#linkage_common">'<tt>common</tt>' Linkage</a></li>
33 <li><a href="#linkage_weak">'<tt>weak</tt>' Linkage</a></li>
34 <li><a href="#linkage_appending">'<tt>appending</tt>' Linkage</a></li>
35 <li><a href="#linkage_externweak">'<tt>extern_weak</tt>' Linkage</a></li>
Chris Lattner80d73c72009-10-10 18:26:06 +0000[diff] [blame]36 <li><a href="#linkage_linkonce_odr">'<tt>linkonce_odr</tt>' Linkage</a></li>
Bill Wendling8693ef82009-07-20 02:41:50 +0000[diff] [blame]37 <li><a href="#linkage_weak">'<tt>weak_odr</tt>' Linkage</a></li>
38 <li><a href="#linkage_external">'<tt>externally visible</tt>' Linkage</a></li>
39 <li><a href="#linkage_dllimport">'<tt>dllimport</tt>' Linkage</a></li>
40 <li><a href="#linkage_dllexport">'<tt>dllexport</tt>' Linkage</a></li>
Bill Wendlinga3c6f6b2009-07-20 01:03:30 +0000[diff] [blame]41 </ol>
42 </li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]43 <li><a href="#callingconv">Calling Conventions</a></li>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]44 <li><a href="#namedtypes">Named Types</a></li>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]45 <li><a href="#globalvars">Global Variables</a></li>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]46 <li><a href="#functionstructure">Functions</a></li>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]47 <li><a href="#aliasstructure">Aliases</a></li>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]48 <li><a href="#namedmetadatastructure">Named Metadata</a></li>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]49 <li><a href="#paramattrs">Parameter Attributes</a></li>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]50 <li><a href="#fnattrs">Function Attributes</a></li>
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]51 <li><a href="#gc">Garbage Collector Names</a></li>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]52 <li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]53 <li><a href="#datalayout">Data Layout</a></li>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]54 <li><a href="#pointeraliasing">Pointer Aliasing Rules</a></li>
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]55 <li><a href="#volatile">Volatile Memory Accesses</a></li>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]56 </ol>
57 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]58 <li><a href="#typesystem">Type System</a>
59 <ol>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]60 <li><a href="#t_classifications">Type Classifications</a></li>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]61 <li><a href="#t_primitive">Primitive Types</a>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]62 <ol>
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]63 <li><a href="#t_integer">Integer Type</a></li>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]64 <li><a href="#t_floating">Floating Point Types</a></li>
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]65 <li><a href="#t_x86mmx">X86mmx Type</a></li>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]66 <li><a href="#t_void">Void Type</a></li>
67 <li><a href="#t_label">Label Type</a></li>
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]68 <li><a href="#t_metadata">Metadata Type</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]69 </ol>
70 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]71 <li><a href="#t_derived">Derived Types</a>
72 <ol>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]73 <li><a href="#t_aggregate">Aggregate Types</a>
74 <ol>
75 <li><a href="#t_array">Array Type</a></li>
76 <li><a href="#t_struct">Structure Type</a></li>
77 <li><a href="#t_pstruct">Packed Structure Type</a></li>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]78 <li><a href="#t_vector">Vector Type</a></li>
79 </ol>
80 </li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]81 <li><a href="#t_function">Function Type</a></li>
82 <li><a href="#t_pointer">Pointer Type</a></li>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]83 <li><a href="#t_opaque">Opaque Type</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]84 </ol>
85 </li>
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]86 <li><a href="#t_uprefs">Type Up-references</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]87 </ol>
88 </li>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]89 <li><a href="#constants">Constants</a>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]90 <ol>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]91 <li><a href="#simpleconstants">Simple Constants</a></li>
Chris Lattner361bfcd2009-02-28 18:32:25 +0000[diff] [blame]92 <li><a href="#complexconstants">Complex Constants</a></li>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]93 <li><a href="#globalconstants">Global Variable and Function Addresses</a></li>
94 <li><a href="#undefvalues">Undefined Values</a></li>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]95 <li><a href="#trapvalues">Trap Values</a></li>
Chris Lattner2bfd3202009-10-27 21:19:13 +0000[diff] [blame]96 <li><a href="#blockaddress">Addresses of Basic Blocks</a></li>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]97 <li><a href="#constantexprs">Constant Expressions</a></li>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]98 </ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]99 </li>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]100 <li><a href="#othervalues">Other Values</a>
101 <ol>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]102 <li><a href="#inlineasm">Inline Assembler Expressions</a></li>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]103 <li><a href="#metadata">Metadata Nodes and Metadata Strings</a></li>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]104 </ol>
105 </li>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]106 <li><a href="#intrinsic_globals">Intrinsic Global Variables</a>
107 <ol>
108 <li><a href="#intg_used">The '<tt>llvm.used</tt>' Global Variable</a></li>
Chris Lattner58f9bb22009-07-20 06:14:25 +0000[diff] [blame]109 <li><a href="#intg_compiler_used">The '<tt>llvm.compiler.used</tt>'
110 Global Variable</a></li>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]111 <li><a href="#intg_global_ctors">The '<tt>llvm.global_ctors</tt>'
112 Global Variable</a></li>
113 <li><a href="#intg_global_dtors">The '<tt>llvm.global_dtors</tt>'
114 Global Variable</a></li>
115 </ol>
116 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]117 <li><a href="#instref">Instruction Reference</a>
118 <ol>
119 <li><a href="#terminators">Terminator Instructions</a>
120 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]121 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
122 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]123 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]124 <li><a href="#i_indirectbr">'<tt>indirectbr</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]125 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]126 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]127 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]128 </ol>
129 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]130 <li><a href="#binaryops">Binary Operations</a>
131 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]132 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]133 <li><a href="#i_fadd">'<tt>fadd</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]134 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]135 <li><a href="#i_fsub">'<tt>fsub</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]136 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]137 <li><a href="#i_fmul">'<tt>fmul</tt>' Instruction</a></li>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]138 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
139 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
140 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]141 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
142 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
143 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]144 </ol>
145 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]146 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
147 <ol>
Reid Spencer2ab01932007-02-02 13:57:07 +0000[diff] [blame]148 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
149 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
150 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]151 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]152 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]153 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]154 </ol>
155 </li>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]156 <li><a href="#vectorops">Vector Operations</a>
157 <ol>
158 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
159 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
160 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]161 </ol>
162 </li>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]163 <li><a href="#aggregateops">Aggregate Operations</a>
164 <ol>
165 <li><a href="#i_extractvalue">'<tt>extractvalue</tt>' Instruction</a></li>
166 <li><a href="#i_insertvalue">'<tt>insertvalue</tt>' Instruction</a></li>
167 </ol>
168 </li>
Chris Lattner6ab66722006-08-15 00:45:58 +0000[diff] [blame]169 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]170 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]171 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
Robert Bocchino820bc75b2006-02-17 21:18:08 +0000[diff] [blame]172 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
173 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
174 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]175 </ol>
176 </li>
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]177 <li><a href="#convertops">Conversion Operations</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]178 <ol>
179 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
180 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
181 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
182 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
183 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]184 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
185 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
186 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
187 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]188 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
189 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]190 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]191 </ol>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]192 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]193 <li><a href="#otherops">Other Operations</a>
194 <ol>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]195 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
196 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]197 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]198 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]199 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Chris Lattner33337472006-01-13 23:26:01 +0000[diff] [blame]200 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]201 </ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]202 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]203 </ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]204 </li>
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]205 <li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]206 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]207 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
208 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]209 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
210 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
211 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]212 </ol>
213 </li>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]214 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
215 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]216 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
217 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
218 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]219 </ol>
220 </li>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]221 <li><a href="#int_codegen">Code Generator Intrinsics</a>
222 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]223 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
224 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
225 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
226 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
227 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
228 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
Dan Gohmane58f7b32010-05-26 21:56:15 +0000[diff] [blame]229 <li><a href="#int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswellaa1c3c12004-04-09 16:43:20 +0000[diff] [blame]230 </ol>
231 </li>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]232 <li><a href="#int_libc">Standard C Library Intrinsics</a>
233 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]234 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
235 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
236 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
237 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
238 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]239 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
240 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
241 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]242 </ol>
243 </li>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]244 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]245 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]246 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattnerb748c672006-01-16 22:34:14 +0000[diff] [blame]247 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
248 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
249 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]250 </ol>
251 </li>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]252 <li><a href="#int_overflow">Arithmetic with Overflow Intrinsics</a>
253 <ol>
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]254 <li><a href="#int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt> Intrinsics</a></li>
255 <li><a href="#int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt> Intrinsics</a></li>
256 <li><a href="#int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt> Intrinsics</a></li>
257 <li><a href="#int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt> Intrinsics</a></li>
258 <li><a href="#int_smul_overflow">'<tt>llvm.smul.with.overflow.*</tt> Intrinsics</a></li>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]259 <li><a href="#int_umul_overflow">'<tt>llvm.umul.with.overflow.*</tt> Intrinsics</a></li>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]260 </ol>
261 </li>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]262 <li><a href="#int_fp16">Half Precision Floating Point Intrinsics</a>
263 <ol>
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]264 <li><a href="#int_convert_to_fp16">'<tt>llvm.convert.to.fp16</tt>' Intrinsic</a></li>
265 <li><a href="#int_convert_from_fp16">'<tt>llvm.convert.from.fp16</tt>' Intrinsic</a></li>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]266 </ol>
267 </li>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]268 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]269 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sands86e01192007-09-11 14:10:23 +0000[diff] [blame]270 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]271 <ol>
272 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]273 </ol>
274 </li>
Bill Wendlingf85850f2008-11-18 22:10:53 +0000[diff] [blame]275 <li><a href="#int_atomics">Atomic intrinsics</a>
276 <ol>
277 <li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
278 <li><a href="#int_atomic_cmp_swap"><tt>llvm.atomic.cmp.swap</tt></a></li>
279 <li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
280 <li><a href="#int_atomic_load_add"><tt>llvm.atomic.load.add</tt></a></li>
281 <li><a href="#int_atomic_load_sub"><tt>llvm.atomic.load.sub</tt></a></li>
282 <li><a href="#int_atomic_load_and"><tt>llvm.atomic.load.and</tt></a></li>
283 <li><a href="#int_atomic_load_nand"><tt>llvm.atomic.load.nand</tt></a></li>
284 <li><a href="#int_atomic_load_or"><tt>llvm.atomic.load.or</tt></a></li>
285 <li><a href="#int_atomic_load_xor"><tt>llvm.atomic.load.xor</tt></a></li>
286 <li><a href="#int_atomic_load_max"><tt>llvm.atomic.load.max</tt></a></li>
287 <li><a href="#int_atomic_load_min"><tt>llvm.atomic.load.min</tt></a></li>
288 <li><a href="#int_atomic_load_umax"><tt>llvm.atomic.load.umax</tt></a></li>
289 <li><a href="#int_atomic_load_umin"><tt>llvm.atomic.load.umin</tt></a></li>
290 </ol>
291 </li>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]292 <li><a href="#int_memorymarkers">Memory Use Markers</a>
293 <ol>
294 <li><a href="#int_lifetime_start"><tt>llvm.lifetime.start</tt></a></li>
295 <li><a href="#int_lifetime_end"><tt>llvm.lifetime.end</tt></a></li>
296 <li><a href="#int_invariant_start"><tt>llvm.invariant.start</tt></a></li>
297 <li><a href="#int_invariant_end"><tt>llvm.invariant.end</tt></a></li>
298 </ol>
299 </li>
Reid Spencer5b2cb0f2007-07-20 19:59:11 +0000[diff] [blame]300 <li><a href="#int_general">General intrinsics</a>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]301 <ol>
Reid Spencer5b2cb0f2007-07-20 19:59:11 +0000[diff] [blame]302 <li><a href="#int_var_annotation">
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]303 '<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]304 <li><a href="#int_annotation">
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]305 '<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]306 <li><a href="#int_trap">
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]307 '<tt>llvm.trap</tt>' Intrinsic</a></li>
308 <li><a href="#int_stackprotector">
309 '<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]310 <li><a href="#int_objectsize">
311 '<tt>llvm.objectsize</tt>' Intrinsic</a></li>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]312 </ol>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]313 </li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]314 </ol>
315 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]316</ol>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]317
318<div class="doc_author">
319 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
320 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]321</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]322
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]323<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]324<h2><a name="abstract">Abstract</a></h2>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]325<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]326
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]327<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]328
329<p>This document is a reference manual for the LLVM assembly language. LLVM is
330 a Static Single Assignment (SSA) based representation that provides type
331 safety, low-level operations, flexibility, and the capability of representing
332 'all' high-level languages cleanly. It is the common code representation
333 used throughout all phases of the LLVM compilation strategy.</p>
334
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]335</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]336
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]337<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]338<h2><a name="introduction">Introduction</a></h2>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]339<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]340
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]341<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]342
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]343<p>The LLVM code representation is designed to be used in three different forms:
344 as an in-memory compiler IR, as an on-disk bitcode representation (suitable
345 for fast loading by a Just-In-Time compiler), and as a human readable
346 assembly language representation. This allows LLVM to provide a powerful
347 intermediate representation for efficient compiler transformations and
348 analysis, while providing a natural means to debug and visualize the
349 transformations. The three different forms of LLVM are all equivalent. This
350 document describes the human readable representation and notation.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]351
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]352<p>The LLVM representation aims to be light-weight and low-level while being
353 expressive, typed, and extensible at the same time. It aims to be a
354 "universal IR" of sorts, by being at a low enough level that high-level ideas
355 may be cleanly mapped to it (similar to how microprocessors are "universal
356 IR's", allowing many source languages to be mapped to them). By providing
357 type information, LLVM can be used as the target of optimizations: for
358 example, through pointer analysis, it can be proven that a C automatic
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]359 variable is never accessed outside of the current function, allowing it to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]360 be promoted to a simple SSA value instead of a memory location.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]361
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]362</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]363
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]364<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]365<h4>
366 <a name="wellformed">Well-Formedness</a>
367</h4>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]368
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]369<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]370
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]371<p>It is important to note that this document describes 'well formed' LLVM
372 assembly language. There is a difference between what the parser accepts and
373 what is considered 'well formed'. For example, the following instruction is
374 syntactically okay, but not well formed:</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]375
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]376<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]377%x = <a href="#i_add">add</a> i32 1, %x
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]378</pre>
379
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]380<p>because the definition of <tt>%x</tt> does not dominate all of its uses. The
381 LLVM infrastructure provides a verification pass that may be used to verify
382 that an LLVM module is well formed. This pass is automatically run by the
383 parser after parsing input assembly and by the optimizer before it outputs
384 bitcode. The violations pointed out by the verifier pass indicate bugs in
385 transformation passes or input to the parser.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]386
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]387</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]388
Chris Lattner87a3dbe2007-10-03 17:34:29 +0000[diff] [blame]389<!-- Describe the typesetting conventions here. -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]390
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]391<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]392<h2><a name="identifiers">Identifiers</a></h2>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]393<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]394
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]395<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]396
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]397<p>LLVM identifiers come in two basic types: global and local. Global
398 identifiers (functions, global variables) begin with the <tt>'@'</tt>
399 character. Local identifiers (register names, types) begin with
400 the <tt>'%'</tt> character. Additionally, there are three different formats
401 for identifiers, for different purposes:</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]402
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]403<ol>
Reid Spencerb23b65f2007-08-07 14:34:28 +0000[diff] [blame]404 <li>Named values are represented as a string of characters with their prefix.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]405 For example, <tt>%foo</tt>, <tt>@DivisionByZero</tt>,
406 <tt>%a.really.long.identifier</tt>. The actual regular expression used is
407 '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'. Identifiers which require
408 other characters in their names can be surrounded with quotes. Special
409 characters may be escaped using <tt>"\xx"</tt> where <tt>xx</tt> is the
410 ASCII code for the character in hexadecimal. In this way, any character
411 can be used in a name value, even quotes themselves.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]412
Reid Spencerb23b65f2007-08-07 14:34:28 +0000[diff] [blame]413 <li>Unnamed values are represented as an unsigned numeric value with their
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]414 prefix. For example, <tt>%12</tt>, <tt>@2</tt>, <tt>%44</tt>.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]415
Reid Spencer8f08d802004-12-09 18:02:53 +0000[diff] [blame]416 <li>Constants, which are described in a <a href="#constants">section about
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]417 constants</a>, below.</li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]418</ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]419
Reid Spencerb23b65f2007-08-07 14:34:28 +0000[diff] [blame]420<p>LLVM requires that values start with a prefix for two reasons: Compilers
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]421 don't need to worry about name clashes with reserved words, and the set of
422 reserved words may be expanded in the future without penalty. Additionally,
423 unnamed identifiers allow a compiler to quickly come up with a temporary
424 variable without having to avoid symbol table conflicts.</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]425
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]426<p>Reserved words in LLVM are very similar to reserved words in other
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]427 languages. There are keywords for different opcodes
428 ('<tt><a href="#i_add">add</a></tt>',
429 '<tt><a href="#i_bitcast">bitcast</a></tt>',
430 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names
431 ('<tt><a href="#t_void">void</a></tt>',
432 '<tt><a href="#t_primitive">i32</a></tt>', etc...), and others. These
433 reserved words cannot conflict with variable names, because none of them
434 start with a prefix character (<tt>'%'</tt> or <tt>'@'</tt>).</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]435
436<p>Here is an example of LLVM code to multiply the integer variable
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]437 '<tt>%X</tt>' by 8:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]438
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]439<p>The easy way:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]440
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]441<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]442%result = <a href="#i_mul">mul</a> i32 %X, 8
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]443</pre>
444
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]445<p>After strength reduction:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]446
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]447<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]448%result = <a href="#i_shl">shl</a> i32 %X, i8 3
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]449</pre>
450
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]451<p>And the hard way:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]452
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]453<pre class="doc_code">
Gabor Greifbd0328f2009-10-28 13:05:07 +0000[diff] [blame]454%0 = <a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
455%1 = <a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]456%result = <a href="#i_add">add</a> i32 %1, %1
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]457</pre>
458
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]459<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several important
460 lexical features of LLVM:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]461
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]462<ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]463 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]464 line.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]465
466 <li>Unnamed temporaries are created when the result of a computation is not
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]467 assigned to a named value.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]468
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]469 <li>Unnamed temporaries are numbered sequentially</li>
470</ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]471
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]472<p>It also shows a convention that we follow in this document. When
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]473 demonstrating instructions, we will follow an instruction with a comment that
474 defines the type and name of value produced. Comments are shown in italic
475 text.</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]476
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]477</div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]478
479<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]480<h2><a name="highlevel">High Level Structure</a></h2>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]481<!-- *********************************************************************** -->
482
483<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]484<h3>
485 <a name="modulestructure">Module Structure</a>
486</h3>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]487
488<div class="doc_text">
489
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]490<p>LLVM programs are composed of "Module"s, each of which is a translation unit
491 of the input programs. Each module consists of functions, global variables,
492 and symbol table entries. Modules may be combined together with the LLVM
493 linker, which merges function (and global variable) definitions, resolves
494 forward declarations, and merges symbol table entries. Here is an example of
495 the "hello world" module:</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]496
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]497<pre class="doc_code">
Chris Lattner54a7be72010-08-17 17:13:42 +0000[diff] [blame]498<i>; Declare the string constant as a global constant.</i>&nbsp;
Nick Lewyckyfea7ddc2011-01-29 01:09:53 +0000[diff] [blame]499<a href="#identifiers">@.LC0</a> = <a href="#linkage_internal">internal</a>&nbsp;<a href="#globalvars">constant</a>&nbsp;<a href="#t_array">[13 x i8]</a> c"hello world\0A\00" <i>; [13 x i8]*</i>&nbsp;
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]500
Chris Lattner54a7be72010-08-17 17:13:42 +0000[diff] [blame]501<i>; External declaration of the puts function</i>&nbsp;
502<a href="#functionstructure">declare</a> i32 @puts(i8*) <i>; i32 (i8*)* </i>&nbsp;
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]503
504<i>; Definition of main function</i>
Chris Lattner54a7be72010-08-17 17:13:42 +0000[diff] [blame]505define i32 @main() { <i>; i32()* </i>&nbsp;
506 <i>; Convert [13 x i8]* to i8 *...</i>&nbsp;
507 %cast210 = <a href="#i_getelementptr">getelementptr</a> [13 x i8]* @.LC0, i64 0, i64 0 <i>; i8*</i>&nbsp;
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]508
Chris Lattner54a7be72010-08-17 17:13:42 +0000[diff] [blame]509 <i>; Call puts function to write out the string to stdout.</i>&nbsp;
510 <a href="#i_call">call</a> i32 @puts(i8* %cast210) <i>; i32</i>&nbsp;
511 <a href="#i_ret">ret</a> i32 0&nbsp;
512}
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]513
514<i>; Named metadata</i>
515!1 = metadata !{i32 41}
516!foo = !{!1, null}
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]517</pre>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]518
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]519<p>This example is made up of a <a href="#globalvars">global variable</a> named
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]520 "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>" function,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]521 a <a href="#functionstructure">function definition</a> for
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]522 "<tt>main</tt>" and <a href="#namedmetadatastructure">named metadata</a>
523 "<tt>foo"</tt>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]524
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]525<p>In general, a module is made up of a list of global values, where both
526 functions and global variables are global values. Global values are
527 represented by a pointer to a memory location (in this case, a pointer to an
528 array of char, and a pointer to a function), and have one of the
529 following <a href="#linkage">linkage types</a>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]530
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]531</div>
532
533<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]534<h3>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]535 <a name="linkage">Linkage Types</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]536</h3>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]537
538<div class="doc_text">
539
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]540<p>All Global Variables and Functions have one of the following types of
541 linkage:</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]542
543<dl>
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]544 <dt><tt><b><a name="linkage_private">private</a></b></tt></dt>
Bill Wendling03bcd6e2010-07-01 21:55:59 +0000[diff] [blame]545 <dd>Global values with "<tt>private</tt>" linkage are only directly accessible
546 by objects in the current module. In particular, linking code into a
547 module with an private global value may cause the private to be renamed as
548 necessary to avoid collisions. Because the symbol is private to the
549 module, all references can be updated. This doesn't show up in any symbol
550 table in the object file.</dd>
Rafael Espindola6de96a12009-01-15 20:18:42 +0000[diff] [blame]551
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]552 <dt><tt><b><a name="linkage_linker_private">linker_private</a></b></tt></dt>
Bill Wendling03bcd6e2010-07-01 21:55:59 +0000[diff] [blame]553 <dd>Similar to <tt>private</tt>, but the symbol is passed through the
554 assembler and evaluated by the linker. Unlike normal strong symbols, they
555 are removed by the linker from the final linked image (executable or
556 dynamic library).</dd>
557
558 <dt><tt><b><a name="linkage_linker_private_weak">linker_private_weak</a></b></tt></dt>
559 <dd>Similar to "<tt>linker_private</tt>", but the symbol is weak. Note that
560 <tt>linker_private_weak</tt> symbols are subject to coalescing by the
561 linker. The symbols are removed by the linker from the final linked image
562 (executable or dynamic library).</dd>
Bill Wendlinga3c6f6b2009-07-20 01:03:30 +0000[diff] [blame]563
Bill Wendling578ee402010-08-20 22:05:50 +0000[diff] [blame]564 <dt><tt><b><a name="linkage_linker_private_weak_def_auto">linker_private_weak_def_auto</a></b></tt></dt>
565 <dd>Similar to "<tt>linker_private_weak</tt>", but it's known that the address
566 of the object is not taken. For instance, functions that had an inline
567 definition, but the compiler decided not to inline it. Note,
568 unlike <tt>linker_private</tt> and <tt>linker_private_weak</tt>,
569 <tt>linker_private_weak_def_auto</tt> may have only <tt>default</tt>
570 visibility. The symbols are removed by the linker from the final linked
571 image (executable or dynamic library).</dd>
572
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]573 <dt><tt><b><a name="linkage_internal">internal</a></b></tt></dt>
Bill Wendling36321712010-06-29 22:34:52 +0000[diff] [blame]574 <dd>Similar to private, but the value shows as a local symbol
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]575 (<tt>STB_LOCAL</tt> in the case of ELF) in the object file. This
576 corresponds to the notion of the '<tt>static</tt>' keyword in C.</dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]577
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]578 <dt><tt><b><a name="linkage_available_externally">available_externally</a></b></tt></dt>
Chris Lattner184f1be2009-04-13 05:44:34 +0000[diff] [blame]579 <dd>Globals with "<tt>available_externally</tt>" linkage are never emitted
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]580 into the object file corresponding to the LLVM module. They exist to
581 allow inlining and other optimizations to take place given knowledge of
582 the definition of the global, which is known to be somewhere outside the
583 module. Globals with <tt>available_externally</tt> linkage are allowed to
584 be discarded at will, and are otherwise the same as <tt>linkonce_odr</tt>.
585 This linkage type is only allowed on definitions, not declarations.</dd>
Chris Lattner184f1be2009-04-13 05:44:34 +0000[diff] [blame]586
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]587 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt></dt>
Chris Lattnere20b4702007-01-14 06:51:48 +0000[diff] [blame]588 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
Chris Lattner0de4caa2010-01-09 19:15:14 +0000[diff] [blame]589 the same name when linkage occurs. This can be used to implement
590 some forms of inline functions, templates, or other code which must be
591 generated in each translation unit that uses it, but where the body may
592 be overridden with a more definitive definition later. Unreferenced
593 <tt>linkonce</tt> globals are allowed to be discarded. Note that
594 <tt>linkonce</tt> linkage does not actually allow the optimizer to
595 inline the body of this function into callers because it doesn't know if
596 this definition of the function is the definitive definition within the
597 program or whether it will be overridden by a stronger definition.
598 To enable inlining and other optimizations, use "<tt>linkonce_odr</tt>"
599 linkage.</dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]600
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]601 <dt><tt><b><a name="linkage_weak">weak</a></b></tt></dt>
Chris Lattnerd0554882009-08-05 05:21:07 +0000[diff] [blame]602 <dd>"<tt>weak</tt>" linkage has the same merging semantics as
603 <tt>linkonce</tt> linkage, except that unreferenced globals with
604 <tt>weak</tt> linkage may not be discarded. This is used for globals that
605 are declared "weak" in C source code.</dd>
606
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]607 <dt><tt><b><a name="linkage_common">common</a></b></tt></dt>
Chris Lattnerd0554882009-08-05 05:21:07 +0000[diff] [blame]608 <dd>"<tt>common</tt>" linkage is most similar to "<tt>weak</tt>" linkage, but
609 they are used for tentative definitions in C, such as "<tt>int X;</tt>" at
610 global scope.
611 Symbols with "<tt>common</tt>" linkage are merged in the same way as
612 <tt>weak symbols</tt>, and they may not be deleted if unreferenced.
Chris Lattner0aff0b22009-08-05 05:41:44 +0000[diff] [blame]613 <tt>common</tt> symbols may not have an explicit section,
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]614 must have a zero initializer, and may not be marked '<a
Chris Lattner0aff0b22009-08-05 05:41:44 +0000[diff] [blame]615 href="#globalvars"><tt>constant</tt></a>'. Functions and aliases may not
616 have common linkage.</dd>
Chris Lattnerd0554882009-08-05 05:21:07 +0000[diff] [blame]617
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]618
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]619 <dt><tt><b><a name="linkage_appending">appending</a></b></tt></dt>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]620 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]621 pointer to array type. When two global variables with appending linkage
622 are linked together, the two global arrays are appended together. This is
623 the LLVM, typesafe, equivalent of having the system linker append together
624 "sections" with identical names when .o files are linked.</dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]625
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]626 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]627 <dd>The semantics of this linkage follow the ELF object file model: the symbol
628 is weak until linked, if not linked, the symbol becomes null instead of
629 being an undefined reference.</dd>
Anton Korobeynikova0554d92007-01-12 19:20:47 +0000[diff] [blame]630
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]631 <dt><tt><b><a name="linkage_linkonce_odr">linkonce_odr</a></b></tt></dt>
632 <dt><tt><b><a name="linkage_weak_odr">weak_odr</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]633 <dd>Some languages allow differing globals to be merged, such as two functions
634 with different semantics. Other languages, such as <tt>C++</tt>, ensure
Bill Wendling03bcd6e2010-07-01 21:55:59 +0000[diff] [blame]635 that only equivalent globals are ever merged (the "one definition rule"
636 &mdash; "ODR"). Such languages can use the <tt>linkonce_odr</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]637 and <tt>weak_odr</tt> linkage types to indicate that the global will only
638 be merged with equivalent globals. These linkage types are otherwise the
639 same as their non-<tt>odr</tt> versions.</dd>
Duncan Sands12da8ce2009-03-07 15:45:40 +0000[diff] [blame]640
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]641 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]642 <dd>If none of the above identifiers are used, the global is externally
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]643 visible, meaning that it participates in linkage and can be used to
644 resolve external symbol references.</dd>
Reid Spencer7972c472007-04-11 23:49:50 +0000[diff] [blame]645</dl>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]646
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]647<p>The next two types of linkage are targeted for Microsoft Windows platform
648 only. They are designed to support importing (exporting) symbols from (to)
649 DLLs (Dynamic Link Libraries).</p>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]650
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]651<dl>
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]652 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt></dt>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]653 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]654 or variable via a global pointer to a pointer that is set up by the DLL
655 exporting the symbol. On Microsoft Windows targets, the pointer name is
656 formed by combining <code>__imp_</code> and the function or variable
657 name.</dd>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]658
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]659 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt></dt>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]660 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]661 pointer to a pointer in a DLL, so that it can be referenced with the
662 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
663 name is formed by combining <code>__imp_</code> and the function or
664 variable name.</dd>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]665</dl>
666
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]667<p>For example, since the "<tt>.LC0</tt>" variable is defined to be internal, if
668 another module defined a "<tt>.LC0</tt>" variable and was linked with this
669 one, one of the two would be renamed, preventing a collision. Since
670 "<tt>main</tt>" and "<tt>puts</tt>" are external (i.e., lacking any linkage
671 declarations), they are accessible outside of the current module.</p>
672
673<p>It is illegal for a function <i>declaration</i> to have any linkage type
674 other than "externally visible", <tt>dllimport</tt>
675 or <tt>extern_weak</tt>.</p>
676
Duncan Sands12da8ce2009-03-07 15:45:40 +0000[diff] [blame]677<p>Aliases can have only <tt>external</tt>, <tt>internal</tt>, <tt>weak</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]678 or <tt>weak_odr</tt> linkages.</p>
679
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]680</div>
681
682<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]683<h3>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]684 <a name="callingconv">Calling Conventions</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]685</h3>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]686
687<div class="doc_text">
688
689<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]690 and <a href="#i_invoke">invokes</a> can all have an optional calling
691 convention specified for the call. The calling convention of any pair of
692 dynamic caller/callee must match, or the behavior of the program is
693 undefined. The following calling conventions are supported by LLVM, and more
694 may be added in the future:</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]695
696<dl>
697 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]698 <dd>This calling convention (the default if no other calling convention is
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]699 specified) matches the target C calling conventions. This calling
700 convention supports varargs function calls and tolerates some mismatch in
701 the declared prototype and implemented declaration of the function (as
702 does normal C).</dd>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]703
704 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]705 <dd>This calling convention attempts to make calls as fast as possible
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]706 (e.g. by passing things in registers). This calling convention allows the
707 target to use whatever tricks it wants to produce fast code for the
708 target, without having to conform to an externally specified ABI
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]709 (Application Binary Interface).
710 <a href="CodeGenerator.html#tailcallopt">Tail calls can only be optimized
Chris Lattnera179e4d2010-03-11 00:22:57 +0000[diff] [blame]711 when this or the GHC convention is used.</a> This calling convention
712 does not support varargs and requires the prototype of all callees to
713 exactly match the prototype of the function definition.</dd>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]714
715 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]716 <dd>This calling convention attempts to make code in the caller as efficient
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]717 as possible under the assumption that the call is not commonly executed.
718 As such, these calls often preserve all registers so that the call does
719 not break any live ranges in the caller side. This calling convention
720 does not support varargs and requires the prototype of all callees to
721 exactly match the prototype of the function definition.</dd>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]722
Chris Lattnera179e4d2010-03-11 00:22:57 +0000[diff] [blame]723 <dt><b>"<tt>cc <em>10</em></tt>" - GHC convention</b>:</dt>
724 <dd>This calling convention has been implemented specifically for use by the
725 <a href="http://www.haskell.org/ghc">Glasgow Haskell Compiler (GHC)</a>.
726 It passes everything in registers, going to extremes to achieve this by
727 disabling callee save registers. This calling convention should not be
728 used lightly but only for specific situations such as an alternative to
729 the <em>register pinning</em> performance technique often used when
730 implementing functional programming languages.At the moment only X86
731 supports this convention and it has the following limitations:
732 <ul>
733 <li>On <em>X86-32</em> only supports up to 4 bit type parameters. No
734 floating point types are supported.</li>
735 <li>On <em>X86-64</em> only supports up to 10 bit type parameters and
736 6 floating point parameters.</li>
737 </ul>
738 This calling convention supports
739 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> but
740 requires both the caller and callee are using it.
741 </dd>
742
Chris Lattner573f64e2005-05-07 01:46:40 +0000[diff] [blame]743 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]744 <dd>Any calling convention may be specified by number, allowing
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]745 target-specific calling conventions to be used. Target specific calling
746 conventions start at 64.</dd>
Chris Lattner573f64e2005-05-07 01:46:40 +0000[diff] [blame]747</dl>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]748
749<p>More calling conventions can be added/defined on an as-needed basis, to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]750 support Pascal conventions or any other well-known target-independent
751 convention.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]752
753</div>
754
755<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]756<h3>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]757 <a name="visibility">Visibility Styles</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]758</h3>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]759
760<div class="doc_text">
761
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]762<p>All Global Variables and Functions have one of the following visibility
763 styles:</p>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]764
765<dl>
766 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]767 <dd>On targets that use the ELF object file format, default visibility means
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]768 that the declaration is visible to other modules and, in shared libraries,
769 means that the declared entity may be overridden. On Darwin, default
770 visibility means that the declaration is visible to other modules. Default
771 visibility corresponds to "external linkage" in the language.</dd>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]772
773 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]774 <dd>Two declarations of an object with hidden visibility refer to the same
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]775 object if they are in the same shared object. Usually, hidden visibility
776 indicates that the symbol will not be placed into the dynamic symbol
777 table, so no other module (executable or shared library) can reference it
778 directly.</dd>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]779
Anton Korobeynikov39f3cff2007-04-29 18:35:00 +0000[diff] [blame]780 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
Anton Korobeynikov39f3cff2007-04-29 18:35:00 +0000[diff] [blame]781 <dd>On ELF, protected visibility indicates that the symbol will be placed in
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]782 the dynamic symbol table, but that references within the defining module
783 will bind to the local symbol. That is, the symbol cannot be overridden by
784 another module.</dd>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]785</dl>
786
787</div>
788
789<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]790<h3>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]791 <a name="namedtypes">Named Types</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]792</h3>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]793
794<div class="doc_text">
795
796<p>LLVM IR allows you to specify name aliases for certain types. This can make
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]797 it easier to read the IR and make the IR more condensed (particularly when
798 recursive types are involved). An example of a name specification is:</p>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]799
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]800<pre class="doc_code">
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]801%mytype = type { %mytype*, i32 }
802</pre>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]803
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]804<p>You may give a name to any <a href="#typesystem">type</a> except
Chris Lattner249b9762010-08-17 23:26:04 +0000[diff] [blame]805 "<a href="#t_void">void</a>". Type name aliases may be used anywhere a type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]806 is expected with the syntax "%mytype".</p>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]807
808<p>Note that type names are aliases for the structural type that they indicate,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]809 and that you can therefore specify multiple names for the same type. This
810 often leads to confusing behavior when dumping out a .ll file. Since LLVM IR
811 uses structural typing, the name is not part of the type. When printing out
812 LLVM IR, the printer will pick <em>one name</em> to render all types of a
813 particular shape. This means that if you have code where two different
814 source types end up having the same LLVM type, that the dumper will sometimes
815 print the "wrong" or unexpected type. This is an important design point and
816 isn't going to change.</p>
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]817
818</div>
819
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]820<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]821<h3>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]822 <a name="globalvars">Global Variables</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]823</h3>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]824
825<div class="doc_text">
826
Chris Lattner5d5aede2005-02-12 19:30:21 +0000[diff] [blame]827<p>Global variables define regions of memory allocated at compilation time
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]828 instead of run-time. Global variables may optionally be initialized, may
829 have an explicit section to be placed in, and may have an optional explicit
830 alignment specified. A variable may be defined as "thread_local", which
831 means that it will not be shared by threads (each thread will have a
832 separated copy of the variable). A variable may be defined as a global
833 "constant," which indicates that the contents of the variable
834 will <b>never</b> be modified (enabling better optimization, allowing the
835 global data to be placed in the read-only section of an executable, etc).
836 Note that variables that need runtime initialization cannot be marked
837 "constant" as there is a store to the variable.</p>
Chris Lattner5d5aede2005-02-12 19:30:21 +0000[diff] [blame]838
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]839<p>LLVM explicitly allows <em>declarations</em> of global variables to be marked
840 constant, even if the final definition of the global is not. This capability
841 can be used to enable slightly better optimization of the program, but
842 requires the language definition to guarantee that optimizations based on the
843 'constantness' are valid for the translation units that do not include the
844 definition.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]845
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]846<p>As SSA values, global variables define pointer values that are in scope
847 (i.e. they dominate) all basic blocks in the program. Global variables
848 always define a pointer to their "content" type because they describe a
849 region of memory, and all memory objects in LLVM are accessed through
850 pointers.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]851
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame]852<p>Global variables can be marked with <tt>unnamed_addr</tt> which indicates
853 that the address is not significant, only the content. Constants marked
Rafael Espindolaf1ed7812011-01-15 08:20:57 +0000[diff] [blame]854 like this can be merged with other constants if they have the same
855 initializer. Note that a constant with significant address <em>can</em>
856 be merged with a <tt>unnamed_addr</tt> constant, the result being a
857 constant whose address is significant.</p>
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame]858
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]859<p>A global variable may be declared to reside in a target-specific numbered
860 address space. For targets that support them, address spaces may affect how
861 optimizations are performed and/or what target instructions are used to
862 access the variable. The default address space is zero. The address space
863 qualifier must precede any other attributes.</p>
Christopher Lamb308121c2007-12-11 09:31:00 +0000[diff] [blame]864
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]865<p>LLVM allows an explicit section to be specified for globals. If the target
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]866 supports it, it will emit globals to the section specified.</p>
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]867
Chris Lattner78e00bc2010-04-28 00:13:42 +0000[diff] [blame]868<p>An explicit alignment may be specified for a global, which must be a power
869 of 2. If not present, or if the alignment is set to zero, the alignment of
870 the global is set by the target to whatever it feels convenient. If an
871 explicit alignment is specified, the global is forced to have exactly that
Chris Lattner4bd85e42010-04-28 00:31:12 +0000[diff] [blame]872 alignment. Targets and optimizers are not allowed to over-align the global
873 if the global has an assigned section. In this case, the extra alignment
874 could be observable: for example, code could assume that the globals are
875 densely packed in their section and try to iterate over them as an array,
876 alignment padding would break this iteration.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]877
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]878<p>For example, the following defines a global in a numbered address space with
879 an initializer, section, and alignment:</p>
Chris Lattner5760c502007-01-14 00:27:09 +0000[diff] [blame]880
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]881<pre class="doc_code">
Dan Gohmanaaa679b2009-01-11 00:40:00 +0000[diff] [blame]882@G = addrspace(5) constant float 1.0, section "foo", align 4
Chris Lattner5760c502007-01-14 00:27:09 +0000[diff] [blame]883</pre>
884
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]885</div>
886
887
888<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]889<h3>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]890 <a name="functionstructure">Functions</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]891</h3>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]892
893<div class="doc_text">
894
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]895<p>LLVM function definitions consist of the "<tt>define</tt>" keyword, an
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]896 optional <a href="#linkage">linkage type</a>, an optional
897 <a href="#visibility">visibility style</a>, an optional
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame]898 <a href="#callingconv">calling convention</a>,
899 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]900 <a href="#paramattrs">parameter attribute</a> for the return type, a function
901 name, a (possibly empty) argument list (each with optional
902 <a href="#paramattrs">parameter attributes</a>), optional
903 <a href="#fnattrs">function attributes</a>, an optional section, an optional
904 alignment, an optional <a href="#gc">garbage collector name</a>, an opening
905 curly brace, a list of basic blocks, and a closing curly brace.</p>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000[diff] [blame]906
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]907<p>LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
908 optional <a href="#linkage">linkage type</a>, an optional
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]909 <a href="#visibility">visibility style</a>, an optional
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame]910 <a href="#callingconv">calling convention</a>,
911 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]912 <a href="#paramattrs">parameter attribute</a> for the return type, a function
913 name, a possibly empty list of arguments, an optional alignment, and an
914 optional <a href="#gc">garbage collector name</a>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]915
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]916<p>A function definition contains a list of basic blocks, forming the CFG
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]917 (Control Flow Graph) for the function. Each basic block may optionally start
918 with a label (giving the basic block a symbol table entry), contains a list
919 of instructions, and ends with a <a href="#terminators">terminator</a>
920 instruction (such as a branch or function return).</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]921
Chris Lattnera59fb102007-06-08 16:52:14 +0000[diff] [blame]922<p>The first basic block in a function is special in two ways: it is immediately
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]923 executed on entrance to the function, and it is not allowed to have
924 predecessor basic blocks (i.e. there can not be any branches to the entry
925 block of a function). Because the block can have no predecessors, it also
926 cannot have any <a href="#i_phi">PHI nodes</a>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]927
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]928<p>LLVM allows an explicit section to be specified for functions. If the target
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]929 supports it, it will emit functions to the section specified.</p>
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]930
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]931<p>An explicit alignment may be specified for a function. If not present, or if
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]932 the alignment is set to zero, the alignment of the function is set by the
933 target to whatever it feels convenient. If an explicit alignment is
934 specified, the function is forced to have at least that much alignment. All
935 alignments must be a power of 2.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]936
Rafael Espindola45e6c192011-01-08 16:42:36 +0000[diff] [blame]937<p>If the <tt>unnamed_addr</tt> attribute is given, the address is know to not
938 be significant and two identical functions can be merged</p>.
939
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]940<h5>Syntax:</h5>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]941<pre class="doc_code">
Chris Lattner0ae02092008-10-13 16:55:18 +0000[diff] [blame]942define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]943 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
944 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
945 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
946 [<a href="#gc">gc</a>] { ... }
947</pre>
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]948
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]949</div>
950
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]951<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]952<h3>
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]953 <a name="aliasstructure">Aliases</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]954</h3>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]955
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]956<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]957
958<p>Aliases act as "second name" for the aliasee value (which can be either
959 function, global variable, another alias or bitcast of global value). Aliases
960 may have an optional <a href="#linkage">linkage type</a>, and an
961 optional <a href="#visibility">visibility style</a>.</p>
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]962
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]963<h5>Syntax:</h5>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]964<pre class="doc_code">
Duncan Sands7e99a942008-09-12 20:48:21 +0000[diff] [blame]965@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Bill Wendling2d8b9a82007-05-29 09:42:13 +0000[diff] [blame]966</pre>
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000[diff] [blame]967
968</div>
969
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]970<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]971<h3>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]972 <a name="namedmetadatastructure">Named Metadata</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]973</h3>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]974
975<div class="doc_text">
976
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]977<p>Named metadata is a collection of metadata. <a href="#metadata">Metadata
Dan Gohman093cb792010-07-21 18:54:18 +0000[diff] [blame]978 nodes</a> (but not metadata strings) are the only valid operands for
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]979 a named metadata.</p>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]980
981<h5>Syntax:</h5>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]982<pre class="doc_code">
Dan Gohman093cb792010-07-21 18:54:18 +0000[diff] [blame]983; Some unnamed metadata nodes, which are referenced by the named metadata.
984!0 = metadata !{metadata !"zero"}
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]985!1 = metadata !{metadata !"one"}
Dan Gohman093cb792010-07-21 18:54:18 +0000[diff] [blame]986!2 = metadata !{metadata !"two"}
Dan Gohman58cd65f2010-07-13 19:48:13 +0000[diff] [blame]987; A named metadata.
Dan Gohman093cb792010-07-21 18:54:18 +0000[diff] [blame]988!name = !{!0, !1, !2}
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]989</pre>
Devang Pateld1a89692010-01-11 19:35:55 +0000[diff] [blame]990
991</div>
992
993<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]994<h3>
995 <a name="paramattrs">Parameter Attributes</a>
996</h3>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]997
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]998<div class="doc_text">
999
1000<p>The return type and each parameter of a function type may have a set of
1001 <i>parameter attributes</i> associated with them. Parameter attributes are
1002 used to communicate additional information about the result or parameters of
1003 a function. Parameter attributes are considered to be part of the function,
1004 not of the function type, so functions with different parameter attributes
1005 can have the same function type.</p>
1006
1007<p>Parameter attributes are simple keywords that follow the type specified. If
1008 multiple parameter attributes are needed, they are space separated. For
1009 example:</p>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]1010
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1011<pre class="doc_code">
Nick Lewyckydac78d82009-02-15 23:06:14 +0000[diff] [blame]1012declare i32 @printf(i8* noalias nocapture, ...)
Chris Lattnerd2597d72008-10-04 18:33:34 +0000[diff] [blame]1013declare i32 @atoi(i8 zeroext)
1014declare signext i8 @returns_signed_char()
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]1015</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]1016
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1017<p>Note that any attributes for the function result (<tt>nounwind</tt>,
1018 <tt>readonly</tt>) come immediately after the argument list.</p>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1019
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1020<p>Currently, only the following parameter attributes are defined:</p>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000[diff] [blame]1021
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1022<dl>
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1023 <dt><tt><b>zeroext</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1024 <dd>This indicates to the code generator that the parameter or return value
Cameron Zwarichac106272011-03-16 22:20:18 +0000[diff] [blame]1025 should be zero-extended to the extent required by the target's ABI (which
1026 is usually 32-bits, but is 8-bits for a i1 on x86-64) by the caller (for a
1027 parameter) or the callee (for a return value).</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000[diff] [blame]1028
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1029 <dt><tt><b>signext</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1030 <dd>This indicates to the code generator that the parameter or return value
Cameron Zwarich341c36d2011-03-17 14:21:58 +0000[diff] [blame]1031 should be sign-extended to the extent required by the target's ABI (which
1032 is usually 32-bits) by the caller (for a parameter) or the callee (for a
1033 return value).</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000[diff] [blame]1034
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1035 <dt><tt><b>inreg</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1036 <dd>This indicates that this parameter or return value should be treated in a
1037 special target-dependent fashion during while emitting code for a function
1038 call or return (usually, by putting it in a register as opposed to memory,
1039 though some targets use it to distinguish between two different kinds of
1040 registers). Use of this attribute is target-specific.</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000[diff] [blame]1041
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1042 <dt><tt><b><a name="byval">byval</a></b></tt></dt>
Chris Lattnerd78dbee2010-11-20 23:49:06 +0000[diff] [blame]1043 <dd><p>This indicates that the pointer parameter should really be passed by
1044 value to the function. The attribute implies that a hidden copy of the
1045 pointee
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1046 is made between the caller and the callee, so the callee is unable to
1047 modify the value in the callee. This attribute is only valid on LLVM
1048 pointer arguments. It is generally used to pass structs and arrays by
1049 value, but is also valid on pointers to scalars. The copy is considered
1050 to belong to the caller not the callee (for example,
1051 <tt><a href="#readonly">readonly</a></tt> functions should not write to
1052 <tt>byval</tt> parameters). This is not a valid attribute for return
Chris Lattnerd78dbee2010-11-20 23:49:06 +0000[diff] [blame]1053 values.</p>
1054
1055 <p>The byval attribute also supports specifying an alignment with
1056 the align attribute. It indicates the alignment of the stack slot to
1057 form and the known alignment of the pointer specified to the call site. If
1058 the alignment is not specified, then the code generator makes a
1059 target-specific assumption.</p></dd>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1060
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]1061 <dt><tt><b><a name="sret">sret</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1062 <dd>This indicates that the pointer parameter specifies the address of a
1063 structure that is the return value of the function in the source program.
1064 This pointer must be guaranteed by the caller to be valid: loads and
1065 stores to the structure may be assumed by the callee to not to trap. This
1066 may only be applied to the first parameter. This is not a valid attribute
1067 for return values. </dd>
1068
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]1069 <dt><tt><b><a name="noalias">noalias</a></b></tt></dt>
Dan Gohmandf12d082010-07-02 18:41:32 +0000[diff] [blame]1070 <dd>This indicates that pointer values
1071 <a href="#pointeraliasing"><i>based</i></a> on the argument or return
Dan Gohmande256292010-07-02 23:46:54 +0000[diff] [blame]1072 value do not alias pointer values which are not <i>based</i> on it,
1073 ignoring certain "irrelevant" dependencies.
1074 For a call to the parent function, dependencies between memory
1075 references from before or after the call and from those during the call
1076 are "irrelevant" to the <tt>noalias</tt> keyword for the arguments and
1077 return value used in that call.
Dan Gohmandf12d082010-07-02 18:41:32 +0000[diff] [blame]1078 The caller shares the responsibility with the callee for ensuring that
1079 these requirements are met.
1080 For further details, please see the discussion of the NoAlias response in
Dan Gohman6c858db2010-07-06 15:26:33 +0000[diff] [blame]1081 <a href="AliasAnalysis.html#MustMayNo">alias analysis</a>.<br>
1082<br>
John McCall72ed8902010-07-06 21:07:14 +0000[diff] [blame]1083 Note that this definition of <tt>noalias</tt> is intentionally
1084 similar to the definition of <tt>restrict</tt> in C99 for function
Chris Lattner5eff9ca2010-07-06 20:51:35 +0000[diff] [blame]1085 arguments, though it is slightly weaker.
Dan Gohman6c858db2010-07-06 15:26:33 +0000[diff] [blame]1086<br>
1087 For function return values, C99's <tt>restrict</tt> is not meaningful,
1088 while LLVM's <tt>noalias</tt> is.
1089 </dd>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1090
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]1091 <dt><tt><b><a name="nocapture">nocapture</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1092 <dd>This indicates that the callee does not make any copies of the pointer
1093 that outlive the callee itself. This is not a valid attribute for return
1094 values.</dd>
1095
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]1096 <dt><tt><b><a name="nest">nest</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1097 <dd>This indicates that the pointer parameter can be excised using the
1098 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
1099 attribute for return values.</dd>
1100</dl>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1101
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1102</div>
1103
1104<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1105<h3>
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1106 <a name="gc">Garbage Collector Names</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1107</h3>
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1108
1109<div class="doc_text">
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1110
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1111<p>Each function may specify a garbage collector name, which is simply a
1112 string:</p>
1113
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1114<pre class="doc_code">
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1115define void @f() gc "name" { ... }
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1116</pre>
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1117
1118<p>The compiler declares the supported values of <i>name</i>. Specifying a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1119 collector which will cause the compiler to alter its output in order to
1120 support the named garbage collection algorithm.</p>
1121
Gordon Henriksen71183b62007-12-10 03:18:06 +0000[diff] [blame]1122</div>
1123
1124<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1125<h3>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1126 <a name="fnattrs">Function Attributes</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1127</h3>
Devang Patelcaacdba2008-09-04 23:05:13 +0000[diff] [blame]1128
1129<div class="doc_text">
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1130
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1131<p>Function attributes are set to communicate additional information about a
1132 function. Function attributes are considered to be part of the function, not
1133 of the function type, so functions with different parameter attributes can
1134 have the same function type.</p>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1135
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1136<p>Function attributes are simple keywords that follow the type specified. If
1137 multiple attributes are needed, they are space separated. For example:</p>
Devang Patelcaacdba2008-09-04 23:05:13 +0000[diff] [blame]1138
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1139<pre class="doc_code">
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1140define void @f() noinline { ... }
1141define void @f() alwaysinline { ... }
1142define void @f() alwaysinline optsize { ... }
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1143define void @f() optsize { ... }
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1144</pre>
Devang Patelcaacdba2008-09-04 23:05:13 +0000[diff] [blame]1145
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1146<dl>
Charles Davisbe5557e2010-02-12 00:31:15 +0000[diff] [blame]1147 <dt><tt><b>alignstack(&lt;<em>n</em>&gt;)</b></tt></dt>
1148 <dd>This attribute indicates that, when emitting the prologue and epilogue,
1149 the backend should forcibly align the stack pointer. Specify the
1150 desired alignment, which must be a power of two, in parentheses.
1151
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1152 <dt><tt><b>alwaysinline</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1153 <dd>This attribute indicates that the inliner should attempt to inline this
1154 function into callers whenever possible, ignoring any active inlining size
1155 threshold for this caller.</dd>
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1156
Charles Davis22fe1862010-10-25 15:37:09 +0000[diff] [blame]1157 <dt><tt><b>hotpatch</b></tt></dt>
Charles Davis1b2d3722010-10-25 16:29:03 +0000[diff] [blame]1158 <dd>This attribute indicates that the function should be 'hotpatchable',
Charles Davis74205252010-10-25 19:07:39 +0000[diff] [blame]1159 meaning the function can be patched and/or hooked even while it is
1160 loaded into memory. On x86, the function prologue will be preceded
1161 by six bytes of padding and will begin with a two-byte instruction.
1162 Most of the functions in the Windows system DLLs in Windows XP SP2 or
1163 higher were compiled in this fashion.</dd>
Charles Davis22fe1862010-10-25 15:37:09 +0000[diff] [blame]1164
Jakob Stoklund Olesen74bb06c2010-02-06 01:16:28 +0000[diff] [blame]1165 <dt><tt><b>inlinehint</b></tt></dt>
1166 <dd>This attribute indicates that the source code contained a hint that inlining
1167 this function is desirable (such as the "inline" keyword in C/C++). It
1168 is just a hint; it imposes no requirements on the inliner.</dd>
1169
Nick Lewycky14b58da2010-07-06 18:24:09 +0000[diff] [blame]1170 <dt><tt><b>naked</b></tt></dt>
1171 <dd>This attribute disables prologue / epilogue emission for the function.
1172 This can have very system-specific consequences.</dd>
1173
1174 <dt><tt><b>noimplicitfloat</b></tt></dt>
1175 <dd>This attributes disables implicit floating point instructions.</dd>
1176
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1177 <dt><tt><b>noinline</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1178 <dd>This attribute indicates that the inliner should never inline this
1179 function in any situation. This attribute may not be used together with
1180 the <tt>alwaysinline</tt> attribute.</dd>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1181
Nick Lewycky14b58da2010-07-06 18:24:09 +0000[diff] [blame]1182 <dt><tt><b>noredzone</b></tt></dt>
1183 <dd>This attribute indicates that the code generator should not use a red
1184 zone, even if the target-specific ABI normally permits it.</dd>
Devang Patel9eb525d2008-09-26 23:51:19 +0000[diff] [blame]1185
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1186 <dt><tt><b>noreturn</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1187 <dd>This function attribute indicates that the function never returns
1188 normally. This produces undefined behavior at runtime if the function
1189 ever does dynamically return.</dd>
Bill Wendlinga8130172008-11-13 01:02:51 +0000[diff] [blame]1190
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1191 <dt><tt><b>nounwind</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1192 <dd>This function attribute indicates that the function never returns with an
1193 unwind or exceptional control flow. If the function does unwind, its
1194 runtime behavior is undefined.</dd>
Bill Wendling0f5541e2008-11-26 19:07:40 +0000[diff] [blame]1195
Nick Lewycky14b58da2010-07-06 18:24:09 +0000[diff] [blame]1196 <dt><tt><b>optsize</b></tt></dt>
1197 <dd>This attribute suggests that optimization passes and code generator passes
1198 make choices that keep the code size of this function low, and otherwise
1199 do optimizations specifically to reduce code size.</dd>
1200
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1201 <dt><tt><b>readnone</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1202 <dd>This attribute indicates that the function computes its result (or decides
1203 to unwind an exception) based strictly on its arguments, without
1204 dereferencing any pointer arguments or otherwise accessing any mutable
1205 state (e.g. memory, control registers, etc) visible to caller functions.
1206 It does not write through any pointer arguments
1207 (including <tt><a href="#byval">byval</a></tt> arguments) and never
1208 changes any state visible to callers. This means that it cannot unwind
1209 exceptions by calling the <tt>C++</tt> exception throwing methods, but
1210 could use the <tt>unwind</tt> instruction.</dd>
Devang Patel310fd4a2009-06-12 19:45:19 +0000[diff] [blame]1211
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1212 <dt><tt><b><a name="readonly">readonly</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1213 <dd>This attribute indicates that the function does not write through any
1214 pointer arguments (including <tt><a href="#byval">byval</a></tt>
1215 arguments) or otherwise modify any state (e.g. memory, control registers,
1216 etc) visible to caller functions. It may dereference pointer arguments
1217 and read state that may be set in the caller. A readonly function always
1218 returns the same value (or unwinds an exception identically) when called
1219 with the same set of arguments and global state. It cannot unwind an
1220 exception by calling the <tt>C++</tt> exception throwing methods, but may
1221 use the <tt>unwind</tt> instruction.</dd>
Anton Korobeynikovc8ce7b082009-07-17 18:07:26 +0000[diff] [blame]1222
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1223 <dt><tt><b><a name="ssp">ssp</a></b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1224 <dd>This attribute indicates that the function should emit a stack smashing
1225 protector. It is in the form of a "canary"&mdash;a random value placed on
1226 the stack before the local variables that's checked upon return from the
1227 function to see if it has been overwritten. A heuristic is used to
1228 determine if a function needs stack protectors or not.<br>
1229<br>
1230 If a function that has an <tt>ssp</tt> attribute is inlined into a
1231 function that doesn't have an <tt>ssp</tt> attribute, then the resulting
1232 function will have an <tt>ssp</tt> attribute.</dd>
1233
Bill Wendling7f4a3362009-11-02 00:24:16 +0000[diff] [blame]1234 <dt><tt><b>sspreq</b></tt></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1235 <dd>This attribute indicates that the function should <em>always</em> emit a
1236 stack smashing protector. This overrides
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1237 the <tt><a href="#ssp">ssp</a></tt> function attribute.<br>
1238<br>
1239 If a function that has an <tt>sspreq</tt> attribute is inlined into a
1240 function that doesn't have an <tt>sspreq</tt> attribute or which has
1241 an <tt>ssp</tt> attribute, then the resulting function will have
1242 an <tt>sspreq</tt> attribute.</dd>
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1243</dl>
1244
Devang Patelcaacdba2008-09-04 23:05:13 +0000[diff] [blame]1245</div>
1246
1247<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1248<h3>
Chris Lattner93564892006-04-08 04:40:53 +0000[diff] [blame]1249 <a name="moduleasm">Module-Level Inline Assembly</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1250</h3>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1251
1252<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1253
1254<p>Modules may contain "module-level inline asm" blocks, which corresponds to
1255 the GCC "file scope inline asm" blocks. These blocks are internally
1256 concatenated by LLVM and treated as a single unit, but may be separated in
1257 the <tt>.ll</tt> file if desired. The syntax is very simple:</p>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1258
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1259<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]1260module asm "inline asm code goes here"
1261module asm "more can go here"
1262</pre>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1263
1264<p>The strings can contain any character by escaping non-printable characters.
1265 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1266 for the number.</p>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1267
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1268<p>The inline asm code is simply printed to the machine code .s file when
1269 assembly code is generated.</p>
1270
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]1271</div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1272
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1273<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1274<h3>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1275 <a name="datalayout">Data Layout</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1276</h3>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1277
1278<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1279
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1280<p>A module may specify a target specific data layout string that specifies how
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1281 data is to be laid out in memory. The syntax for the data layout is
1282 simply:</p>
1283
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]1284<pre class="doc_code">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1285target datalayout = "<i>layout specification</i>"
1286</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1287
1288<p>The <i>layout specification</i> consists of a list of specifications
1289 separated by the minus sign character ('-'). Each specification starts with
1290 a letter and may include other information after the letter to define some
1291 aspect of the data layout. The specifications accepted are as follows:</p>
1292
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1293<dl>
1294 <dt><tt>E</tt></dt>
1295 <dd>Specifies that the target lays out data in big-endian form. That is, the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1296 bits with the most significance have the lowest address location.</dd>
1297
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1298 <dt><tt>e</tt></dt>
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]1299 <dd>Specifies that the target lays out data in little-endian form. That is,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1300 the bits with the least significance have the lowest address
1301 location.</dd>
1302
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1303 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1304 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1305 <i>preferred</i> alignments. All sizes are in bits. Specifying
1306 the <i>pref</i> alignment is optional. If omitted, the
1307 preceding <tt>:</tt> should be omitted too.</dd>
1308
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1309 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1310 <dd>This specifies the alignment for an integer type of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1311 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1312
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1313 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1314 <dd>This specifies the alignment for a vector type of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1315 <i>size</i>.</dd>
1316
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1317 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1318 <dd>This specifies the alignment for a floating point type of a given bit
Dale Johannesence522852010-05-28 18:54:47 +0000[diff] [blame]1319 <i>size</i>. Only values of <i>size</i> that are supported by the target
1320 will work. 32 (float) and 64 (double) are supported on all targets;
1321 80 or 128 (different flavors of long double) are also supported on some
1322 targets.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1323
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1324 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1325 <dd>This specifies the alignment for an aggregate type of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1326 <i>size</i>.</dd>
1327
Daniel Dunbar7921a592009-06-08 22:17:53 +0000[diff] [blame]1328 <dt><tt>s<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1329 <dd>This specifies the alignment for a stack object of a given bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1330 <i>size</i>.</dd>
Chris Lattnera381eff2009-11-07 09:35:34 +0000[diff] [blame]1331
1332 <dt><tt>n<i>size1</i>:<i>size2</i>:<i>size3</i>...</tt></dt>
1333 <dd>This specifies a set of native integer widths for the target CPU
1334 in bits. For example, it might contain "n32" for 32-bit PowerPC,
1335 "n32:64" for PowerPC 64, or "n8:16:32:64" for X86-64. Elements of
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1336 this set are considered to support most general arithmetic
Chris Lattnera381eff2009-11-07 09:35:34 +0000[diff] [blame]1337 operations efficiently.</dd>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1338</dl>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1339
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1340<p>When constructing the data layout for a given target, LLVM starts with a
Dan Gohman61110ae2010-04-28 00:36:01 +0000[diff] [blame]1341 default set of specifications which are then (possibly) overridden by the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1342 specifications in the <tt>datalayout</tt> keyword. The default specifications
1343 are given in this list:</p>
1344
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1345<ul>
1346 <li><tt>E</tt> - big endian</li>
Dan Gohman8ad777d2010-02-23 02:44:03 +0000[diff] [blame]1347 <li><tt>p:64:64:64</tt> - 64-bit pointers with 64-bit alignment</li>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1348 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1349 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1350 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1351 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]1352 <li><tt>i64:32:64</tt> - i64 has ABI alignment of 32-bits but preferred
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1353 alignment of 64-bits</li>
1354 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1355 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1356 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1357 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1358 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
Daniel Dunbar7921a592009-06-08 22:17:53 +0000[diff] [blame]1359 <li><tt>s0:64:64</tt> - stack objects are 64-bit aligned</li>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1360</ul>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1361
1362<p>When LLVM is determining the alignment for a given type, it uses the
1363 following rules:</p>
1364
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1365<ol>
1366 <li>If the type sought is an exact match for one of the specifications, that
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1367 specification is used.</li>
1368
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1369 <li>If no match is found, and the type sought is an integer type, then the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1370 smallest integer type that is larger than the bitwidth of the sought type
1371 is used. If none of the specifications are larger than the bitwidth then
1372 the the largest integer type is used. For example, given the default
1373 specifications above, the i7 type will use the alignment of i8 (next
1374 largest) while both i65 and i256 will use the alignment of i64 (largest
1375 specified).</li>
1376
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1377 <li>If no match is found, and the type sought is a vector type, then the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1378 largest vector type that is smaller than the sought vector type will be
1379 used as a fall back. This happens because &lt;128 x double&gt; can be
1380 implemented in terms of 64 &lt;2 x double&gt;, for example.</li>
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1381</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1382
Reid Spencer50c723a2007-02-19 23:54:10 +0000[diff] [blame]1383</div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1384
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1385<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1386<h3>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1387 <a name="pointeraliasing">Pointer Aliasing Rules</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1388</h3>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1389
1390<div class="doc_text">
1391
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1392<p>Any memory access must be done through a pointer value associated
Andreas Bolkae39f0332009-07-27 20:37:10 +0000[diff] [blame]1393with an address range of the memory access, otherwise the behavior
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1394is undefined. Pointer values are associated with address ranges
1395according to the following rules:</p>
1396
1397<ul>
Dan Gohmandf12d082010-07-02 18:41:32 +0000[diff] [blame]1398 <li>A pointer value is associated with the addresses associated with
1399 any value it is <i>based</i> on.
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1400 <li>An address of a global variable is associated with the address
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1401 range of the variable's storage.</li>
1402 <li>The result value of an allocation instruction is associated with
1403 the address range of the allocated storage.</li>
1404 <li>A null pointer in the default address-space is associated with
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1405 no address.</li>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1406 <li>An integer constant other than zero or a pointer value returned
1407 from a function not defined within LLVM may be associated with address
1408 ranges allocated through mechanisms other than those provided by
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1409 LLVM. Such ranges shall not overlap with any ranges of addresses
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1410 allocated by mechanisms provided by LLVM.</li>
Dan Gohmandf12d082010-07-02 18:41:32 +0000[diff] [blame]1411</ul>
1412
1413<p>A pointer value is <i>based</i> on another pointer value according
1414 to the following rules:</p>
1415
1416<ul>
1417 <li>A pointer value formed from a
1418 <tt><a href="#i_getelementptr">getelementptr</a></tt> operation
1419 is <i>based</i> on the first operand of the <tt>getelementptr</tt>.</li>
1420 <li>The result value of a
1421 <tt><a href="#i_bitcast">bitcast</a></tt> is <i>based</i> on the operand
1422 of the <tt>bitcast</tt>.</li>
1423 <li>A pointer value formed by an
1424 <tt><a href="#i_inttoptr">inttoptr</a></tt> is <i>based</i> on all
1425 pointer values that contribute (directly or indirectly) to the
1426 computation of the pointer's value.</li>
1427 <li>The "<i>based</i> on" relationship is transitive.</li>
1428</ul>
1429
1430<p>Note that this definition of <i>"based"</i> is intentionally
1431 similar to the definition of <i>"based"</i> in C99, though it is
1432 slightly weaker.</p>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1433
1434<p>LLVM IR does not associate types with memory. The result type of a
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1435<tt><a href="#i_load">load</a></tt> merely indicates the size and
1436alignment of the memory from which to load, as well as the
Dan Gohman4eb47192010-06-17 19:23:50 +0000[diff] [blame]1437interpretation of the value. The first operand type of a
Andreas Bolka8ae4e242009-07-29 00:02:05 +0000[diff] [blame]1438<tt><a href="#i_store">store</a></tt> similarly only indicates the size
1439and alignment of the store.</p>
Dan Gohman6154a012009-07-27 18:07:55 +0000[diff] [blame]1440
1441<p>Consequently, type-based alias analysis, aka TBAA, aka
1442<tt>-fstrict-aliasing</tt>, is not applicable to general unadorned
1443LLVM IR. <a href="#metadata">Metadata</a> may be used to encode
1444additional information which specialized optimization passes may use
1445to implement type-based alias analysis.</p>
1446
1447</div>
1448
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]1449<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1450<h3>
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]1451 <a name="volatile">Volatile Memory Accesses</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1452</h3>
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]1453
1454<div class="doc_text">
1455
1456<p>Certain memory accesses, such as <a href="#i_load"><tt>load</tt></a>s, <a
1457href="#i_store"><tt>store</tt></a>s, and <a
1458href="#int_memcpy"><tt>llvm.memcpy</tt></a>s may be marked <tt>volatile</tt>.
1459The optimizers must not change the number of volatile operations or change their
1460order of execution relative to other volatile operations. The optimizers
1461<i>may</i> change the order of volatile operations relative to non-volatile
1462operations. This is not Java's "volatile" and has no cross-thread
1463synchronization behavior.</p>
1464
1465</div>
1466
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1467<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1468<h2><a name="typesystem">Type System</a></h2>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1469<!-- *********************************************************************** -->
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1470
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1471<div class="doc_text">
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1472
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1473<p>The LLVM type system is one of the most important features of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1474 intermediate representation. Being typed enables a number of optimizations
1475 to be performed on the intermediate representation directly, without having
1476 to do extra analyses on the side before the transformation. A strong type
1477 system makes it easier to read the generated code and enables novel analyses
1478 and transformations that are not feasible to perform on normal three address
1479 code representations.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]1480
1481</div>
1482
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1483<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1484<h3>
1485 <a name="t_classifications">Type Classifications</a>
1486</h3>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1487
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1488<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1489
1490<p>The types fall into a few useful classifications:</p>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]1491
1492<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1493 <tbody>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1494 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1495 <tr>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1496 <td><a href="#t_integer">integer</a></td>
Reid Spencer138249b2007-05-16 18:44:01 +0000[diff] [blame]1497 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1498 </tr>
1499 <tr>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1500 <td><a href="#t_floating">floating point</a></td>
1501 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1502 </tr>
1503 <tr>
1504 <td><a name="t_firstclass">first class</a></td>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1505 <td><a href="#t_integer">integer</a>,
1506 <a href="#t_floating">floating point</a>,
1507 <a href="#t_pointer">pointer</a>,
Dan Gohman08783a882008-06-18 18:42:13 +0000[diff] [blame]1508 <a href="#t_vector">vector</a>,
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]1509 <a href="#t_struct">structure</a>,
1510 <a href="#t_array">array</a>,
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1511 <a href="#t_label">label</a>,
1512 <a href="#t_metadata">metadata</a>.
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1513 </td>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1514 </tr>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1515 <tr>
1516 <td><a href="#t_primitive">primitive</a></td>
1517 <td><a href="#t_label">label</a>,
1518 <a href="#t_void">void</a>,
Tobias Grosser4c8c95b2010-12-28 20:29:31 +0000[diff] [blame]1519 <a href="#t_integer">integer</a>,
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1520 <a href="#t_floating">floating point</a>,
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]1521 <a href="#t_x86mmx">x86mmx</a>,
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1522 <a href="#t_metadata">metadata</a>.</td>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1523 </tr>
1524 <tr>
1525 <td><a href="#t_derived">derived</a></td>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1526 <td><a href="#t_array">array</a>,
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1527 <a href="#t_function">function</a>,
1528 <a href="#t_pointer">pointer</a>,
1529 <a href="#t_struct">structure</a>,
1530 <a href="#t_pstruct">packed structure</a>,
1531 <a href="#t_vector">vector</a>,
1532 <a href="#t_opaque">opaque</a>.
Dan Gohman93bf60d2008-10-14 16:32:04 +0000[diff] [blame]1533 </td>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1534 </tr>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1535 </tbody>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1536</table>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]1537
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1538<p>The <a href="#t_firstclass">first class</a> types are perhaps the most
1539 important. Values of these types are the only ones which can be produced by
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1540 instructions.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1541
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1542</div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1543
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1544<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1545<h3>
1546 <a name="t_primitive">Primitive Types</a>
1547</h3>
Chris Lattner43542b32008-01-04 04:34:14 +0000[diff] [blame]1548
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1549<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1550
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1551<p>The primitive types are the fundamental building blocks of the LLVM
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1552 system.</p>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1553
Chris Lattner43542b32008-01-04 04:34:14 +0000[diff] [blame]1554</div>
1555
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1556<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1557<h4>
1558 <a name="t_integer">Integer Type</a>
1559</h4>
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1560
1561<div class="doc_text">
1562
1563<h5>Overview:</h5>
1564<p>The integer type is a very simple type that simply specifies an arbitrary
1565 bit width for the integer type desired. Any bit width from 1 bit to
1566 2<sup>23</sup>-1 (about 8 million) can be specified.</p>
1567
1568<h5>Syntax:</h5>
1569<pre>
1570 iN
1571</pre>
1572
1573<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1574 value.</p>
1575
1576<h5>Examples:</h5>
1577<table class="layout">
1578 <tr class="layout">
1579 <td class="left"><tt>i1</tt></td>
1580 <td class="left">a single-bit integer.</td>
1581 </tr>
1582 <tr class="layout">
1583 <td class="left"><tt>i32</tt></td>
1584 <td class="left">a 32-bit integer.</td>
1585 </tr>
1586 <tr class="layout">
1587 <td class="left"><tt>i1942652</tt></td>
1588 <td class="left">a really big integer of over 1 million bits.</td>
1589 </tr>
1590</table>
1591
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1592</div>
1593
1594<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1595<h4>
1596 <a name="t_floating">Floating Point Types</a>
1597</h4>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1598
1599<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1600
1601<table>
1602 <tbody>
1603 <tr><th>Type</th><th>Description</th></tr>
1604 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1605 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1606 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1607 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1608 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1609 </tbody>
1610</table>
1611
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1612</div>
1613
1614<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1615<h4>
1616 <a name="t_x86mmx">X86mmx Type</a>
1617</h4>
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]1618
1619<div class="doc_text">
1620
1621<h5>Overview:</h5>
1622<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>
1623
1624<h5>Syntax:</h5>
1625<pre>
Dale Johannesenb1f0ff12010-10-01 01:07:02 +0000[diff] [blame]1626 x86mmx
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]1627</pre>
1628
1629</div>
1630
1631<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1632<h4>
1633 <a name="t_void">Void Type</a>
1634</h4>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1635
1636<div class="doc_text">
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1637
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1638<h5>Overview:</h5>
1639<p>The void type does not represent any value and has no size.</p>
1640
1641<h5>Syntax:</h5>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1642<pre>
1643 void
1644</pre>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1645
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1646</div>
1647
1648<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1649<h4>
1650 <a name="t_label">Label Type</a>
1651</h4>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1652
1653<div class="doc_text">
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1654
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1655<h5>Overview:</h5>
1656<p>The label type represents code labels.</p>
1657
1658<h5>Syntax:</h5>
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1659<pre>
1660 label
1661</pre>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1662
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1663</div>
1664
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1665<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1666<h4>
1667 <a name="t_metadata">Metadata Type</a>
1668</h4>
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1669
1670<div class="doc_text">
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1671
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1672<h5>Overview:</h5>
Nick Lewycky93e06a52009-09-27 23:27:42 +0000[diff] [blame]1673<p>The metadata type represents embedded metadata. No derived types may be
1674 created from metadata except for <a href="#t_function">function</a>
1675 arguments.
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1676
1677<h5>Syntax:</h5>
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1678<pre>
1679 metadata
1680</pre>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1681
Nick Lewyckyadbc2842009-05-30 05:06:04 +0000[diff] [blame]1682</div>
1683
Chris Lattner7824d182008-01-04 04:32:38 +0000[diff] [blame]1684
1685<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1686<h3>
1687 <a name="t_derived">Derived Types</a>
1688</h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1689
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1690<div class="doc_text">
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1691
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1692<p>The real power in LLVM comes from the derived types in the system. This is
1693 what allows a programmer to represent arrays, functions, pointers, and other
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]1694 useful types. Each of these types contain one or more element types which
1695 may be a primitive type, or another derived type. For example, it is
1696 possible to have a two dimensional array, using an array as the element type
1697 of another array.</p>
Dan Gohman142ccc02009-01-24 15:58:40 +0000[diff] [blame]1698
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1699
1700</div>
1701
1702<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1703<h4>
1704 <a name="t_aggregate">Aggregate Types</a>
1705</h4>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1706
1707<div class="doc_text">
1708
1709<p>Aggregate Types are a subset of derived types that can contain multiple
1710 member types. <a href="#t_array">Arrays</a>,
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]1711 <a href="#t_struct">structs</a>, and <a href="#t_vector">vectors</a> are
1712 aggregate types.</p>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]1713
1714</div>
1715
Reid Spencer138249b2007-05-16 18:44:01 +0000[diff] [blame]1716<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1717<h4>
1718 <a name="t_array">Array Type</a>
1719</h4>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1720
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1721<div class="doc_text">
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1722
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1723<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1724<p>The array type is a very simple derived type that arranges elements
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1725 sequentially in memory. The array type requires a size (number of elements)
1726 and an underlying data type.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1727
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]1728<h5>Syntax:</h5>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1729<pre>
1730 [&lt;# elements&gt; x &lt;elementtype&gt;]
1731</pre>
1732
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1733<p>The number of elements is a constant integer value; <tt>elementtype</tt> may
1734 be any type with a size.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1735
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]1736<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1737<table class="layout">
1738 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1739 <td class="left"><tt>[40 x i32]</tt></td>
1740 <td class="left">Array of 40 32-bit integer values.</td>
1741 </tr>
1742 <tr class="layout">
1743 <td class="left"><tt>[41 x i32]</tt></td>
1744 <td class="left">Array of 41 32-bit integer values.</td>
1745 </tr>
1746 <tr class="layout">
1747 <td class="left"><tt>[4 x i8]</tt></td>
1748 <td class="left">Array of 4 8-bit integer values.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1749 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1750</table>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1751<p>Here are some examples of multidimensional arrays:</p>
1752<table class="layout">
1753 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1754 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1755 <td class="left">3x4 array of 32-bit integer values.</td>
1756 </tr>
1757 <tr class="layout">
1758 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1759 <td class="left">12x10 array of single precision floating point values.</td>
1760 </tr>
1761 <tr class="layout">
1762 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1763 <td class="left">2x3x4 array of 16-bit integer values.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1764 </tr>
1765</table>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +0000[diff] [blame]1766
Dan Gohmanc74bc282009-11-09 19:01:53 +0000[diff] [blame]1767<p>There is no restriction on indexing beyond the end of the array implied by
1768 a static type (though there are restrictions on indexing beyond the bounds
1769 of an allocated object in some cases). This means that single-dimension
1770 'variable sized array' addressing can be implemented in LLVM with a zero
1771 length array type. An implementation of 'pascal style arrays' in LLVM could
1772 use the type "<tt>{ i32, [0 x float]}</tt>", for example.</p>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +0000[diff] [blame]1773
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1774</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1775
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1776<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1777<h4>
1778 <a name="t_function">Function Type</a>
1779</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1780
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1781<div class="doc_text">
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]1782
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1783<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1784<p>The function type can be thought of as a function signature. It consists of
1785 a return type and a list of formal parameter types. The return type of a
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]1786 function type is a first class type or a void type.</p>
Devang Pateld6cff512008-03-10 20:49:15 +0000[diff] [blame]1787
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1788<h5>Syntax:</h5>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]1789<pre>
Nick Lewycky14d1ccc2009-09-27 07:55:32 +0000[diff] [blame]1790 &lt;returntype&gt; (&lt;parameter list&gt;)
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]1791</pre>
1792
John Criswell4c0cf7f2005-10-24 16:17:18 +0000[diff] [blame]1793<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1794 specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1795 which indicates that the function takes a variable number of arguments.
1796 Variable argument functions can access their arguments with
1797 the <a href="#int_varargs">variable argument handling intrinsic</a>
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]1798 functions. '<tt>&lt;returntype&gt;</tt>' is any type except
Nick Lewycky93e06a52009-09-27 23:27:42 +0000[diff] [blame]1799 <a href="#t_label">label</a>.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]1800
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1801<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1802<table class="layout">
1803 <tr class="layout">
Reid Spencer58c08712006-12-31 07:18:34 +0000[diff] [blame]1804 <td class="left"><tt>i32 (i32)</tt></td>
1805 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1806 </td>
Reid Spencer58c08712006-12-31 07:18:34 +0000[diff] [blame]1807 </tr><tr class="layout">
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]1808 <td class="left"><tt>float&nbsp;(i16,&nbsp;i32&nbsp;*)&nbsp;*
Reid Spencer655dcc62006-12-31 07:20:23 +0000[diff] [blame]1809 </tt></td>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1810 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]1811 an <tt>i16</tt> and a <a href="#t_pointer">pointer</a> to <tt>i32</tt>,
1812 returning <tt>float</tt>.
Reid Spencer58c08712006-12-31 07:18:34 +0000[diff] [blame]1813 </td>
1814 </tr><tr class="layout">
1815 <td class="left"><tt>i32 (i8*, ...)</tt></td>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]1816 <td class="left">A vararg function that takes at least one
1817 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1818 which returns an integer. This is the signature for <tt>printf</tt> in
Reid Spencer58c08712006-12-31 07:18:34 +0000[diff] [blame]1819 LLVM.
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1820 </td>
Devang Patele3dfc1c2008-03-24 05:35:41 +0000[diff] [blame]1821 </tr><tr class="layout">
1822 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Nick Lewycky14d1ccc2009-09-27 07:55:32 +0000[diff] [blame]1823 <td class="left">A function taking an <tt>i32</tt>, returning a
1824 <a href="#t_struct">structure</a> containing two <tt>i32</tt> values
Devang Patele3dfc1c2008-03-24 05:35:41 +0000[diff] [blame]1825 </td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1826 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1827</table>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]1828
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1829</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1830
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1831<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1832<h4>
1833 <a name="t_struct">Structure Type</a>
1834</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1835
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1836<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1837
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1838<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1839<p>The structure type is used to represent a collection of data members together
1840 in memory. The packing of the field types is defined to match the ABI of the
1841 underlying processor. The elements of a structure may be any type that has a
1842 size.</p>
1843
Jeffrey Yasskinf991bbb2010-01-11 19:19:26 +0000[diff] [blame]1844<p>Structures in memory are accessed using '<tt><a href="#i_load">load</a></tt>'
1845 and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field
1846 with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
1847 Structures in registers are accessed using the
1848 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' and
1849 '<tt><a href="#i_insertvalue">insertvalue</a></tt>' instructions.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1850<h5>Syntax:</h5>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1851<pre>
1852 { &lt;type list&gt; }
1853</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1854
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1855<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1856<table class="layout">
1857 <tr class="layout">
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]1858 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1859 <td class="left">A triple of three <tt>i32</tt> values</td>
1860 </tr><tr class="layout">
1861 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1862 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1863 second element is a <a href="#t_pointer">pointer</a> to a
1864 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1865 an <tt>i32</tt>.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1866 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1867</table>
Dan Gohman142ccc02009-01-24 15:58:40 +0000[diff] [blame]1868
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1869</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1870
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1871<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1872<h4>
1873 <a name="t_pstruct">Packed Structure Type</a>
1874</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1875
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]1876<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1877
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]1878<h5>Overview:</h5>
1879<p>The packed structure type is used to represent a collection of data members
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1880 together in memory. There is no padding between fields. Further, the
1881 alignment of a packed structure is 1 byte. The elements of a packed
1882 structure may be any type that has a size.</p>
1883
1884<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt> and
1885 '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field with
1886 the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
1887
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]1888<h5>Syntax:</h5>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1889<pre>
1890 &lt; { &lt;type list&gt; } &gt;
1891</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1892
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]1893<h5>Examples:</h5>
1894<table class="layout">
1895 <tr class="layout">
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]1896 <td class="left"><tt>&lt; { i32, i32, i32 } &gt;</tt></td>
1897 <td class="left">A triple of three <tt>i32</tt> values</td>
1898 </tr><tr class="layout">
Bill Wendlingb175fa42008-09-07 10:26:33 +0000[diff] [blame]1899 <td class="left">
1900<tt>&lt;&nbsp;{&nbsp;float,&nbsp;i32&nbsp;(i32)*&nbsp;}&nbsp;&gt;</tt></td>
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]1901 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1902 second element is a <a href="#t_pointer">pointer</a> to a
1903 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1904 an <tt>i32</tt>.</td>
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]1905 </tr>
1906</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1907
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame]1908</div>
1909
1910<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1911<h4>
1912 <a name="t_pointer">Pointer Type</a>
1913</h4>
Chris Lattner4a67c912009-02-08 19:53:29 +0000[diff] [blame]1914
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1915<div class="doc_text">
1916
1917<h5>Overview:</h5>
Dan Gohman88481112010-02-25 16:50:07 +0000[diff] [blame]1918<p>The pointer type is used to specify memory locations.
1919 Pointers are commonly used to reference objects in memory.</p>
1920
1921<p>Pointer types may have an optional address space attribute defining the
1922 numbered address space where the pointed-to object resides. The default
1923 address space is number zero. The semantics of non-zero address
1924 spaces are target-specific.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1925
1926<p>Note that LLVM does not permit pointers to void (<tt>void*</tt>) nor does it
1927 permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</tt> instead.</p>
Chris Lattner4a67c912009-02-08 19:53:29 +0000[diff] [blame]1928
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]1929<h5>Syntax:</h5>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]1930<pre>
1931 &lt;type&gt; *
1932</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1933
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]1934<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1935<table class="layout">
1936 <tr class="layout">
Dan Gohman623806e2009-01-04 23:44:43 +0000[diff] [blame]1937 <td class="left"><tt>[4 x i32]*</tt></td>
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1938 <td class="left">A <a href="#t_pointer">pointer</a> to <a
1939 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
1940 </tr>
1941 <tr class="layout">
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]1942 <td class="left"><tt>i32 (i32*) *</tt></td>
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1943 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]1944 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1945 <tt>i32</tt>.</td>
1946 </tr>
1947 <tr class="layout">
1948 <td class="left"><tt>i32 addrspace(5)*</tt></td>
1949 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
1950 that resides in address space #5.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1951 </tr>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1952</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1953
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1954</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1955
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]1956<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1957<h4>
1958 <a name="t_vector">Vector Type</a>
1959</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1960
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1961<div class="doc_text">
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1962
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]1963<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1964<p>A vector type is a simple derived type that represents a vector of elements.
1965 Vector types are used when multiple primitive data are operated in parallel
1966 using a single instruction (SIMD). A vector type requires a size (number of
Duncan Sands31c0e0e2009-11-27 13:38:03 +0000[diff] [blame]1967 elements) and an underlying primitive data type. Vector types are considered
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]1968 <a href="#t_firstclass">first class</a>.</p>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1969
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]1970<h5>Syntax:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1971<pre>
1972 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1973</pre>
1974
Chris Lattnerf11031a2010-10-10 18:20:35 +0000[diff] [blame]1975<p>The number of elements is a constant integer value larger than 0; elementtype
1976 may be any integer or floating point type. Vectors of size zero are not
1977 allowed, and pointers are not allowed as the element type.</p>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1978
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]1979<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1980<table class="layout">
1981 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]1982 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
1983 <td class="left">Vector of 4 32-bit integer values.</td>
1984 </tr>
1985 <tr class="layout">
1986 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
1987 <td class="left">Vector of 8 32-bit floating-point values.</td>
1988 </tr>
1989 <tr class="layout">
1990 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
1991 <td class="left">Vector of 2 64-bit integer values.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1992 </tr>
1993</table>
Dan Gohman142ccc02009-01-24 15:58:40 +0000[diff] [blame]1994
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1995</div>
1996
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1997<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]1998<h4>
1999 <a name="t_opaque">Opaque Type</a>
2000</h4>
2001
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]2002<div class="doc_text">
2003
2004<h5>Overview:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]2005<p>Opaque types are used to represent unknown types in the system. This
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2006 corresponds (for example) to the C notion of a forward declared structure
2007 type. In LLVM, opaque types can eventually be resolved to any type (not just
2008 a structure type).</p>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]2009
2010<h5>Syntax:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]2011<pre>
2012 opaque
2013</pre>
2014
2015<h5>Examples:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]2016<table class="layout">
2017 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +0000[diff] [blame]2018 <td class="left"><tt>opaque</tt></td>
2019 <td class="left">An opaque type.</td>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]2020 </tr>
2021</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2022
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]2023</div>
2024
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]2025<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2026<h3>
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]2027 <a name="t_uprefs">Type Up-references</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2028</h3>
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]2029
2030<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2031
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]2032<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2033<p>An "up reference" allows you to refer to a lexically enclosing type without
2034 requiring it to have a name. For instance, a structure declaration may
2035 contain a pointer to any of the types it is lexically a member of. Example
2036 of up references (with their equivalent as named type declarations)
2037 include:</p>
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]2038
2039<pre>
Chris Lattnerbf1d5452009-02-09 10:00:56 +0000[diff] [blame]2040 { \2 * } %x = type { %x* }
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]2041 { \2 }* %y = type { %y }*
2042 \1* %z = type %z*
2043</pre>
2044
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2045<p>An up reference is needed by the asmprinter for printing out cyclic types
2046 when there is no declared name for a type in the cycle. Because the
2047 asmprinter does not want to print out an infinite type string, it needs a
2048 syntax to handle recursive types that have no names (all names are optional
2049 in llvm IR).</p>
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]2050
2051<h5>Syntax:</h5>
2052<pre>
2053 \&lt;level&gt;
2054</pre>
2055
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2056<p>The level is the count of the lexical type that is being referred to.</p>
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]2057
2058<h5>Examples:</h5>
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]2059<table class="layout">
2060 <tr class="layout">
2061 <td class="left"><tt>\1*</tt></td>
2062 <td class="left">Self-referential pointer.</td>
2063 </tr>
2064 <tr class="layout">
2065 <td class="left"><tt>{ { \3*, i8 }, i32 }</tt></td>
2066 <td class="left">Recursive structure where the upref refers to the out-most
2067 structure.</td>
2068 </tr>
2069</table>
Chris Lattnercf7a5842009-02-02 07:32:36 +0000[diff] [blame]2070
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2071</div>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]2072
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2073<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2074<h2><a name="constants">Constants</a></h2>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2075<!-- *********************************************************************** -->
2076
2077<div class="doc_text">
2078
2079<p>LLVM has several different basic types of constants. This section describes
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2080 them all and their syntax.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2081
2082</div>
2083
2084<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2085<h3>
2086 <a name="simpleconstants">Simple Constants</a>
2087</h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2088
2089<div class="doc_text">
2090
2091<dl>
2092 <dt><b>Boolean constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2093 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]2094 constants of the <tt><a href="#t_integer">i1</a></tt> type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2095
2096 <dt><b>Integer constants</b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2097 <dd>Standard integers (such as '4') are constants of
2098 the <a href="#t_integer">integer</a> type. Negative numbers may be used
2099 with integer types.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2100
2101 <dt><b>Floating point constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2102 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2103 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
2104 notation (see below). The assembler requires the exact decimal value of a
2105 floating-point constant. For example, the assembler accepts 1.25 but
2106 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
2107 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2108
2109 <dt><b>Null pointer constants</b></dt>
John Criswelldfe6a862004-12-10 15:51:16 +0000[diff] [blame]2110 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2111 and must be of <a href="#t_pointer">pointer type</a>.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2112</dl>
2113
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2114<p>The one non-intuitive notation for constants is the hexadecimal form of
2115 floating point constants. For example, the form '<tt>double
2116 0x432ff973cafa8000</tt>' is equivalent to (but harder to read than)
2117 '<tt>double 4.5e+15</tt>'. The only time hexadecimal floating point
2118 constants are required (and the only time that they are generated by the
2119 disassembler) is when a floating point constant must be emitted but it cannot
2120 be represented as a decimal floating point number in a reasonable number of
2121 digits. For example, NaN's, infinities, and other special values are
2122 represented in their IEEE hexadecimal format so that assembly and disassembly
2123 do not cause any bits to change in the constants.</p>
2124
Dale Johannesencd4a3012009-02-11 22:14:51 +0000[diff] [blame]2125<p>When using the hexadecimal form, constants of types float and double are
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2126 represented using the 16-digit form shown above (which matches the IEEE754
2127 representation for double); float values must, however, be exactly
2128 representable as IEE754 single precision. Hexadecimal format is always used
2129 for long double, and there are three forms of long double. The 80-bit format
2130 used by x86 is represented as <tt>0xK</tt> followed by 20 hexadecimal digits.
2131 The 128-bit format used by PowerPC (two adjacent doubles) is represented
2132 by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit format
2133 is represented by <tt>0xL</tt> followed by 32 hexadecimal digits; no
2134 currently supported target uses this format. Long doubles will only work if
2135 they match the long double format on your target. All hexadecimal formats
2136 are big-endian (sign bit at the left).</p>
2137
Dale Johannesen33e5c352010-10-01 00:48:59 +0000[diff] [blame]2138<p>There are no constants of type x86mmx.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2139</div>
2140
2141<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2142<h3>
Bill Wendling972b7202009-07-20 02:32:41 +0000[diff] [blame]2143<a name="aggregateconstants"></a> <!-- old anchor -->
2144<a name="complexconstants">Complex Constants</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2145</h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2146
2147<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2148
Chris Lattner361bfcd2009-02-28 18:32:25 +0000[diff] [blame]2149<p>Complex constants are a (potentially recursive) combination of simple
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2150 constants and smaller complex constants.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2151
2152<dl>
2153 <dt><b>Structure constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2154 <dd>Structure constants are represented with notation similar to structure
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2155 type definitions (a comma separated list of elements, surrounded by braces
2156 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
2157 where "<tt>@G</tt>" is declared as "<tt>@G = external global i32</tt>".
2158 Structure constants must have <a href="#t_struct">structure type</a>, and
2159 the number and types of elements must match those specified by the
2160 type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2161
2162 <dt><b>Array constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2163 <dd>Array constants are represented with notation similar to array type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2164 definitions (a comma separated list of elements, surrounded by square
2165 brackets (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74
2166 ]</tt>". Array constants must have <a href="#t_array">array type</a>, and
2167 the number and types of elements must match those specified by the
2168 type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2169
Reid Spencer404a3252007-02-15 03:07:05 +0000[diff] [blame]2170 <dt><b>Vector constants</b></dt>
Reid Spencer404a3252007-02-15 03:07:05 +0000[diff] [blame]2171 <dd>Vector constants are represented with notation similar to vector type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2172 definitions (a comma separated list of elements, surrounded by
2173 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32
2174 42, i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must
2175 have <a href="#t_vector">vector type</a>, and the number and types of
2176 elements must match those specified by the type.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2177
2178 <dt><b>Zero initialization</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2179 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]2180 value to zero of <em>any</em> type, including scalar and
2181 <a href="#t_aggregate">aggregate</a> types.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2182 This is often used to avoid having to print large zero initializers
2183 (e.g. for large arrays) and is always exactly equivalent to using explicit
2184 zero initializers.</dd>
Nick Lewycky49f89192009-04-04 07:22:01 +0000[diff] [blame]2185
2186 <dt><b>Metadata node</b></dt>
Nick Lewycky8e2c4f42009-05-30 16:08:30 +0000[diff] [blame]2187 <dd>A metadata node is a structure-like constant with
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2188 <a href="#t_metadata">metadata type</a>. For example: "<tt>metadata !{
2189 i32 0, metadata !"test" }</tt>". Unlike other constants that are meant to
2190 be interpreted as part of the instruction stream, metadata is a place to
2191 attach additional information such as debug info.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2192</dl>
2193
2194</div>
2195
2196<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2197<h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2198 <a name="globalconstants">Global Variable and Function Addresses</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2199</h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2200
2201<div class="doc_text">
2202
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2203<p>The addresses of <a href="#globalvars">global variables</a>
2204 and <a href="#functionstructure">functions</a> are always implicitly valid
2205 (link-time) constants. These constants are explicitly referenced when
2206 the <a href="#identifiers">identifier for the global</a> is used and always
2207 have <a href="#t_pointer">pointer</a> type. For example, the following is a
2208 legal LLVM file:</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2209
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2210<pre class="doc_code">
Chris Lattner00538a12007-06-06 18:28:13 +0000[diff] [blame]2211@X = global i32 17
2212@Y = global i32 42
2213@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2214</pre>
2215
2216</div>
2217
2218<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2219<h3>
2220 <a name="undefvalues">Undefined Values</a>
2221</h3>
2222
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2223<div class="doc_text">
2224
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2225<p>The string '<tt>undef</tt>' can be used anywhere a constant is expected, and
Benjamin Kramer0f420382009-10-12 14:46:08 +0000[diff] [blame]2226 indicates that the user of the value may receive an unspecified bit-pattern.
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2227 Undefined values may be of any type (other than '<tt>label</tt>'
2228 or '<tt>void</tt>') and be used anywhere a constant is permitted.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2229
Chris Lattner92ada5d2009-09-11 01:49:31 +0000[diff] [blame]2230<p>Undefined values are useful because they indicate to the compiler that the
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2231 program is well defined no matter what value is used. This gives the
2232 compiler more freedom to optimize. Here are some examples of (potentially
2233 surprising) transformations that are valid (in pseudo IR):</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2234
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2235
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2236<pre class="doc_code">
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2237 %A = add %X, undef
2238 %B = sub %X, undef
2239 %C = xor %X, undef
2240Safe:
2241 %A = undef
2242 %B = undef
2243 %C = undef
2244</pre>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2245
2246<p>This is safe because all of the output bits are affected by the undef bits.
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2247 Any output bit can have a zero or one depending on the input bits.</p>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2248
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2249<pre class="doc_code">
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2250 %A = or %X, undef
2251 %B = and %X, undef
2252Safe:
2253 %A = -1
2254 %B = 0
2255Unsafe:
2256 %A = undef
2257 %B = undef
2258</pre>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2259
2260<p>These logical operations have bits that are not always affected by the input.
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2261 For example, if <tt>%X</tt> has a zero bit, then the output of the
2262 '<tt>and</tt>' operation will always be a zero for that bit, no matter what
2263 the corresponding bit from the '<tt>undef</tt>' is. As such, it is unsafe to
2264 optimize or assume that the result of the '<tt>and</tt>' is '<tt>undef</tt>'.
2265 However, it is safe to assume that all bits of the '<tt>undef</tt>' could be
2266 0, and optimize the '<tt>and</tt>' to 0. Likewise, it is safe to assume that
2267 all the bits of the '<tt>undef</tt>' operand to the '<tt>or</tt>' could be
2268 set, allowing the '<tt>or</tt>' to be folded to -1.</p>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2269
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2270<pre class="doc_code">
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2271 %A = select undef, %X, %Y
2272 %B = select undef, 42, %Y
2273 %C = select %X, %Y, undef
2274Safe:
2275 %A = %X (or %Y)
2276 %B = 42 (or %Y)
2277 %C = %Y
2278Unsafe:
2279 %A = undef
2280 %B = undef
2281 %C = undef
2282</pre>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2283
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2284<p>This set of examples shows that undefined '<tt>select</tt>' (and conditional
2285 branch) conditions can go <em>either way</em>, but they have to come from one
2286 of the two operands. In the <tt>%A</tt> example, if <tt>%X</tt> and
2287 <tt>%Y</tt> were both known to have a clear low bit, then <tt>%A</tt> would
2288 have to have a cleared low bit. However, in the <tt>%C</tt> example, the
2289 optimizer is allowed to assume that the '<tt>undef</tt>' operand could be the
2290 same as <tt>%Y</tt>, allowing the whole '<tt>select</tt>' to be
2291 eliminated.</p>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2292
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2293<pre class="doc_code">
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2294 %A = xor undef, undef
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2295
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2296 %B = undef
2297 %C = xor %B, %B
2298
2299 %D = undef
2300 %E = icmp lt %D, 4
2301 %F = icmp gte %D, 4
2302
2303Safe:
2304 %A = undef
2305 %B = undef
2306 %C = undef
2307 %D = undef
2308 %E = undef
2309 %F = undef
2310</pre>
Chris Lattnerec72b9b2009-09-07 22:52:39 +0000[diff] [blame]2311
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2312<p>This example points out that two '<tt>undef</tt>' operands are not
2313 necessarily the same. This can be surprising to people (and also matches C
2314 semantics) where they assume that "<tt>X^X</tt>" is always zero, even
2315 if <tt>X</tt> is undefined. This isn't true for a number of reasons, but the
2316 short answer is that an '<tt>undef</tt>' "variable" can arbitrarily change
2317 its value over its "live range". This is true because the variable doesn't
2318 actually <em>have a live range</em>. Instead, the value is logically read
2319 from arbitrary registers that happen to be around when needed, so the value
2320 is not necessarily consistent over time. In fact, <tt>%A</tt> and <tt>%C</tt>
2321 need to have the same semantics or the core LLVM "replace all uses with"
2322 concept would not hold.</p>
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2323
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2324<pre class="doc_code">
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2325 %A = fdiv undef, %X
2326 %B = fdiv %X, undef
2327Safe:
2328 %A = undef
2329b: unreachable
2330</pre>
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2331
2332<p>These examples show the crucial difference between an <em>undefined
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2333 value</em> and <em>undefined behavior</em>. An undefined value (like
2334 '<tt>undef</tt>') is allowed to have an arbitrary bit-pattern. This means that
2335 the <tt>%A</tt> operation can be constant folded to '<tt>undef</tt>', because
2336 the '<tt>undef</tt>' could be an SNaN, and <tt>fdiv</tt> is not (currently)
2337 defined on SNaN's. However, in the second example, we can make a more
2338 aggressive assumption: because the <tt>undef</tt> is allowed to be an
2339 arbitrary value, we are allowed to assume that it could be zero. Since a
2340 divide by zero has <em>undefined behavior</em>, we are allowed to assume that
2341 the operation does not execute at all. This allows us to delete the divide and
2342 all code after it. Because the undefined operation "can't happen", the
2343 optimizer can assume that it occurs in dead code.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2344
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2345<pre class="doc_code">
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2346a: store undef -> %X
2347b: store %X -> undef
2348Safe:
2349a: &lt;deleted&gt;
2350b: unreachable
2351</pre>
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2352
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2353<p>These examples reiterate the <tt>fdiv</tt> example: a store <em>of</em> an
2354 undefined value can be assumed to not have any effect; we can assume that the
2355 value is overwritten with bits that happen to match what was already there.
2356 However, a store <em>to</em> an undefined location could clobber arbitrary
2357 memory, therefore, it has undefined behavior.</p>
Chris Lattnera34a7182009-09-07 23:33:52 +0000[diff] [blame]2358
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2359</div>
2360
2361<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2362<h3>
2363 <a name="trapvalues">Trap Values</a>
2364</h3>
2365
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2366<div class="doc_text">
2367
Dan Gohmanb2a709b2010-04-26 20:21:21 +0000[diff] [blame]2368<p>Trap values are similar to <a href="#undefvalues">undef values</a>, however
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2369 instead of representing an unspecified bit pattern, they represent the
2370 fact that an instruction or constant expression which cannot evoke side
2371 effects has nevertheless detected a condition which results in undefined
Dan Gohmanb2a709b2010-04-26 20:21:21 +0000[diff] [blame]2372 behavior.</p>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2373
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2374<p>There is currently no way of representing a trap value in the IR; they
Dan Gohmanac355aa2010-05-03 14:51:43 +0000[diff] [blame]2375 only exist when produced by operations such as
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2376 <a href="#i_add"><tt>add</tt></a> with the <tt>nsw</tt> flag.</p>
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2377
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2378<p>Trap value behavior is defined in terms of value <i>dependence</i>:</p>
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2379
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2380<ul>
2381<li>Values other than <a href="#i_phi"><tt>phi</tt></a> nodes depend on
2382 their operands.</li>
2383
2384<li><a href="#i_phi"><tt>Phi</tt></a> nodes depend on the operand corresponding
2385 to their dynamic predecessor basic block.</li>
2386
2387<li>Function arguments depend on the corresponding actual argument values in
2388 the dynamic callers of their functions.</li>
2389
2390<li><a href="#i_call"><tt>Call</tt></a> instructions depend on the
2391 <a href="#i_ret"><tt>ret</tt></a> instructions that dynamically transfer
2392 control back to them.</li>
2393
Dan Gohman7292a752010-05-03 14:55:22 +0000[diff] [blame]2394<li><a href="#i_invoke"><tt>Invoke</tt></a> instructions depend on the
2395 <a href="#i_ret"><tt>ret</tt></a>, <a href="#i_unwind"><tt>unwind</tt></a>,
2396 or exception-throwing call instructions that dynamically transfer control
2397 back to them.</li>
2398
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2399<li>Non-volatile loads and stores depend on the most recent stores to all of the
2400 referenced memory addresses, following the order in the IR
2401 (including loads and stores implied by intrinsics such as
2402 <a href="#int_memcpy"><tt>@llvm.memcpy</tt></a>.)</li>
2403
Dan Gohman3513ea52010-05-03 14:59:34 +0000[diff] [blame]2404<!-- TODO: In the case of multiple threads, this only applies if the store
2405 "happens-before" the load or store. -->
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2406
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2407<!-- TODO: floating-point exception state -->
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2408
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2409<li>An instruction with externally visible side effects depends on the most
2410 recent preceding instruction with externally visible side effects, following
Dan Gohman6c858db2010-07-06 15:26:33 +0000[diff] [blame]2411 the order in the IR. (This includes
2412 <a href="#volatile">volatile operations</a>.)</li>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2413
Dan Gohman7292a752010-05-03 14:55:22 +0000[diff] [blame]2414<li>An instruction <i>control-depends</i> on a
2415 <a href="#terminators">terminator instruction</a>
2416 if the terminator instruction has multiple successors and the instruction
2417 is always executed when control transfers to one of the successors, and
Chris Lattner0ab5e2c2011-04-15 05:18:47 +0000[diff] [blame]2418 may not be executed when control is transferred to another.</li>
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2419
Dan Gohmanc8454ee2011-04-12 23:05:59 +0000[diff] [blame]2420<li>Additionally, an instruction also <i>control-depends</i> on a terminator
2421 instruction if the set of instructions it otherwise depends on would be
Chris Lattner0ab5e2c2011-04-15 05:18:47 +0000[diff] [blame]2422 different if the terminator had transferred control to a different
Dan Gohmanc8454ee2011-04-12 23:05:59 +0000[diff] [blame]2423 successor.</li>
2424
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2425<li>Dependence is transitive.</li>
2426
2427</ul>
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2428
2429<p>Whenever a trap value is generated, all values which depend on it evaluate
2430 to trap. If they have side effects, the evoke their side effects as if each
2431 operand with a trap value were undef. If they have externally-visible side
2432 effects, the behavior is undefined.</p>
2433
2434<p>Here are some examples:</p>
Dan Gohman48a25882010-04-26 20:54:53 +0000[diff] [blame]2435
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2436<pre class="doc_code">
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2437entry:
2438 %trap = sub nuw i32 0, 1 ; Results in a trap value.
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2439 %still_trap = and i32 %trap, 0 ; Whereas (and i32 undef, 0) would return 0.
2440 %trap_yet_again = getelementptr i32* @h, i32 %still_trap
2441 store i32 0, i32* %trap_yet_again ; undefined behavior
2442
2443 store i32 %trap, i32* @g ; Trap value conceptually stored to memory.
2444 %trap2 = load i32* @g ; Returns a trap value, not just undef.
2445
2446 volatile store i32 %trap, i32* @g ; External observation; undefined behavior.
2447
2448 %narrowaddr = bitcast i32* @g to i16*
2449 %wideaddr = bitcast i32* @g to i64*
2450 %trap3 = load 16* %narrowaddr ; Returns a trap value.
2451 %trap4 = load i64* %widaddr ; Returns a trap value.
2452
2453 %cmp = icmp i32 slt %trap, 0 ; Returns a trap value.
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2454 %br i1 %cmp, %true, %end ; Branch to either destination.
2455
2456true:
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2457 volatile store i32 0, i32* @g ; This is control-dependent on %cmp, so
2458 ; it has undefined behavior.
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2459 br label %end
2460
2461end:
2462 %p = phi i32 [ 0, %entry ], [ 1, %true ]
2463 ; Both edges into this PHI are
2464 ; control-dependent on %cmp, so this
Dan Gohman2f1ae062010-04-28 00:49:41 +0000[diff] [blame]2465 ; always results in a trap value.
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2466
Dan Gohmanc8454ee2011-04-12 23:05:59 +0000[diff] [blame]2467 volatile store i32 0, i32* @g ; This would depend on the store in %true
2468 ; if %cmp is true, or the store in %entry
2469 ; otherwise, so this is undefined behavior.
2470
2471 %br i1 %cmp, %second_true, %second_end
2472 ; The same branch again, but this time the
2473 ; true block doesn't have side effects.
2474
2475second_true:
2476 ; No side effects!
2477 br label %end
2478
2479second_end:
2480 volatile store i32 0, i32* @g ; This time, the instruction always depends
2481 ; on the store in %end. Also, it is
2482 ; control-equivalent to %end, so this is
2483 ; well- defined (again, ignoring earlier
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2484 ; undefined behavior in this example).
Dan Gohmanb8b85c12010-04-26 23:36:52 +0000[diff] [blame]2485</pre>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2486
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]2487</div>
2488
2489<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2490<h3>
2491 <a name="blockaddress">Addresses of Basic Blocks</a>
2492</h3>
2493
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2494<div class="doc_text">
2495
Chris Lattneraa99c942009-11-01 01:27:45 +0000[diff] [blame]2496<p><b><tt>blockaddress(@function, %block)</tt></b></p>
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2497
2498<p>The '<tt>blockaddress</tt>' constant computes the address of the specified
Chris Lattner5c5f0ac2009-10-27 21:49:40 +0000[diff] [blame]2499 basic block in the specified function, and always has an i8* type. Taking
Chris Lattneraa99c942009-11-01 01:27:45 +0000[diff] [blame]2500 the address of the entry block is illegal.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2501
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2502<p>This value only has defined behavior when used as an operand to the
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2503 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>' instruction, or for
2504 comparisons against null. Pointer equality tests between labels addresses
2505 results in undefined behavior &mdash; though, again, comparison against null
2506 is ok, and no label is equal to the null pointer. This may be passed around
2507 as an opaque pointer sized value as long as the bits are not inspected. This
2508 allows <tt>ptrtoint</tt> and arithmetic to be performed on these values so
2509 long as the original value is reconstituted before the <tt>indirectbr</tt>
2510 instruction.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]2511
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2512<p>Finally, some targets may provide defined semantics when using the value as
2513 the operand to an inline assembly, but that is target specific.</p>
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]2514
2515</div>
2516
2517
2518<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2519<h3>
2520 <a name="constantexprs">Constant Expressions</a>
2521</h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2522
2523<div class="doc_text">
2524
2525<p>Constant expressions are used to allow expressions involving other constants
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2526 to be used as constants. Constant expressions may be of
2527 any <a href="#t_firstclass">first class</a> type and may involve any LLVM
2528 operation that does not have side effects (e.g. load and call are not
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]2529 supported). The following is the syntax for constant expressions:</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2530
2531<dl>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2532 <dt><b><tt>trunc (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2533 <dd>Truncate a constant to another type. The bit size of CST must be larger
2534 than the bit size of TYPE. Both types must be integers.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2535
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2536 <dt><b><tt>zext (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2537 <dd>Zero extend a constant to another type. The bit size of CST must be
Duncan Sandsa522e562010-07-13 12:06:14 +0000[diff] [blame]2538 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2539
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2540 <dt><b><tt>sext (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2541 <dd>Sign extend a constant to another type. The bit size of CST must be
Duncan Sandsa522e562010-07-13 12:06:14 +0000[diff] [blame]2542 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2543
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2544 <dt><b><tt>fptrunc (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2545 <dd>Truncate a floating point constant to another floating point type. The
2546 size of CST must be larger than the size of TYPE. Both types must be
2547 floating point.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2548
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2549 <dt><b><tt>fpext (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2550 <dd>Floating point extend a constant to another type. The size of CST must be
2551 smaller or equal to the size of TYPE. Both types must be floating
2552 point.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2553
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2554 <dt><b><tt>fptoui (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2555 <dd>Convert a floating point constant to the corresponding unsigned integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2556 constant. TYPE must be a scalar or vector integer type. CST must be of
2557 scalar or vector floating point type. Both CST and TYPE must be scalars,
2558 or vectors of the same number of elements. If the value won't fit in the
2559 integer type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2560
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2561 <dt><b><tt>fptosi (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2562 <dd>Convert a floating point constant to the corresponding signed integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2563 constant. TYPE must be a scalar or vector integer type. CST must be of
2564 scalar or vector floating point type. Both CST and TYPE must be scalars,
2565 or vectors of the same number of elements. If the value won't fit in the
2566 integer type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2567
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2568 <dt><b><tt>uitofp (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2569 <dd>Convert an unsigned integer constant to the corresponding floating point
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2570 constant. TYPE must be a scalar or vector floating point type. CST must be
2571 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2572 vectors of the same number of elements. If the value won't fit in the
2573 floating point type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2574
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2575 <dt><b><tt>sitofp (CST to TYPE)</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2576 <dd>Convert a signed integer constant to the corresponding floating point
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2577 constant. TYPE must be a scalar or vector floating point type. CST must be
2578 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2579 vectors of the same number of elements. If the value won't fit in the
2580 floating point type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2581
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2582 <dt><b><tt>ptrtoint (CST to TYPE)</tt></b></dt>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]2583 <dd>Convert a pointer typed constant to the corresponding integer constant
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2584 <tt>TYPE</tt> must be an integer type. <tt>CST</tt> must be of pointer
2585 type. The <tt>CST</tt> value is zero extended, truncated, or unchanged to
2586 make it fit in <tt>TYPE</tt>.</dd>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]2587
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2588 <dt><b><tt>inttoptr (CST to TYPE)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2589 <dd>Convert a integer constant to a pointer constant. TYPE must be a pointer
2590 type. CST must be of integer type. The CST value is zero extended,
2591 truncated, or unchanged to make it fit in a pointer size. This one is
2592 <i>really</i> dangerous!</dd>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]2593
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2594 <dt><b><tt>bitcast (CST to TYPE)</tt></b></dt>
Chris Lattner789dee32009-02-28 18:27:03 +0000[diff] [blame]2595 <dd>Convert a constant, CST, to another TYPE. The constraints of the operands
2596 are the same as those for the <a href="#i_bitcast">bitcast
2597 instruction</a>.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2598
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2599 <dt><b><tt>getelementptr (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
2600 <dt><b><tt>getelementptr inbounds (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2601 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2602 constants. As with the <a href="#i_getelementptr">getelementptr</a>
2603 instruction, the index list may have zero or more indexes, which are
2604 required to make sense for the type of "CSTPTR".</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2605
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2606 <dt><b><tt>select (COND, VAL1, VAL2)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2607 <dd>Perform the <a href="#i_select">select operation</a> on constants.</dd>
Reid Spencer9965ee72006-12-04 19:23:19 +0000[diff] [blame]2608
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2609 <dt><b><tt>icmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer9965ee72006-12-04 19:23:19 +0000[diff] [blame]2610 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
2611
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2612 <dt><b><tt>fcmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer9965ee72006-12-04 19:23:19 +0000[diff] [blame]2613 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
Robert Bocchino7e97a6d2006-01-10 19:31:34 +0000[diff] [blame]2614
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2615 <dt><b><tt>extractelement (VAL, IDX)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2616 <dd>Perform the <a href="#i_extractelement">extractelement operation</a> on
2617 constants.</dd>
Robert Bocchino7e97a6d2006-01-10 19:31:34 +0000[diff] [blame]2618
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2619 <dt><b><tt>insertelement (VAL, ELT, IDX)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2620 <dd>Perform the <a href="#i_insertelement">insertelement operation</a> on
2621 constants.</dd>
Chris Lattner016a0e52006-04-08 00:13:41 +0000[diff] [blame]2622
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2623 <dt><b><tt>shufflevector (VEC1, VEC2, IDXMASK)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2624 <dd>Perform the <a href="#i_shufflevector">shufflevector operation</a> on
2625 constants.</dd>
Chris Lattner016a0e52006-04-08 00:13:41 +0000[diff] [blame]2626
Nick Lewycky9ab9a7f2010-05-29 06:44:15 +0000[diff] [blame]2627 <dt><b><tt>extractvalue (VAL, IDX0, IDX1, ...)</tt></b></dt>
2628 <dd>Perform the <a href="#i_extractvalue">extractvalue operation</a> on
2629 constants. The index list is interpreted in a similar manner as indices in
2630 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2631 index value must be specified.</dd>
2632
2633 <dt><b><tt>insertvalue (VAL, ELT, IDX0, IDX1, ...)</tt></b></dt>
2634 <dd>Perform the <a href="#i_insertvalue">insertvalue operation</a> on
2635 constants. The index list is interpreted in a similar manner as indices in
2636 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2637 index value must be specified.</dd>
2638
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]2639 <dt><b><tt>OPCODE (LHS, RHS)</tt></b></dt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2640 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
2641 be any of the <a href="#binaryops">binary</a>
2642 or <a href="#bitwiseops">bitwise binary</a> operations. The constraints
2643 on operands are the same as those for the corresponding instruction
2644 (e.g. no bitwise operations on floating point values are allowed).</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2645</dl>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2646
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2647</div>
Chris Lattnerb1652612004-03-08 16:49:10 +0000[diff] [blame]2648
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2649<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2650<h2><a name="othervalues">Other Values</a></h2>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2651<!-- *********************************************************************** -->
2652
2653<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2654<h3>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2655<a name="inlineasm">Inline Assembler Expressions</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2656</h3>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2657
2658<div class="doc_text">
2659
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2660<p>LLVM supports inline assembler expressions (as opposed
2661 to <a href="#moduleasm"> Module-Level Inline Assembly</a>) through the use of
2662 a special value. This value represents the inline assembler as a string
2663 (containing the instructions to emit), a list of operand constraints (stored
Dale Johannesen63c94fe2009-10-13 21:56:55 +0000[diff] [blame]2664 as a string), a flag that indicates whether or not the inline asm
Dale Johannesen1cfb9582009-10-21 23:28:00 +0000[diff] [blame]2665 expression has side effects, and a flag indicating whether the function
2666 containing the asm needs to align its stack conservatively. An example
2667 inline assembler expression is:</p>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2668
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2669<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2670i32 (i32) asm "bswap $0", "=r,r"
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2671</pre>
2672
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2673<p>Inline assembler expressions may <b>only</b> be used as the callee operand of
2674 a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we
2675 have:</p>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2676
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2677<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2678%X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2679</pre>
2680
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2681<p>Inline asms with side effects not visible in the constraint list must be
2682 marked as having side effects. This is done through the use of the
2683 '<tt>sideeffect</tt>' keyword, like so:</p>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2684
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2685<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]2686call void asm sideeffect "eieio", ""()
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2687</pre>
2688
Dale Johannesen1cfb9582009-10-21 23:28:00 +0000[diff] [blame]2689<p>In some cases inline asms will contain code that will not work unless the
2690 stack is aligned in some way, such as calls or SSE instructions on x86,
2691 yet will not contain code that does that alignment within the asm.
2692 The compiler should make conservative assumptions about what the asm might
2693 contain and should generate its usual stack alignment code in the prologue
2694 if the '<tt>alignstack</tt>' keyword is present:</p>
Dale Johannesen63c94fe2009-10-13 21:56:55 +0000[diff] [blame]2695
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2696<pre class="doc_code">
Dale Johannesen1cfb9582009-10-21 23:28:00 +0000[diff] [blame]2697call void asm alignstack "eieio", ""()
Dale Johannesen63c94fe2009-10-13 21:56:55 +0000[diff] [blame]2698</pre>
Dale Johannesen63c94fe2009-10-13 21:56:55 +0000[diff] [blame]2699
2700<p>If both keywords appear the '<tt>sideeffect</tt>' keyword must come
2701 first.</p>
2702
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2703<p>TODO: The format of the asm and constraints string still need to be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2704 documented here. Constraints on what can be done (e.g. duplication, moving,
2705 etc need to be documented). This is probably best done by reference to
2706 another document that covers inline asm from a holistic perspective.</p>
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2707</div>
2708
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2709<h4>
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2710<a name="inlineasm_md">Inline Asm Metadata</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2711</h4>
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2712
2713<div class="doc_text">
2714
2715<p>The call instructions that wrap inline asm nodes may have a "!srcloc" MDNode
Chris Lattner79ffdc72010-11-17 08:20:42 +0000[diff] [blame]2716 attached to it that contains a list of constant integers. If present, the
2717 code generator will use the integer as the location cookie value when report
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2718 errors through the LLVMContext error reporting mechanisms. This allows a
Dan Gohman61110ae2010-04-28 00:36:01 +0000[diff] [blame]2719 front-end to correlate backend errors that occur with inline asm back to the
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2720 source code that produced it. For example:</p>
2721
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2722<pre class="doc_code">
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2723call void asm sideeffect "something bad", ""()<b>, !srcloc !42</b>
2724...
2725!42 = !{ i32 1234567 }
2726</pre>
Chris Lattner51065562010-04-07 05:38:05 +0000[diff] [blame]2727
2728<p>It is up to the front-end to make sense of the magic numbers it places in the
Chris Lattner79ffdc72010-11-17 08:20:42 +0000[diff] [blame]2729 IR. If the MDNode contains multiple constants, the code generator will use
2730 the one that corresponds to the line of the asm that the error occurs on.</p>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2731
2732</div>
2733
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]2734<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2735<h3>
2736 <a name="metadata">Metadata Nodes and Metadata Strings</a>
2737</h3>
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]2738
2739<div class="doc_text">
2740
2741<p>LLVM IR allows metadata to be attached to instructions in the program that
2742 can convey extra information about the code to the optimizers and code
2743 generator. One example application of metadata is source-level debug
2744 information. There are two metadata primitives: strings and nodes. All
2745 metadata has the <tt>metadata</tt> type and is identified in syntax by a
2746 preceding exclamation point ('<tt>!</tt>').</p>
2747
2748<p>A metadata string is a string surrounded by double quotes. It can contain
2749 any character by escaping non-printable characters with "\xx" where "xx" is
2750 the two digit hex code. For example: "<tt>!"test\00"</tt>".</p>
2751
2752<p>Metadata nodes are represented with notation similar to structure constants
2753 (a comma separated list of elements, surrounded by braces and preceded by an
2754 exclamation point). For example: "<tt>!{ metadata !"test\00", i32
2755 10}</tt>". Metadata nodes can have any values as their operand.</p>
2756
2757<p>A <a href="#namedmetadatastructure">named metadata</a> is a collection of
2758 metadata nodes, which can be looked up in the module symbol table. For
2759 example: "<tt>!foo = metadata !{!4, !3}</tt>".
2760
Devang Patel9984bd62010-03-04 23:44:48 +0000[diff] [blame]2761<p>Metadata can be used as function arguments. Here <tt>llvm.dbg.value</tt>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2762 function is using two metadata arguments.</p>
Devang Patel9984bd62010-03-04 23:44:48 +0000[diff] [blame]2763
Bill Wendlingc0e10672011-03-02 02:17:11 +0000[diff] [blame]2764<div class="doc_code">
2765<pre>
2766call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2767</pre>
2768</div>
Devang Patel9984bd62010-03-04 23:44:48 +0000[diff] [blame]2769
2770<p>Metadata can be attached with an instruction. Here metadata <tt>!21</tt> is
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]2771 attached with <tt>add</tt> instruction using <tt>!dbg</tt> identifier.</p>
Devang Patel9984bd62010-03-04 23:44:48 +0000[diff] [blame]2772
Bill Wendlingc0e10672011-03-02 02:17:11 +0000[diff] [blame]2773<div class="doc_code">
2774<pre>
2775%indvar.next = add i64 %indvar, 1, !dbg !21
2776</pre>
2777</div>
2778
Chris Lattnerc2f8f162010-01-15 21:50:19 +0000[diff] [blame]2779</div>
2780
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2781
2782<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2783<h2>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2784 <a name="intrinsic_globals">Intrinsic Global Variables</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2785</h2>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2786<!-- *********************************************************************** -->
2787
2788<p>LLVM has a number of "magic" global variables that contain data that affect
2789code generation or other IR semantics. These are documented here. All globals
Chris Lattner58f9bb22009-07-20 06:14:25 +0000[diff] [blame]2790of this sort should have a section specified as "<tt>llvm.metadata</tt>". This
2791section and all globals that start with "<tt>llvm.</tt>" are reserved for use
2792by LLVM.</p>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2793
2794<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2795<h3>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2796<a name="intg_used">The '<tt>llvm.used</tt>' Global Variable</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2797</h3>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2798
2799<div class="doc_text">
2800
2801<p>The <tt>@llvm.used</tt> global is an array with i8* element type which has <a
2802href="#linkage_appending">appending linkage</a>. This array contains a list of
2803pointers to global variables and functions which may optionally have a pointer
2804cast formed of bitcast or getelementptr. For example, a legal use of it is:</p>
2805
2806<pre>
2807 @X = global i8 4
2808 @Y = global i32 123
2809
2810 @llvm.used = appending global [2 x i8*] [
2811 i8* @X,
2812 i8* bitcast (i32* @Y to i8*)
2813 ], section "llvm.metadata"
2814</pre>
2815
2816<p>If a global variable appears in the <tt>@llvm.used</tt> list, then the
2817compiler, assembler, and linker are required to treat the symbol as if there is
2818a reference to the global that it cannot see. For example, if a variable has
2819internal linkage and no references other than that from the <tt>@llvm.used</tt>
2820list, it cannot be deleted. This is commonly used to represent references from
2821inline asms and other things the compiler cannot "see", and corresponds to
2822"attribute((used))" in GNU C.</p>
2823
2824<p>On some targets, the code generator must emit a directive to the assembler or
2825object file to prevent the assembler and linker from molesting the symbol.</p>
2826
2827</div>
2828
2829<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2830<h3>
2831 <a name="intg_compiler_used">
2832 The '<tt>llvm.compiler.used</tt>' Global Variable
2833 </a>
2834</h3>
Chris Lattner58f9bb22009-07-20 06:14:25 +0000[diff] [blame]2835
2836<div class="doc_text">
2837
2838<p>The <tt>@llvm.compiler.used</tt> directive is the same as the
2839<tt>@llvm.used</tt> directive, except that it only prevents the compiler from
2840touching the symbol. On targets that support it, this allows an intelligent
2841linker to optimize references to the symbol without being impeded as it would be
2842by <tt>@llvm.used</tt>.</p>
2843
2844<p>This is a rare construct that should only be used in rare circumstances, and
2845should not be exposed to source languages.</p>
2846
2847</div>
2848
2849<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2850<h3>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2851<a name="intg_global_ctors">The '<tt>llvm.global_ctors</tt>' Global Variable</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2852</h3>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2853
2854<div class="doc_text">
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]2855<pre>
2856%0 = type { i32, void ()* }
David Chisnallb492b812010-04-30 19:27:35 +0000[diff] [blame]2857@llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor }]
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]2858</pre>
2859<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.
2860</p>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2861
2862</div>
2863
2864<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2865<h3>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2866<a name="intg_global_dtors">The '<tt>llvm.global_dtors</tt>' Global Variable</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2867</h3>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2868
2869<div class="doc_text">
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]2870<pre>
2871%0 = type { i32, void ()* }
David Chisnallb492b812010-04-30 19:27:35 +0000[diff] [blame]2872@llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor }]
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]2873</pre>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2874
David Chisnalla9d4a6f2010-04-30 19:23:49 +0000[diff] [blame]2875<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.
2876</p>
Chris Lattnerae76db52009-07-20 05:55:19 +0000[diff] [blame]2877
2878</div>
2879
2880
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]2881<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2882<h2><a name="instref">Instruction Reference</a></h2>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2883<!-- *********************************************************************** -->
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2884
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2885<div class="doc_text">
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2886
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2887<p>The LLVM instruction set consists of several different classifications of
2888 instructions: <a href="#terminators">terminator
2889 instructions</a>, <a href="#binaryops">binary instructions</a>,
2890 <a href="#bitwiseops">bitwise binary instructions</a>,
2891 <a href="#memoryops">memory instructions</a>, and
2892 <a href="#otherops">other instructions</a>.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2893
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2894</div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2895
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2896<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2897<h3>
2898 <a name="terminators">Terminator Instructions</a>
2899</h3>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2900
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2901<div class="doc_text">
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2902
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2903<p>As mentioned <a href="#functionstructure">previously</a>, every basic block
2904 in a program ends with a "Terminator" instruction, which indicates which
2905 block should be executed after the current block is finished. These
2906 terminator instructions typically yield a '<tt>void</tt>' value: they produce
2907 control flow, not values (the one exception being the
2908 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
2909
Duncan Sands626b0242010-04-15 20:35:54 +0000[diff] [blame]2910<p>There are seven different terminator instructions: the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2911 '<a href="#i_ret"><tt>ret</tt></a>' instruction, the
2912 '<a href="#i_br"><tt>br</tt></a>' instruction, the
2913 '<a href="#i_switch"><tt>switch</tt></a>' instruction, the
Bill Wendling33fef7e2009-11-02 00:25:26 +0000[diff] [blame]2914 '<a href="#i_indirectbr">'<tt>indirectbr</tt></a>' Instruction, the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2915 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the
2916 '<a href="#i_unwind"><tt>unwind</tt></a>' instruction, and the
2917 '<a href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2918
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2919</div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]2920
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2921<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2922<h4>
2923 <a name="i_ret">'<tt>ret</tt>' Instruction</a>
2924</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2925
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2926<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2927
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2928<h5>Syntax:</h5>
Dan Gohmancc3132e2008-10-04 19:00:07 +0000[diff] [blame]2929<pre>
2930 ret &lt;type&gt; &lt;value&gt; <i>; Return a value from a non-void function</i>
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]2931 ret void <i>; Return from void function</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2932</pre>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]2933
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2934<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2935<p>The '<tt>ret</tt>' instruction is used to return control flow (and optionally
2936 a value) from a function back to the caller.</p>
2937
2938<p>There are two forms of the '<tt>ret</tt>' instruction: one that returns a
2939 value and then causes control flow, and one that just causes control flow to
2940 occur.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]2941
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2942<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2943<p>The '<tt>ret</tt>' instruction optionally accepts a single argument, the
2944 return value. The type of the return value must be a
2945 '<a href="#t_firstclass">first class</a>' type.</p>
Dan Gohmancc3132e2008-10-04 19:00:07 +0000[diff] [blame]2946
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2947<p>A function is not <a href="#wellformed">well formed</a> if it it has a
2948 non-void return type and contains a '<tt>ret</tt>' instruction with no return
2949 value or a return value with a type that does not match its type, or if it
2950 has a void return type and contains a '<tt>ret</tt>' instruction with a
2951 return value.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]2952
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2953<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2954<p>When the '<tt>ret</tt>' instruction is executed, control flow returns back to
2955 the calling function's context. If the caller is a
2956 "<a href="#i_call"><tt>call</tt></a>" instruction, execution continues at the
2957 instruction after the call. If the caller was an
2958 "<a href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues at
2959 the beginning of the "normal" destination block. If the instruction returns
2960 a value, that value shall set the call or invoke instruction's return
2961 value.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]2962
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2963<h5>Example:</h5>
Chris Lattnerda508ac2008-04-23 04:59:35 +0000[diff] [blame]2964<pre>
2965 ret i32 5 <i>; Return an integer value of 5</i>
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]2966 ret void <i>; Return from a void function</i>
Bill Wendling050ee8f2009-02-28 22:12:54 +0000[diff] [blame]2967 ret { i32, i8 } { i32 4, i8 2 } <i>; Return a struct of values 4 and 2</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2968</pre>
Dan Gohman3065b612009-01-12 23:12:39 +0000[diff] [blame]2969
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2970</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2971<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]2972<h4>
2973 <a name="i_br">'<tt>br</tt>' Instruction</a>
2974</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2975
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2976<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2977
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2978<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2979<pre>
2980 br i1 &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;<br> br label &lt;dest&gt; <i>; Unconditional branch</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2981</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2982
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2983<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2984<p>The '<tt>br</tt>' instruction is used to cause control flow to transfer to a
2985 different basic block in the current function. There are two forms of this
2986 instruction, corresponding to a conditional branch and an unconditional
2987 branch.</p>
2988
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2989<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2990<p>The conditional branch form of the '<tt>br</tt>' instruction takes a single
2991 '<tt>i1</tt>' value and two '<tt>label</tt>' values. The unconditional form
2992 of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>' value as a
2993 target.</p>
2994
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2995<h5>Semantics:</h5>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]2996<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]2997 argument is evaluated. If the value is <tt>true</tt>, control flows to the
2998 '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
2999 control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
3000
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3001<h5>Example:</h5>
Bill Wendling30235112009-07-20 02:39:26 +0000[diff] [blame]3002<pre>
3003Test:
3004 %cond = <a href="#i_icmp">icmp</a> eq i32 %a, %b
3005 br i1 %cond, label %IfEqual, label %IfUnequal
3006IfEqual:
3007 <a href="#i_ret">ret</a> i32 1
3008IfUnequal:
3009 <a href="#i_ret">ret</a> i32 0
3010</pre>
3011
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3012</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3013
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3014<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3015<h4>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3016 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3017</h4>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3018
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3019<div class="doc_text">
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3020
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3021<h5>Syntax:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3022<pre>
3023 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
3024</pre>
3025
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3026<h5>Overview:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3027<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3028 several different places. It is a generalization of the '<tt>br</tt>'
3029 instruction, allowing a branch to occur to one of many possible
3030 destinations.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3031
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3032<h5>Arguments:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3033<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3034 comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination,
3035 and an array of pairs of comparison value constants and '<tt>label</tt>'s.
3036 The table is not allowed to contain duplicate constant entries.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3037
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3038<h5>Semantics:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3039<p>The <tt>switch</tt> instruction specifies a table of values and
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3040 destinations. When the '<tt>switch</tt>' instruction is executed, this table
3041 is searched for the given value. If the value is found, control flow is
Benjamin Kramer0f420382009-10-12 14:46:08 +0000[diff] [blame]3042 transferred to the corresponding destination; otherwise, control flow is
3043 transferred to the default destination.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3044
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3045<h5>Implementation:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3046<p>Depending on properties of the target machine and the particular
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3047 <tt>switch</tt> instruction, this instruction may be code generated in
3048 different ways. For example, it could be generated as a series of chained
3049 conditional branches or with a lookup table.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3050
3051<h5>Example:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3052<pre>
3053 <i>; Emulate a conditional br instruction</i>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]3054 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Dan Gohman623806e2009-01-04 23:44:43 +0000[diff] [blame]3055 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3056
3057 <i>; Emulate an unconditional br instruction</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]3058 switch i32 0, label %dest [ ]
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]3059
3060 <i>; Implement a jump table:</i>
Dan Gohman623806e2009-01-04 23:44:43 +0000[diff] [blame]3061 switch i32 %val, label %otherwise [ i32 0, label %onzero
3062 i32 1, label %onone
3063 i32 2, label %ontwo ]
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3064</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3065
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3066</div>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3067
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3068
3069<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3070<h4>
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]3071 <a name="i_indirectbr">'<tt>indirectbr</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3072</h4>
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3073
3074<div class="doc_text">
3075
3076<h5>Syntax:</h5>
3077<pre>
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]3078 indirectbr &lt;somety&gt;* &lt;address&gt;, [ label &lt;dest1&gt;, label &lt;dest2&gt;, ... ]
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3079</pre>
3080
3081<h5>Overview:</h5>
3082
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]3083<p>The '<tt>indirectbr</tt>' instruction implements an indirect branch to a label
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3084 within the current function, whose address is specified by
Chris Lattnere4801f72009-10-27 21:01:34 +0000[diff] [blame]3085 "<tt>address</tt>". Address must be derived from a <a
3086 href="#blockaddress">blockaddress</a> constant.</p>
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3087
3088<h5>Arguments:</h5>
3089
3090<p>The '<tt>address</tt>' argument is the address of the label to jump to. The
3091 rest of the arguments indicate the full set of possible destinations that the
3092 address may point to. Blocks are allowed to occur multiple times in the
3093 destination list, though this isn't particularly useful.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3094
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3095<p>This destination list is required so that dataflow analysis has an accurate
3096 understanding of the CFG.</p>
3097
3098<h5>Semantics:</h5>
3099
3100<p>Control transfers to the block specified in the address argument. All
3101 possible destination blocks must be listed in the label list, otherwise this
3102 instruction has undefined behavior. This implies that jumps to labels
3103 defined in other functions have undefined behavior as well.</p>
3104
3105<h5>Implementation:</h5>
3106
3107<p>This is typically implemented with a jump through a register.</p>
3108
3109<h5>Example:</h5>
3110<pre>
Chris Lattnerd04cb6d2009-10-28 00:19:10 +0000[diff] [blame]3111 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
Chris Lattner3ed871f2009-10-27 19:13:16 +0000[diff] [blame]3112</pre>
3113
3114</div>
3115
3116
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3117<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3118<h4>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3119 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3120</h4>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3121
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3122<div class="doc_text">
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3123
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3124<h5>Syntax:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3125<pre>
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]3126 &lt;result&gt; = invoke [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>] &lt;ptr to function ty&gt; &lt;function ptr val&gt;(&lt;function args&gt;) [<a href="#fnattrs">fn attrs</a>]
Chris Lattner6b7a0082006-05-14 18:23:06 +0000[diff] [blame]3127 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3128</pre>
3129
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]3130<h5>Overview:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3131<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3132 function, with the possibility of control flow transfer to either the
3133 '<tt>normal</tt>' label or the '<tt>exception</tt>' label. If the callee
3134 function returns with the "<tt><a href="#i_ret">ret</a></tt>" instruction,
3135 control flow will return to the "normal" label. If the callee (or any
3136 indirect callees) returns with the "<a href="#i_unwind"><tt>unwind</tt></a>"
3137 instruction, control is interrupted and continued at the dynamically nearest
3138 "exception" label.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3139
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3140<h5>Arguments:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3141<p>This instruction requires several arguments:</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3142
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3143<ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3144 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
3145 convention</a> the call should use. If none is specified, the call
3146 defaults to using C calling conventions.</li>
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]3147
3148 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3149 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
3150 '<tt>inreg</tt>' attributes are valid here.</li>
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]3151
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3152 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3153 function value being invoked. In most cases, this is a direct function
3154 invocation, but indirect <tt>invoke</tt>s are just as possible, branching
3155 off an arbitrary pointer to function value.</li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3156
3157 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3158 function to be invoked. </li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3159
3160 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]3161 signature argument types and parameter attributes. All arguments must be
3162 of <a href="#t_firstclass">first class</a> type. If the function
3163 signature indicates the function accepts a variable number of arguments,
3164 the extra arguments can be specified.</li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3165
3166 <li>'<tt>normal label</tt>': the label reached when the called function
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3167 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3168
3169 <li>'<tt>exception label</tt>': the label reached when a callee returns with
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3170 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3171
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]3172 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3173 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
3174 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3175</ol>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3176
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3177<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3178<p>This instruction is designed to operate as a standard
3179 '<tt><a href="#i_call">call</a></tt>' instruction in most regards. The
3180 primary difference is that it establishes an association with a label, which
3181 is used by the runtime library to unwind the stack.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3182
3183<p>This instruction is used in languages with destructors to ensure that proper
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3184 cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
3185 exception. Additionally, this is important for implementation of
3186 '<tt>catch</tt>' clauses in high-level languages that support them.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3187
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3188<p>For the purposes of the SSA form, the definition of the value returned by the
3189 '<tt>invoke</tt>' instruction is deemed to occur on the edge from the current
3190 block to the "normal" label. If the callee unwinds then no return value is
3191 available.</p>
Dan Gohman9069d892009-05-22 21:47:08 +0000[diff] [blame]3192
Chris Lattner97257f82010-01-15 18:08:37 +0000[diff] [blame]3193<p>Note that the code generator does not yet completely support unwind, and
3194that the invoke/unwind semantics are likely to change in future versions.</p>
3195
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3196<h5>Example:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3197<pre>
Nick Lewycky084ab472008-03-16 07:18:12 +0000[diff] [blame]3198 %retval = invoke i32 @Test(i32 15) to label %Continue
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]3199 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewycky084ab472008-03-16 07:18:12 +0000[diff] [blame]3200 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]3201 unwind label %TestCleanup <i>; {i32}:retval set</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3202</pre>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3203
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3204</div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3205
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]3206<!-- _______________________________________________________________________ -->
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3207
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3208<h4>
3209 <a name="i_unwind">'<tt>unwind</tt>' Instruction</a>
3210</h4>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3211
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3212<div class="doc_text">
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3213
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]3214<h5>Syntax:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3215<pre>
3216 unwind
3217</pre>
3218
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]3219<h5>Overview:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3220<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3221 at the first callee in the dynamic call stack which used
3222 an <a href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call.
3223 This is primarily used to implement exception handling.</p>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3224
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]3225<h5>Semantics:</h5>
Chris Lattnerfe8519c2008-04-19 21:01:16 +0000[diff] [blame]3226<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3227 immediately halt. The dynamic call stack is then searched for the
3228 first <a href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack.
3229 Once found, execution continues at the "exceptional" destination block
3230 specified by the <tt>invoke</tt> instruction. If there is no <tt>invoke</tt>
3231 instruction in the dynamic call chain, undefined behavior results.</p>
3232
Chris Lattner97257f82010-01-15 18:08:37 +0000[diff] [blame]3233<p>Note that the code generator does not yet completely support unwind, and
3234that the invoke/unwind semantics are likely to change in future versions.</p>
3235
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3236</div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3237
3238<!-- _______________________________________________________________________ -->
3239
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3240<h4>
3241 <a name="i_unreachable">'<tt>unreachable</tt>' Instruction</a>
3242</h4>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3243
3244<div class="doc_text">
3245
3246<h5>Syntax:</h5>
3247<pre>
3248 unreachable
3249</pre>
3250
3251<h5>Overview:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3252<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3253 instruction is used to inform the optimizer that a particular portion of the
3254 code is not reachable. This can be used to indicate that the code after a
3255 no-return function cannot be reached, and other facts.</p>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3256
3257<h5>Semantics:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3258<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3259
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]3260</div>
3261
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3262<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3263<h3>
3264 <a name="binaryops">Binary Operations</a>
3265</h3>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3266
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3267<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3268
3269<p>Binary operators are used to do most of the computation in a program. They
3270 require two operands of the same type, execute an operation on them, and
3271 produce a single value. The operands might represent multiple data, as is
3272 the case with the <a href="#t_vector">vector</a> data type. The result value
3273 has the same type as its operands.</p>
3274
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3275<p>There are several different binary operators:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3276
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3277</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3278
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3279<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3280<h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3281 <a name="i_add">'<tt>add</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3282</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3283
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3284<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3285
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3286<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3287<pre>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3288 &lt;result&gt; = add &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohman957b1312009-09-02 17:31:42 +0000[diff] [blame]3289 &lt;result&gt; = add nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3290 &lt;result&gt; = add nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3291 &lt;result&gt; = add nuw nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3292</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3293
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3294<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3295<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3296
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3297<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3298<p>The two arguments to the '<tt>add</tt>' instruction must
3299 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3300 integer values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3301
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3302<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3303<p>The value produced is the integer sum of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3304
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3305<p>If the sum has unsigned overflow, the result returned is the mathematical
3306 result modulo 2<sup>n</sup>, where n is the bit width of the result.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3307
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3308<p>Because LLVM integers use a two's complement representation, this instruction
3309 is appropriate for both signed and unsigned integers.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3310
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3311<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3312 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3313 <tt>nsw</tt> keywords are present, the result value of the <tt>add</tt>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]3314 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3315 respectively, occurs.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3316
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3317<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3318<pre>
3319 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3320</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3321
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3322</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3323
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3324<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3325<h4>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3326 <a name="i_fadd">'<tt>fadd</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3327</h4>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3328
3329<div class="doc_text">
3330
3331<h5>Syntax:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3332<pre>
3333 &lt;result&gt; = fadd &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3334</pre>
3335
3336<h5>Overview:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3337<p>The '<tt>fadd</tt>' instruction returns the sum of its two operands.</p>
3338
3339<h5>Arguments:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3340<p>The two arguments to the '<tt>fadd</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3341 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3342 floating point values. Both arguments must have identical types.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3343
3344<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3345<p>The value produced is the floating point sum of the two operands.</p>
3346
3347<h5>Example:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3348<pre>
3349 &lt;result&gt; = fadd float 4.0, %var <i>; yields {float}:result = 4.0 + %var</i>
3350</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3351
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3352</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3353
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3354<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3355<h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3356 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3357</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3358
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3359<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3360
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3361<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3362<pre>
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3363 &lt;result&gt; = sub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohman957b1312009-09-02 17:31:42 +0000[diff] [blame]3364 &lt;result&gt; = sub nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3365 &lt;result&gt; = sub nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3366 &lt;result&gt; = sub nuw nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3367</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3368
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3369<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3370<p>The '<tt>sub</tt>' instruction returns the difference of its two
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3371 operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3372
3373<p>Note that the '<tt>sub</tt>' instruction is used to represent the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3374 '<tt>neg</tt>' instruction present in most other intermediate
3375 representations.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3376
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3377<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3378<p>The two arguments to the '<tt>sub</tt>' instruction must
3379 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3380 integer values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3381
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3382<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3383<p>The value produced is the integer difference of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3384
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3385<p>If the difference has unsigned overflow, the result returned is the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3386 mathematical result modulo 2<sup>n</sup>, where n is the bit width of the
3387 result.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3388
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3389<p>Because LLVM integers use a two's complement representation, this instruction
3390 is appropriate for both signed and unsigned integers.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3391
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3392<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3393 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3394 <tt>nsw</tt> keywords are present, the result value of the <tt>sub</tt>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]3395 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3396 respectively, occurs.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3397
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3398<h5>Example:</h5>
Bill Wendling2d8b9a82007-05-29 09:42:13 +0000[diff] [blame]3399<pre>
3400 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]3401 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3402</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3403
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3404</div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3405
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3406<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3407<h4>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3408 <a name="i_fsub">'<tt>fsub</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3409</h4>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3410
3411<div class="doc_text">
3412
3413<h5>Syntax:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3414<pre>
3415 &lt;result&gt; = fsub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3416</pre>
3417
3418<h5>Overview:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3419<p>The '<tt>fsub</tt>' instruction returns the difference of its two
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3420 operands.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3421
3422<p>Note that the '<tt>fsub</tt>' instruction is used to represent the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3423 '<tt>fneg</tt>' instruction present in most other intermediate
3424 representations.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3425
3426<h5>Arguments:</h5>
Bill Wendling972b7202009-07-20 02:32:41 +0000[diff] [blame]3427<p>The two arguments to the '<tt>fsub</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3428 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3429 floating point values. Both arguments must have identical types.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3430
3431<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3432<p>The value produced is the floating point difference of the two operands.</p>
3433
3434<h5>Example:</h5>
3435<pre>
3436 &lt;result&gt; = fsub float 4.0, %var <i>; yields {float}:result = 4.0 - %var</i>
3437 &lt;result&gt; = fsub float -0.0, %val <i>; yields {float}:result = -%var</i>
3438</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3439
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3440</div>
3441
3442<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3443<h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3444 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3445</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3446
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3447<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3448
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3449<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3450<pre>
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3451 &lt;result&gt; = mul &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohman957b1312009-09-02 17:31:42 +0000[diff] [blame]3452 &lt;result&gt; = mul nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3453 &lt;result&gt; = mul nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3454 &lt;result&gt; = mul nuw nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3455</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3456
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3457<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3458<p>The '<tt>mul</tt>' instruction returns the product of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3459
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3460<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3461<p>The two arguments to the '<tt>mul</tt>' instruction must
3462 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3463 integer values. Both arguments must have identical types.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3464
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3465<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3466<p>The value produced is the integer product of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3467
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3468<p>If the result of the multiplication has unsigned overflow, the result
3469 returned is the mathematical result modulo 2<sup>n</sup>, where n is the bit
3470 width of the result.</p>
3471
3472<p>Because LLVM integers use a two's complement representation, and the result
3473 is the same width as the operands, this instruction returns the correct
3474 result for both signed and unsigned integers. If a full product
3475 (e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands should
3476 be sign-extended or zero-extended as appropriate to the width of the full
3477 product.</p>
3478
Dan Gohman902dfff2009-07-22 22:44:56 +0000[diff] [blame]3479<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3480 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3481 <tt>nsw</tt> keywords are present, the result value of the <tt>mul</tt>
Dan Gohmanffc9a6b2010-04-22 23:14:21 +0000[diff] [blame]3482 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3483 respectively, occurs.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3484
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3485<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3486<pre>
3487 &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3488</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3489
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3490</div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3491
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3492<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3493<h4>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3494 <a name="i_fmul">'<tt>fmul</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3495</h4>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3496
3497<div class="doc_text">
3498
3499<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3500<pre>
3501 &lt;result&gt; = fmul &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3502</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3503
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3504<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3505<p>The '<tt>fmul</tt>' instruction returns the product of its two operands.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3506
3507<h5>Arguments:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3508<p>The two arguments to the '<tt>fmul</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3509 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3510 floating point values. Both arguments must have identical types.</p>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3511
3512<h5>Semantics:</h5>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3513<p>The value produced is the floating point product of the two operands.</p>
3514
3515<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3516<pre>
3517 &lt;result&gt; = fmul float 4.0, %var <i>; yields {float}:result = 4.0 * %var</i>
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3518</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3519
Dan Gohmana5b96452009-06-04 22:49:04 +0000[diff] [blame]3520</div>
3521
3522<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3523<h4>
3524 <a name="i_udiv">'<tt>udiv</tt>' Instruction</a>
3525</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3526
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3527<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3528
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3529<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3530<pre>
Chris Lattner35315d02011-02-06 21:44:57 +0000[diff] [blame]3531 &lt;result&gt; = udiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3532 &lt;result&gt; = udiv exact &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3533</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3534
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3535<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3536<p>The '<tt>udiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3537
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3538<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3539<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3540 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3541 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3542
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3543<h5>Semantics:</h5>
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3544<p>The value produced is the unsigned integer quotient of the two operands.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3545
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3546<p>Note that unsigned integer division and signed integer division are distinct
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3547 operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
3548
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3549<p>Division by zero leads to undefined behavior.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3550
Chris Lattner35315d02011-02-06 21:44:57 +0000[diff] [blame]3551<p>If the <tt>exact</tt> keyword is present, the result value of the
3552 <tt>udiv</tt> is a <a href="#trapvalues">trap value</a> if %op1 is not a
3553 multiple of %op2 (as such, "((a udiv exact b) mul b) == a").</p>
3554
3555
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3556<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3557<pre>
3558 &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3559</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3560
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3561</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3562
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3563<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3564<h4>
3565 <a name="i_sdiv">'<tt>sdiv</tt>' Instruction</a>
3566</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3567
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3568<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3569
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3570<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3571<pre>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3572 &lt;result&gt; = sdiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohman957b1312009-09-02 17:31:42 +0000[diff] [blame]3573 &lt;result&gt; = sdiv exact &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3574</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3575
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3576<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3577<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3578
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3579<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3580<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3581 <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 +0000[diff] [blame]3583
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3584<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3585<p>The value produced is the signed integer quotient of the two operands rounded
3586 towards zero.</p>
3587
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3588<p>Note that signed integer division and unsigned integer division are distinct
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3589 operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
3590
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3591<p>Division by zero leads to undefined behavior. Overflow also leads to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3592 undefined behavior; this is a rare case, but can occur, for example, by doing
3593 a 32-bit division of -2147483648 by -1.</p>
3594
Dan Gohman71dfd782009-07-22 00:04:19 +0000[diff] [blame]3595<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohman57255802010-04-23 15:23:32 +0000[diff] [blame]3596 <tt>sdiv</tt> is a <a href="#trapvalues">trap value</a> if the result would
Dan Gohmane501ff72010-07-11 00:08:34 +0000[diff] [blame]3597 be rounded.</p>
Dan Gohmanb07de442009-07-20 22:41:19 +0000[diff] [blame]3598
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3599<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3600<pre>
3601 &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3602</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3603
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3604</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3605
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3606<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3607<h4>
3608 <a name="i_fdiv">'<tt>fdiv</tt>' Instruction</a>
3609</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3610
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3611<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3612
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3613<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3614<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +0000[diff] [blame]3615 &lt;result&gt; = fdiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3616</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3617
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3618<h5>Overview:</h5>
3619<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3620
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3621<h5>Arguments:</h5>
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]3622<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3623 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3624 floating point values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3625
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3626<h5>Semantics:</h5>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]3627<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3628
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3629<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3630<pre>
3631 &lt;result&gt; = fdiv float 4.0, %var <i>; yields {float}:result = 4.0 / %var</i>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3632</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3633
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3634</div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3635
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3636<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3637<h4>
3638 <a name="i_urem">'<tt>urem</tt>' Instruction</a>
3639</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3640
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3641<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3642
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3643<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3644<pre>
3645 &lt;result&gt; = urem &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3646</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3647
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3648<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3649<p>The '<tt>urem</tt>' instruction returns the remainder from the unsigned
3650 division of its two arguments.</p>
3651
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3652<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3653<p>The two arguments to the '<tt>urem</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3654 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3655 values. Both arguments must have identical types.</p>
3656
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3657<h5>Semantics:</h5>
3658<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3659 This instruction always performs an unsigned division to get the
3660 remainder.</p>
3661
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3662<p>Note that unsigned integer remainder and signed integer remainder are
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3663 distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
3664
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3665<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3666
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3667<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3668<pre>
3669 &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3670</pre>
3671
3672</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3673
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3674<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3675<h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3676 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3677</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3678
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3679<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3680
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3681<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3682<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +0000[diff] [blame]3683 &lt;result&gt; = srem &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3684</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3685
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3686<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3687<p>The '<tt>srem</tt>' instruction returns the remainder from the signed
3688 division of its two operands. This instruction can also take
3689 <a href="#t_vector">vector</a> versions of the values in which case the
3690 elements must be integers.</p>
Chris Lattnerb8f816e2008-01-04 04:33:49 +0000[diff] [blame]3691
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3692<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3693<p>The two arguments to the '<tt>srem</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3694 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3695 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3696
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3697<h5>Semantics:</h5>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3698<p>This instruction returns the <i>remainder</i> of a division (where the result
Duncan Sands2769c6e2011-03-07 09:12:24 +0000[diff] [blame]3699 is either zero or has the same sign as the dividend, <tt>op1</tt>), not the
3700 <i>modulo</i> operator (where the result is either zero or has the same sign
3701 as the divisor, <tt>op2</tt>) of a value.
3702 For more information about the difference,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3703 see <a href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
3704 Math Forum</a>. For a table of how this is implemented in various languages,
3705 please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
3706 Wikipedia: modulo operation</a>.</p>
3707
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3708<p>Note that signed integer remainder and unsigned integer remainder are
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3709 distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
3710
Chris Lattner2f2427e2008-01-28 00:36:27 +0000[diff] [blame]3711<p>Taking the remainder of a division by zero leads to undefined behavior.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3712 Overflow also leads to undefined behavior; this is a rare case, but can
3713 occur, for example, by taking the remainder of a 32-bit division of
3714 -2147483648 by -1. (The remainder doesn't actually overflow, but this rule
3715 lets srem be implemented using instructions that return both the result of
3716 the division and the remainder.)</p>
3717
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3718<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3719<pre>
3720 &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3721</pre>
3722
3723</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3724
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3725<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3726<h4>
3727 <a name="i_frem">'<tt>frem</tt>' Instruction</a>
3728</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3729
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3730<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3731
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3732<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3733<pre>
3734 &lt;result&gt; = frem &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3735</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3736
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3737<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3738<p>The '<tt>frem</tt>' instruction returns the remainder from the division of
3739 its two operands.</p>
3740
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3741<h5>Arguments:</h5>
3742<p>The two arguments to the '<tt>frem</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3743 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3744 floating point values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3745
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3746<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3747<p>This instruction returns the <i>remainder</i> of a division. The remainder
3748 has the same sign as the dividend.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3749
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]3750<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3751<pre>
3752 &lt;result&gt; = frem float 4.0, %var <i>; yields {float}:result = 4.0 % %var</i>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3753</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3754
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3755</div>
Robert Bocchino820bc75b2006-02-17 21:18:08 +0000[diff] [blame]3756
Reid Spencer2ab01932007-02-02 13:57:07 +0000[diff] [blame]3757<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3758<h3>
3759 <a name="bitwiseops">Bitwise Binary Operations</a>
3760</h3>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3761
Reid Spencer2ab01932007-02-02 13:57:07 +0000[diff] [blame]3762<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3763
3764<p>Bitwise binary operators are used to do various forms of bit-twiddling in a
3765 program. They are generally very efficient instructions and can commonly be
3766 strength reduced from other instructions. They require two operands of the
3767 same type, execute an operation on them, and produce a single value. The
3768 resulting value is the same type as its operands.</p>
3769
Reid Spencer2ab01932007-02-02 13:57:07 +0000[diff] [blame]3770</div>
3771
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3772<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3773<h4>
3774 <a name="i_shl">'<tt>shl</tt>' Instruction</a>
3775</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3776
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3777<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3778
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3779<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3780<pre>
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]3781 &lt;result&gt; = shl &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3782 &lt;result&gt; = shl nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3783 &lt;result&gt; = shl nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3784 &lt;result&gt; = shl nuw nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3785</pre>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3786
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3787<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3788<p>The '<tt>shl</tt>' instruction returns the first operand shifted to the left
3789 a specified number of bits.</p>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3790
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3791<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3792<p>Both arguments to the '<tt>shl</tt>' instruction must be the
3793 same <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3794 integer type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3795
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3796<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3797<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod
3798 2<sup>n</sup>, where <tt>n</tt> is the width of the result. If <tt>op2</tt>
3799 is (statically or dynamically) negative or equal to or larger than the number
3800 of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3801 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3802 shift amount in <tt>op2</tt>.</p>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3803
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]3804<p>If the <tt>nuw</tt> keyword is present, then the shift produces a
3805 <a href="#trapvalues">trap value</a> if it shifts out any non-zero bits. If
Chris Lattnerf10dfdc2011-02-09 16:44:44 +0000[diff] [blame]3806 the <tt>nsw</tt> keyword is present, then the shift produces a
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]3807 <a href="#trapvalues">trap value</a> if it shifts out any bits that disagree
3808 with the resultant sign bit. As such, NUW/NSW have the same semantics as
3809 they would if the shift were expressed as a mul instruction with the same
3810 nsw/nuw bits in (mul %op1, (shl 1, %op2)).</p>
3811
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3812<h5>Example:</h5>
3813<pre>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3814 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
3815 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
3816 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3817 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Mon P Wang4dd832d2008-12-09 05:46:39 +0000[diff] [blame]3818 &lt;result&gt; = shl &lt;2 x i32&gt; &lt; i32 1, i32 1&gt;, &lt; i32 1, i32 2&gt; <i>; yields: result=&lt;2 x i32&gt; &lt; i32 2, i32 4&gt;</i>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3819</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3820
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3821</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3822
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3823<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3824<h4>
3825 <a name="i_lshr">'<tt>lshr</tt>' Instruction</a>
3826</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3827
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3828<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3829
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3830<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3831<pre>
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]3832 &lt;result&gt; = lshr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3833 &lt;result&gt; = lshr exact &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3834</pre>
3835
3836<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3837<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
3838 operand shifted to the right a specified number of bits with zero fill.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3839
3840<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3841<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3842 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3843 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3844
3845<h5>Semantics:</h5>
3846<p>This instruction always performs a logical shift right operation. The most
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3847 significant bits of the result will be filled with zero bits after the shift.
3848 If <tt>op2</tt> is (statically or dynamically) equal to or larger than the
3849 number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3850 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3851 shift amount in <tt>op2</tt>.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3852
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]3853<p>If the <tt>exact</tt> keyword is present, the result value of the
3854 <tt>lshr</tt> is a <a href="#trapvalues">trap value</a> if any of the bits
3855 shifted out are non-zero.</p>
3856
3857
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3858<h5>Example:</h5>
3859<pre>
3860 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
3861 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
3862 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
3863 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3864 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Mon P Wang4dd832d2008-12-09 05:46:39 +0000[diff] [blame]3865 &lt;result&gt; = lshr &lt;2 x i32&gt; &lt; i32 -2, i32 4&gt;, &lt; i32 1, i32 2&gt; <i>; yields: result=&lt;2 x i32&gt; &lt; i32 0x7FFFFFFF, i32 1&gt;</i>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3866</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3867
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3868</div>
3869
Reid Spencer2ab01932007-02-02 13:57:07 +0000[diff] [blame]3870<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3871<h4>
3872 <a name="i_ashr">'<tt>ashr</tt>' Instruction</a>
3873</h4>
3874
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3875<div class="doc_text">
3876
3877<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3878<pre>
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]3879 &lt;result&gt; = ashr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3880 &lt;result&gt; = ashr exact &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3881</pre>
3882
3883<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3884<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
3885 operand shifted to the right a specified number of bits with sign
3886 extension.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3887
3888<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3889<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3890 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3891 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3892
3893<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3894<p>This instruction always performs an arithmetic shift right operation, The
3895 most significant bits of the result will be filled with the sign bit
3896 of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
3897 larger than the number of bits in <tt>op1</tt>, the result is undefined. If
3898 the arguments are vectors, each vector element of <tt>op1</tt> is shifted by
3899 the corresponding shift amount in <tt>op2</tt>.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3900
Chris Lattnera676c0f2011-02-07 16:40:21 +0000[diff] [blame]3901<p>If the <tt>exact</tt> keyword is present, the result value of the
3902 <tt>ashr</tt> is a <a href="#trapvalues">trap value</a> if any of the bits
3903 shifted out are non-zero.</p>
3904
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3905<h5>Example:</h5>
3906<pre>
3907 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
3908 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
3909 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
3910 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattnerf0e50112007-10-03 21:01:14 +0000[diff] [blame]3911 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Mon P Wang4dd832d2008-12-09 05:46:39 +0000[diff] [blame]3912 &lt;result&gt; = ashr &lt;2 x i32&gt; &lt; i32 -2, i32 4&gt;, &lt; i32 1, i32 3&gt; <i>; yields: result=&lt;2 x i32&gt; &lt; i32 -1, i32 0&gt;</i>
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3913</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3914
Reid Spencer04e259b2007-01-31 21:39:12 +0000[diff] [blame]3915</div>
3916
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3917<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3918<h4>
3919 <a name="i_and">'<tt>and</tt>' Instruction</a>
3920</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3921
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3922<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3923
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3924<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3925<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +0000[diff] [blame]3926 &lt;result&gt; = and &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3927</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3928
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3929<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3930<p>The '<tt>and</tt>' instruction returns the bitwise logical and of its two
3931 operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3932
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3933<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3934<p>The two arguments to the '<tt>and</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3935 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3936 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3937
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3938<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3939<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3940
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3941<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3942 <tbody>
3943 <tr>
3944 <td>In0</td>
3945 <td>In1</td>
3946 <td>Out</td>
3947 </tr>
3948 <tr>
3949 <td>0</td>
3950 <td>0</td>
3951 <td>0</td>
3952 </tr>
3953 <tr>
3954 <td>0</td>
3955 <td>1</td>
3956 <td>0</td>
3957 </tr>
3958 <tr>
3959 <td>1</td>
3960 <td>0</td>
3961 <td>0</td>
3962 </tr>
3963 <tr>
3964 <td>1</td>
3965 <td>1</td>
3966 <td>1</td>
3967 </tr>
3968 </tbody>
3969</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3970
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3971<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3972<pre>
3973 &lt;result&gt; = and i32 4, %var <i>; yields {i32}:result = 4 &amp; %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]3974 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
3975 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3976</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3977</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3978<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]3979<h4>
3980 <a name="i_or">'<tt>or</tt>' Instruction</a>
3981</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]3982
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3983<div class="doc_text">
3984
3985<h5>Syntax:</h5>
3986<pre>
3987 &lt;result&gt; = or &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3988</pre>
3989
3990<h5>Overview:</h5>
3991<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive or of its
3992 two operands.</p>
3993
3994<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]3995<p>The two arguments to the '<tt>or</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]3996 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3997 values. Both arguments must have identical types.</p>
3998
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3999<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4000<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4001
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]4002<table border="1" cellspacing="0" cellpadding="4">
4003 <tbody>
4004 <tr>
4005 <td>In0</td>
4006 <td>In1</td>
4007 <td>Out</td>
4008 </tr>
4009 <tr>
4010 <td>0</td>
4011 <td>0</td>
4012 <td>0</td>
4013 </tr>
4014 <tr>
4015 <td>0</td>
4016 <td>1</td>
4017 <td>1</td>
4018 </tr>
4019 <tr>
4020 <td>1</td>
4021 <td>0</td>
4022 <td>1</td>
4023 </tr>
4024 <tr>
4025 <td>1</td>
4026 <td>1</td>
4027 <td>1</td>
4028 </tr>
4029 </tbody>
4030</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4031
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4032<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4033<pre>
4034 &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4035 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
4036 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4037</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4038
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4039</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4040
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4041<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4042<h4>
4043 <a name="i_xor">'<tt>xor</tt>' Instruction</a>
4044</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4045
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4046<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4047
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4048<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4049<pre>
4050 &lt;result&gt; = xor &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4051</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4052
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4053<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4054<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive or of
4055 its two operands. The <tt>xor</tt> is used to implement the "one's
4056 complement" operation, which is the "~" operator in C.</p>
4057
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4058<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4059<p>The two arguments to the '<tt>xor</tt>' instruction must be
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4060 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4061 values. Both arguments must have identical types.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]4062
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4063<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4064<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4065
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]4066<table border="1" cellspacing="0" cellpadding="4">
4067 <tbody>
4068 <tr>
4069 <td>In0</td>
4070 <td>In1</td>
4071 <td>Out</td>
4072 </tr>
4073 <tr>
4074 <td>0</td>
4075 <td>0</td>
4076 <td>0</td>
4077 </tr>
4078 <tr>
4079 <td>0</td>
4080 <td>1</td>
4081 <td>1</td>
4082 </tr>
4083 <tr>
4084 <td>1</td>
4085 <td>0</td>
4086 <td>1</td>
4087 </tr>
4088 <tr>
4089 <td>1</td>
4090 <td>1</td>
4091 <td>0</td>
4092 </tr>
4093 </tbody>
4094</table>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4095
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4096<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4097<pre>
4098 &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4099 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
4100 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
4101 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4102</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4103
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4104</div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4105
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4106<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4107<h3>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4108 <a name="vectorops">Vector Operations</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4109</h3>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4110
4111<div class="doc_text">
4112
4113<p>LLVM supports several instructions to represent vector operations in a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4114 target-independent manner. These instructions cover the element-access and
4115 vector-specific operations needed to process vectors effectively. While LLVM
4116 does directly support these vector operations, many sophisticated algorithms
4117 will want to use target-specific intrinsics to take full advantage of a
4118 specific target.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4119
4120</div>
4121
4122<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4123<h4>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4124 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4125</h4>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4126
4127<div class="doc_text">
4128
4129<h5>Syntax:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4130<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4131 &lt;result&gt; = extractelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, i32 &lt;idx&gt; <i>; yields &lt;ty&gt;</i>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4132</pre>
4133
4134<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4135<p>The '<tt>extractelement</tt>' instruction extracts a single scalar element
4136 from a vector at a specified index.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4137
4138
4139<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4140<p>The first operand of an '<tt>extractelement</tt>' instruction is a value
4141 of <a href="#t_vector">vector</a> type. The second operand is an index
4142 indicating the position from which to extract the element. The index may be
4143 a variable.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4144
4145<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4146<p>The result is a scalar of the same type as the element type of
4147 <tt>val</tt>. Its value is the value at position <tt>idx</tt> of
4148 <tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4149 results are undefined.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4150
4151<h5>Example:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4152<pre>
Gabor Greif03ab4dc2009-10-28 13:14:50 +0000[diff] [blame]4153 &lt;result&gt; = extractelement &lt;4 x i32&gt; %vec, i32 0 <i>; yields i32</i>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4154</pre>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4155
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4156</div>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4157
4158<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4159<h4>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4160 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4161</h4>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4162
4163<div class="doc_text">
4164
4165<h5>Syntax:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4166<pre>
Dan Gohman43ba0672008-05-12 23:38:42 +0000[diff] [blame]4167 &lt;result&gt; = insertelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, &lt;ty&gt; &lt;elt&gt;, i32 &lt;idx&gt; <i>; yields &lt;n x &lt;ty&gt;&gt;</i>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4168</pre>
4169
4170<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4171<p>The '<tt>insertelement</tt>' instruction inserts a scalar element into a
4172 vector at a specified index.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4173
4174<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4175<p>The first operand of an '<tt>insertelement</tt>' instruction is a value
4176 of <a href="#t_vector">vector</a> type. The second operand is a scalar value
4177 whose type must equal the element type of the first operand. The third
4178 operand is an index indicating the position at which to insert the value.
4179 The index may be a variable.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4180
4181<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4182<p>The result is a vector of the same type as <tt>val</tt>. Its element values
4183 are those of <tt>val</tt> except at position <tt>idx</tt>, where it gets the
4184 value <tt>elt</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4185 results are undefined.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4186
4187<h5>Example:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4188<pre>
Gabor Greif03ab4dc2009-10-28 13:14:50 +0000[diff] [blame]4189 &lt;result&gt; = insertelement &lt;4 x i32&gt; %vec, i32 1, i32 0 <i>; yields &lt;4 x i32&gt;</i>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4190</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4191
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4192</div>
4193
4194<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4195<h4>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4196 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4197</h4>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4198
4199<div class="doc_text">
4200
4201<h5>Syntax:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4202<pre>
Mon P Wang25f01062008-11-10 04:46:22 +0000[diff] [blame]4203 &lt;result&gt; = shufflevector &lt;n x &lt;ty&gt;&gt; &lt;v1&gt;, &lt;n x &lt;ty&gt;&gt; &lt;v2&gt;, &lt;m x i32&gt; &lt;mask&gt; <i>; yields &lt;m x &lt;ty&gt;&gt;</i>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4204</pre>
4205
4206<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4207<p>The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
4208 from two input vectors, returning a vector with the same element type as the
4209 input and length that is the same as the shuffle mask.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4210
4211<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4212<p>The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
4213 with types that match each other. The third argument is a shuffle mask whose
4214 element type is always 'i32'. The result of the instruction is a vector
4215 whose length is the same as the shuffle mask and whose element type is the
4216 same as the element type of the first two operands.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4217
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4218<p>The shuffle mask operand is required to be a constant vector with either
4219 constant integer or undef values.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4220
4221<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4222<p>The elements of the two input vectors are numbered from left to right across
4223 both of the vectors. The shuffle mask operand specifies, for each element of
4224 the result vector, which element of the two input vectors the result element
4225 gets. The element selector may be undef (meaning "don't care") and the
4226 second operand may be undef if performing a shuffle from only one vector.</p>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4227
4228<h5>Example:</h5>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4229<pre>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4230 &lt;result&gt; = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]4231 &lt;4 x i32&gt; &lt;i32 0, i32 4, i32 1, i32 5&gt; <i>; yields &lt;4 x i32&gt;</i>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4232 &lt;result&gt; = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; undef,
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4233 &lt;4 x i32&gt; &lt;i32 0, i32 1, i32 2, i32 3&gt; <i>; yields &lt;4 x i32&gt;</i> - Identity shuffle.
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4234 &lt;result&gt; = shufflevector &lt;8 x i32&gt; %v1, &lt;8 x i32&gt; undef,
Mon P Wang25f01062008-11-10 04:46:22 +0000[diff] [blame]4235 &lt;4 x i32&gt; &lt;i32 0, i32 1, i32 2, i32 3&gt; <i>; yields &lt;4 x i32&gt;</i>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4236 &lt;result&gt; = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
Mon P Wang25f01062008-11-10 04:46:22 +0000[diff] [blame]4237 &lt;8 x i32&gt; &lt;i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7 &gt; <i>; yields &lt;8 x i32&gt;</i>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4238</pre>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4239
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4240</div>
Tanya Lattnerb138bbe2006-04-14 19:24:33 +0000[diff] [blame]4241
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]4242<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4243<h3>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4244 <a name="aggregateops">Aggregate Operations</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4245</h3>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4246
4247<div class="doc_text">
4248
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4249<p>LLVM supports several instructions for working with
4250 <a href="#t_aggregate">aggregate</a> values.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4251
4252</div>
4253
4254<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4255<h4>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4256 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4257</h4>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4258
4259<div class="doc_text">
4260
4261<h5>Syntax:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4262<pre>
4263 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
4264</pre>
4265
4266<h5>Overview:</h5>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4267<p>The '<tt>extractvalue</tt>' instruction extracts the value of a member field
4268 from an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4269
4270<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4271<p>The first operand of an '<tt>extractvalue</tt>' instruction is a value
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]4272 of <a href="#t_struct">struct</a> or
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4273 <a href="#t_array">array</a> type. The operands are constant indices to
4274 specify which value to extract in a similar manner as indices in a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4275 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
Frits van Bommel7cf63ac2010-12-05 20:54:38 +0000[diff] [blame]4276 <p>The major differences to <tt>getelementptr</tt> indexing are:</p>
4277 <ul>
4278 <li>Since the value being indexed is not a pointer, the first index is
4279 omitted and assumed to be zero.</li>
4280 <li>At least one index must be specified.</li>
4281 <li>Not only struct indices but also array indices must be in
4282 bounds.</li>
4283 </ul>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4284
4285<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4286<p>The result is the value at the position in the aggregate specified by the
4287 index operands.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4288
4289<h5>Example:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4290<pre>
Gabor Greif03ab4dc2009-10-28 13:14:50 +0000[diff] [blame]4291 &lt;result&gt; = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4292</pre>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4293
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4294</div>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4295
4296<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4297<h4>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4298 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4299</h4>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4300
4301<div class="doc_text">
4302
4303<h5>Syntax:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4304<pre>
Jeffrey Yasskinf991bbb2010-01-11 19:19:26 +0000[diff] [blame]4305 &lt;result&gt; = insertvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;ty&gt; &lt;elt&gt;, &lt;idx&gt; <i>; yields &lt;aggregate type&gt;</i>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4306</pre>
4307
4308<h5>Overview:</h5>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4309<p>The '<tt>insertvalue</tt>' instruction inserts a value into a member field
4310 in an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4311
4312<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4313<p>The first operand of an '<tt>insertvalue</tt>' instruction is a value
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]4314 of <a href="#t_struct">struct</a> or
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4315 <a href="#t_array">array</a> type. The second operand is a first-class
4316 value to insert. The following operands are constant indices indicating
4317 the position at which to insert the value in a similar manner as indices in a
Frits van Bommel7cf63ac2010-12-05 20:54:38 +0000[diff] [blame]4318 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' instruction. The
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4319 value to insert must have the same type as the value identified by the
4320 indices.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4321
4322<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4323<p>The result is an aggregate of the same type as <tt>val</tt>. Its value is
4324 that of <tt>val</tt> except that the value at the position specified by the
4325 indices is that of <tt>elt</tt>.</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4326
4327<h5>Example:</h5>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4328<pre>
Jeffrey Yasskinf991bbb2010-01-11 19:19:26 +0000[diff] [blame]4329 %agg1 = insertvalue {i32, float} undef, i32 1, 0 <i>; yields {i32 1, float undef}</i>
4330 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 <i>; yields {i32 1, float %val}</i>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4331</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4332
Dan Gohmanb9d66602008-05-12 23:51:09 +0000[diff] [blame]4333</div>
4334
4335
4336<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4337<h3>
Chris Lattner6ab66722006-08-15 00:45:58 +0000[diff] [blame]4338 <a name="memoryops">Memory Access and Addressing Operations</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4339</h3>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4340
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4341<div class="doc_text">
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4342
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4343<p>A key design point of an SSA-based representation is how it represents
4344 memory. In LLVM, no memory locations are in SSA form, which makes things
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]4345 very simple. This section describes how to read, write, and allocate
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4346 memory in LLVM.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4347
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4348</div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4349
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4350<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4351<h4>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4352 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4353</h4>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4354
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4355<div class="doc_text">
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4356
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4357<h5>Syntax:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4358<pre>
Dan Gohman2140a742010-05-28 01:14:11 +0000[diff] [blame]4359 &lt;result&gt; = alloca &lt;type&gt;[, &lt;ty&gt; &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4360</pre>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4361
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4362<h5>Overview:</h5>
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]4363<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4364 currently executing function, to be automatically released when this function
4365 returns to its caller. The object is always allocated in the generic address
4366 space (address space zero).</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4367
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4368<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4369<p>The '<tt>alloca</tt>' instruction
4370 allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt> bytes of memory on the
4371 runtime stack, returning a pointer of the appropriate type to the program.
4372 If "NumElements" is specified, it is the number of elements allocated,
4373 otherwise "NumElements" is defaulted to be one. If a constant alignment is
4374 specified, the value result of the allocation is guaranteed to be aligned to
4375 at least that boundary. If not specified, or if zero, the target can choose
4376 to align the allocation on any convenient boundary compatible with the
4377 type.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4378
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4379<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4380
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4381<h5>Semantics:</h5>
Bill Wendling9ee6a312009-05-08 20:49:29 +0000[diff] [blame]4382<p>Memory is allocated; a pointer is returned. The operation is undefined if
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4383 there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
4384 memory is automatically released when the function returns. The
4385 '<tt>alloca</tt>' instruction is commonly used to represent automatic
4386 variables that must have an address available. When the function returns
4387 (either with the <tt><a href="#i_ret">ret</a></tt>
4388 or <tt><a href="#i_unwind">unwind</a></tt> instructions), the memory is
4389 reclaimed. Allocating zero bytes is legal, but the result is undefined.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4390
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4391<h5>Example:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4392<pre>
Dan Gohman7a5acb52009-01-04 23:49:44 +0000[diff] [blame]4393 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
4394 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
4395 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
4396 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4397</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4398
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4399</div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]4400
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4401<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4402<h4>
4403 <a name="i_load">'<tt>load</tt>' Instruction</a>
4404</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4405
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4406<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4407
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4408<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4409<pre>
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4410 &lt;result&gt; = load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;]
4411 &lt;result&gt; = volatile load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;]
4412 !&lt;index&gt; = !{ i32 1 }
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4413</pre>
4414
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4415<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4416<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4417
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4418<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4419<p>The argument to the '<tt>load</tt>' instruction specifies the memory address
4420 from which to load. The pointer must point to
4421 a <a href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
4422 marked as <tt>volatile</tt>, then the optimizer is not allowed to modify the
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]4423 number or order of execution of this <tt>load</tt> with other <a
4424 href="#volatile">volatile operations</a>.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4425
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4426<p>The optional constant <tt>align</tt> argument specifies the alignment of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4427 operation (that is, the alignment of the memory address). A value of 0 or an
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4428 omitted <tt>align</tt> argument means that the operation has the preferential
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4429 alignment for the target. It is the responsibility of the code emitter to
4430 ensure that the alignment information is correct. Overestimating the
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4431 alignment results in undefined behavior. Underestimating the alignment may
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4432 produce less efficient code. An alignment of 1 is always safe.</p>
4433
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4434<p>The optional <tt>!nontemporal</tt> metadata must reference a single
4435 metatadata name &lt;index&gt; corresponding to a metadata node with
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]4436 one <tt>i32</tt> entry of value 1. The existence of
Bill Wendlingbc4024f2010-02-25 21:23:24 +0000[diff] [blame]4437 the <tt>!nontemporal</tt> metatadata on the instruction tells the optimizer
4438 and code generator that this load is not expected to be reused in the cache.
4439 The code generator may select special instructions to save cache bandwidth,
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]4440 such as the <tt>MOVNT</tt> instruction on x86.</p>
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4441
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4442<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4443<p>The location of memory pointed to is loaded. If the value being loaded is of
4444 scalar type then the number of bytes read does not exceed the minimum number
4445 of bytes needed to hold all bits of the type. For example, loading an
4446 <tt>i24</tt> reads at most three bytes. When loading a value of a type like
4447 <tt>i20</tt> with a size that is not an integral number of bytes, the result
4448 is undefined if the value was not originally written using a store of the
4449 same type.</p>
4450
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4451<h5>Examples:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4452<pre>
4453 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
4454 <a href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4455 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4456</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4457
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4458</div>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4459
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4460<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4461<h4>
4462 <a name="i_store">'<tt>store</tt>' Instruction</a>
4463</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4464
Reid Spencera89fb182006-11-09 21:18:01 +0000[diff] [blame]4465<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4466
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4467<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4468<pre>
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]4469 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>
4470 volatile store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;] <i>; yields {void}</i>
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4471</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4472
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4473<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4474<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4475
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4476<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4477<p>There are two arguments to the '<tt>store</tt>' instruction: a value to store
4478 and an address at which to store it. The type of the
4479 '<tt>&lt;pointer&gt;</tt>' operand must be a pointer to
4480 the <a href="#t_firstclass">first class</a> type of the
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]4481 '<tt>&lt;value&gt;</tt>' operand. If the <tt>store</tt> is marked as
4482 <tt>volatile</tt>, then the optimizer is not allowed to modify the number or
4483 order of execution of this <tt>store</tt> with other <a
4484 href="#volatile">volatile operations</a>.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4485
4486<p>The optional constant "align" argument specifies the alignment of the
4487 operation (that is, the alignment of the memory address). A value of 0 or an
4488 omitted "align" argument means that the operation has the preferential
4489 alignment for the target. It is the responsibility of the code emitter to
4490 ensure that the alignment information is correct. Overestimating the
4491 alignment results in an undefined behavior. Underestimating the alignment may
4492 produce less efficient code. An alignment of 1 is always safe.</p>
4493
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4494<p>The optional !nontemporal metadata must reference a single metatadata
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]4495 name &lt;index&gt; corresponding to a metadata node with one i32 entry of
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]4496 value 1. The existence of the !nontemporal metatadata on the
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4497 instruction tells the optimizer and code generator that this load is
4498 not expected to be reused in the cache. The code generator may
4499 select special instructions to save cache bandwidth, such as the
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]4500 MOVNT instruction on x86.</p>
David Greene9641d062010-02-16 20:50:18 +0000[diff] [blame]4501
4502
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]4503<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4504<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>' at the
4505 location specified by the '<tt>&lt;pointer&gt;</tt>' operand. If
4506 '<tt>&lt;value&gt;</tt>' is of scalar type then the number of bytes written
4507 does not exceed the minimum number of bytes needed to hold all bits of the
4508 type. For example, storing an <tt>i24</tt> writes at most three bytes. When
4509 writing a value of a type like <tt>i20</tt> with a size that is not an
4510 integral number of bytes, it is unspecified what happens to the extra bits
4511 that do not belong to the type, but they will typically be overwritten.</p>
4512
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4513<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4514<pre>
4515 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling8830ffe2007-10-22 05:10:05 +0000[diff] [blame]4516 store i32 3, i32* %ptr <i>; yields {void}</i>
4517 %val = <a href="#i_load">load</a> i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4518</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4519
Reid Spencer443460a2006-11-09 21:15:49 +0000[diff] [blame]4520</div>
4521
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]4522<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4523<h4>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4524 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4525</h4>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4526
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4527<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4528
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]4529<h5>Syntax:</h5>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4530<pre>
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]4531 &lt;result&gt; = getelementptr &lt;pty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
Dan Gohman1639c392009-07-27 21:53:46 +0000[diff] [blame]4532 &lt;result&gt; = getelementptr inbounds &lt;pty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4533</pre>
4534
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]4535<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4536<p>The '<tt>getelementptr</tt>' instruction is used to get the address of a
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4537 subelement of an <a href="#t_aggregate">aggregate</a> data structure.
4538 It performs address calculation only and does not access memory.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4539
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]4540<h5>Arguments:</h5>
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]4541<p>The first argument is always a pointer, and forms the basis of the
Chris Lattnera40b9122009-07-29 06:44:13 +0000[diff] [blame]4542 calculation. The remaining arguments are indices that indicate which of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4543 elements of the aggregate object are indexed. The interpretation of each
4544 index is dependent on the type being indexed into. The first index always
4545 indexes the pointer value given as the first argument, the second index
4546 indexes a value of the type pointed to (not necessarily the value directly
4547 pointed to, since the first index can be non-zero), etc. The first type
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4548 indexed into must be a pointer value, subsequent types can be arrays,
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]4549 vectors, and structs. Note that subsequent types being indexed into
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4550 can never be pointers, since that would require loading the pointer before
4551 continuing calculation.</p>
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]4552
4553<p>The type of each index argument depends on the type it is indexing into.
Chris Lattner13ee7952010-08-28 04:09:24 +0000[diff] [blame]4554 When indexing into a (optionally packed) structure, only <tt>i32</tt>
Chris Lattner392be582010-02-12 20:49:41 +0000[diff] [blame]4555 integer <b>constants</b> are allowed. When indexing into an array, pointer
4556 or vector, integers of any width are allowed, and they are not required to be
Chris Lattnera40b9122009-07-29 06:44:13 +0000[diff] [blame]4557 constant.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4558
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4559<p>For example, let's consider a C code fragment and how it gets compiled to
4560 LLVM:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4561
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]4562<pre class="doc_code">
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4563struct RT {
4564 char A;
Chris Lattnera446f1b2007-05-29 15:43:56 +0000[diff] [blame]4565 int B[10][20];
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4566 char C;
4567};
4568struct ST {
Chris Lattnera446f1b2007-05-29 15:43:56 +0000[diff] [blame]4569 int X;
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4570 double Y;
4571 struct RT Z;
4572};
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4573
Chris Lattnera446f1b2007-05-29 15:43:56 +0000[diff] [blame]4574int *foo(struct ST *s) {
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4575 return &amp;s[1].Z.B[5][13];
4576}
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4577</pre>
4578
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4579<p>The LLVM code generated by the GCC frontend is:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4580
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]4581<pre class="doc_code">
Chris Lattnerbc088212009-01-11 20:53:49 +0000[diff] [blame]4582%RT = <a href="#namedtypes">type</a> { i8 , [10 x [20 x i32]], i8 }
4583%ST = <a href="#namedtypes">type</a> { i32, double, %RT }
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4584
Dan Gohman6b867702009-07-25 02:23:48 +0000[diff] [blame]4585define i32* @foo(%ST* %s) {
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]4586entry:
4587 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
4588 ret i32* %reg
4589}
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4590</pre>
4591
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]4592<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4593<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4594 type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
4595 }</tt>' type, a structure. The second index indexes into the third element
4596 of the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
4597 i8 }</tt>' type, another structure. The third index indexes into the second
4598 element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
4599 array. The two dimensions of the array are subscripted into, yielding an
4600 '<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a
4601 pointer to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4602
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4603<p>Note that it is perfectly legal to index partially through a structure,
4604 returning a pointer to an inner element. Because of this, the LLVM code for
4605 the given testcase is equivalent to:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4606
4607<pre>
Dan Gohman6b867702009-07-25 02:23:48 +0000[diff] [blame]4608 define i32* @foo(%ST* %s) {
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4609 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
Jeff Cohen5819f182007-04-22 01:17:39 +0000[diff] [blame]4610 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
4611 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4612 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
4613 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
4614 ret i32* %t5
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4615 }
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]4616</pre>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +0000[diff] [blame]4617
Dan Gohman1639c392009-07-27 21:53:46 +0000[diff] [blame]4618<p>If the <tt>inbounds</tt> keyword is present, the result value of the
Dan Gohman57255802010-04-23 15:23:32 +0000[diff] [blame]4619 <tt>getelementptr</tt> is a <a href="#trapvalues">trap value</a> if the
4620 base pointer is not an <i>in bounds</i> address of an allocated object,
4621 or if any of the addresses that would be formed by successive addition of
4622 the offsets implied by the indices to the base address with infinitely
4623 precise arithmetic are not an <i>in bounds</i> address of that allocated
4624 object. The <i>in bounds</i> addresses for an allocated object are all
4625 the addresses that point into the object, plus the address one byte past
4626 the end.</p>
Dan Gohman1639c392009-07-27 21:53:46 +0000[diff] [blame]4627
4628<p>If the <tt>inbounds</tt> keyword is not present, the offsets are added to
4629 the base address with silently-wrapping two's complement arithmetic, and
4630 the result value of the <tt>getelementptr</tt> may be outside the object
4631 pointed to by the base pointer. The result value may not necessarily be
4632 used to access memory though, even if it happens to point into allocated
4633 storage. See the <a href="#pointeraliasing">Pointer Aliasing Rules</a>
4634 section for more information.</p>
4635
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4636<p>The getelementptr instruction is often confusing. For some more insight into
4637 how it works, see <a href="GetElementPtr.html">the getelementptr FAQ</a>.</p>
Chris Lattner6ab66722006-08-15 00:45:58 +0000[diff] [blame]4638
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]4639<h5>Example:</h5>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4640<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4641 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]4642 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
4643 <i>; yields i8*:vptr</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]4644 %vptr = getelementptr {i32, &lt;2 x i8&gt;}* %svptr, i64 0, i32 1, i32 1
Matthijs Kooijman0e268272008-10-13 13:44:15 +0000[diff] [blame]4645 <i>; yields i8*:eptr</i>
4646 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Sanjiv Gupta0c155e62009-04-25 07:27:44 +0000[diff] [blame]4647 <i>; yields i32*:iptr</i>
Sanjiv Gupta77abea02009-04-24 16:38:13 +0000[diff] [blame]4648 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4649</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4650
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]4651</div>
Reid Spencer443460a2006-11-09 21:15:49 +0000[diff] [blame]4652
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4653<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4654<h3>
4655 <a name="convertops">Conversion Operations</a>
4656</h3>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4657
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4658<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4659
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]4660<p>The instructions in this category are the conversion instructions (casting)
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4661 which all take a single operand and a type. They perform various bit
4662 conversions on the operand.</p>
4663
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4664</div>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]4665
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]4666<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4667<h4>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4668 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4669</h4>
4670
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4671<div class="doc_text">
4672
4673<h5>Syntax:</h5>
4674<pre>
4675 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4676</pre>
4677
4678<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4679<p>The '<tt>trunc</tt>' instruction truncates its operand to the
4680 type <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4681
4682<h5>Arguments:</h5>
Nadav Rotem502f1b92011-02-24 21:01:34 +0000[diff] [blame]4683<p>The '<tt>trunc</tt>' instruction takes a value to trunc, and a type to trunc it to.
4684 Both types must be of <a href="#t_integer">integer</a> types, or vectors
4685 of the same number of integers.
4686 The bit size of the <tt>value</tt> must be larger than
4687 the bit size of the destination type, <tt>ty2</tt>.
4688 Equal sized types are not allowed.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4689
4690<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4691<p>The '<tt>trunc</tt>' instruction truncates the high order bits
4692 in <tt>value</tt> and converts the remaining bits to <tt>ty2</tt>. Since the
4693 source size must be larger than the destination size, <tt>trunc</tt> cannot
4694 be a <i>no-op cast</i>. It will always truncate bits.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4695
4696<h5>Example:</h5>
4697<pre>
Nadav Rotem502f1b92011-02-24 21:01:34 +0000[diff] [blame]4698 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
4699 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
4700 %Z = trunc i32 122 to i1 <i>; yields i1:false</i>
4701 %W = trunc &lt;2 x i16&gt; &lt;i16 8, i16 7&gt; to &lt;2 x i8&gt; <i>; yields &lt;i8 8, i8 7&gt;</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4702</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4703
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4704</div>
4705
4706<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4707<h4>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4708 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4709</h4>
4710
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4711<div class="doc_text">
4712
4713<h5>Syntax:</h5>
4714<pre>
4715 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4716</pre>
4717
4718<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4719<p>The '<tt>zext</tt>' instruction zero extends its operand to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4720 <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4721
4722
4723<h5>Arguments:</h5>
Nadav Rotem25f2ac92011-02-20 12:37:50 +0000[diff] [blame]4724<p>The '<tt>zext</tt>' instruction takes a value to cast, and a type to cast it to.
4725 Both types must be of <a href="#t_integer">integer</a> types, or vectors
4726 of the same number of integers.
4727 The bit size of the <tt>value</tt> must be smaller than
4728 the bit size of the destination type,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4729 <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4730
4731<h5>Semantics:</h5>
4732<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4733 bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4734
Reid Spencer07c9c682007-01-12 15:46:11 +0000[diff] [blame]4735<p>When zero extending from i1, the result will always be either 0 or 1.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4736
4737<h5>Example:</h5>
4738<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4739 %X = zext i32 257 to i64 <i>; yields i64:257</i>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]4740 %Y = zext i1 true to i32 <i>; yields i32:1</i>
Nadav Rotem25f2ac92011-02-20 12:37:50 +0000[diff] [blame]4741 %Z = zext &lt;2 x i16&gt; &lt;i16 8, i16 7&gt; to &lt;2 x i32&gt; <i>; yields &lt;i32 8, i32 7&gt;</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4742</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4743
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4744</div>
4745
4746<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4747<h4>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4748 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4749</h4>
4750
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4751<div class="doc_text">
4752
4753<h5>Syntax:</h5>
4754<pre>
4755 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4756</pre>
4757
4758<h5>Overview:</h5>
4759<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
4760
4761<h5>Arguments:</h5>
Nadav Rotem502f1b92011-02-24 21:01:34 +0000[diff] [blame]4762<p>The '<tt>sext</tt>' instruction takes a value to cast, and a type to cast it to.
4763 Both types must be of <a href="#t_integer">integer</a> types, or vectors
4764 of the same number of integers.
4765 The bit size of the <tt>value</tt> must be smaller than
4766 the bit size of the destination type,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4767 <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4768
4769<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4770<p>The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
4771 bit (highest order bit) of the <tt>value</tt> until it reaches the bit size
4772 of the type <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4773
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]4774<p>When sign extending from i1, the extension always results in -1 or 0.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4775
4776<h5>Example:</h5>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4777<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4778 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]4779 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
Nadav Rotem502f1b92011-02-24 21:01:34 +0000[diff] [blame]4780 %Z = sext &lt;2 x i16&gt; &lt;i16 8, i16 7&gt; to &lt;2 x i32&gt; <i>; yields &lt;i32 8, i32 7&gt;</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4781</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4782
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4783</div>
4784
4785<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4786<h4>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4787 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4788</h4>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4789
4790<div class="doc_text">
4791
4792<h5>Syntax:</h5>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4793<pre>
4794 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4795</pre>
4796
4797<h5>Overview:</h5>
4798<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4799 <tt>ty2</tt>.</p>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4800
4801<h5>Arguments:</h5>
4802<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4803 point</a> value to cast and a <a href="#t_floating">floating point</a> type
4804 to cast it to. The size of <tt>value</tt> must be larger than the size of
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4805 <tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4806 <i>no-op cast</i>.</p>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4807
4808<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4809<p>The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4810 <a href="#t_floating">floating point</a> type to a smaller
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4811 <a href="#t_floating">floating point</a> type. If the value cannot fit
4812 within the destination type, <tt>ty2</tt>, then the results are
4813 undefined.</p>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4814
4815<h5>Example:</h5>
4816<pre>
4817 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
4818 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
4819</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4820
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]4821</div>
4822
4823<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4824<h4>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4825 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4826</h4>
4827
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4828<div class="doc_text">
4829
4830<h5>Syntax:</h5>
4831<pre>
4832 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4833</pre>
4834
4835<h5>Overview:</h5>
4836<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4837 floating point value.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4838
4839<h5>Arguments:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4840<p>The '<tt>fpext</tt>' instruction takes a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4841 <a href="#t_floating">floating point</a> <tt>value</tt> to cast, and
4842 a <a href="#t_floating">floating point</a> type to cast it to. The source
4843 type must be smaller than the destination type.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4844
4845<h5>Semantics:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4846<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4847 <a href="#t_floating">floating point</a> type to a larger
4848 <a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
4849 used to make a <i>no-op cast</i> because it always changes bits. Use
4850 <tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4851
4852<h5>Example:</h5>
4853<pre>
Nick Lewycky9feca672011-03-31 18:20:19 +0000[diff] [blame]4854 %X = fpext float 3.125 to double <i>; yields double:3.125000e+00</i>
4855 %Y = fpext double %X to fp128 <i>; yields fp128:0xL00000000000000004000900000000000</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4856</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4857
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4858</div>
4859
4860<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4861<h4>
Reid Spencer2eadb532007-01-21 00:29:26 +0000[diff] [blame]4862 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4863</h4>
4864
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4865<div class="doc_text">
4866
4867<h5>Syntax:</h5>
4868<pre>
Reid Spencer753163d2007-07-31 14:40:14 +0000[diff] [blame]4869 &lt;result&gt; = fptoui &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4870</pre>
4871
4872<h5>Overview:</h5>
Reid Spencer753163d2007-07-31 14:40:14 +0000[diff] [blame]4873<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4874 unsigned integer equivalent of type <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4875
4876<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4877<p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
4878 scalar or vector <a href="#t_floating">floating point</a> value, and a type
4879 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4880 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4881 vector integer type with the same number of elements as <tt>ty</tt></p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4882
4883<h5>Semantics:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4884<p>The '<tt>fptoui</tt>' instruction converts its
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4885 <a href="#t_floating">floating point</a> operand into the nearest (rounding
4886 towards zero) unsigned integer value. If the value cannot fit
4887 in <tt>ty2</tt>, the results are undefined.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4888
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4889<h5>Example:</h5>
4890<pre>
Reid Spencer753163d2007-07-31 14:40:14 +0000[diff] [blame]4891 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner5b95a172007-09-22 03:17:52 +0000[diff] [blame]4892 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Gabor Greiff50fd572009-10-28 09:21:30 +0000[diff] [blame]4893 %Z = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4894</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4895
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4896</div>
4897
4898<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4899<h4>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4900 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4901</h4>
4902
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4903<div class="doc_text">
4904
4905<h5>Syntax:</h5>
4906<pre>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4907 &lt;result&gt; = fptosi &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4908</pre>
4909
4910<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4911<p>The '<tt>fptosi</tt>' instruction converts
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4912 <a href="#t_floating">floating point</a> <tt>value</tt> to
4913 type <tt>ty2</tt>.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]4914
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]4915<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4916<p>The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
4917 scalar or vector <a href="#t_floating">floating point</a> value, and a type
4918 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4919 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4920 vector integer type with the same number of elements as <tt>ty</tt></p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]4921
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]4922<h5>Semantics:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]4923<p>The '<tt>fptosi</tt>' instruction converts its
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4924 <a href="#t_floating">floating point</a> operand into the nearest (rounding
4925 towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
4926 the results are undefined.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]4927
Chris Lattner70de6632001-07-09 00:26:23 +0000[diff] [blame]4928<h5>Example:</h5>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]4929<pre>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]4930 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner5b95a172007-09-22 03:17:52 +0000[diff] [blame]4931 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Gabor Greiff50fd572009-10-28 09:21:30 +0000[diff] [blame]4932 %Z = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4933</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4934
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4935</div>
4936
4937<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4938<h4>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4939 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4940</h4>
4941
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4942<div class="doc_text">
4943
4944<h5>Syntax:</h5>
4945<pre>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4946 &lt;result&gt; = uitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4947</pre>
4948
4949<h5>Overview:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4950<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4951 integer and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4952
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4953<h5>Arguments:</h5>
Nate Begemand4d45c22007-11-17 03:58:34 +0000[diff] [blame]4954<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4955 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
4956 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4957 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4958 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4959
4960<h5>Semantics:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4961<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4962 integer quantity and converts it to the corresponding floating point
4963 value. If the value cannot fit in the floating point value, the results are
4964 undefined.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4965
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4966<h5>Example:</h5>
4967<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]4968 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]4969 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4970</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4971
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4972</div>
4973
4974<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4975<h4>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4976 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]4977</h4>
4978
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4979<div class="doc_text">
4980
4981<h5>Syntax:</h5>
4982<pre>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]4983 &lt;result&gt; = sitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4984</pre>
4985
4986<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4987<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed integer
4988 and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4989
4990<h5>Arguments:</h5>
Nate Begemand4d45c22007-11-17 03:58:34 +0000[diff] [blame]4991<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be a
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4992 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
4993 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4994 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4995 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]4996
4997<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]4998<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed integer
4999 quantity and converts it to the corresponding floating point value. If the
5000 value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5001
5002<h5>Example:</h5>
5003<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5004 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]5005 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5006</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5007
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5008</div>
5009
5010<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5011<h4>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5012 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5013</h4>
5014
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5015<div class="doc_text">
5016
5017<h5>Syntax:</h5>
5018<pre>
5019 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5020</pre>
5021
5022<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5023<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
5024 the integer type <tt>ty2</tt>.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5025
5026<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5027<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
5028 must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
5029 <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> type.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5030
5031<h5>Semantics:</h5>
5032<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5033 <tt>ty2</tt> by interpreting the pointer value as an integer and either
5034 truncating or zero extending that value to the size of the integer type. If
5035 <tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
5036 <tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
5037 are the same size, then nothing is done (<i>no-op cast</i>) other than a type
5038 change.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5039
5040<h5>Example:</h5>
5041<pre>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5042 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
5043 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5044</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5045
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5046</div>
5047
5048<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5049<h4>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5050 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5051</h4>
5052
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5053<div class="doc_text">
5054
5055<h5>Syntax:</h5>
5056<pre>
5057 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5058</pre>
5059
5060<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5061<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to a
5062 pointer type, <tt>ty2</tt>.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5063
5064<h5>Arguments:</h5>
Duncan Sands16f122e2007-03-30 12:22:09 +0000[diff] [blame]5065<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5066 value to cast, and a type to cast it to, which must be a
5067 <a href="#t_pointer">pointer</a> type.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5068
5069<h5>Semantics:</h5>
5070<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5071 <tt>ty2</tt> by applying either a zero extension or a truncation depending on
5072 the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
5073 size of a pointer then a truncation is done. If <tt>value</tt> is smaller
5074 than the size of a pointer then a zero extension is done. If they are the
5075 same size, nothing is done (<i>no-op cast</i>).</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5076
5077<h5>Example:</h5>
5078<pre>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5079 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
Gabor Greiff50fd572009-10-28 09:21:30 +0000[diff] [blame]5080 %Y = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
5081 %Z = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5082</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5083
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]5084</div>
5085
5086<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5087<h4>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]5088 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5089</h4>
5090
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5091<div class="doc_text">
5092
5093<h5>Syntax:</h5>
5094<pre>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]5095 &lt;result&gt; = bitcast &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5096</pre>
5097
5098<h5>Overview:</h5>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]5099<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5100 <tt>ty2</tt> without changing any bits.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5101
5102<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5103<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be a
5104 non-aggregate first class value, and a type to cast it to, which must also be
5105 a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes
5106 of <tt>value</tt> and the destination type, <tt>ty2</tt>, must be
5107 identical. If the source type is a pointer, the destination type must also be
5108 a pointer. This instruction supports bitwise conversion of vectors to
5109 integers and to vectors of other types (as long as they have the same
5110 size).</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5111
5112<h5>Semantics:</h5>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]5113<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5114 <tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
5115 this conversion. The conversion is done as if the <tt>value</tt> had been
5116 stored to memory and read back as type <tt>ty2</tt>. Pointer types may only
5117 be converted to other pointer types with this instruction. To convert
5118 pointers to other types, use the <a href="#i_inttoptr">inttoptr</a> or
5119 <a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]5120
5121<h5>Example:</h5>
5122<pre>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5123 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5124 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5125 %Z = bitcast &lt;2 x int&gt; %V to i64; <i>; yields i64: %V</i>
Chris Lattner70de6632001-07-09 00:26:23 +0000[diff] [blame]5126</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5127
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5128</div>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5129
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5130<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5131<h3>
5132 <a name="otherops">Other Operations</a>
5133</h3>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5134
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5135<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5136
5137<p>The instructions in this category are the "miscellaneous" instructions, which
5138 defy better classification.</p>
5139
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5140</div>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5141
5142<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5143<h4>
5144 <a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
5145</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5146
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5147<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5148
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5149<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5150<pre>
5151 &lt;result&gt; = icmp &lt;cond&gt; &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {i1} or {&lt;N x i1&gt;}:result</i>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5152</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5153
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5154<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5155<p>The '<tt>icmp</tt>' instruction returns a boolean value or a vector of
5156 boolean values based on comparison of its two integer, integer vector, or
5157 pointer operands.</p>
5158
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5159<h5>Arguments:</h5>
5160<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5161 the condition code indicating the kind of comparison to perform. It is not a
5162 value, just a keyword. The possible condition code are:</p>
5163
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5164<ol>
5165 <li><tt>eq</tt>: equal</li>
5166 <li><tt>ne</tt>: not equal </li>
5167 <li><tt>ugt</tt>: unsigned greater than</li>
5168 <li><tt>uge</tt>: unsigned greater or equal</li>
5169 <li><tt>ult</tt>: unsigned less than</li>
5170 <li><tt>ule</tt>: unsigned less or equal</li>
5171 <li><tt>sgt</tt>: signed greater than</li>
5172 <li><tt>sge</tt>: signed greater or equal</li>
5173 <li><tt>slt</tt>: signed less than</li>
5174 <li><tt>sle</tt>: signed less or equal</li>
5175</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5176
Chris Lattnerc0f423a2007-01-15 01:54:13 +0000[diff] [blame]5177<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5178 <a href="#t_pointer">pointer</a> or integer <a href="#t_vector">vector</a>
5179 typed. They must also be identical types.</p>
5180
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5181<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5182<p>The '<tt>icmp</tt>' compares <tt>op1</tt> and <tt>op2</tt> according to the
5183 condition code given as <tt>cond</tt>. The comparison performed always yields
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]5184 either an <a href="#t_integer"><tt>i1</tt></a> or vector of <tt>i1</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5185 result, as follows:</p>
5186
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5187<ol>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5188 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5189 <tt>false</tt> otherwise. No sign interpretation is necessary or
5190 performed.</li>
5191
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5192 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5193 <tt>false</tt> otherwise. No sign interpretation is necessary or
5194 performed.</li>
5195
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5196 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5197 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5198
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5199 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5200 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5201 to <tt>op2</tt>.</li>
5202
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5203 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5204 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5205
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5206 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5207 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5208
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5209 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5210 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5211
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5212 <li><tt>sge</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5213 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5214 to <tt>op2</tt>.</li>
5215
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5216 <li><tt>slt</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5217 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5218
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5219 <li><tt>sle</tt>: interprets the operands as signed values and yields
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5220 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5221</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5222
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5223<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5224 values are compared as if they were integers.</p>
5225
5226<p>If the operands are integer vectors, then they are compared element by
5227 element. The result is an <tt>i1</tt> vector with the same number of elements
5228 as the values being compared. Otherwise, the result is an <tt>i1</tt>.</p>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5229
5230<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5231<pre>
5232 &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5233 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
5234 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
5235 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
5236 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
5237 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5238</pre>
Dan Gohmana5127ab2009-01-22 01:39:38 +0000[diff] [blame]5239
5240<p>Note that the code generator does not yet support vector types with
5241 the <tt>icmp</tt> instruction.</p>
5242
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5243</div>
5244
5245<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5246<h4>
5247 <a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
5248</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5249
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5250<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5251
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5252<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5253<pre>
5254 &lt;result&gt; = fcmp &lt;cond&gt; &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {i1} or {&lt;N x i1&gt;}:result</i>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5255</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5256
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5257<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5258<p>The '<tt>fcmp</tt>' instruction returns a boolean value or vector of boolean
5259 values based on comparison of its operands.</p>
5260
5261<p>If the operands are floating point scalars, then the result type is a boolean
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]5262(<a href="#t_integer"><tt>i1</tt></a>).</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5263
5264<p>If the operands are floating point vectors, then the result type is a vector
5265 of boolean with the same number of elements as the operands being
5266 compared.</p>
5267
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5268<h5>Arguments:</h5>
5269<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5270 the condition code indicating the kind of comparison to perform. It is not a
5271 value, just a keyword. The possible condition code are:</p>
5272
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5273<ol>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5274 <li><tt>false</tt>: no comparison, always returns false</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5275 <li><tt>oeq</tt>: ordered and equal</li>
5276 <li><tt>ogt</tt>: ordered and greater than </li>
5277 <li><tt>oge</tt>: ordered and greater than or equal</li>
5278 <li><tt>olt</tt>: ordered and less than </li>
5279 <li><tt>ole</tt>: ordered and less than or equal</li>
5280 <li><tt>one</tt>: ordered and not equal</li>
5281 <li><tt>ord</tt>: ordered (no nans)</li>
5282 <li><tt>ueq</tt>: unordered or equal</li>
5283 <li><tt>ugt</tt>: unordered or greater than </li>
5284 <li><tt>uge</tt>: unordered or greater than or equal</li>
5285 <li><tt>ult</tt>: unordered or less than </li>
5286 <li><tt>ule</tt>: unordered or less than or equal</li>
5287 <li><tt>une</tt>: unordered or not equal</li>
5288 <li><tt>uno</tt>: unordered (either nans)</li>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5289 <li><tt>true</tt>: no comparison, always returns true</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5290</ol>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5291
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5292<p><i>Ordered</i> means that neither operand is a QNAN while
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5293 <i>unordered</i> means that either operand may be a QNAN.</p>
5294
5295<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be either
5296 a <a href="#t_floating">floating point</a> type or
5297 a <a href="#t_vector">vector</a> of floating point type. They must have
5298 identical types.</p>
5299
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5300<h5>Semantics:</h5>
Gabor Greif0f75ad02008-08-07 21:46:00 +0000[diff] [blame]5301<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5302 according to the condition code given as <tt>cond</tt>. If the operands are
5303 vectors, then the vectors are compared element by element. Each comparison
Nick Lewycky84a1eeb2009-09-27 00:45:11 +0000[diff] [blame]5304 performed always yields an <a href="#t_integer">i1</a> result, as
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5305 follows:</p>
5306
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5307<ol>
5308 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5309
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5310 <li><tt>oeq</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5311 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5312
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5313 <li><tt>ogt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]5314 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5315
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5316 <li><tt>oge</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5317 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5318
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5319 <li><tt>olt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5320 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5321
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5322 <li><tt>ole</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5323 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5324
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5325 <li><tt>one</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5326 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5327
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5328 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5329
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5330 <li><tt>ueq</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5331 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5332
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5333 <li><tt>ugt</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5334 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5335
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5336 <li><tt>uge</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5337 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5338
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5339 <li><tt>ult</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5340 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5341
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5342 <li><tt>ule</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5343 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5344
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5345 <li><tt>une</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5346 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5347
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]5348 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5349
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5350 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
5351</ol>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5352
5353<h5>Example:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5354<pre>
5355 &lt;result&gt; = fcmp oeq float 4.0, 5.0 <i>; yields: result=false</i>
Dan Gohmanc579d972008-09-09 01:02:47 +0000[diff] [blame]5356 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
5357 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
5358 &lt;result&gt; = fcmp ueq double 1.0, 2.0 <i>; yields: result=false</i>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5359</pre>
Dan Gohmana5127ab2009-01-22 01:39:38 +0000[diff] [blame]5360
5361<p>Note that the code generator does not yet support vector types with
5362 the <tt>fcmp</tt> instruction.</p>
5363
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]5364</div>
5365
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5366<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5367<h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5368 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5369</h4>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5370
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5371<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5372
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5373<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5374<pre>
5375 &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...
5376</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5377
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5378<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5379<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in the
5380 SSA graph representing the function.</p>
5381
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5382<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5383<p>The type of the incoming values is specified with the first type field. After
5384 this, the '<tt>phi</tt>' instruction takes a list of pairs as arguments, with
5385 one pair for each predecessor basic block of the current block. Only values
5386 of <a href="#t_firstclass">first class</a> type may be used as the value
5387 arguments to the PHI node. Only labels may be used as the label
5388 arguments.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5389
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5390<p>There must be no non-phi instructions between the start of a basic block and
5391 the PHI instructions: i.e. PHI instructions must be first in a basic
5392 block.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5393
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5394<p>For the purposes of the SSA form, the use of each incoming value is deemed to
5395 occur on the edge from the corresponding predecessor block to the current
5396 block (but after any definition of an '<tt>invoke</tt>' instruction's return
5397 value on the same edge).</p>
Jay Foad1a4eea52009-06-03 10:20:10 +0000[diff] [blame]5398
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5399<h5>Semantics:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5400<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5401 specified by the pair corresponding to the predecessor basic block that
5402 executed just prior to the current block.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5403
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5404<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +0000[diff] [blame]5405<pre>
5406Loop: ; Infinite loop that counts from 0 on up...
5407 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
5408 %nextindvar = add i32 %indvar, 1
5409 br label %Loop
5410</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5411
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]5412</div>
5413
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5414<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5415<h4>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5416 <a name="i_select">'<tt>select</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5417</h4>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5418
5419<div class="doc_text">
5420
5421<h5>Syntax:</h5>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5422<pre>
Dan Gohmanc579d972008-09-09 01:02:47 +0000[diff] [blame]5423 &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>
5424
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]5425 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5426</pre>
5427
5428<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5429<p>The '<tt>select</tt>' instruction is used to choose one value based on a
5430 condition, without branching.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5431
5432
5433<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5434<p>The '<tt>select</tt>' instruction requires an 'i1' value or a vector of 'i1'
5435 values indicating the condition, and two values of the
5436 same <a href="#t_firstclass">first class</a> type. If the val1/val2 are
5437 vectors and the condition is a scalar, then entire vectors are selected, not
5438 individual elements.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5439
5440<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5441<p>If the condition is an i1 and it evaluates to 1, the instruction returns the
5442 first value argument; otherwise, it returns the second value argument.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5443
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5444<p>If the condition is a vector of i1, then the value arguments must be vectors
5445 of the same size, and the selection is done element by element.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5446
5447<h5>Example:</h5>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5448<pre>
Reid Spencer36a15422007-01-12 03:35:51 +0000[diff] [blame]5449 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5450</pre>
Dan Gohmana5127ab2009-01-22 01:39:38 +0000[diff] [blame]5451
5452<p>Note that the code generator does not yet support conditions
5453 with vector type.</p>
5454
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]5455</div>
5456
Robert Bocchinof72fdfe2006-01-15 20:48:27 +0000[diff] [blame]5457<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5458<h4>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5459 <a name="i_call">'<tt>call</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5460</h4>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5461
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5462<div class="doc_text">
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5463
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5464<h5>Syntax:</h5>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5465<pre>
Devang Patel02256232008-10-07 17:48:33 +0000[diff] [blame]5466 &lt;result&gt; = [tail] call [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>] &lt;ty&gt; [&lt;fnty&gt;*] &lt;fnptrval&gt;(&lt;function args&gt;) [<a href="#fnattrs">fn attrs</a>]
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5467</pre>
5468
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5469<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5470<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5471
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5472<h5>Arguments:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5473<p>This instruction requires several arguments:</p>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5474
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]5475<ol>
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]5476 <li>The optional "tail" marker indicates that the callee function does not
5477 access any allocas or varargs in the caller. Note that calls may be
5478 marked "tail" even if they do not occur before
5479 a <a href="#i_ret"><tt>ret</tt></a> instruction. If the "tail" marker is
5480 present, the function call is eligible for tail call optimization,
5481 but <a href="CodeGenerator.html#tailcallopt">might not in fact be
Evan Cheng59676492010-03-08 21:05:02 +0000[diff] [blame]5482 optimized into a jump</a>. The code generator may optimize calls marked
5483 "tail" with either 1) automatic <a href="CodeGenerator.html#sibcallopt">
5484 sibling call optimization</a> when the caller and callee have
5485 matching signatures, or 2) forced tail call optimization when the
5486 following extra requirements are met:
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]5487 <ul>
5488 <li>Caller and callee both have the calling
5489 convention <tt>fastcc</tt>.</li>
5490 <li>The call is in tail position (ret immediately follows call and ret
5491 uses value of call or is void).</li>
5492 <li>Option <tt>-tailcallopt</tt> is enabled,
Dan Gohman6232f732010-03-02 01:08:11 +0000[diff] [blame]5493 or <code>llvm::GuaranteedTailCallOpt</code> is <code>true</code>.</li>
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]5494 <li><a href="CodeGenerator.html#tailcallopt">Platform specific
5495 constraints are met.</a></li>
5496 </ul>
5497 </li>
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]5498
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5499 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
5500 convention</a> the call should use. If none is specified, the call
Jeffrey Yasskinb8677462010-01-09 19:44:16 +0000[diff] [blame]5501 defaults to using C calling conventions. The calling convention of the
5502 call must match the calling convention of the target function, or else the
5503 behavior is undefined.</li>
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]5504
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5505 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
5506 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
5507 '<tt>inreg</tt>' attributes are valid here.</li>
5508
5509 <li>'<tt>ty</tt>': the type of the call instruction itself which is also the
5510 type of the return value. Functions that return no value are marked
5511 <tt><a href="#t_void">void</a></tt>.</li>
5512
5513 <li>'<tt>fnty</tt>': shall be the signature of the pointer to function value
5514 being invoked. The argument types must match the types implied by this
5515 signature. This type can be omitted if the function is not varargs and if
5516 the function type does not return a pointer to a function.</li>
5517
5518 <li>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
5519 be invoked. In most cases, this is a direct function invocation, but
5520 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
5521 to function value.</li>
5522
5523 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner47f2a832010-03-02 06:36:51 +0000[diff] [blame]5524 signature argument types and parameter attributes. All arguments must be
5525 of <a href="#t_firstclass">first class</a> type. If the function
5526 signature indicates the function accepts a variable number of arguments,
5527 the extra arguments can be specified.</li>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5528
5529 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
5530 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
5531 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]5532</ol>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5533
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5534<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5535<p>The '<tt>call</tt>' instruction is used to cause control flow to transfer to
5536 a specified function, with its incoming arguments bound to the specified
5537 values. Upon a '<tt><a href="#i_ret">ret</a></tt>' instruction in the called
5538 function, control flow continues with the instruction after the function
5539 call, and the return value of the function is bound to the result
5540 argument.</p>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5541
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]5542<h5>Example:</h5>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5543<pre>
Nick Lewyckya9b13d52007-09-08 13:57:50 +0000[diff] [blame]5544 %retval = call i32 @test(i32 %argc)
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]5545 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) <i>; yields i32</i>
Chris Lattnerfb7c88d2008-03-21 17:24:17 +0000[diff] [blame]5546 %X = tail call i32 @foo() <i>; yields i32</i>
5547 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
5548 call void %foo(i8 97 signext)
Devang Pateld6cff512008-03-10 20:49:15 +0000[diff] [blame]5549
5550 %struct.A = type { i32, i8 }
Devang Patel7e9b05e2008-10-06 18:50:38 +0000[diff] [blame]5551 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohmancc3132e2008-10-04 19:00:07 +0000[diff] [blame]5552 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
5553 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattner6cbe8e92008-10-08 06:26:11 +0000[diff] [blame]5554 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijmaneefa7df2008-10-07 10:03:45 +0000[diff] [blame]5555 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]5556</pre>
5557
Dale Johannesen68f971b2009-09-24 18:38:21 +0000[diff] [blame]5558<p>llvm treats calls to some functions with names and arguments that match the
Dale Johannesen722212d2009-09-25 17:04:42 +0000[diff] [blame]5559standard C99 library as being the C99 library functions, and may perform
5560optimizations or generate code for them under that assumption. This is
5561something we'd like to change in the future to provide better support for
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]5562freestanding environments and non-C-based languages.</p>
Dale Johannesen68f971b2009-09-24 18:38:21 +0000[diff] [blame]5563
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5564</div>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5565
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5566<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5567<h4>
Chris Lattner33337472006-01-13 23:26:01 +0000[diff] [blame]5568 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5569</h4>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5570
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5571<div class="doc_text">
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5572
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]5573<h5>Syntax:</h5>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5574<pre>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]5575 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5576</pre>
5577
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]5578<h5>Overview:</h5>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]5579<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5580 the "variable argument" area of a function call. It is used to implement the
5581 <tt>va_arg</tt> macro in C.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5582
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]5583<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5584<p>This instruction takes a <tt>va_list*</tt> value and the type of the
5585 argument. It returns a value of the specified argument type and increments
5586 the <tt>va_list</tt> to point to the next argument. The actual type
5587 of <tt>va_list</tt> is target specific.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5588
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]5589<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5590<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified type
5591 from the specified <tt>va_list</tt> and causes the <tt>va_list</tt> to point
5592 to the next argument. For more information, see the variable argument
5593 handling <a href="#int_varargs">Intrinsic Functions</a>.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5594
5595<p>It is legal for this instruction to be called in a function which does not
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5596 take a variable number of arguments, for example, the <tt>vfprintf</tt>
5597 function.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5598
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5599<p><tt>va_arg</tt> is an LLVM instruction instead of
5600 an <a href="#intrinsics">intrinsic function</a> because it takes a type as an
5601 argument.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5602
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]5603<h5>Example:</h5>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]5604<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
5605
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5606<p>Note that the code generator does not yet fully support va_arg on many
5607 targets. Also, it does not currently support va_arg with aggregate types on
5608 any target.</p>
Dan Gohman3065b612009-01-12 23:12:39 +0000[diff] [blame]5609
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5610</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5611
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5612<!-- *********************************************************************** -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5613<h2><a name="intrinsics">Intrinsic Functions</a></h2>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]5614<!-- *********************************************************************** -->
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5615
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5616<div class="doc_text">
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5617
5618<p>LLVM supports the notion of an "intrinsic function". These functions have
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5619 well known names and semantics and are required to follow certain
5620 restrictions. Overall, these intrinsics represent an extension mechanism for
5621 the LLVM language that does not require changing all of the transformations
5622 in LLVM when adding to the language (or the bitcode reader/writer, the
5623 parser, etc...).</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5624
John Criswell88190562005-05-16 16:17:45 +0000[diff] [blame]5625<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5626 prefix is reserved in LLVM for intrinsic names; thus, function names may not
5627 begin with this prefix. Intrinsic functions must always be external
5628 functions: you cannot define the body of intrinsic functions. Intrinsic
5629 functions may only be used in call or invoke instructions: it is illegal to
5630 take the address of an intrinsic function. Additionally, because intrinsic
5631 functions are part of the LLVM language, it is required if any are added that
5632 they be documented here.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5633
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5634<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents a
5635 family of functions that perform the same operation but on different data
5636 types. Because LLVM can represent over 8 million different integer types,
5637 overloading is used commonly to allow an intrinsic function to operate on any
5638 integer type. One or more of the argument types or the result type can be
5639 overloaded to accept any integer type. Argument types may also be defined as
5640 exactly matching a previous argument's type or the result type. This allows
5641 an intrinsic function which accepts multiple arguments, but needs all of them
5642 to be of the same type, to only be overloaded with respect to a single
5643 argument or the result.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5644
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5645<p>Overloaded intrinsics will have the names of its overloaded argument types
5646 encoded into its function name, each preceded by a period. Only those types
5647 which are overloaded result in a name suffix. Arguments whose type is matched
5648 against another type do not. For example, the <tt>llvm.ctpop</tt> function
5649 can take an integer of any width and returns an integer of exactly the same
5650 integer width. This leads to a family of functions such as
5651 <tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29
5652 %val)</tt>. Only one type, the return type, is overloaded, and only one type
5653 suffix is required. Because the argument's type is matched against the return
5654 type, it does not require its own name suffix.</p>
Reid Spencer4eefaab2007-04-01 08:04:23 +0000[diff] [blame]5655
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]5656<p>To learn how to add an intrinsic function, please see the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5657 <a href="ExtendingLLVM.html">Extending LLVM Guide</a>.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5658
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5659</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5660
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5661<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5662<h3>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5663 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5664</h3>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5665
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5666<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5667
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5668<p>Variable argument support is defined in LLVM with
5669 the <a href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
5670 intrinsic functions. These functions are related to the similarly named
5671 macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5672
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5673<p>All of these functions operate on arguments that use a target-specific value
5674 type "<tt>va_list</tt>". The LLVM assembly language reference manual does
5675 not define what this type is, so all transformations should be prepared to
5676 handle these functions regardless of the type used.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5677
Chris Lattner30b868d2006-05-15 17:26:46 +0000[diff] [blame]5678<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5679 instruction and the variable argument handling intrinsic functions are
5680 used.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5681
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]5682<pre class="doc_code">
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]5683define i32 @test(i32 %X, ...) {
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5684 ; Initialize variable argument processing
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5685 %ap = alloca i8*
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]5686 %ap2 = bitcast i8** %ap to i8*
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]5687 call void @llvm.va_start(i8* %ap2)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5688
5689 ; Read a single integer argument
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5690 %tmp = va_arg i8** %ap, i32
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5691
5692 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5693 %aq = alloca i8*
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]5694 %aq2 = bitcast i8** %aq to i8*
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5695 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]5696 call void @llvm.va_end(i8* %aq2)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5697
5698 ; Stop processing of arguments.
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]5699 call void @llvm.va_end(i8* %ap2)
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000[diff] [blame]5700 ret i32 %tmp
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5701}
Anton Korobeynikov640bbe02007-03-21 23:58:04 +0000[diff] [blame]5702
5703declare void @llvm.va_start(i8*)
5704declare void @llvm.va_copy(i8*, i8*)
5705declare void @llvm.va_end(i8*)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5706</pre>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5707
Bill Wendling3716c5d2007-05-29 09:04:49 +0000[diff] [blame]5708</div>
5709
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5710<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5711<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5712 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5713</h4>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5714
5715
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5716<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5717
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5718<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5719<pre>
5720 declare void %llvm.va_start(i8* &lt;arglist&gt;)
5721</pre>
5722
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5723<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5724<p>The '<tt>llvm.va_start</tt>' intrinsic initializes <tt>*&lt;arglist&gt;</tt>
5725 for subsequent use by <tt><a href="#i_va_arg">va_arg</a></tt>.</p>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]5726
5727<h5>Arguments:</h5>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]5728<p>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]5729
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5730<h5>Semantics:</h5>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000[diff] [blame]5731<p>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5732 macro available in C. In a target-dependent way, it initializes
5733 the <tt>va_list</tt> element to which the argument points, so that the next
5734 call to <tt>va_arg</tt> will produce the first variable argument passed to
5735 the function. Unlike the C <tt>va_start</tt> macro, this intrinsic does not
5736 need to know the last argument of the function as the compiler can figure
5737 that out.</p>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]5738
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5739</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5740
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5741<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5742<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5743 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5744</h4>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5745
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5746<div class="doc_text">
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]5747
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5748<h5>Syntax:</h5>
5749<pre>
5750 declare void @llvm.va_end(i8* &lt;arglist&gt;)
5751</pre>
5752
5753<h5>Overview:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5754<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5755 which has been initialized previously
5756 with <tt><a href="#int_va_start">llvm.va_start</a></tt>
5757 or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]5758
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5759<h5>Arguments:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5760<p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]5761
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5762<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5763<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5764 macro available in C. In a target-dependent way, it destroys
5765 the <tt>va_list</tt> element to which the argument points. Calls
5766 to <a href="#int_va_start"><tt>llvm.va_start</tt></a>
5767 and <a href="#int_va_copy"> <tt>llvm.va_copy</tt></a> must be matched exactly
5768 with calls to <tt>llvm.va_end</tt>.</p>
Chris Lattnerdb0790c2007-01-08 07:55:15 +0000[diff] [blame]5769
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5770</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5771
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5772<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5773<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5774 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5775</h4>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5776
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5777<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5778
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5779<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5780<pre>
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]5781 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5782</pre>
5783
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5784<h5>Overview:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5785<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5786 from the source argument list to the destination argument list.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5787
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5788<h5>Arguments:</h5>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]5789<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5790 The second argument is a pointer to a <tt>va_list</tt> element to copy
5791 from.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5792
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]5793<h5>Semantics:</h5>
Jeff Cohen222a8a42007-04-29 01:07:00 +0000[diff] [blame]5794<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5795 macro available in C. In a target-dependent way, it copies the
5796 source <tt>va_list</tt> element into the destination <tt>va_list</tt>
5797 element. This intrinsic is necessary because
5798 the <tt><a href="#int_va_start"> llvm.va_start</a></tt> intrinsic may be
5799 arbitrarily complex and require, for example, memory allocation.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5800
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]5801</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]5802
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]5803<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5804<h3>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5805 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5806</h3>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5807
5808<div class="doc_text">
5809
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5810<p>LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattner67c37d12008-08-05 18:29:16 +0000[diff] [blame]5811Collection</a> (GC) requires the implementation and generation of these
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5812intrinsics. These intrinsics allow identification of <a href="#int_gcroot">GC
5813roots on the stack</a>, as well as garbage collector implementations that
5814require <a href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a>
5815barriers. Front-ends for type-safe garbage collected languages should generate
5816these intrinsics to make use of the LLVM garbage collectors. For more details,
5817see <a href="GarbageCollection.html">Accurate Garbage Collection with
5818LLVM</a>.</p>
Christopher Lamb55c6d4f2007-12-17 01:00:21 +0000[diff] [blame]5819
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5820<p>The garbage collection intrinsics only operate on objects in the generic
5821 address space (address space zero).</p>
Christopher Lamb55c6d4f2007-12-17 01:00:21 +0000[diff] [blame]5822
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5823</div>
5824
5825<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5826<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5827 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5828</h4>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5829
5830<div class="doc_text">
5831
5832<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5833<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]5834 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5835</pre>
5836
5837<h5>Overview:</h5>
John Criswelldfe6a862004-12-10 15:51:16 +0000[diff] [blame]5838<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5839 the code generator, and allows some metadata to be associated with it.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5840
5841<h5>Arguments:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5842<p>The first argument specifies the address of a stack object that contains the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5843 root pointer. The second pointer (which must be either a constant or a
5844 global value address) contains the meta-data to be associated with the
5845 root.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5846
5847<h5>Semantics:</h5>
Chris Lattner851b7712008-04-24 05:59:56 +0000[diff] [blame]5848<p>At runtime, a call to this intrinsic stores a null pointer into the "ptrloc"
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5849 location. At compile-time, the code generator generates information to allow
5850 the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
5851 intrinsic may only be used in a function which <a href="#gc">specifies a GC
5852 algorithm</a>.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5853
5854</div>
5855
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5856<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5857<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5858 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5859</h4>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5860
5861<div class="doc_text">
5862
5863<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5864<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]5865 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5866</pre>
5867
5868<h5>Overview:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5869<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5870 locations, allowing garbage collector implementations that require read
5871 barriers.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5872
5873<h5>Arguments:</h5>
Chris Lattnerf9228072006-03-14 20:02:51 +0000[diff] [blame]5874<p>The second argument is the address to read from, which should be an address
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5875 allocated from the garbage collector. The first object is a pointer to the
5876 start of the referenced object, if needed by the language runtime (otherwise
5877 null).</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5878
5879<h5>Semantics:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5880<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5881 instruction, but may be replaced with substantially more complex code by the
5882 garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
5883 may only be used in a function which <a href="#gc">specifies a GC
5884 algorithm</a>.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5885
5886</div>
5887
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5888<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5889<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5890 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5891</h4>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5892
5893<div class="doc_text">
5894
5895<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5896<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]5897 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5898</pre>
5899
5900<h5>Overview:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5901<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5902 locations, allowing garbage collector implementations that require write
5903 barriers (such as generational or reference counting collectors).</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5904
5905<h5>Arguments:</h5>
Chris Lattnerf9228072006-03-14 20:02:51 +0000[diff] [blame]5906<p>The first argument is the reference to store, the second is the start of the
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5907 object to store it to, and the third is the address of the field of Obj to
5908 store to. If the runtime does not require a pointer to the object, Obj may
5909 be null.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5910
5911<h5>Semantics:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5912<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5913 instruction, but may be replaced with substantially more complex code by the
5914 garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
5915 may only be used in a function which <a href="#gc">specifies a GC
5916 algorithm</a>.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5917
5918</div>
5919
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]5920<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5921<h3>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5922 <a name="int_codegen">Code Generator Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5923</h3>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5924
5925<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5926
5927<p>These intrinsics are provided by LLVM to expose special features that may
5928 only be implemented with code generator support.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5929
5930</div>
5931
5932<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5933<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5934 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5935</h4>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5936
5937<div class="doc_text">
5938
5939<h5>Syntax:</h5>
5940<pre>
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]5941 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5942</pre>
5943
5944<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5945<p>The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
5946 target-specific value indicating the return address of the current function
5947 or one of its callers.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5948
5949<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5950<p>The argument to this intrinsic indicates which function to return the address
5951 for. Zero indicates the calling function, one indicates its caller, etc.
5952 The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5953
5954<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5955<p>The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer
5956 indicating the return address of the specified call frame, or zero if it
5957 cannot be identified. The value returned by this intrinsic is likely to be
5958 incorrect or 0 for arguments other than zero, so it should only be used for
5959 debugging purposes.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5960
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5961<p>Note that calling this intrinsic does not prevent function inlining or other
5962 aggressive transformations, so the value returned may not be that of the
5963 obvious source-language caller.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5964
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5965</div>
5966
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5967<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5968<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]5969 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]5970</h4>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5971
5972<div class="doc_text">
5973
5974<h5>Syntax:</h5>
5975<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]5976 declare i8* @llvm.frameaddress(i32 &lt;level&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5977</pre>
5978
5979<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5980<p>The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
5981 target-specific frame pointer value for the specified stack frame.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5982
5983<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5984<p>The argument to this intrinsic indicates which function to return the frame
5985 pointer for. Zero indicates the calling function, one indicates its caller,
5986 etc. The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5987
5988<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5989<p>The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer
5990 indicating the frame address of the specified call frame, or zero if it
5991 cannot be identified. The value returned by this intrinsic is likely to be
5992 incorrect or 0 for arguments other than zero, so it should only be used for
5993 debugging purposes.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5994
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]5995<p>Note that calling this intrinsic does not prevent function inlining or other
5996 aggressive transformations, so the value returned may not be that of the
5997 obvious source-language caller.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5998
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]5999</div>
6000
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6001<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6002<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6003 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6004</h4>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6005
6006<div class="doc_text">
6007
6008<h5>Syntax:</h5>
6009<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6010 declare i8* @llvm.stacksave()
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6011</pre>
6012
6013<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6014<p>The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state
6015 of the function stack, for use
6016 with <a href="#int_stackrestore"> <tt>llvm.stackrestore</tt></a>. This is
6017 useful for implementing language features like scoped automatic variable
6018 sized arrays in C99.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6019
6020<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6021<p>This intrinsic returns a opaque pointer value that can be passed
6022 to <a href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When
6023 an <tt>llvm.stackrestore</tt> intrinsic is executed with a value saved
6024 from <tt>llvm.stacksave</tt>, it effectively restores the state of the stack
6025 to the state it was in when the <tt>llvm.stacksave</tt> intrinsic executed.
6026 In practice, this pops any <a href="#i_alloca">alloca</a> blocks from the
6027 stack that were allocated after the <tt>llvm.stacksave</tt> was executed.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6028
6029</div>
6030
6031<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6032<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6033 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6034</h4>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6035
6036<div class="doc_text">
6037
6038<h5>Syntax:</h5>
6039<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6040 declare void @llvm.stackrestore(i8* %ptr)
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6041</pre>
6042
6043<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6044<p>The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
6045 the function stack to the state it was in when the
6046 corresponding <a href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic
6047 executed. This is useful for implementing language features like scoped
6048 automatic variable sized arrays in C99.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6049
6050<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6051<p>See the description
6052 for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.</p>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6053
6054</div>
6055
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]6056<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6057<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6058 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6059</h4>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6060
6061<div class="doc_text">
6062
6063<h5>Syntax:</h5>
6064<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]6065 declare void @llvm.prefetch(i8* &lt;address&gt;, i32 &lt;rw&gt;, i32 &lt;locality&gt;)
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6066</pre>
6067
6068<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6069<p>The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to
6070 insert a prefetch instruction if supported; otherwise, it is a noop.
6071 Prefetches have no effect on the behavior of the program but can change its
6072 performance characteristics.</p>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6073
6074<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6075<p><tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the
6076 specifier determining if the fetch should be for a read (0) or write (1),
6077 and <tt>locality</tt> is a temporal locality specifier ranging from (0) - no
6078 locality, to (3) - extremely local keep in cache. The <tt>rw</tt>
6079 and <tt>locality</tt> arguments must be constant integers.</p>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6080
6081<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6082<p>This intrinsic does not modify the behavior of the program. In particular,
6083 prefetches cannot trap and do not produce a value. On targets that support
6084 this intrinsic, the prefetch can provide hints to the processor cache for
6085 better performance.</p>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]6086
6087</div>
6088
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6089<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6090<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6091 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6092</h4>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6093
6094<div class="doc_text">
6095
6096<h5>Syntax:</h5>
6097<pre>
Chris Lattner12477732007-09-21 17:30:40 +0000[diff] [blame]6098 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6099</pre>
6100
6101<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6102<p>The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program
6103 Counter (PC) in a region of code to simulators and other tools. The method
6104 is target specific, but it is expected that the marker will use exported
6105 symbols to transmit the PC of the marker. The marker makes no guarantees
6106 that it will remain with any specific instruction after optimizations. It is
6107 possible that the presence of a marker will inhibit optimizations. The
6108 intended use is to be inserted after optimizations to allow correlations of
6109 simulation runs.</p>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6110
6111<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6112<p><tt>id</tt> is a numerical id identifying the marker.</p>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6113
6114<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6115<p>This intrinsic does not modify the behavior of the program. Backends that do
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]6116 not support this intrinsic may ignore it.</p>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]6117
6118</div>
6119
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6120<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6121<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6122 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6123</h4>
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6124
6125<div class="doc_text">
6126
6127<h5>Syntax:</h5>
6128<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]6129 declare i64 @llvm.readcyclecounter()
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6130</pre>
6131
6132<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6133<p>The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
6134 counter register (or similar low latency, high accuracy clocks) on those
6135 targets that support it. On X86, it should map to RDTSC. On Alpha, it
6136 should map to RPCC. As the backing counters overflow quickly (on the order
6137 of 9 seconds on alpha), this should only be used for small timings.</p>
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6138
6139<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6140<p>When directly supported, reading the cycle counter should not modify any
6141 memory. Implementations are allowed to either return a application specific
6142 value or a system wide value. On backends without support, this is lowered
6143 to a constant 0.</p>
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]6144
6145</div>
6146
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6147<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6148<h3>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6149 <a name="int_libc">Standard C Library Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6150</h3>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6151
6152<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6153
6154<p>LLVM provides intrinsics for a few important standard C library functions.
6155 These intrinsics allow source-language front-ends to pass information about
6156 the alignment of the pointer arguments to the code generator, providing
6157 opportunity for more efficient code generation.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6158
6159</div>
6160
6161<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6162<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6163 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6164</h4>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6165
6166<div class="doc_text">
6167
6168<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6169<p>This is an overloaded intrinsic. You can use <tt>llvm.memcpy</tt> on any
Mon P Wang508127b2010-04-07 06:35:53 +0000[diff] [blame]6170 integer bit width and for different address spaces. Not all targets support
6171 all bit widths however.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6172
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6173<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6174 declare void @llvm.memcpy.p0i8.p0i8.i32(i8* &lt;dest&gt;, i8* &lt;src&gt;,
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6175 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6176 declare void @llvm.memcpy.p0i8.p0i8.i64(i8* &lt;dest&gt;, i8* &lt;src&gt;,
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6177 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6178</pre>
6179
6180<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6181<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6182 source location to the destination location.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6183
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6184<p>Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6185 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6186 and the pointers can be in specified address spaces.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6187
6188<h5>Arguments:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6189
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6190<p>The first argument is a pointer to the destination, the second is a pointer
6191 to the source. The third argument is an integer argument specifying the
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6192 number of bytes to copy, the fourth argument is the alignment of the
6193 source and destination locations, and the fifth is a boolean indicating a
6194 volatile access.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6195
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]6196<p>If the call to this intrinsic has an alignment value that is not 0 or 1,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6197 then the caller guarantees that both the source and destination pointers are
6198 aligned to that boundary.</p>
Chris Lattner4c67c482004-02-12 21:18:15 +0000[diff] [blame]6199
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]6200<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6201 <tt>llvm.memcpy</tt> call is a <a href="#volatile">volatile operation</a>.
6202 The detailed access behavior is not very cleanly specified and it is unwise
6203 to depend on it.</p>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6204
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6205<h5>Semantics:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6206
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6207<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6208 source location to the destination location, which are not allowed to
6209 overlap. It copies "len" bytes of memory over. If the argument is known to
6210 be aligned to some boundary, this can be specified as the fourth argument,
6211 otherwise it should be set to 0 or 1.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6212
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]6213</div>
6214
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6215<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6216<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6217 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6218</h4>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6219
6220<div class="doc_text">
6221
6222<h5>Syntax:</h5>
Chris Lattnerdd708342008-11-21 16:42:48 +0000[diff] [blame]6223<p>This is an overloaded intrinsic. You can use llvm.memmove on any integer bit
Mon P Wang508127b2010-04-07 06:35:53 +0000[diff] [blame]6224 width and for different address space. Not all targets support all bit
6225 widths however.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6226
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6227<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6228 declare void @llvm.memmove.p0i8.p0i8.i32(i8* &lt;dest&gt;, i8* &lt;src&gt;,
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6229 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6230 declare void @llvm.memmove.p0i8.p0i8.i64(i8* &lt;dest&gt;, i8* &lt;src&gt;,
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6231 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6232</pre>
6233
6234<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6235<p>The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the
6236 source location to the destination location. It is similar to the
6237 '<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to
6238 overlap.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6239
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6240<p>Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6241 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6242 and the pointers can be in specified address spaces.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6243
6244<h5>Arguments:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6245
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6246<p>The first argument is a pointer to the destination, the second is a pointer
6247 to the source. The third argument is an integer argument specifying the
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6248 number of bytes to copy, the fourth argument is the alignment of the
6249 source and destination locations, and the fifth is a boolean indicating a
6250 volatile access.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6251
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]6252<p>If the call to this intrinsic has an alignment value that is not 0 or 1,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6253 then the caller guarantees that the source and destination pointers are
6254 aligned to that boundary.</p>
Chris Lattner4c67c482004-02-12 21:18:15 +0000[diff] [blame]6255
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]6256<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6257 <tt>llvm.memmove</tt> call is a <a href="#volatile">volatile operation</a>.
6258 The detailed access behavior is not very cleanly specified and it is unwise
6259 to depend on it.</p>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6260
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6261<h5>Semantics:</h5>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6262
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6263<p>The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the
6264 source location to the destination location, which may overlap. It copies
6265 "len" bytes of memory over. If the argument is known to be aligned to some
6266 boundary, this can be specified as the fourth argument, otherwise it should
6267 be set to 0 or 1.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6268
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]6269</div>
6270
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6271<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6272<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6273 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6274</h4>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6275
6276<div class="doc_text">
6277
6278<h5>Syntax:</h5>
Chris Lattnerdd708342008-11-21 16:42:48 +0000[diff] [blame]6279<p>This is an overloaded intrinsic. You can use llvm.memset on any integer bit
John Criswellad05ae42010-07-30 16:30:28 +0000[diff] [blame]6280 width and for different address spaces. However, not all targets support all
6281 bit widths.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6282
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6283<pre>
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6284 declare void @llvm.memset.p0i8.i32(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattner685db9d2010-04-08 00:54:34 +0000[diff] [blame]6285 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanaabfdb32010-05-28 17:13:49 +0000[diff] [blame]6286 declare void @llvm.memset.p0i8.i64(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattner685db9d2010-04-08 00:54:34 +0000[diff] [blame]6287 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6288</pre>
6289
6290<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6291<p>The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a
6292 particular byte value.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6293
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6294<p>Note that, unlike the standard libc function, the <tt>llvm.memset</tt>
John Criswellad05ae42010-07-30 16:30:28 +0000[diff] [blame]6295 intrinsic does not return a value and takes extra alignment/volatile
6296 arguments. Also, the destination can be in an arbitrary address space.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6297
6298<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6299<p>The first argument is a pointer to the destination to fill, the second is the
John Criswellad05ae42010-07-30 16:30:28 +0000[diff] [blame]6300 byte value with which to fill it, the third argument is an integer argument
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6301 specifying the number of bytes to fill, and the fourth argument is the known
John Criswellad05ae42010-07-30 16:30:28 +0000[diff] [blame]6302 alignment of the destination location.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6303
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]6304<p>If the call to this intrinsic has an alignment value that is not 0 or 1,
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6305 then the caller guarantees that the destination pointer is aligned to that
6306 boundary.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6307
Jeffrey Yasskin5d284ae2010-04-26 21:21:24 +0000[diff] [blame]6308<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6309 <tt>llvm.memset</tt> call is a <a href="#volatile">volatile operation</a>.
6310 The detailed access behavior is not very cleanly specified and it is unwise
6311 to depend on it.</p>
Chris Lattnerbd4ca622010-04-08 00:53:57 +0000[diff] [blame]6312
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6313<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6314<p>The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting
6315 at the destination location. If the argument is known to be aligned to some
6316 boundary, this can be specified as the fourth argument, otherwise it should
6317 be set to 0 or 1.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6318
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]6319</div>
6320
Chris Lattner3b4f4372004-06-11 02:28:03 +0000[diff] [blame]6321<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6322<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6323 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6324</h4>
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6325
6326<div class="doc_text">
6327
6328<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6329<p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
6330 floating point or vector of floating point type. Not all targets support all
6331 types however.</p>
6332
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6333<pre>
Dale Johannesendd89d272007-10-02 17:47:38 +0000[diff] [blame]6334 declare float @llvm.sqrt.f32(float %Val)
6335 declare double @llvm.sqrt.f64(double %Val)
6336 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
6337 declare fp128 @llvm.sqrt.f128(fp128 %Val)
6338 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6339</pre>
6340
6341<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6342<p>The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
6343 returning the same value as the libm '<tt>sqrt</tt>' functions would.
6344 Unlike <tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined
6345 behavior for negative numbers other than -0.0 (which allows for better
6346 optimization, because there is no need to worry about errno being
6347 set). <tt>llvm.sqrt(-0.0)</tt> is defined to return -0.0 like IEEE sqrt.</p>
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6348
6349<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6350<p>The argument and return value are floating point numbers of the same
6351 type.</p>
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6352
6353<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6354<p>This function returns the sqrt of the specified operand if it is a
6355 nonnegative floating point number.</p>
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6356
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]6357</div>
6358
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6359<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6360<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6361 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6362</h4>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6363
6364<div class="doc_text">
6365
6366<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6367<p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
6368 floating point or vector of floating point type. Not all targets support all
6369 types however.</p>
6370
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6371<pre>
Dale Johannesendd89d272007-10-02 17:47:38 +0000[diff] [blame]6372 declare float @llvm.powi.f32(float %Val, i32 %power)
6373 declare double @llvm.powi.f64(double %Val, i32 %power)
6374 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
6375 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
6376 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6377</pre>
6378
6379<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6380<p>The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
6381 specified (positive or negative) power. The order of evaluation of
6382 multiplications is not defined. When a vector of floating point type is
6383 used, the second argument remains a scalar integer value.</p>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6384
6385<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6386<p>The second argument is an integer power, and the first is a value to raise to
6387 that power.</p>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6388
6389<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6390<p>This function returns the first value raised to the second power with an
6391 unspecified sequence of rounding operations.</p>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6392
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]6393</div>
6394
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6395<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6396<h4>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6397 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6398</h4>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6399
6400<div class="doc_text">
6401
6402<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6403<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
6404 floating point or vector of floating point type. Not all targets support all
6405 types however.</p>
6406
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6407<pre>
6408 declare float @llvm.sin.f32(float %Val)
6409 declare double @llvm.sin.f64(double %Val)
6410 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
6411 declare fp128 @llvm.sin.f128(fp128 %Val)
6412 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
6413</pre>
6414
6415<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6416<p>The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6417
6418<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6419<p>The argument and return value are floating point numbers of the same
6420 type.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6421
6422<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6423<p>This function returns the sine of the specified operand, returning the same
6424 values as the libm <tt>sin</tt> functions would, and handles error conditions
6425 in the same way.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6426
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6427</div>
6428
6429<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6430<h4>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6431 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6432</h4>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6433
6434<div class="doc_text">
6435
6436<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6437<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
6438 floating point or vector of floating point type. Not all targets support all
6439 types however.</p>
6440
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6441<pre>
6442 declare float @llvm.cos.f32(float %Val)
6443 declare double @llvm.cos.f64(double %Val)
6444 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
6445 declare fp128 @llvm.cos.f128(fp128 %Val)
6446 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
6447</pre>
6448
6449<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6450<p>The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6451
6452<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6453<p>The argument and return value are floating point numbers of the same
6454 type.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6455
6456<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6457<p>This function returns the cosine of the specified operand, returning the same
6458 values as the libm <tt>cos</tt> functions would, and handles error conditions
6459 in the same way.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6460
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6461</div>
6462
6463<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6464<h4>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6465 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6466</h4>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6467
6468<div class="doc_text">
6469
6470<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6471<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
6472 floating point or vector of floating point type. Not all targets support all
6473 types however.</p>
6474
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6475<pre>
6476 declare float @llvm.pow.f32(float %Val, float %Power)
6477 declare double @llvm.pow.f64(double %Val, double %Power)
6478 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
6479 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
6480 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
6481</pre>
6482
6483<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6484<p>The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
6485 specified (positive or negative) power.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6486
6487<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6488<p>The second argument is a floating point power, and the first is a value to
6489 raise to that power.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6490
6491<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6492<p>This function returns the first value raised to the second power, returning
6493 the same values as the libm <tt>pow</tt> functions would, and handles error
6494 conditions in the same way.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6495
Dan Gohmanb6324c12007-10-15 20:30:11 +0000[diff] [blame]6496</div>
6497
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6498<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6499<h3>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6500 <a name="int_manip">Bit Manipulation Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6501</h3>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6502
6503<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6504
6505<p>LLVM provides intrinsics for a few important bit manipulation operations.
6506 These allow efficient code generation for some algorithms.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6507
6508</div>
6509
6510<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6511<h4>
Reid Spencer96a5f022007-04-04 02:42:35 +0000[diff] [blame]6512 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6513</h4>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6514
6515<div class="doc_text">
6516
6517<h5>Syntax:</h5>
Reid Spencer4eefaab2007-04-01 08:04:23 +0000[diff] [blame]6518<p>This is an overloaded intrinsic function. You can use bswap on any integer
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6519 type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
6520
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6521<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6522 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
6523 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
6524 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6525</pre>
6526
6527<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6528<p>The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
6529 values with an even number of bytes (positive multiple of 16 bits). These
6530 are useful for performing operations on data that is not in the target's
6531 native byte order.</p>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6532
6533<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6534<p>The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
6535 and low byte of the input i16 swapped. Similarly,
6536 the <tt>llvm.bswap.i32</tt> intrinsic returns an i32 value that has the four
6537 bytes of the input i32 swapped, so that if the input bytes are numbered 0, 1,
6538 2, 3 then the returned i32 will have its bytes in 3, 2, 1, 0 order.
6539 The <tt>llvm.bswap.i48</tt>, <tt>llvm.bswap.i64</tt> and other intrinsics
6540 extend this concept to additional even-byte lengths (6 bytes, 8 bytes and
6541 more, respectively).</p>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]6542
6543</div>
6544
6545<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6546<h4>
Reid Spencerb4f9a6f2006-01-16 21:12:35 +0000[diff] [blame]6547 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6548</h4>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6549
6550<div class="doc_text">
6551
6552<h5>Syntax:</h5>
Reid Spencer4eefaab2007-04-01 08:04:23 +0000[diff] [blame]6553<p>This is an overloaded intrinsic. You can use llvm.ctpop on any integer bit
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6554 width. Not all targets support all bit widths however.</p>
6555
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6556<pre>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6557 declare i8 @llvm.ctpop.i8(i8 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6558 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]6559 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6560 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
6561 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6562</pre>
6563
6564<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6565<p>The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set
6566 in a value.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6567
6568<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6569<p>The only argument is the value to be counted. The argument may be of any
6570 integer type. The return type must match the argument type.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6571
6572<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6573<p>The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6574
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6575</div>
6576
6577<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6578<h4>
Chris Lattnerb748c672006-01-16 22:34:14 +0000[diff] [blame]6579 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6580</h4>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6581
6582<div class="doc_text">
6583
6584<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6585<p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
6586 integer bit width. Not all targets support all bit widths however.</p>
6587
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6588<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6589 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
6590 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]6591 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6592 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
6593 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6594</pre>
6595
6596<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6597<p>The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
6598 leading zeros in a variable.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6599
6600<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6601<p>The only argument is the value to be counted. The argument may be of any
6602 integer type. The return type must match the argument type.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6603
6604<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6605<p>The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant)
6606 zeros in a variable. If the src == 0 then the result is the size in bits of
6607 the type of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6608
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]6609</div>
Chris Lattner3b4f4372004-06-11 02:28:03 +0000[diff] [blame]6610
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6611<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6612<h4>
Chris Lattnerb748c672006-01-16 22:34:14 +0000[diff] [blame]6613 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6614</h4>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6615
6616<div class="doc_text">
6617
6618<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6619<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
6620 integer bit width. Not all targets support all bit widths however.</p>
6621
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6622<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6623 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
6624 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +0000[diff] [blame]6625 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +0000[diff] [blame]6626 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
6627 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6628</pre>
6629
6630<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6631<p>The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
6632 trailing zeros.</p>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6633
6634<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6635<p>The only argument is the value to be counted. The argument may be of any
6636 integer type. The return type must match the argument type.</p>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6637
6638<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6639<p>The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant)
6640 zeros in a variable. If the src == 0 then the result is the size in bits of
6641 the type of src. For example, <tt>llvm.cttz(2) = 1</tt>.</p>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6642
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]6643</div>
6644
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]6645<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6646<h3>
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]6647 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6648</h3>
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]6649
6650<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6651
6652<p>LLVM provides intrinsics for some arithmetic with overflow operations.</p>
Bill Wendlingfd2bd722009-02-08 04:04:40 +0000[diff] [blame]6653
6654</div>
6655
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6656<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6657<h4>
6658 <a name="int_sadd_overflow">
6659 '<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics
6660 </a>
6661</h4>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6662
6663<div class="doc_text">
6664
6665<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6666<p>This is an overloaded intrinsic. You can use <tt>llvm.sadd.with.overflow</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6667 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6668
6669<pre>
6670 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
6671 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6672 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
6673</pre>
6674
6675<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6676<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6677 a signed addition of the two arguments, and indicate whether an overflow
6678 occurred during the signed summation.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6679
6680<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6681<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6682 be of integer types of any bit width, but they must have the same bit
6683 width. The second element of the result structure must be of
6684 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6685 undergo signed addition.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6686
6687<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6688<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6689 a signed addition of the two variables. They return a structure &mdash; the
6690 first element of which is the signed summation, and the second element of
6691 which is a bit specifying if the signed summation resulted in an
6692 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6693
6694<h5>Examples:</h5>
6695<pre>
6696 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6697 %sum = extractvalue {i32, i1} %res, 0
6698 %obit = extractvalue {i32, i1} %res, 1
6699 br i1 %obit, label %overflow, label %normal
6700</pre>
6701
6702</div>
6703
6704<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6705<h4>
6706 <a name="int_uadd_overflow">
6707 '<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics
6708 </a>
6709</h4>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6710
6711<div class="doc_text">
6712
6713<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6714<p>This is an overloaded intrinsic. You can use <tt>llvm.uadd.with.overflow</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6715 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6716
6717<pre>
6718 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
6719 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6720 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
6721</pre>
6722
6723<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6724<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6725 an unsigned addition of the two arguments, and indicate whether a carry
6726 occurred during the unsigned summation.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6727
6728<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6729<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6730 be of integer types of any bit width, but they must have the same bit
6731 width. The second element of the result structure must be of
6732 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6733 undergo unsigned addition.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6734
6735<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6736<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6737 an unsigned addition of the two arguments. They return a structure &mdash;
6738 the first element of which is the sum, and the second element of which is a
6739 bit specifying if the unsigned summation resulted in a carry.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6740
6741<h5>Examples:</h5>
6742<pre>
6743 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6744 %sum = extractvalue {i32, i1} %res, 0
6745 %obit = extractvalue {i32, i1} %res, 1
6746 br i1 %obit, label %carry, label %normal
6747</pre>
6748
6749</div>
6750
6751<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6752<h4>
6753 <a name="int_ssub_overflow">
6754 '<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics
6755 </a>
6756</h4>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6757
6758<div class="doc_text">
6759
6760<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6761<p>This is an overloaded intrinsic. You can use <tt>llvm.ssub.with.overflow</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6762 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6763
6764<pre>
6765 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
6766 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6767 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
6768</pre>
6769
6770<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6771<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6772 a signed subtraction of the two arguments, and indicate whether an overflow
6773 occurred during the signed subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6774
6775<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6776<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6777 be of integer types of any bit width, but they must have the same bit
6778 width. The second element of the result structure must be of
6779 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6780 undergo signed subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6781
6782<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6783<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6784 a signed subtraction of the two arguments. They return a structure &mdash;
6785 the first element of which is the subtraction, and the second element of
6786 which is a bit specifying if the signed subtraction resulted in an
6787 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6788
6789<h5>Examples:</h5>
6790<pre>
6791 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6792 %sum = extractvalue {i32, i1} %res, 0
6793 %obit = extractvalue {i32, i1} %res, 1
6794 br i1 %obit, label %overflow, label %normal
6795</pre>
6796
6797</div>
6798
6799<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6800<h4>
6801 <a name="int_usub_overflow">
6802 '<tt>llvm.usub.with.overflow.*</tt>' Intrinsics
6803 </a>
6804</h4>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6805
6806<div class="doc_text">
6807
6808<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6809<p>This is an overloaded intrinsic. You can use <tt>llvm.usub.with.overflow</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6810 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6811
6812<pre>
6813 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
6814 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6815 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
6816</pre>
6817
6818<h5>Overview:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6819<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6820 an unsigned subtraction of the two arguments, and indicate whether an
6821 overflow occurred during the unsigned subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6822
6823<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6824<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6825 be of integer types of any bit width, but they must have the same bit
6826 width. The second element of the result structure must be of
6827 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6828 undergo unsigned subtraction.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6829
6830<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6831<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6832 an unsigned subtraction of the two arguments. They return a structure &mdash;
6833 the first element of which is the subtraction, and the second element of
6834 which is a bit specifying if the unsigned subtraction resulted in an
6835 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6836
6837<h5>Examples:</h5>
6838<pre>
6839 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6840 %sum = extractvalue {i32, i1} %res, 0
6841 %obit = extractvalue {i32, i1} %res, 1
6842 br i1 %obit, label %overflow, label %normal
6843</pre>
6844
6845</div>
6846
6847<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6848<h4>
6849 <a name="int_smul_overflow">
6850 '<tt>llvm.smul.with.overflow.*</tt>' Intrinsics
6851 </a>
6852</h4>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6853
6854<div class="doc_text">
6855
6856<h5>Syntax:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6857<p>This is an overloaded intrinsic. You can use <tt>llvm.smul.with.overflow</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6858 on any integer bit width.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6859
6860<pre>
6861 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
6862 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6863 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
6864</pre>
6865
6866<h5>Overview:</h5>
6867
6868<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6869 a signed multiplication of the two arguments, and indicate whether an
6870 overflow occurred during the signed multiplication.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6871
6872<h5>Arguments:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6873<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6874 be of integer types of any bit width, but they must have the same bit
6875 width. The second element of the result structure must be of
6876 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6877 undergo signed multiplication.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6878
6879<h5>Semantics:</h5>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6880<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6881 a signed multiplication of the two arguments. They return a structure &mdash;
6882 the first element of which is the multiplication, and the second element of
6883 which is a bit specifying if the signed multiplication resulted in an
6884 overflow.</p>
Bill Wendlingf4d70622009-02-08 01:40:31 +0000[diff] [blame]6885
6886<h5>Examples:</h5>
6887<pre>
6888 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6889 %sum = extractvalue {i32, i1} %res, 0
6890 %obit = extractvalue {i32, i1} %res, 1
6891 br i1 %obit, label %overflow, label %normal
6892</pre>
6893
Reid Spencer5bf54c82007-04-11 23:23:49 +0000[diff] [blame]6894</div>
6895
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6896<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6897<h4>
6898 <a name="int_umul_overflow">
6899 '<tt>llvm.umul.with.overflow.*</tt>' Intrinsics
6900 </a>
6901</h4>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6902
6903<div class="doc_text">
6904
6905<h5>Syntax:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6906<p>This is an overloaded intrinsic. You can use <tt>llvm.umul.with.overflow</tt>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6907 on any integer bit width.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6908
6909<pre>
6910 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
6911 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6912 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
6913</pre>
6914
6915<h5>Overview:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6916<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6917 a unsigned multiplication of the two arguments, and indicate whether an
6918 overflow occurred during the unsigned multiplication.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6919
6920<h5>Arguments:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6921<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6922 be of integer types of any bit width, but they must have the same bit
6923 width. The second element of the result structure must be of
6924 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6925 undergo unsigned multiplication.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6926
6927<h5>Semantics:</h5>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6928<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]6929 an unsigned multiplication of the two arguments. They return a structure
6930 &mdash; the first element of which is the multiplication, and the second
6931 element of which is a bit specifying if the unsigned multiplication resulted
6932 in an overflow.</p>
Bill Wendlingb9a73272009-02-08 23:00:09 +0000[diff] [blame]6933
6934<h5>Examples:</h5>
6935<pre>
6936 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6937 %sum = extractvalue {i32, i1} %res, 0
6938 %obit = extractvalue {i32, i1} %res, 1
6939 br i1 %obit, label %overflow, label %normal
6940</pre>
6941
6942</div>
6943
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]6944<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6945<h3>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]6946 <a name="int_fp16">Half Precision Floating Point Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6947</h3>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]6948
6949<div class="doc_text">
6950
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]6951<p>Half precision floating point is a storage-only format. This means that it is
6952 a dense encoding (in memory) but does not support computation in the
6953 format.</p>
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]6954
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]6955<p>This means that code must first load the half-precision floating point
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]6956 value as an i16, then convert it to float with <a
6957 href="#int_convert_from_fp16"><tt>llvm.convert.from.fp16</tt></a>.
6958 Computation can then be performed on the float value (including extending to
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]6959 double etc). To store the value back to memory, it is first converted to
6960 float if needed, then converted to i16 with
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]6961 <a href="#int_convert_to_fp16"><tt>llvm.convert.to.fp16</tt></a>, then
6962 storing as an i16 value.</p>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]6963</div>
6964
6965<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]6966<h4>
6967 <a name="int_convert_to_fp16">
6968 '<tt>llvm.convert.to.fp16</tt>' Intrinsic
6969 </a>
6970</h4>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]6971
6972<div class="doc_text">
6973
6974<h5>Syntax:</h5>
6975<pre>
6976 declare i16 @llvm.convert.to.fp16(f32 %a)
6977</pre>
6978
6979<h5>Overview:</h5>
6980<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
6981 a conversion from single precision floating point format to half precision
6982 floating point format.</p>
6983
6984<h5>Arguments:</h5>
6985<p>The intrinsic function contains single argument - the value to be
6986 converted.</p>
6987
6988<h5>Semantics:</h5>
6989<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
6990 a conversion from single precision floating point format to half precision
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]6991 floating point format. The return value is an <tt>i16</tt> which
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]6992 contains the converted number.</p>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]6993
6994<h5>Examples:</h5>
6995<pre>
6996 %res = call i16 @llvm.convert.to.fp16(f32 %a)
6997 store i16 %res, i16* @x, align 2
6998</pre>
6999
7000</div>
7001
7002<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7003<h4>
7004 <a name="int_convert_from_fp16">
7005 '<tt>llvm.convert.from.fp16</tt>' Intrinsic
7006 </a>
7007</h4>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7008
7009<div class="doc_text">
7010
7011<h5>Syntax:</h5>
7012<pre>
7013 declare f32 @llvm.convert.from.fp16(i16 %a)
7014</pre>
7015
7016<h5>Overview:</h5>
7017<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs
7018 a conversion from half precision floating point format to single precision
7019 floating point format.</p>
7020
7021<h5>Arguments:</h5>
7022<p>The intrinsic function contains single argument - the value to be
7023 converted.</p>
7024
7025<h5>Semantics:</h5>
7026<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs a
Chris Lattner022a9fb2010-03-15 04:12:21 +0000[diff] [blame]7027 conversion from half single precision floating point format to single
Chris Lattnerbbd8bd32010-03-14 23:03:31 +0000[diff] [blame]7028 precision floating point format. The input half-float value is represented by
7029 an <tt>i16</tt> value.</p>
Anton Korobeynikovcd4dd9c2010-03-14 18:42:47 +0000[diff] [blame]7030
7031<h5>Examples:</h5>
7032<pre>
7033 %a = load i16* @x, align 2
7034 %res = call f32 @llvm.convert.from.fp16(i16 %a)
7035</pre>
7036
7037</div>
7038
7039<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7040<h3>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]7041 <a name="int_debugger">Debugger Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7042</h3>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]7043
7044<div class="doc_text">
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]7045
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7046<p>The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt>
7047 prefix), are described in
7048 the <a href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source
7049 Level Debugging</a> document.</p>
7050
7051</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]7052
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]7053<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7054<h3>
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]7055 <a name="int_eh">Exception Handling Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7056</h3>
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]7057
7058<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7059
7060<p>The LLVM exception handling intrinsics (which all start with
7061 <tt>llvm.eh.</tt> prefix), are described in
7062 the <a href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
7063 Handling</a> document.</p>
7064
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]7065</div>
7066
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7067<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7068<h3>
Duncan Sands86e01192007-09-11 14:10:23 +0000[diff] [blame]7069 <a name="int_trampoline">Trampoline Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7070</h3>
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7071
7072<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7073
7074<p>This intrinsic makes it possible to excise one parameter, marked with
Dan Gohman3770af52010-07-02 23:18:08 +0000[diff] [blame]7075 the <a href="#nest"><tt>nest</tt></a> attribute, from a function.
7076 The result is a callable
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7077 function pointer lacking the nest parameter - the caller does not need to
7078 provide a value for it. Instead, the value to use is stored in advance in a
7079 "trampoline", a block of memory usually allocated on the stack, which also
7080 contains code to splice the nest value into the argument list. This is used
7081 to implement the GCC nested function address extension.</p>
7082
7083<p>For example, if the function is
7084 <tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
7085 pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as
7086 follows:</p>
7087
Benjamin Kramer79698be2010-07-13 12:26:09 +0000[diff] [blame]7088<pre class="doc_code">
Duncan Sands86e01192007-09-11 14:10:23 +0000[diff] [blame]7089 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
7090 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7091 %p = call i8* @llvm.init.trampoline(i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8*), i8* %nval)
Duncan Sands86e01192007-09-11 14:10:23 +0000[diff] [blame]7092 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7093</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7094
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7095<p>The call <tt>%val = call i32 %fp(i32 %x, i32 %y)</tt> is then equivalent
7096 to <tt>%val = call i32 %f(i8* %nval, i32 %x, i32 %y)</tt>.</p>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7097
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7098</div>
7099
7100<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7101<h4>
7102 <a name="int_it">
7103 '<tt>llvm.init.trampoline</tt>' Intrinsic
7104 </a>
7105</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7106
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7107<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7108
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7109<h5>Syntax:</h5>
7110<pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7111 declare i8* @llvm.init.trampoline(i8* &lt;tramp&gt;, i8* &lt;func&gt;, i8* &lt;nval&gt;)
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7112</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7113
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7114<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7115<p>This fills the memory pointed to by <tt>tramp</tt> with code and returns a
7116 function pointer suitable for executing it.</p>
7117
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7118<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7119<p>The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
7120 pointers. The <tt>tramp</tt> argument must point to a sufficiently large and
7121 sufficiently aligned block of memory; this memory is written to by the
7122 intrinsic. Note that the size and the alignment are target-specific - LLVM
7123 currently provides no portable way of determining them, so a front-end that
7124 generates this intrinsic needs to have some target-specific knowledge.
7125 The <tt>func</tt> argument must hold a function bitcast to
7126 an <tt>i8*</tt>.</p>
7127
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7128<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7129<p>The block of memory pointed to by <tt>tramp</tt> is filled with target
7130 dependent code, turning it into a function. A pointer to this function is
7131 returned, but needs to be bitcast to an <a href="#int_trampoline">appropriate
7132 function pointer type</a> before being called. The new function's signature
7133 is the same as that of <tt>func</tt> with any arguments marked with
7134 the <tt>nest</tt> attribute removed. At most one such <tt>nest</tt> argument
7135 is allowed, and it must be of pointer type. Calling the new function is
7136 equivalent to calling <tt>func</tt> with the same argument list, but
7137 with <tt>nval</tt> used for the missing <tt>nest</tt> argument. If, after
7138 calling <tt>llvm.init.trampoline</tt>, the memory pointed to
7139 by <tt>tramp</tt> is modified, then the effect of any later call to the
7140 returned function pointer is undefined.</p>
7141
Duncan Sands644f9172007-07-27 12:58:54 +0000[diff] [blame]7142</div>
7143
7144<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7145<h3>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7146 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7147</h3>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7148
7149<div class="doc_text">
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7150
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7151<p>These intrinsic functions expand the "universal IR" of LLVM to represent
7152 hardware constructs for atomic operations and memory synchronization. This
7153 provides an interface to the hardware, not an interface to the programmer. It
7154 is aimed at a low enough level to allow any programming models or APIs
7155 (Application Programming Interfaces) which need atomic behaviors to map
7156 cleanly onto it. It is also modeled primarily on hardware behavior. Just as
7157 hardware provides a "universal IR" for source languages, it also provides a
7158 starting point for developing a "universal" atomic operation and
7159 synchronization IR.</p>
7160
7161<p>These do <em>not</em> form an API such as high-level threading libraries,
7162 software transaction memory systems, atomic primitives, and intrinsic
7163 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
7164 application libraries. The hardware interface provided by LLVM should allow
7165 a clean implementation of all of these APIs and parallel programming models.
7166 No one model or paradigm should be selected above others unless the hardware
7167 itself ubiquitously does so.</p>
7168
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7169</div>
7170
7171<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7172<h4>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7173 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7174</h4>
7175
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7176<div class="doc_text">
7177<h5>Syntax:</h5>
7178<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7179 declare void @llvm.memory.barrier(i1 &lt;ll&gt;, i1 &lt;ls&gt;, i1 &lt;sl&gt;, i1 &lt;ss&gt;, i1 &lt;device&gt;)
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7180</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7181
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7182<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7183<p>The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
7184 specific pairs of memory access types.</p>
7185
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7186<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7187<p>The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
7188 The first four arguments enables a specific barrier as listed below. The
Dan Gohmana269a0a2010-03-01 17:41:39 +0000[diff] [blame]7189 fifth argument specifies that the barrier applies to io or device or uncached
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7190 memory.</p>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7191
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7192<ul>
7193 <li><tt>ll</tt>: load-load barrier</li>
7194 <li><tt>ls</tt>: load-store barrier</li>
7195 <li><tt>sl</tt>: store-load barrier</li>
7196 <li><tt>ss</tt>: store-store barrier</li>
7197 <li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
7198</ul>
7199
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7200<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7201<p>This intrinsic causes the system to enforce some ordering constraints upon
7202 the loads and stores of the program. This barrier does not
7203 indicate <em>when</em> any events will occur, it only enforces
7204 an <em>order</em> in which they occur. For any of the specified pairs of load
7205 and store operations (f.ex. load-load, or store-load), all of the first
7206 operations preceding the barrier will complete before any of the second
7207 operations succeeding the barrier begin. Specifically the semantics for each
7208 pairing is as follows:</p>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7209
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7210<ul>
7211 <li><tt>ll</tt>: All loads before the barrier must complete before any load
7212 after the barrier begins.</li>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7213 <li><tt>ls</tt>: All loads before the barrier must complete before any
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7214 store after the barrier begins.</li>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7215 <li><tt>ss</tt>: All stores before the barrier must complete before any
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7216 store after the barrier begins.</li>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7217 <li><tt>sl</tt>: All stores before the barrier must complete before any
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7218 load after the barrier begins.</li>
7219</ul>
7220
7221<p>These semantics are applied with a logical "and" behavior when more than one
7222 is enabled in a single memory barrier intrinsic.</p>
7223
7224<p>Backends may implement stronger barriers than those requested when they do
7225 not support as fine grained a barrier as requested. Some architectures do
7226 not need all types of barriers and on such architectures, these become
7227 noops.</p>
7228
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7229<h5>Example:</h5>
7230<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7231%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7232%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7233 store i32 4, %ptr
7234
7235%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7236 call void @llvm.memory.barrier(i1 false, i1 true, i1 false, i1 false)
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7237 <i>; guarantee the above finishes</i>
7238 store i32 8, %ptr <i>; before this begins</i>
7239</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7240
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7241</div>
7242
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7243<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7244<h4>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7245 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7246</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7247
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7248<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7249
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7250<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7251<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
7252 any integer bit width and for different address spaces. Not all targets
7253 support all bit widths however.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7254
7255<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7256 declare i8 @llvm.atomic.cmp.swap.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt;)
7257 declare i16 @llvm.atomic.cmp.swap.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt;)
7258 declare i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt;)
7259 declare i64 @llvm.atomic.cmp.swap.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;cmp&gt;, i64 &lt;val&gt;)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7260</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7261
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7262<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7263<p>This loads a value in memory and compares it to a given value. If they are
7264 equal, it stores a new value into the memory.</p>
7265
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7266<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7267<p>The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result
7268 as well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
7269 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
7270 this integer type. While any bit width integer may be used, targets may only
7271 lower representations they support in hardware.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7272
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7273<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7274<p>This entire intrinsic must be executed atomically. It first loads the value
7275 in memory pointed to by <tt>ptr</tt> and compares it with the
7276 value <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the
7277 memory. The loaded value is yielded in all cases. This provides the
7278 equivalent of an atomic compare-and-swap operation within the SSA
7279 framework.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7280
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7281<h5>Examples:</h5>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7282<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7283%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7284%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7285 store i32 4, %ptr
7286
7287%val1 = add i32 4, 4
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7288%result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %ptr, i32 4, %val1)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7289 <i>; yields {i32}:result1 = 4</i>
7290%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
7291%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
7292
7293%val2 = add i32 1, 1
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7294%result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %ptr, i32 5, %val2)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7295 <i>; yields {i32}:result2 = 8</i>
7296%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
7297
7298%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
7299</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7300
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7301</div>
7302
7303<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7304<h4>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7305 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7306</h4>
7307
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7308<div class="doc_text">
7309<h5>Syntax:</h5>
7310
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7311<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
7312 integer bit width. Not all targets support all bit widths however.</p>
7313
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7314<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7315 declare i8 @llvm.atomic.swap.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;val&gt;)
7316 declare i16 @llvm.atomic.swap.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;val&gt;)
7317 declare i32 @llvm.atomic.swap.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;val&gt;)
7318 declare i64 @llvm.atomic.swap.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;val&gt;)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7319</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7320
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7321<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7322<p>This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
7323 the value from memory. It then stores the value in <tt>val</tt> in the memory
7324 at <tt>ptr</tt>.</p>
7325
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7326<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7327<p>The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both
7328 the <tt>val</tt> argument and the result must be integers of the same bit
7329 width. The first argument, <tt>ptr</tt>, must be a pointer to a value of this
7330 integer type. The targets may only lower integer representations they
7331 support.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7332
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7333<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7334<p>This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
7335 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
7336 equivalent of an atomic swap operation within the SSA framework.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7337
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7338<h5>Examples:</h5>
7339<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7340%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7341%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7342 store i32 4, %ptr
7343
7344%val1 = add i32 4, 4
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7345%result1 = call i32 @llvm.atomic.swap.i32.p0i32(i32* %ptr, i32 %val1)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7346 <i>; yields {i32}:result1 = 4</i>
7347%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
7348%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
7349
7350%val2 = add i32 1, 1
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7351%result2 = call i32 @llvm.atomic.swap.i32.p0i32(i32* %ptr, i32 %val2)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7352 <i>; yields {i32}:result2 = 8</i>
7353
7354%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
7355%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
7356</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7357
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7358</div>
7359
7360<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7361<h4>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7362 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7363</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7364
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7365<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7366
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7367<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7368<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.add</tt> on
7369 any integer bit width. Not all targets support all bit widths however.</p>
7370
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7371<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7372 declare i8 @llvm.atomic.load.add.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7373 declare i16 @llvm.atomic.load.add.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7374 declare i32 @llvm.atomic.load.add.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7375 declare i64 @llvm.atomic.load.add.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7376</pre>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7377
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7378<h5>Overview:</h5>
7379<p>This intrinsic adds <tt>delta</tt> to the value stored in memory
7380 at <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
7381
7382<h5>Arguments:</h5>
7383<p>The intrinsic takes two arguments, the first a pointer to an integer value
7384 and the second an integer value. The result is also an integer value. These
7385 integer types can have any bit width, but they must all have the same bit
7386 width. The targets may only lower integer representations they support.</p>
7387
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7388<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7389<p>This intrinsic does a series of operations atomically. It first loads the
7390 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
7391 to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7392
7393<h5>Examples:</h5>
7394<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7395%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7396%ptr = bitcast i8* %mallocP to i32*
7397 store i32 4, %ptr
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7398%result1 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 4)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7399 <i>; yields {i32}:result1 = 4</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7400%result2 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 2)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7401 <i>; yields {i32}:result2 = 8</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7402%result3 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 5)
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7403 <i>; yields {i32}:result3 = 10</i>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7404%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7405</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7406
Andrew Lenharth95528942008-02-21 06:45:13 +0000[diff] [blame]7407</div>
7408
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7409<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7410<h4>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7411 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7412</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7413
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7414<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7415
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7416<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7417<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.sub</tt> on
7418 any integer bit width and for different address spaces. Not all targets
7419 support all bit widths however.</p>
7420
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7421<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7422 declare i8 @llvm.atomic.load.sub.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7423 declare i16 @llvm.atomic.load.sub.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7424 declare i32 @llvm.atomic.load.sub.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7425 declare i64 @llvm.atomic.load.sub.i64.p0i32(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7426</pre>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7427
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7428<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7429<p>This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7430 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
7431
7432<h5>Arguments:</h5>
7433<p>The intrinsic takes two arguments, the first a pointer to an integer value
7434 and the second an integer value. The result is also an integer value. These
7435 integer types can have any bit width, but they must all have the same bit
7436 width. The targets may only lower integer representations they support.</p>
7437
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7438<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7439<p>This intrinsic does a series of operations atomically. It first loads the
7440 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
7441 result to <tt>ptr</tt>. It yields the original value stored
7442 at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7443
7444<h5>Examples:</h5>
7445<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7446%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7447%ptr = bitcast i8* %mallocP to i32*
7448 store i32 8, %ptr
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7449%result1 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 4)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7450 <i>; yields {i32}:result1 = 8</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7451%result2 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 2)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7452 <i>; yields {i32}:result2 = 4</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7453%result3 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 5)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7454 <i>; yields {i32}:result3 = 2</i>
7455%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
7456</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7457
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7458</div>
7459
7460<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7461<h4>
7462 <a name="int_atomic_load_and">
7463 '<tt>llvm.atomic.load.and.*</tt>' Intrinsic
7464 </a>
7465 <br>
7466 <a name="int_atomic_load_nand">
7467 '<tt>llvm.atomic.load.nand.*</tt>' Intrinsic
7468 </a>
7469 <br>
7470 <a name="int_atomic_load_or">
7471 '<tt>llvm.atomic.load.or.*</tt>' Intrinsic
7472 </a>
7473 <br>
7474 <a name="int_atomic_load_xor">
7475 '<tt>llvm.atomic.load.xor.*</tt>' Intrinsic
7476 </a>
7477</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7478
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7479<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7480
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7481<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7482<p>These are overloaded intrinsics. You can
7483 use <tt>llvm.atomic.load_and</tt>, <tt>llvm.atomic.load_nand</tt>,
7484 <tt>llvm.atomic.load_or</tt>, and <tt>llvm.atomic.load_xor</tt> on any integer
7485 bit width and for different address spaces. Not all targets support all bit
7486 widths however.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7487
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7488<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7489 declare i8 @llvm.atomic.load.and.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7490 declare i16 @llvm.atomic.load.and.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7491 declare i32 @llvm.atomic.load.and.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7492 declare i64 @llvm.atomic.load.and.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7493</pre>
7494
7495<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7496 declare i8 @llvm.atomic.load.or.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7497 declare i16 @llvm.atomic.load.or.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7498 declare i32 @llvm.atomic.load.or.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7499 declare i64 @llvm.atomic.load.or.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7500</pre>
7501
7502<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7503 declare i8 @llvm.atomic.load.nand.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7504 declare i16 @llvm.atomic.load.nand.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7505 declare i32 @llvm.atomic.load.nand.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7506 declare i64 @llvm.atomic.load.nand.i64.p0i32(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7507</pre>
7508
7509<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7510 declare i8 @llvm.atomic.load.xor.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7511 declare i16 @llvm.atomic.load.xor.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7512 declare i32 @llvm.atomic.load.xor.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7513 declare i64 @llvm.atomic.load.xor.i64.p0i32(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7514</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7515
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7516<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7517<p>These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
7518 the value stored in memory at <tt>ptr</tt>. It yields the original value
7519 at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7520
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7521<h5>Arguments:</h5>
7522<p>These intrinsics take two arguments, the first a pointer to an integer value
7523 and the second an integer value. The result is also an integer value. These
7524 integer types can have any bit width, but they must all have the same bit
7525 width. The targets may only lower integer representations they support.</p>
7526
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7527<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7528<p>These intrinsics does a series of operations atomically. They first load the
7529 value stored at <tt>ptr</tt>. They then do the bitwise
7530 operation <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the
7531 original value stored at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7532
7533<h5>Examples:</h5>
7534<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7535%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7536%ptr = bitcast i8* %mallocP to i32*
7537 store i32 0x0F0F, %ptr
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7538%result0 = call i32 @llvm.atomic.load.nand.i32.p0i32(i32* %ptr, i32 0xFF)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7539 <i>; yields {i32}:result0 = 0x0F0F</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7540%result1 = call i32 @llvm.atomic.load.and.i32.p0i32(i32* %ptr, i32 0xFF)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7541 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7542%result2 = call i32 @llvm.atomic.load.or.i32.p0i32(i32* %ptr, i32 0F)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7543 <i>; yields {i32}:result2 = 0xF0</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7544%result3 = call i32 @llvm.atomic.load.xor.i32.p0i32(i32* %ptr, i32 0F)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7545 <i>; yields {i32}:result3 = FF</i>
7546%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
7547</pre>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7548
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7549</div>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7550
7551<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7552<h4>
7553 <a name="int_atomic_load_max">
7554 '<tt>llvm.atomic.load.max.*</tt>' Intrinsic
7555 </a>
7556 <br>
7557 <a name="int_atomic_load_min">
7558 '<tt>llvm.atomic.load.min.*</tt>' Intrinsic
7559 </a>
7560 <br>
7561 <a name="int_atomic_load_umax">
7562 '<tt>llvm.atomic.load.umax.*</tt>' Intrinsic
7563 </a>
7564 <br>
7565 <a name="int_atomic_load_umin">
7566 '<tt>llvm.atomic.load.umin.*</tt>' Intrinsic
7567 </a>
7568</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7569
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7570<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7571
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7572<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7573<p>These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
7574 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
7575 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
7576 address spaces. Not all targets support all bit widths however.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7577
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7578<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7579 declare i8 @llvm.atomic.load.max.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7580 declare i16 @llvm.atomic.load.max.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7581 declare i32 @llvm.atomic.load.max.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7582 declare i64 @llvm.atomic.load.max.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7583</pre>
7584
7585<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7586 declare i8 @llvm.atomic.load.min.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7587 declare i16 @llvm.atomic.load.min.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7588 declare i32 @llvm.atomic.load.min.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7589 declare i64 @llvm.atomic.load.min.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7590</pre>
7591
7592<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7593 declare i8 @llvm.atomic.load.umax.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7594 declare i16 @llvm.atomic.load.umax.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7595 declare i32 @llvm.atomic.load.umax.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7596 declare i64 @llvm.atomic.load.umax.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7597</pre>
7598
7599<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7600 declare i8 @llvm.atomic.load.umin.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7601 declare i16 @llvm.atomic.load.umin.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7602 declare i32 @llvm.atomic.load.umin.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7603 declare i64 @llvm.atomic.load.umin.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7604</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7605
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7606<h5>Overview:</h5>
Eric Christopher455c5772009-12-05 02:46:03 +0000[diff] [blame]7607<p>These intrinsics takes the signed or unsigned minimum or maximum of
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7608 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
7609 original value at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7610
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7611<h5>Arguments:</h5>
7612<p>These intrinsics take two arguments, the first a pointer to an integer value
7613 and the second an integer value. The result is also an integer value. These
7614 integer types can have any bit width, but they must all have the same bit
7615 width. The targets may only lower integer representations they support.</p>
7616
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7617<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7618<p>These intrinsics does a series of operations atomically. They first load the
7619 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or
7620 max <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They
7621 yield the original value stored at <tt>ptr</tt>.</p>
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7622
7623<h5>Examples:</h5>
7624<pre>
Victor Hernandeza70c6df2009-10-26 23:44:29 +0000[diff] [blame]7625%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7626%ptr = bitcast i8* %mallocP to i32*
7627 store i32 7, %ptr
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7628%result0 = call i32 @llvm.atomic.load.min.i32.p0i32(i32* %ptr, i32 -2)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7629 <i>; yields {i32}:result0 = 7</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7630%result1 = call i32 @llvm.atomic.load.max.i32.p0i32(i32* %ptr, i32 8)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7631 <i>; yields {i32}:result1 = -2</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7632%result2 = call i32 @llvm.atomic.load.umin.i32.p0i32(i32* %ptr, i32 10)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7633 <i>; yields {i32}:result2 = 8</i>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7634%result3 = call i32 @llvm.atomic.load.umax.i32.p0i32(i32* %ptr, i32 30)
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7635 <i>; yields {i32}:result3 = 8</i>
7636%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
7637</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7638
Mon P Wang6a490372008-06-25 08:15:39 +0000[diff] [blame]7639</div>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7640
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7641
7642<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7643<h3>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7644 <a name="int_memorymarkers">Memory Use Markers</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7645</h3>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7646
7647<div class="doc_text">
7648
7649<p>This class of intrinsics exists to information about the lifetime of memory
7650 objects and ranges where variables are immutable.</p>
7651
7652</div>
7653
7654<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7655<h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7656 <a name="int_lifetime_start">'<tt>llvm.lifetime.start</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7657</h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7658
7659<div class="doc_text">
7660
7661<h5>Syntax:</h5>
7662<pre>
7663 declare void @llvm.lifetime.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7664</pre>
7665
7666<h5>Overview:</h5>
7667<p>The '<tt>llvm.lifetime.start</tt>' intrinsic specifies the start of a memory
7668 object's lifetime.</p>
7669
7670<h5>Arguments:</h5>
Nick Lewycky9bc89042009-10-13 07:57:33 +0000[diff] [blame]7671<p>The first argument is a constant integer representing the size of the
7672 object, or -1 if it is variable sized. The second argument is a pointer to
7673 the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7674
7675<h5>Semantics:</h5>
7676<p>This intrinsic indicates that before this point in the code, the value of the
7677 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
Nick Lewyckyd20fd592009-10-27 16:56:58 +0000[diff] [blame]7678 never be used and has an undefined value. A load from the pointer that
7679 precedes this intrinsic can be replaced with
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7680 <tt>'<a href="#undefvalues">undef</a>'</tt>.</p>
7681
7682</div>
7683
7684<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7685<h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7686 <a name="int_lifetime_end">'<tt>llvm.lifetime.end</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7687</h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7688
7689<div class="doc_text">
7690
7691<h5>Syntax:</h5>
7692<pre>
7693 declare void @llvm.lifetime.end(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7694</pre>
7695
7696<h5>Overview:</h5>
7697<p>The '<tt>llvm.lifetime.end</tt>' intrinsic specifies the end of a memory
7698 object's lifetime.</p>
7699
7700<h5>Arguments:</h5>
Nick Lewycky9bc89042009-10-13 07:57:33 +0000[diff] [blame]7701<p>The first argument is a constant integer representing the size of the
7702 object, or -1 if it is variable sized. The second argument is a pointer to
7703 the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7704
7705<h5>Semantics:</h5>
7706<p>This intrinsic indicates that after this point in the code, the value of the
7707 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
7708 never be used and has an undefined value. Any stores into the memory object
7709 following this intrinsic may be removed as dead.
7710
7711</div>
7712
7713<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7714<h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7715 <a name="int_invariant_start">'<tt>llvm.invariant.start</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7716</h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7717
7718<div class="doc_text">
7719
7720<h5>Syntax:</h5>
7721<pre>
Nick Lewycky2965d3e2010-11-30 04:13:41 +0000[diff] [blame]7722 declare {}* @llvm.invariant.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7723</pre>
7724
7725<h5>Overview:</h5>
7726<p>The '<tt>llvm.invariant.start</tt>' intrinsic specifies that the contents of
7727 a memory object will not change.</p>
7728
7729<h5>Arguments:</h5>
Nick Lewycky9bc89042009-10-13 07:57:33 +0000[diff] [blame]7730<p>The first argument is a constant integer representing the size of the
7731 object, or -1 if it is variable sized. The second argument is a pointer to
7732 the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7733
7734<h5>Semantics:</h5>
7735<p>This intrinsic indicates that until an <tt>llvm.invariant.end</tt> that uses
7736 the return value, the referenced memory location is constant and
7737 unchanging.</p>
7738
7739</div>
7740
7741<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7742<h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7743 <a name="int_invariant_end">'<tt>llvm.invariant.end</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7744</h4>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7745
7746<div class="doc_text">
7747
7748<h5>Syntax:</h5>
7749<pre>
7750 declare void @llvm.invariant.end({}* &lt;start&gt;, i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7751</pre>
7752
7753<h5>Overview:</h5>
7754<p>The '<tt>llvm.invariant.end</tt>' intrinsic specifies that the contents of
7755 a memory object are mutable.</p>
7756
7757<h5>Arguments:</h5>
7758<p>The first argument is the matching <tt>llvm.invariant.start</tt> intrinsic.
Nick Lewycky9bc89042009-10-13 07:57:33 +0000[diff] [blame]7759 The second argument is a constant integer representing the size of the
7760 object, or -1 if it is variable sized and the third argument is a pointer
7761 to the object.</p>
Nick Lewycky6f7d8342009-10-13 07:03:23 +0000[diff] [blame]7762
7763<h5>Semantics:</h5>
7764<p>This intrinsic indicates that the memory is mutable again.</p>
7765
7766</div>
7767
Andrew Lenharth9b254ee2008-02-16 01:24:58 +0000[diff] [blame]7768<!-- ======================================================================= -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7769<h3>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7770 <a name="int_general">General Intrinsics</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7771</h3>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7772
7773<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7774
7775<p>This class of intrinsics is designed to be generic and has no specific
7776 purpose.</p>
7777
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7778</div>
7779
7780<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7781<h4>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7782 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7783</h4>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7784
7785<div class="doc_text">
7786
7787<h5>Syntax:</h5>
7788<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7789 declare void @llvm.var.annotation(i8* &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7790</pre>
7791
7792<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7793<p>The '<tt>llvm.var.annotation</tt>' intrinsic.</p>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7794
7795<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7796<p>The first argument is a pointer to a value, the second is a pointer to a
7797 global string, the third is a pointer to a global string which is the source
7798 file name, and the last argument is the line number.</p>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7799
7800<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7801<p>This intrinsic allows annotation of local variables with arbitrary strings.
7802 This can be useful for special purpose optimizations that want to look for
7803 these annotations. These have no other defined use, they are ignored by code
7804 generation and optimization.</p>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7805
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000[diff] [blame]7806</div>
7807
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7808<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7809<h4>
Tanya Lattner0186a652007-09-21 23:57:59 +0000[diff] [blame]7810 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7811</h4>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7812
7813<div class="doc_text">
7814
7815<h5>Syntax:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7816<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
7817 any integer bit width.</p>
7818
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7819<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7820 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7821 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7822 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7823 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7824 declare i256 @llvm.annotation.i256(i256 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7825</pre>
7826
7827<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7828<p>The '<tt>llvm.annotation</tt>' intrinsic.</p>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7829
7830<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7831<p>The first argument is an integer value (result of some expression), the
7832 second is a pointer to a global string, the third is a pointer to a global
7833 string which is the source file name, and the last argument is the line
7834 number. It returns the value of the first argument.</p>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7835
7836<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7837<p>This intrinsic allows annotations to be put on arbitrary expressions with
7838 arbitrary strings. This can be useful for special purpose optimizations that
7839 want to look for these annotations. These have no other defined use, they
7840 are ignored by code generation and optimization.</p>
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7841
Tanya Lattner293c0372007-09-21 22:59:12 +0000[diff] [blame]7842</div>
Jim Laskey2211f492007-03-14 19:31:19 +0000[diff] [blame]7843
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]7844<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7845<h4>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]7846 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7847</h4>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]7848
7849<div class="doc_text">
7850
7851<h5>Syntax:</h5>
7852<pre>
7853 declare void @llvm.trap()
7854</pre>
7855
7856<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7857<p>The '<tt>llvm.trap</tt>' intrinsic.</p>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]7858
7859<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7860<p>None.</p>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]7861
7862<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7863<p>This intrinsics is lowered to the target dependent trap instruction. If the
7864 target does not have a trap instruction, this intrinsic will be lowered to
7865 the call of the <tt>abort()</tt> function.</p>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]7866
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000[diff] [blame]7867</div>
7868
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7869<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7870<h4>
Misha Brukman50de2b22008-11-22 23:55:29 +0000[diff] [blame]7871 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7872</h4>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7873
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7874<div class="doc_text">
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7875
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7876<h5>Syntax:</h5>
7877<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7878 declare void @llvm.stackprotector(i8* &lt;guard&gt;, i8** &lt;slot&gt;)
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7879</pre>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7880
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7881<h5>Overview:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7882<p>The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and
7883 stores it onto the stack at <tt>slot</tt>. The stack slot is adjusted to
7884 ensure that it is placed on the stack before local variables.</p>
7885
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7886<h5>Arguments:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7887<p>The <tt>llvm.stackprotector</tt> intrinsic requires two pointer
7888 arguments. The first argument is the value loaded from the stack
7889 guard <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt>
7890 that has enough space to hold the value of the guard.</p>
7891
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7892<h5>Semantics:</h5>
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7893<p>This intrinsic causes the prologue/epilogue inserter to force the position of
7894 the <tt>AllocaInst</tt> stack slot to be before local variables on the
7895 stack. This is to ensure that if a local variable on the stack is
7896 overwritten, it will destroy the value of the guard. When the function exits,
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]7897 the guard on the stack is checked against the original guard. If they are
Bill Wendlingd9a66f72009-07-20 02:29:24 +0000[diff] [blame]7898 different, then the program aborts by calling the <tt>__stack_chk_fail()</tt>
7899 function.</p>
7900
Bill Wendling14313312008-11-19 05:56:17 +0000[diff] [blame]7901</div>
7902
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]7903<!-- _______________________________________________________________________ -->
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7904<h4>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]7905 <a name="int_objectsize">'<tt>llvm.objectsize</tt>' Intrinsic</a>
NAKAMURA Takumifc8d9302011-04-18 23:59:50 +0000[diff] [blame]7906</h4>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]7907
7908<div class="doc_text">
7909
7910<h5>Syntax:</h5>
7911<pre>
Dan Gohmand6a6f612010-05-28 17:07:41 +0000[diff] [blame]7912 declare i32 @llvm.objectsize.i32(i8* &lt;object&gt;, i1 &lt;type&gt;)
7913 declare i64 @llvm.objectsize.i64(i8* &lt;object&gt;, i1 &lt;type&gt;)
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]7914</pre>
7915
7916<h5>Overview:</h5>
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]7917<p>The <tt>llvm.objectsize</tt> intrinsic is designed to provide information to
7918 the optimizers to determine at compile time whether a) an operation (like
7919 memcpy) will overflow a buffer that corresponds to an object, or b) that a
7920 runtime check for overflow isn't necessary. An object in this context means
7921 an allocation of a specific class, structure, array, or other object.</p>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]7922
7923<h5>Arguments:</h5>
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]7924<p>The <tt>llvm.objectsize</tt> intrinsic takes two arguments. The first
Eric Christopher31e39bd2009-12-23 00:29:49 +0000[diff] [blame]7925 argument is a pointer to or into the <tt>object</tt>. The second argument
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]7926 is a boolean 0 or 1. This argument determines whether you want the
7927 maximum (0) or minimum (1) bytes remaining. This needs to be a literal 0 or
Eric Christopher31e39bd2009-12-23 00:29:49 +0000[diff] [blame]7928 1, variables are not allowed.</p>
7929
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]7930<h5>Semantics:</h5>
7931<p>The <tt>llvm.objectsize</tt> intrinsic is lowered to either a constant
Bill Wendling6bbe0912010-10-27 01:07:41 +0000[diff] [blame]7932 representing the size of the object concerned, or <tt>i32/i64 -1 or 0</tt>,
7933 depending on the <tt>type</tt> argument, if the size cannot be determined at
7934 compile time.</p>
Eric Christopher73484322009-11-30 08:03:53 +0000[diff] [blame]7935
7936</div>
7937
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]7938<!-- *********************************************************************** -->
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]7939<hr>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]7940<address>
7941 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman86242e12008-12-11 17:34:48 +0000[diff] [blame]7942 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]7943 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman86242e12008-12-11 17:34:48 +0000[diff] [blame]7944 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]7945
7946 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
NAKAMURA Takumica46f5a2011-04-09 02:13:37 +0000[diff] [blame]7947 <a href="http://llvm.org/">The LLVM Compiler Infrastructure</a><br>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]7948 Last modified: $Date$
7949</address>
Chris Lattnerb8f816e2008-01-04 04:33:49 +0000[diff] [blame]7950
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]7951</body>
7952</html>