Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / 5aa5aab112164bae0cafa5f17c18618ed2c3123b / . / docs / LangRef.html
blob: 5fe2316f8283fe18200d7d3a6edef8cde8afdead [file] [log] [blame]
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
3<html>
4<head>
5 <title>LLVM Assembly Language Reference Manual</title>
6 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
7 <meta name="author" content="Chris Lattner">
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]8 <meta name="description"
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]9 content="LLVM Assembly Language Reference Manual.">
10 <link rel="stylesheet" href="llvm.css" type="text/css">
11</head>
12
13<body>
14
15<div class="doc_title"> LLVM Language Reference Manual </div>
16<ol>
17 <li><a href="#abstract">Abstract</a></li>
18 <li><a href="#introduction">Introduction</a></li>
19 <li><a href="#identifiers">Identifiers</a></li>
20 <li><a href="#highlevel">High Level Structure</a>
21 <ol>
22 <li><a href="#modulestructure">Module Structure</a></li>
Bill Wendling41a07852009-07-20 01:03:30 +0000[diff] [blame]23 <li><a href="#linkage">Linkage Types</a>
24 <ol>
Bill Wendlinge2753242009-07-20 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>
27 <li><a href="#linkage_internal">'<tt>internal</tt>' Linkage</a></li>
28 <li><a href="#linkage_available_externally">'<tt>available_externally</tt>' Linkage</a></li>
29 <li><a href="#linkage_linkonce">'<tt>linkonce</tt>' Linkage</a></li>
30 <li><a href="#linkage_common">'<tt>common</tt>' Linkage</a></li>
31 <li><a href="#linkage_weak">'<tt>weak</tt>' Linkage</a></li>
32 <li><a href="#linkage_appending">'<tt>appending</tt>' Linkage</a></li>
33 <li><a href="#linkage_externweak">'<tt>extern_weak</tt>' Linkage</a></li>
Chris Lattner0fee5c22009-10-10 18:26:06 +0000[diff] [blame]34 <li><a href="#linkage_linkonce_odr">'<tt>linkonce_odr</tt>' Linkage</a></li>
Bill Wendlinge2753242009-07-20 02:41:50 +0000[diff] [blame]35 <li><a href="#linkage_weak">'<tt>weak_odr</tt>' Linkage</a></li>
36 <li><a href="#linkage_external">'<tt>externally visible</tt>' Linkage</a></li>
37 <li><a href="#linkage_dllimport">'<tt>dllimport</tt>' Linkage</a></li>
38 <li><a href="#linkage_dllexport">'<tt>dllexport</tt>' Linkage</a></li>
Bill Wendling41a07852009-07-20 01:03:30 +0000[diff] [blame]39 </ol>
40 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]41 <li><a href="#callingconv">Calling Conventions</a></li>
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]42 <li><a href="#namedtypes">Named Types</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]43 <li><a href="#globalvars">Global Variables</a></li>
44 <li><a href="#functionstructure">Functions</a></li>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]45 <li><a href="#aliasstructure">Aliases</a></li>
Devang Patela4bb6792010-01-11 19:35:55 +0000[diff] [blame]46 <li><a href="#namedmetadatastructure">Named Metadata</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]47 <li><a href="#paramattrs">Parameter Attributes</a></li>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]48 <li><a href="#fnattrs">Function Attributes</a></li>
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]49 <li><a href="#gc">Garbage Collector Names</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]50 <li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
51 <li><a href="#datalayout">Data Layout</a></li>
Dan Gohman27b47012009-07-27 18:07:55 +0000[diff] [blame]52 <li><a href="#pointeraliasing">Pointer Aliasing Rules</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]53 </ol>
54 </li>
55 <li><a href="#typesystem">Type System</a>
56 <ol>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]57 <li><a href="#t_classifications">Type Classifications</a></li>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]58 <li><a href="#t_primitive">Primitive Types</a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]59 <ol>
Nick Lewycky244cf482009-09-27 00:45:11 +0000[diff] [blame]60 <li><a href="#t_integer">Integer Type</a></li>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]61 <li><a href="#t_floating">Floating Point Types</a></li>
62 <li><a href="#t_void">Void Type</a></li>
63 <li><a href="#t_label">Label Type</a></li>
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]64 <li><a href="#t_metadata">Metadata Type</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]65 </ol>
66 </li>
67 <li><a href="#t_derived">Derived Types</a>
68 <ol>
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame]69 <li><a href="#t_aggregate">Aggregate Types</a>
70 <ol>
71 <li><a href="#t_array">Array Type</a></li>
72 <li><a href="#t_struct">Structure Type</a></li>
73 <li><a href="#t_pstruct">Packed Structure Type</a></li>
74 <li><a href="#t_union">Union Type</a></li>
75 <li><a href="#t_vector">Vector Type</a></li>
76 </ol>
77 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]78 <li><a href="#t_function">Function Type</a></li>
79 <li><a href="#t_pointer">Pointer Type</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]80 <li><a href="#t_opaque">Opaque Type</a></li>
81 </ol>
82 </li>
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]83 <li><a href="#t_uprefs">Type Up-references</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]84 </ol>
85 </li>
86 <li><a href="#constants">Constants</a>
87 <ol>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]88 <li><a href="#simpleconstants">Simple Constants</a></li>
Chris Lattner97063852009-02-28 18:32:25 +0000[diff] [blame]89 <li><a href="#complexconstants">Complex Constants</a></li>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]90 <li><a href="#globalconstants">Global Variable and Function Addresses</a></li>
91 <li><a href="#undefvalues">Undefined Values</a></li>
Chris Lattner29246b52009-10-27 21:19:13 +0000[diff] [blame]92 <li><a href="#blockaddress">Addresses of Basic Blocks</a></li>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]93 <li><a href="#constantexprs">Constant Expressions</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]94 </ol>
95 </li>
96 <li><a href="#othervalues">Other Values</a>
97 <ol>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]98 <li><a href="#inlineasm">Inline Assembler Expressions</a></li>
Devang Patela4bb6792010-01-11 19:35:55 +0000[diff] [blame]99 <li><a href="#metadata">Metadata Nodes and Metadata Strings</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]100 </ol>
101 </li>
Chris Lattner75c24e02009-07-20 05:55:19 +0000[diff] [blame]102 <li><a href="#intrinsic_globals">Intrinsic Global Variables</a>
103 <ol>
104 <li><a href="#intg_used">The '<tt>llvm.used</tt>' Global Variable</a></li>
Chris Lattner1e0e0d12009-07-20 06:14:25 +0000[diff] [blame]105 <li><a href="#intg_compiler_used">The '<tt>llvm.compiler.used</tt>'
106 Global Variable</a></li>
Chris Lattner75c24e02009-07-20 05:55:19 +0000[diff] [blame]107 <li><a href="#intg_global_ctors">The '<tt>llvm.global_ctors</tt>'
108 Global Variable</a></li>
109 <li><a href="#intg_global_dtors">The '<tt>llvm.global_dtors</tt>'
110 Global Variable</a></li>
111 </ol>
112 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]113 <li><a href="#instref">Instruction Reference</a>
114 <ol>
115 <li><a href="#terminators">Terminator Instructions</a>
116 <ol>
117 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
118 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
119 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
Chris Lattner4c3800f2009-10-28 00:19:10 +0000[diff] [blame]120 <li><a href="#i_indirectbr">'<tt>indirectbr</tt>' Instruction</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]121 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
122 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
123 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
124 </ol>
125 </li>
126 <li><a href="#binaryops">Binary Operations</a>
127 <ol>
128 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]129 <li><a href="#i_fadd">'<tt>fadd</tt>' Instruction</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]130 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]131 <li><a href="#i_fsub">'<tt>fsub</tt>' Instruction</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]132 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]133 <li><a href="#i_fmul">'<tt>fmul</tt>' Instruction</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]134 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
135 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
136 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
137 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
138 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
139 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
140 </ol>
141 </li>
142 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
143 <ol>
144 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
145 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
146 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
147 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
148 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
149 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
150 </ol>
151 </li>
152 <li><a href="#vectorops">Vector Operations</a>
153 <ol>
154 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
155 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
156 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
157 </ol>
158 </li>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]159 <li><a href="#aggregateops">Aggregate Operations</a>
160 <ol>
161 <li><a href="#i_extractvalue">'<tt>extractvalue</tt>' Instruction</a></li>
162 <li><a href="#i_insertvalue">'<tt>insertvalue</tt>' Instruction</a></li>
163 </ol>
164 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]165 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
166 <ol>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]167 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
168 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
169 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
170 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
171 </ol>
172 </li>
173 <li><a href="#convertops">Conversion Operations</a>
174 <ol>
175 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
176 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
177 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
178 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
179 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
180 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
181 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
182 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
183 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
184 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
185 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
186 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
187 </ol>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]188 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]189 <li><a href="#otherops">Other Operations</a>
190 <ol>
191 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
192 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
193 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
194 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
195 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
196 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
197 </ol>
198 </li>
199 </ol>
200 </li>
201 <li><a href="#intrinsics">Intrinsic Functions</a>
202 <ol>
203 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
204 <ol>
205 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
206 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
207 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
208 </ol>
209 </li>
210 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
211 <ol>
212 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
213 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
214 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
215 </ol>
216 </li>
217 <li><a href="#int_codegen">Code Generator Intrinsics</a>
218 <ol>
219 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
220 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
221 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
222 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
223 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
224 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
225 <li><a href="#int_readcyclecounter"><tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
226 </ol>
227 </li>
228 <li><a href="#int_libc">Standard C Library Intrinsics</a>
229 <ol>
230 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
231 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
232 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
233 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
234 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]235 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
236 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
237 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]238 </ol>
239 </li>
240 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
241 <ol>
242 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
243 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
244 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
245 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]246 </ol>
247 </li>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]248 <li><a href="#int_overflow">Arithmetic with Overflow Intrinsics</a>
249 <ol>
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]250 <li><a href="#int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt> Intrinsics</a></li>
251 <li><a href="#int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt> Intrinsics</a></li>
252 <li><a href="#int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt> Intrinsics</a></li>
253 <li><a href="#int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt> Intrinsics</a></li>
254 <li><a href="#int_smul_overflow">'<tt>llvm.smul.with.overflow.*</tt> Intrinsics</a></li>
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]255 <li><a href="#int_umul_overflow">'<tt>llvm.umul.with.overflow.*</tt> Intrinsics</a></li>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]256 </ol>
257 </li>
Anton Korobeynikov8c98a892010-03-14 18:42:47 +0000[diff] [blame]258 <li><a href="#int_fp16">Half Precision Floating Point Intrinsics</a>
259 <ol>
Chris Lattnerebc48e52010-03-14 23:03:31 +0000[diff] [blame]260 <li><a href="#int_convert_to_fp16">'<tt>llvm.convert.to.fp16</tt>' Intrinsic</a></li>
261 <li><a href="#int_convert_from_fp16">'<tt>llvm.convert.from.fp16</tt>' Intrinsic</a></li>
Anton Korobeynikov8c98a892010-03-14 18:42:47 +0000[diff] [blame]262 </ol>
263 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]264 <li><a href="#int_debugger">Debugger intrinsics</a></li>
265 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]266 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]267 <ol>
268 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]269 </ol>
270 </li>
Bill Wendling9127adb2008-11-18 22:10:53 +0000[diff] [blame]271 <li><a href="#int_atomics">Atomic intrinsics</a>
272 <ol>
273 <li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
274 <li><a href="#int_atomic_cmp_swap"><tt>llvm.atomic.cmp.swap</tt></a></li>
275 <li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
276 <li><a href="#int_atomic_load_add"><tt>llvm.atomic.load.add</tt></a></li>
277 <li><a href="#int_atomic_load_sub"><tt>llvm.atomic.load.sub</tt></a></li>
278 <li><a href="#int_atomic_load_and"><tt>llvm.atomic.load.and</tt></a></li>
279 <li><a href="#int_atomic_load_nand"><tt>llvm.atomic.load.nand</tt></a></li>
280 <li><a href="#int_atomic_load_or"><tt>llvm.atomic.load.or</tt></a></li>
281 <li><a href="#int_atomic_load_xor"><tt>llvm.atomic.load.xor</tt></a></li>
282 <li><a href="#int_atomic_load_max"><tt>llvm.atomic.load.max</tt></a></li>
283 <li><a href="#int_atomic_load_min"><tt>llvm.atomic.load.min</tt></a></li>
284 <li><a href="#int_atomic_load_umax"><tt>llvm.atomic.load.umax</tt></a></li>
285 <li><a href="#int_atomic_load_umin"><tt>llvm.atomic.load.umin</tt></a></li>
286 </ol>
287 </li>
Nick Lewyckyc888d352009-10-13 07:03:23 +0000[diff] [blame]288 <li><a href="#int_memorymarkers">Memory Use Markers</a>
289 <ol>
290 <li><a href="#int_lifetime_start"><tt>llvm.lifetime.start</tt></a></li>
291 <li><a href="#int_lifetime_end"><tt>llvm.lifetime.end</tt></a></li>
292 <li><a href="#int_invariant_start"><tt>llvm.invariant.start</tt></a></li>
293 <li><a href="#int_invariant_end"><tt>llvm.invariant.end</tt></a></li>
294 </ol>
295 </li>
Reid Spencerb043f672007-07-20 19:59:11 +0000[diff] [blame]296 <li><a href="#int_general">General intrinsics</a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]297 <ol>
Reid Spencerb043f672007-07-20 19:59:11 +0000[diff] [blame]298 <li><a href="#int_var_annotation">
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]299 '<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]300 <li><a href="#int_annotation">
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]301 '<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]302 <li><a href="#int_trap">
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]303 '<tt>llvm.trap</tt>' Intrinsic</a></li>
304 <li><a href="#int_stackprotector">
305 '<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
Eric Christopher767a3722009-11-30 08:03:53 +0000[diff] [blame]306 <li><a href="#int_objectsize">
307 '<tt>llvm.objectsize</tt>' Intrinsic</a></li>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]308 </ol>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]309 </li>
310 </ol>
311 </li>
312</ol>
313
314<div class="doc_author">
315 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
316 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
317</div>
318
319<!-- *********************************************************************** -->
320<div class="doc_section"> <a name="abstract">Abstract </a></div>
321<!-- *********************************************************************** -->
322
323<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]324
325<p>This document is a reference manual for the LLVM assembly language. LLVM is
326 a Static Single Assignment (SSA) based representation that provides type
327 safety, low-level operations, flexibility, and the capability of representing
328 'all' high-level languages cleanly. It is the common code representation
329 used throughout all phases of the LLVM compilation strategy.</p>
330
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]331</div>
332
333<!-- *********************************************************************** -->
334<div class="doc_section"> <a name="introduction">Introduction</a> </div>
335<!-- *********************************************************************** -->
336
337<div class="doc_text">
338
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]339<p>The LLVM code representation is designed to be used in three different forms:
340 as an in-memory compiler IR, as an on-disk bitcode representation (suitable
341 for fast loading by a Just-In-Time compiler), and as a human readable
342 assembly language representation. This allows LLVM to provide a powerful
343 intermediate representation for efficient compiler transformations and
344 analysis, while providing a natural means to debug and visualize the
345 transformations. The three different forms of LLVM are all equivalent. This
346 document describes the human readable representation and notation.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]347
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]348<p>The LLVM representation aims to be light-weight and low-level while being
349 expressive, typed, and extensible at the same time. It aims to be a
350 "universal IR" of sorts, by being at a low enough level that high-level ideas
351 may be cleanly mapped to it (similar to how microprocessors are "universal
352 IR's", allowing many source languages to be mapped to them). By providing
353 type information, LLVM can be used as the target of optimizations: for
354 example, through pointer analysis, it can be proven that a C automatic
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]355 variable is never accessed outside of the current function, allowing it to
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]356 be promoted to a simple SSA value instead of a memory location.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]357
358</div>
359
360<!-- _______________________________________________________________________ -->
361<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
362
363<div class="doc_text">
364
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]365<p>It is important to note that this document describes 'well formed' LLVM
366 assembly language. There is a difference between what the parser accepts and
367 what is considered 'well formed'. For example, the following instruction is
368 syntactically okay, but not well formed:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]369
370<div class="doc_code">
371<pre>
372%x = <a href="#i_add">add</a> i32 1, %x
373</pre>
374</div>
375
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]376<p>because the definition of <tt>%x</tt> does not dominate all of its uses. The
377 LLVM infrastructure provides a verification pass that may be used to verify
378 that an LLVM module is well formed. This pass is automatically run by the
379 parser after parsing input assembly and by the optimizer before it outputs
380 bitcode. The violations pointed out by the verifier pass indicate bugs in
381 transformation passes or input to the parser.</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]382
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]383</div>
384
Chris Lattnera83fdc02007-10-03 17:34:29 +0000[diff] [blame]385<!-- Describe the typesetting conventions here. -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]386
387<!-- *********************************************************************** -->
388<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
389<!-- *********************************************************************** -->
390
391<div class="doc_text">
392
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]393<p>LLVM identifiers come in two basic types: global and local. Global
394 identifiers (functions, global variables) begin with the <tt>'@'</tt>
395 character. Local identifiers (register names, types) begin with
396 the <tt>'%'</tt> character. Additionally, there are three different formats
397 for identifiers, for different purposes:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]398
399<ol>
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]400 <li>Named values are represented as a string of characters with their prefix.
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]401 For example, <tt>%foo</tt>, <tt>@DivisionByZero</tt>,
402 <tt>%a.really.long.identifier</tt>. The actual regular expression used is
403 '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'. Identifiers which require
404 other characters in their names can be surrounded with quotes. Special
405 characters may be escaped using <tt>"\xx"</tt> where <tt>xx</tt> is the
406 ASCII code for the character in hexadecimal. In this way, any character
407 can be used in a name value, even quotes themselves.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]408
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]409 <li>Unnamed values are represented as an unsigned numeric value with their
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]410 prefix. For example, <tt>%12</tt>, <tt>@2</tt>, <tt>%44</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]411
412 <li>Constants, which are described in a <a href="#constants">section about
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]413 constants</a>, below.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]414</ol>
415
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]416<p>LLVM requires that values start with a prefix for two reasons: Compilers
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]417 don't need to worry about name clashes with reserved words, and the set of
418 reserved words may be expanded in the future without penalty. Additionally,
419 unnamed identifiers allow a compiler to quickly come up with a temporary
420 variable without having to avoid symbol table conflicts.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]421
422<p>Reserved words in LLVM are very similar to reserved words in other
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]423 languages. There are keywords for different opcodes
424 ('<tt><a href="#i_add">add</a></tt>',
425 '<tt><a href="#i_bitcast">bitcast</a></tt>',
426 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names
427 ('<tt><a href="#t_void">void</a></tt>',
428 '<tt><a href="#t_primitive">i32</a></tt>', etc...), and others. These
429 reserved words cannot conflict with variable names, because none of them
430 start with a prefix character (<tt>'%'</tt> or <tt>'@'</tt>).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]431
432<p>Here is an example of LLVM code to multiply the integer variable
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]433 '<tt>%X</tt>' by 8:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]434
435<p>The easy way:</p>
436
437<div class="doc_code">
438<pre>
439%result = <a href="#i_mul">mul</a> i32 %X, 8
440</pre>
441</div>
442
443<p>After strength reduction:</p>
444
445<div class="doc_code">
446<pre>
447%result = <a href="#i_shl">shl</a> i32 %X, i8 3
448</pre>
449</div>
450
451<p>And the hard way:</p>
452
453<div class="doc_code">
454<pre>
Gabor Greifc0ea7672009-10-28 13:05:07 +0000[diff] [blame]455%0 = <a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
456%1 = <a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]457%result = <a href="#i_add">add</a> i32 %1, %1
458</pre>
459</div>
460
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]461<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several important
462 lexical features of LLVM:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]463
464<ol>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]465 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]466 line.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]467
468 <li>Unnamed temporaries are created when the result of a computation is not
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]469 assigned to a named value.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]470
471 <li>Unnamed temporaries are numbered sequentially</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]472</ol>
473
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]474<p>It also shows a convention that we follow in this document. When
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]475 demonstrating instructions, we will follow an instruction with a comment that
476 defines the type and name of value produced. Comments are shown in italic
477 text.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]478
479</div>
480
481<!-- *********************************************************************** -->
482<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
483<!-- *********************************************************************** -->
484
485<!-- ======================================================================= -->
486<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
487</div>
488
489<div class="doc_text">
490
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]491<p>LLVM programs are composed of "Module"s, each of which is a translation unit
492 of the input programs. Each module consists of functions, global variables,
493 and symbol table entries. Modules may be combined together with the LLVM
494 linker, which merges function (and global variable) definitions, resolves
495 forward declarations, and merges symbol table entries. Here is an example of
496 the "hello world" module:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]497
498<div class="doc_code">
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]499<pre>
500<i>; Declare the string constant as a global constant.</i>
501<a href="#identifiers">@.LC0</a> = <a href="#linkage_internal">internal</a> <a href="#globalvars">constant</a> <a href="#t_array">[13 x i8]</a> c"hello world\0A\00" <i>; [13 x i8]*</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]502
503<i>; External declaration of the puts function</i>
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]504<a href="#functionstructure">declare</a> i32 @puts(i8 *) <i>; i32(i8 *)* </i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]505
506<i>; Definition of main function</i>
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]507define i32 @main() { <i>; i32()* </i>
508 <i>; Convert [13 x i8]* to i8 *...</i>
509 %cast210 = <a href="#i_getelementptr">getelementptr</a> [13 x i8]* @.LC0, i64 0, i64 0 <i>; i8 *</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]510
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]511 <i>; Call puts function to write out the string to stdout.</i>
512 <a href="#i_call">call</a> i32 @puts(i8 * %cast210) <i>; i32</i>
Devang Patela4bb6792010-01-11 19:35:55 +0000[diff] [blame]513 <a href="#i_ret">ret</a> i32 0<br>}
514
515<i>; Named metadata</i>
516!1 = metadata !{i32 41}
517!foo = !{!1, null}
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]518</pre>
519</div>
520
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]521<p>This example is made up of a <a href="#globalvars">global variable</a> named
Devang Patela4bb6792010-01-11 19:35:55 +0000[diff] [blame]522 "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>" function,
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]523 a <a href="#functionstructure">function definition</a> for
Devang Patela4bb6792010-01-11 19:35:55 +0000[diff] [blame]524 "<tt>main</tt>" and <a href="#namedmetadatastructure">named metadata</a>
525 "<tt>foo"</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]526
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]527<p>In general, a module is made up of a list of global values, where both
528 functions and global variables are global values. Global values are
529 represented by a pointer to a memory location (in this case, a pointer to an
530 array of char, and a pointer to a function), and have one of the
531 following <a href="#linkage">linkage types</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]532
533</div>
534
535<!-- ======================================================================= -->
536<div class="doc_subsection">
537 <a name="linkage">Linkage Types</a>
538</div>
539
540<div class="doc_text">
541
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]542<p>All Global Variables and Functions have one of the following types of
543 linkage:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]544
545<dl>
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]546 <dt><tt><b><a name="linkage_private">private</a></b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]547 <dd>Global values with private linkage are only directly accessible by objects
548 in the current module. In particular, linking code into a module with an
549 private global value may cause the private to be renamed as necessary to
550 avoid collisions. Because the symbol is private to the module, all
551 references can be updated. This doesn't show up in any symbol table in the
552 object file.</dd>
Rafael Espindolaa168fc92009-01-15 20:18:42 +0000[diff] [blame]553
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]554 <dt><tt><b><a name="linkage_linker_private">linker_private</a></b></tt></dt>
Bill Wendling41a07852009-07-20 01:03:30 +0000[diff] [blame]555 <dd>Similar to private, but the symbol is passed through the assembler and
Chris Lattner06c1ecc2009-08-24 04:32:16 +0000[diff] [blame]556 removed by the linker after evaluation. Note that (unlike private
557 symbols) linker_private symbols are subject to coalescing by the linker:
558 weak symbols get merged and redefinitions are rejected. However, unlike
559 normal strong symbols, they are removed by the linker from the final
560 linked image (executable or dynamic library).</dd>
Bill Wendling41a07852009-07-20 01:03:30 +0000[diff] [blame]561
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]562 <dt><tt><b><a name="linkage_internal">internal</a></b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]563 <dd>Similar to private, but the value shows as a local symbol
564 (<tt>STB_LOCAL</tt> in the case of ELF) in the object file. This
565 corresponds to the notion of the '<tt>static</tt>' keyword in C.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]566
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]567 <dt><tt><b><a name="linkage_available_externally">available_externally</a></b></tt></dt>
Chris Lattner68433442009-04-13 05:44:34 +0000[diff] [blame]568 <dd>Globals with "<tt>available_externally</tt>" linkage are never emitted
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]569 into the object file corresponding to the LLVM module. They exist to
570 allow inlining and other optimizations to take place given knowledge of
571 the definition of the global, which is known to be somewhere outside the
572 module. Globals with <tt>available_externally</tt> linkage are allowed to
573 be discarded at will, and are otherwise the same as <tt>linkonce_odr</tt>.
574 This linkage type is only allowed on definitions, not declarations.</dd>
Chris Lattner68433442009-04-13 05:44:34 +0000[diff] [blame]575
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]576 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]577 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
Chris Lattnerd16c5512010-01-09 19:15:14 +0000[diff] [blame]578 the same name when linkage occurs. This can be used to implement
579 some forms of inline functions, templates, or other code which must be
580 generated in each translation unit that uses it, but where the body may
581 be overridden with a more definitive definition later. Unreferenced
582 <tt>linkonce</tt> globals are allowed to be discarded. Note that
583 <tt>linkonce</tt> linkage does not actually allow the optimizer to
584 inline the body of this function into callers because it doesn't know if
585 this definition of the function is the definitive definition within the
586 program or whether it will be overridden by a stronger definition.
587 To enable inlining and other optimizations, use "<tt>linkonce_odr</tt>"
588 linkage.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]589
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]590 <dt><tt><b><a name="linkage_weak">weak</a></b></tt></dt>
Chris Lattner7215c7f2009-08-05 05:21:07 +0000[diff] [blame]591 <dd>"<tt>weak</tt>" linkage has the same merging semantics as
592 <tt>linkonce</tt> linkage, except that unreferenced globals with
593 <tt>weak</tt> linkage may not be discarded. This is used for globals that
594 are declared "weak" in C source code.</dd>
595
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]596 <dt><tt><b><a name="linkage_common">common</a></b></tt></dt>
Chris Lattner7215c7f2009-08-05 05:21:07 +0000[diff] [blame]597 <dd>"<tt>common</tt>" linkage is most similar to "<tt>weak</tt>" linkage, but
598 they are used for tentative definitions in C, such as "<tt>int X;</tt>" at
599 global scope.
600 Symbols with "<tt>common</tt>" linkage are merged in the same way as
601 <tt>weak symbols</tt>, and they may not be deleted if unreferenced.
Chris Lattner83c4a712009-08-05 05:41:44 +0000[diff] [blame]602 <tt>common</tt> symbols may not have an explicit section,
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]603 must have a zero initializer, and may not be marked '<a
Chris Lattner83c4a712009-08-05 05:41:44 +0000[diff] [blame]604 href="#globalvars"><tt>constant</tt></a>'. Functions and aliases may not
605 have common linkage.</dd>
Chris Lattner7215c7f2009-08-05 05:21:07 +0000[diff] [blame]606
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]607
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]608 <dt><tt><b><a name="linkage_appending">appending</a></b></tt></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]609 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]610 pointer to array type. When two global variables with appending linkage
611 are linked together, the two global arrays are appended together. This is
612 the LLVM, typesafe, equivalent of having the system linker append together
613 "sections" with identical names when .o files are linked.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]614
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]615 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]616 <dd>The semantics of this linkage follow the ELF object file model: the symbol
617 is weak until linked, if not linked, the symbol becomes null instead of
618 being an undefined reference.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]619
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]620 <dt><tt><b><a name="linkage_linkonce_odr">linkonce_odr</a></b></tt></dt>
621 <dt><tt><b><a name="linkage_weak_odr">weak_odr</a></b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]622 <dd>Some languages allow differing globals to be merged, such as two functions
623 with different semantics. Other languages, such as <tt>C++</tt>, ensure
624 that only equivalent globals are ever merged (the "one definition rule" -
625 "ODR"). Such languages can use the <tt>linkonce_odr</tt>
626 and <tt>weak_odr</tt> linkage types to indicate that the global will only
627 be merged with equivalent globals. These linkage types are otherwise the
628 same as their non-<tt>odr</tt> versions.</dd>
Duncan Sands19d161f2009-03-07 15:45:40 +0000[diff] [blame]629
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]630 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]631 <dd>If none of the above identifiers are used, the global is externally
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]632 visible, meaning that it participates in linkage and can be used to
633 resolve external symbol references.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]634</dl>
635
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]636<p>The next two types of linkage are targeted for Microsoft Windows platform
637 only. They are designed to support importing (exporting) symbols from (to)
638 DLLs (Dynamic Link Libraries).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]639
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]640<dl>
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]641 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]642 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]643 or variable via a global pointer to a pointer that is set up by the DLL
644 exporting the symbol. On Microsoft Windows targets, the pointer name is
645 formed by combining <code>__imp_</code> and the function or variable
646 name.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]647
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]648 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]649 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]650 pointer to a pointer in a DLL, so that it can be referenced with the
651 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
652 name is formed by combining <code>__imp_</code> and the function or
653 variable name.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]654</dl>
655
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]656<p>For example, since the "<tt>.LC0</tt>" variable is defined to be internal, if
657 another module defined a "<tt>.LC0</tt>" variable and was linked with this
658 one, one of the two would be renamed, preventing a collision. Since
659 "<tt>main</tt>" and "<tt>puts</tt>" are external (i.e., lacking any linkage
660 declarations), they are accessible outside of the current module.</p>
661
662<p>It is illegal for a function <i>declaration</i> to have any linkage type
663 other than "externally visible", <tt>dllimport</tt>
664 or <tt>extern_weak</tt>.</p>
665
Duncan Sands19d161f2009-03-07 15:45:40 +0000[diff] [blame]666<p>Aliases can have only <tt>external</tt>, <tt>internal</tt>, <tt>weak</tt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]667 or <tt>weak_odr</tt> linkages.</p>
668
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]669</div>
670
671<!-- ======================================================================= -->
672<div class="doc_subsection">
673 <a name="callingconv">Calling Conventions</a>
674</div>
675
676<div class="doc_text">
677
678<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]679 and <a href="#i_invoke">invokes</a> can all have an optional calling
680 convention specified for the call. The calling convention of any pair of
681 dynamic caller/callee must match, or the behavior of the program is
682 undefined. The following calling conventions are supported by LLVM, and more
683 may be added in the future:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]684
685<dl>
686 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]687 <dd>This calling convention (the default if no other calling convention is
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]688 specified) matches the target C calling conventions. This calling
689 convention supports varargs function calls and tolerates some mismatch in
690 the declared prototype and implemented declaration of the function (as
691 does normal C).</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]692
693 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]694 <dd>This calling convention attempts to make calls as fast as possible
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]695 (e.g. by passing things in registers). This calling convention allows the
696 target to use whatever tricks it wants to produce fast code for the
697 target, without having to conform to an externally specified ABI
Jeffrey Yasskin3e1a3a02010-01-09 19:44:16 +0000[diff] [blame]698 (Application Binary Interface).
699 <a href="CodeGenerator.html#tailcallopt">Tail calls can only be optimized
Chris Lattnerac9a9392010-03-11 00:22:57 +0000[diff] [blame]700 when this or the GHC convention is used.</a> This calling convention
701 does not support varargs and requires the prototype of all callees to
702 exactly match the prototype of the function definition.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]703
704 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]705 <dd>This calling convention attempts to make code in the caller as efficient
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]706 as possible under the assumption that the call is not commonly executed.
707 As such, these calls often preserve all registers so that the call does
708 not break any live ranges in the caller side. This calling convention
709 does not support varargs and requires the prototype of all callees to
710 exactly match the prototype of the function definition.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]711
Chris Lattnerac9a9392010-03-11 00:22:57 +0000[diff] [blame]712 <dt><b>"<tt>cc <em>10</em></tt>" - GHC convention</b>:</dt>
713 <dd>This calling convention has been implemented specifically for use by the
714 <a href="http://www.haskell.org/ghc">Glasgow Haskell Compiler (GHC)</a>.
715 It passes everything in registers, going to extremes to achieve this by
716 disabling callee save registers. This calling convention should not be
717 used lightly but only for specific situations such as an alternative to
718 the <em>register pinning</em> performance technique often used when
719 implementing functional programming languages.At the moment only X86
720 supports this convention and it has the following limitations:
721 <ul>
722 <li>On <em>X86-32</em> only supports up to 4 bit type parameters. No
723 floating point types are supported.</li>
724 <li>On <em>X86-64</em> only supports up to 10 bit type parameters and
725 6 floating point parameters.</li>
726 </ul>
727 This calling convention supports
728 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> but
729 requires both the caller and callee are using it.
730 </dd>
731
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]732 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]733 <dd>Any calling convention may be specified by number, allowing
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]734 target-specific calling conventions to be used. Target specific calling
735 conventions start at 64.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]736</dl>
737
738<p>More calling conventions can be added/defined on an as-needed basis, to
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]739 support Pascal conventions or any other well-known target-independent
740 convention.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]741
742</div>
743
744<!-- ======================================================================= -->
745<div class="doc_subsection">
746 <a name="visibility">Visibility Styles</a>
747</div>
748
749<div class="doc_text">
750
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]751<p>All Global Variables and Functions have one of the following visibility
752 styles:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]753
754<dl>
755 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]756 <dd>On targets that use the ELF object file format, default visibility means
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]757 that the declaration is visible to other modules and, in shared libraries,
758 means that the declared entity may be overridden. On Darwin, default
759 visibility means that the declaration is visible to other modules. Default
760 visibility corresponds to "external linkage" in the language.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]761
762 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]763 <dd>Two declarations of an object with hidden visibility refer to the same
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]764 object if they are in the same shared object. Usually, hidden visibility
765 indicates that the symbol will not be placed into the dynamic symbol
766 table, so no other module (executable or shared library) can reference it
767 directly.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]768
769 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]770 <dd>On ELF, protected visibility indicates that the symbol will be placed in
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]771 the dynamic symbol table, but that references within the defining module
772 will bind to the local symbol. That is, the symbol cannot be overridden by
773 another module.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]774</dl>
775
776</div>
777
778<!-- ======================================================================= -->
779<div class="doc_subsection">
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]780 <a name="namedtypes">Named Types</a>
781</div>
782
783<div class="doc_text">
784
785<p>LLVM IR allows you to specify name aliases for certain types. This can make
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]786 it easier to read the IR and make the IR more condensed (particularly when
787 recursive types are involved). An example of a name specification is:</p>
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]788
789<div class="doc_code">
790<pre>
791%mytype = type { %mytype*, i32 }
792</pre>
793</div>
794
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]795<p>You may give a name to any <a href="#typesystem">type</a> except
796 "<a href="t_void">void</a>". Type name aliases may be used anywhere a type
797 is expected with the syntax "%mytype".</p>
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]798
799<p>Note that type names are aliases for the structural type that they indicate,
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]800 and that you can therefore specify multiple names for the same type. This
801 often leads to confusing behavior when dumping out a .ll file. Since LLVM IR
802 uses structural typing, the name is not part of the type. When printing out
803 LLVM IR, the printer will pick <em>one name</em> to render all types of a
804 particular shape. This means that if you have code where two different
805 source types end up having the same LLVM type, that the dumper will sometimes
806 print the "wrong" or unexpected type. This is an important design point and
807 isn't going to change.</p>
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]808
809</div>
810
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]811<!-- ======================================================================= -->
812<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]813 <a name="globalvars">Global Variables</a>
814</div>
815
816<div class="doc_text">
817
818<p>Global variables define regions of memory allocated at compilation time
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]819 instead of run-time. Global variables may optionally be initialized, may
820 have an explicit section to be placed in, and may have an optional explicit
821 alignment specified. A variable may be defined as "thread_local", which
822 means that it will not be shared by threads (each thread will have a
823 separated copy of the variable). A variable may be defined as a global
824 "constant," which indicates that the contents of the variable
825 will <b>never</b> be modified (enabling better optimization, allowing the
826 global data to be placed in the read-only section of an executable, etc).
827 Note that variables that need runtime initialization cannot be marked
828 "constant" as there is a store to the variable.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]829
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]830<p>LLVM explicitly allows <em>declarations</em> of global variables to be marked
831 constant, even if the final definition of the global is not. This capability
832 can be used to enable slightly better optimization of the program, but
833 requires the language definition to guarantee that optimizations based on the
834 'constantness' are valid for the translation units that do not include the
835 definition.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]836
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]837<p>As SSA values, global variables define pointer values that are in scope
838 (i.e. they dominate) all basic blocks in the program. Global variables
839 always define a pointer to their "content" type because they describe a
840 region of memory, and all memory objects in LLVM are accessed through
841 pointers.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]842
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]843<p>A global variable may be declared to reside in a target-specific numbered
844 address space. For targets that support them, address spaces may affect how
845 optimizations are performed and/or what target instructions are used to
846 access the variable. The default address space is zero. The address space
847 qualifier must precede any other attributes.</p>
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]848
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]849<p>LLVM allows an explicit section to be specified for globals. If the target
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]850 supports it, it will emit globals to the section specified.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]851
852<p>An explicit alignment may be specified for a global. If not present, or if
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]853 the alignment is set to zero, the alignment of the global is set by the
854 target to whatever it feels convenient. If an explicit alignment is
855 specified, the global is forced to have at least that much alignment. All
856 alignments must be a power of 2.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]857
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]858<p>For example, the following defines a global in a numbered address space with
859 an initializer, section, and alignment:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]860
861<div class="doc_code">
862<pre>
Dan Gohman21ef02c2009-01-11 00:40:00 +0000[diff] [blame]863@G = addrspace(5) constant float 1.0, section "foo", align 4
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]864</pre>
865</div>
866
867</div>
868
869
870<!-- ======================================================================= -->
871<div class="doc_subsection">
872 <a name="functionstructure">Functions</a>
873</div>
874
875<div class="doc_text">
876
Dan Gohman22dc6682010-03-01 17:41:39 +0000[diff] [blame]877<p>LLVM function definitions consist of the "<tt>define</tt>" keyword, an
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]878 optional <a href="#linkage">linkage type</a>, an optional
879 <a href="#visibility">visibility style</a>, an optional
880 <a href="#callingconv">calling convention</a>, a return type, an optional
881 <a href="#paramattrs">parameter attribute</a> for the return type, a function
882 name, a (possibly empty) argument list (each with optional
883 <a href="#paramattrs">parameter attributes</a>), optional
884 <a href="#fnattrs">function attributes</a>, an optional section, an optional
885 alignment, an optional <a href="#gc">garbage collector name</a>, an opening
886 curly brace, a list of basic blocks, and a closing curly brace.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]887
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]888<p>LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
889 optional <a href="#linkage">linkage type</a>, an optional
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]890 <a href="#visibility">visibility style</a>, an optional
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]891 <a href="#callingconv">calling convention</a>, a return type, an optional
892 <a href="#paramattrs">parameter attribute</a> for the return type, a function
893 name, a possibly empty list of arguments, an optional alignment, and an
894 optional <a href="#gc">garbage collector name</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]895
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]896<p>A function definition contains a list of basic blocks, forming the CFG
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]897 (Control Flow Graph) for the function. Each basic block may optionally start
898 with a label (giving the basic block a symbol table entry), contains a list
899 of instructions, and ends with a <a href="#terminators">terminator</a>
900 instruction (such as a branch or function return).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]901
902<p>The first basic block in a function is special in two ways: it is immediately
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]903 executed on entrance to the function, and it is not allowed to have
904 predecessor basic blocks (i.e. there can not be any branches to the entry
905 block of a function). Because the block can have no predecessors, it also
906 cannot have any <a href="#i_phi">PHI nodes</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]907
908<p>LLVM allows an explicit section to be specified for functions. If the target
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]909 supports it, it will emit functions to the section specified.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]910
911<p>An explicit alignment may be specified for a function. If not present, or if
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]912 the alignment is set to zero, the alignment of the function is set by the
913 target to whatever it feels convenient. If an explicit alignment is
914 specified, the function is forced to have at least that much alignment. All
915 alignments must be a power of 2.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]916
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]917<h5>Syntax:</h5>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]918<div class="doc_code">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]919<pre>
Chris Lattner1e5c5cd02008-10-13 16:55:18 +0000[diff] [blame]920define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]921 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
922 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
923 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
924 [<a href="#gc">gc</a>] { ... }
925</pre>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]926</div>
927
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]928</div>
929
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]930<!-- ======================================================================= -->
931<div class="doc_subsection">
932 <a name="aliasstructure">Aliases</a>
933</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]934
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]935<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]936
937<p>Aliases act as "second name" for the aliasee value (which can be either
938 function, global variable, another alias or bitcast of global value). Aliases
939 may have an optional <a href="#linkage">linkage type</a>, and an
940 optional <a href="#visibility">visibility style</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]941
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]942<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]943<div class="doc_code">
944<pre>
Duncan Sandsd7bfabf2008-09-12 20:48:21 +0000[diff] [blame]945@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]946</pre>
947</div>
948
949</div>
950
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]951<!-- ======================================================================= -->
Devang Patela4bb6792010-01-11 19:35:55 +0000[diff] [blame]952<div class="doc_subsection">
953 <a name="namedmetadatastructure">Named Metadata</a>
954</div>
955
956<div class="doc_text">
957
Chris Lattnerd0d96292010-01-15 21:50:19 +0000[diff] [blame]958<p>Named metadata is a collection of metadata. <a href="#metadata">Metadata
959 nodes</a> (but not metadata strings) and null are the only valid operands for
960 a named metadata.</p>
Devang Patela4bb6792010-01-11 19:35:55 +0000[diff] [blame]961
962<h5>Syntax:</h5>
963<div class="doc_code">
964<pre>
965!1 = metadata !{metadata !"one"}
966!name = !{null, !1}
967</pre>
968</div>
969
970</div>
971
972<!-- ======================================================================= -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]973<div class="doc_subsection"><a name="paramattrs">Parameter Attributes</a></div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]974
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]975<div class="doc_text">
976
977<p>The return type and each parameter of a function type may have a set of
978 <i>parameter attributes</i> associated with them. Parameter attributes are
979 used to communicate additional information about the result or parameters of
980 a function. Parameter attributes are considered to be part of the function,
981 not of the function type, so functions with different parameter attributes
982 can have the same function type.</p>
983
984<p>Parameter attributes are simple keywords that follow the type specified. If
985 multiple parameter attributes are needed, they are space separated. For
986 example:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]987
988<div class="doc_code">
989<pre>
Nick Lewycky3022a742009-02-15 23:06:14 +0000[diff] [blame]990declare i32 @printf(i8* noalias nocapture, ...)
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]991declare i32 @atoi(i8 zeroext)
992declare signext i8 @returns_signed_char()
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]993</pre>
994</div>
995
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]996<p>Note that any attributes for the function result (<tt>nounwind</tt>,
997 <tt>readonly</tt>) come immediately after the argument list.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]998
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]999<p>Currently, only the following parameter attributes are defined:</p>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]1000
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1001<dl>
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1002 <dt><tt><b>zeroext</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1003 <dd>This indicates to the code generator that the parameter or return value
1004 should be zero-extended to a 32-bit value by the caller (for a parameter)
1005 or the callee (for a return value).</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]1006
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1007 <dt><tt><b>signext</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1008 <dd>This indicates to the code generator that the parameter or return value
1009 should be sign-extended to a 32-bit value by the caller (for a parameter)
1010 or the callee (for a return value).</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]1011
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1012 <dt><tt><b>inreg</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1013 <dd>This indicates that this parameter or return value should be treated in a
1014 special target-dependent fashion during while emitting code for a function
1015 call or return (usually, by putting it in a register as opposed to memory,
1016 though some targets use it to distinguish between two different kinds of
1017 registers). Use of this attribute is target-specific.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]1018
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1019 <dt><tt><b><a name="byval">byval</a></b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1020 <dd>This indicates that the pointer parameter should really be passed by value
1021 to the function. The attribute implies that a hidden copy of the pointee
1022 is made between the caller and the callee, so the callee is unable to
1023 modify the value in the callee. This attribute is only valid on LLVM
1024 pointer arguments. It is generally used to pass structs and arrays by
1025 value, but is also valid on pointers to scalars. The copy is considered
1026 to belong to the caller not the callee (for example,
1027 <tt><a href="#readonly">readonly</a></tt> functions should not write to
1028 <tt>byval</tt> parameters). This is not a valid attribute for return
1029 values. The byval attribute also supports specifying an alignment with
1030 the align attribute. This has a target-specific effect on the code
1031 generator that usually indicates a desired alignment for the synthesized
1032 stack slot.</dd>
1033
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1034 <dt><tt><b>sret</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1035 <dd>This indicates that the pointer parameter specifies the address of a
1036 structure that is the return value of the function in the source program.
1037 This pointer must be guaranteed by the caller to be valid: loads and
1038 stores to the structure may be assumed by the callee to not to trap. This
1039 may only be applied to the first parameter. This is not a valid attribute
1040 for return values. </dd>
1041
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1042 <dt><tt><b>noalias</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1043 <dd>This indicates that the pointer does not alias any global or any other
1044 parameter. The caller is responsible for ensuring that this is the
1045 case. On a function return value, <tt>noalias</tt> additionally indicates
1046 that the pointer does not alias any other pointers visible to the
1047 caller. For further details, please see the discussion of the NoAlias
1048 response in
1049 <a href="http://llvm.org/docs/AliasAnalysis.html#MustMayNo">alias
1050 analysis</a>.</dd>
1051
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1052 <dt><tt><b>nocapture</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1053 <dd>This indicates that the callee does not make any copies of the pointer
1054 that outlive the callee itself. This is not a valid attribute for return
1055 values.</dd>
1056
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1057 <dt><tt><b>nest</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1058 <dd>This indicates that the pointer parameter can be excised using the
1059 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
1060 attribute for return values.</dd>
1061</dl>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1062
1063</div>
1064
1065<!-- ======================================================================= -->
1066<div class="doc_subsection">
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]1067 <a name="gc">Garbage Collector Names</a>
1068</div>
1069
1070<div class="doc_text">
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]1071
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1072<p>Each function may specify a garbage collector name, which is simply a
1073 string:</p>
1074
1075<div class="doc_code">
1076<pre>
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1077define void @f() gc "name" { ... }
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1078</pre>
1079</div>
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]1080
1081<p>The compiler declares the supported values of <i>name</i>. Specifying a
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1082 collector which will cause the compiler to alter its output in order to
1083 support the named garbage collection algorithm.</p>
1084
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]1085</div>
1086
1087<!-- ======================================================================= -->
1088<div class="doc_subsection">
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1089 <a name="fnattrs">Function Attributes</a>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1090</div>
1091
1092<div class="doc_text">
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1093
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1094<p>Function attributes are set to communicate additional information about a
1095 function. Function attributes are considered to be part of the function, not
1096 of the function type, so functions with different parameter attributes can
1097 have the same function type.</p>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1098
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1099<p>Function attributes are simple keywords that follow the type specified. If
1100 multiple attributes are needed, they are space separated. For example:</p>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1101
1102<div class="doc_code">
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1103<pre>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1104define void @f() noinline { ... }
1105define void @f() alwaysinline { ... }
1106define void @f() alwaysinline optsize { ... }
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1107define void @f() optsize { ... }
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1108</pre>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1109</div>
1110
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1111<dl>
Charles Davisfaa8f752010-02-12 00:31:15 +0000[diff] [blame]1112 <dt><tt><b>alignstack(&lt;<em>n</em>&gt;)</b></tt></dt>
1113 <dd>This attribute indicates that, when emitting the prologue and epilogue,
1114 the backend should forcibly align the stack pointer. Specify the
1115 desired alignment, which must be a power of two, in parentheses.
1116
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1117 <dt><tt><b>alwaysinline</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1118 <dd>This attribute indicates that the inliner should attempt to inline this
1119 function into callers whenever possible, ignoring any active inlining size
1120 threshold for this caller.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1121
Jakob Stoklund Olesen77180732010-02-06 01:16:28 +0000[diff] [blame]1122 <dt><tt><b>inlinehint</b></tt></dt>
1123 <dd>This attribute indicates that the source code contained a hint that inlining
1124 this function is desirable (such as the "inline" keyword in C/C++). It
1125 is just a hint; it imposes no requirements on the inliner.</dd>
1126
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1127 <dt><tt><b>noinline</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1128 <dd>This attribute indicates that the inliner should never inline this
1129 function in any situation. This attribute may not be used together with
1130 the <tt>alwaysinline</tt> attribute.</dd>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1131
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1132 <dt><tt><b>optsize</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1133 <dd>This attribute suggests that optimization passes and code generator passes
1134 make choices that keep the code size of this function low, and otherwise
1135 do optimizations specifically to reduce code size.</dd>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1136
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1137 <dt><tt><b>noreturn</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1138 <dd>This function attribute indicates that the function never returns
1139 normally. This produces undefined behavior at runtime if the function
1140 ever does dynamically return.</dd>
Bill Wendlingdfaabba2008-11-13 01:02:51 +0000[diff] [blame]1141
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1142 <dt><tt><b>nounwind</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1143 <dd>This function attribute indicates that the function never returns with an
1144 unwind or exceptional control flow. If the function does unwind, its
1145 runtime behavior is undefined.</dd>
Bill Wendlingbe9ec3f2008-11-26 19:07:40 +0000[diff] [blame]1146
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1147 <dt><tt><b>readnone</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1148 <dd>This attribute indicates that the function computes its result (or decides
1149 to unwind an exception) based strictly on its arguments, without
1150 dereferencing any pointer arguments or otherwise accessing any mutable
1151 state (e.g. memory, control registers, etc) visible to caller functions.
1152 It does not write through any pointer arguments
1153 (including <tt><a href="#byval">byval</a></tt> arguments) and never
1154 changes any state visible to callers. This means that it cannot unwind
1155 exceptions by calling the <tt>C++</tt> exception throwing methods, but
1156 could use the <tt>unwind</tt> instruction.</dd>
Devang Patela2f9f412009-06-12 19:45:19 +0000[diff] [blame]1157
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1158 <dt><tt><b><a name="readonly">readonly</a></b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1159 <dd>This attribute indicates that the function does not write through any
1160 pointer arguments (including <tt><a href="#byval">byval</a></tt>
1161 arguments) or otherwise modify any state (e.g. memory, control registers,
1162 etc) visible to caller functions. It may dereference pointer arguments
1163 and read state that may be set in the caller. A readonly function always
1164 returns the same value (or unwinds an exception identically) when called
1165 with the same set of arguments and global state. It cannot unwind an
1166 exception by calling the <tt>C++</tt> exception throwing methods, but may
1167 use the <tt>unwind</tt> instruction.</dd>
Anton Korobeynikovedd7d112009-07-17 18:07:26 +0000[diff] [blame]1168
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1169 <dt><tt><b><a name="ssp">ssp</a></b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1170 <dd>This attribute indicates that the function should emit a stack smashing
1171 protector. It is in the form of a "canary"&mdash;a random value placed on
1172 the stack before the local variables that's checked upon return from the
1173 function to see if it has been overwritten. A heuristic is used to
1174 determine if a function needs stack protectors or not.<br>
1175<br>
1176 If a function that has an <tt>ssp</tt> attribute is inlined into a
1177 function that doesn't have an <tt>ssp</tt> attribute, then the resulting
1178 function will have an <tt>ssp</tt> attribute.</dd>
1179
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1180 <dt><tt><b>sspreq</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1181 <dd>This attribute indicates that the function should <em>always</em> emit a
1182 stack smashing protector. This overrides
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]1183 the <tt><a href="#ssp">ssp</a></tt> function attribute.<br>
1184<br>
1185 If a function that has an <tt>sspreq</tt> attribute is inlined into a
1186 function that doesn't have an <tt>sspreq</tt> attribute or which has
1187 an <tt>ssp</tt> attribute, then the resulting function will have
1188 an <tt>sspreq</tt> attribute.</dd>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1189
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1190 <dt><tt><b>noredzone</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1191 <dd>This attribute indicates that the code generator should not use a red
1192 zone, even if the target-specific ABI normally permits it.</dd>
1193
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1194 <dt><tt><b>noimplicitfloat</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1195 <dd>This attributes disables implicit floating point instructions.</dd>
1196
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1197 <dt><tt><b>naked</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1198 <dd>This attribute disables prologue / epilogue emission for the function.
1199 This can have very system-specific consequences.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1200</dl>
1201
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1202</div>
1203
1204<!-- ======================================================================= -->
1205<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1206 <a name="moduleasm">Module-Level Inline Assembly</a>
1207</div>
1208
1209<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1210
1211<p>Modules may contain "module-level inline asm" blocks, which corresponds to
1212 the GCC "file scope inline asm" blocks. These blocks are internally
1213 concatenated by LLVM and treated as a single unit, but may be separated in
1214 the <tt>.ll</tt> file if desired. The syntax is very simple:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1215
1216<div class="doc_code">
1217<pre>
1218module asm "inline asm code goes here"
1219module asm "more can go here"
1220</pre>
1221</div>
1222
1223<p>The strings can contain any character by escaping non-printable characters.
1224 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1225 for the number.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1226
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1227<p>The inline asm code is simply printed to the machine code .s file when
1228 assembly code is generated.</p>
1229
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1230</div>
1231
1232<!-- ======================================================================= -->
1233<div class="doc_subsection">
1234 <a name="datalayout">Data Layout</a>
1235</div>
1236
1237<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1238
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1239<p>A module may specify a target specific data layout string that specifies how
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1240 data is to be laid out in memory. The syntax for the data layout is
1241 simply:</p>
1242
1243<div class="doc_code">
1244<pre>
1245target datalayout = "<i>layout specification</i>"
1246</pre>
1247</div>
1248
1249<p>The <i>layout specification</i> consists of a list of specifications
1250 separated by the minus sign character ('-'). Each specification starts with
1251 a letter and may include other information after the letter to define some
1252 aspect of the data layout. The specifications accepted are as follows:</p>
1253
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1254<dl>
1255 <dt><tt>E</tt></dt>
1256 <dd>Specifies that the target lays out data in big-endian form. That is, the
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1257 bits with the most significance have the lowest address location.</dd>
1258
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1259 <dt><tt>e</tt></dt>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1260 <dd>Specifies that the target lays out data in little-endian form. That is,
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1261 the bits with the least significance have the lowest address
1262 location.</dd>
1263
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1264 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]1265 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1266 <i>preferred</i> alignments. All sizes are in bits. Specifying
1267 the <i>pref</i> alignment is optional. If omitted, the
1268 preceding <tt>:</tt> should be omitted too.</dd>
1269
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1270 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1271 <dd>This specifies the alignment for an integer type of a given bit
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1272 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1273
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1274 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]1275 <dd>This specifies the alignment for a vector type of a given bit
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1276 <i>size</i>.</dd>
1277
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1278 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]1279 <dd>This specifies the alignment for a floating point type of a given bit
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1280 <i>size</i>. The value of <i>size</i> must be either 32 (float) or 64
1281 (double).</dd>
1282
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1283 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1284 <dd>This specifies the alignment for an aggregate type of a given bit
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1285 <i>size</i>.</dd>
1286
Daniel Dunbard88a97b2009-06-08 22:17:53 +0000[diff] [blame]1287 <dt><tt>s<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1288 <dd>This specifies the alignment for a stack object of a given bit
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1289 <i>size</i>.</dd>
Chris Lattnerece41802009-11-07 09:35:34 +0000[diff] [blame]1290
1291 <dt><tt>n<i>size1</i>:<i>size2</i>:<i>size3</i>...</tt></dt>
1292 <dd>This specifies a set of native integer widths for the target CPU
1293 in bits. For example, it might contain "n32" for 32-bit PowerPC,
1294 "n32:64" for PowerPC 64, or "n8:16:32:64" for X86-64. Elements of
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]1295 this set are considered to support most general arithmetic
Chris Lattnerece41802009-11-07 09:35:34 +0000[diff] [blame]1296 operations efficiently.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1297</dl>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1298
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1299<p>When constructing the data layout for a given target, LLVM starts with a
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1300 default set of specifications which are then (possibly) overriden by the
1301 specifications in the <tt>datalayout</tt> keyword. The default specifications
1302 are given in this list:</p>
1303
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1304<ul>
1305 <li><tt>E</tt> - big endian</li>
Dan Gohmane78194f2010-02-23 02:44:03 +0000[diff] [blame]1306 <li><tt>p:64:64:64</tt> - 64-bit pointers with 64-bit alignment</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1307 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1308 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1309 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1310 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1311 <li><tt>i64:32:64</tt> - i64 has ABI alignment of 32-bits but preferred
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1312 alignment of 64-bits</li>
1313 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1314 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1315 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1316 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1317 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
Daniel Dunbard88a97b2009-06-08 22:17:53 +0000[diff] [blame]1318 <li><tt>s0:64:64</tt> - stack objects are 64-bit aligned</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1319</ul>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1320
1321<p>When LLVM is determining the alignment for a given type, it uses the
1322 following rules:</p>
1323
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1324<ol>
1325 <li>If the type sought is an exact match for one of the specifications, that
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1326 specification is used.</li>
1327
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1328 <li>If no match is found, and the type sought is an integer type, then the
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1329 smallest integer type that is larger than the bitwidth of the sought type
1330 is used. If none of the specifications are larger than the bitwidth then
1331 the the largest integer type is used. For example, given the default
1332 specifications above, the i7 type will use the alignment of i8 (next
1333 largest) while both i65 and i256 will use the alignment of i64 (largest
1334 specified).</li>
1335
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1336 <li>If no match is found, and the type sought is a vector type, then the
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1337 largest vector type that is smaller than the sought vector type will be
1338 used as a fall back. This happens because &lt;128 x double&gt; can be
1339 implemented in terms of 64 &lt;2 x double&gt;, for example.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1340</ol>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1341
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1342</div>
1343
Dan Gohman27b47012009-07-27 18:07:55 +0000[diff] [blame]1344<!-- ======================================================================= -->
1345<div class="doc_subsection">
1346 <a name="pointeraliasing">Pointer Aliasing Rules</a>
1347</div>
1348
1349<div class="doc_text">
1350
Andreas Bolka11fbf432009-07-29 00:02:05 +0000[diff] [blame]1351<p>Any memory access must be done through a pointer value associated
Andreas Bolka23bece42009-07-27 20:37:10 +0000[diff] [blame]1352with an address range of the memory access, otherwise the behavior
Dan Gohman27b47012009-07-27 18:07:55 +0000[diff] [blame]1353is undefined. Pointer values are associated with address ranges
1354according to the following rules:</p>
1355
1356<ul>
Andreas Bolka11fbf432009-07-29 00:02:05 +0000[diff] [blame]1357 <li>A pointer value formed from a
1358 <tt><a href="#i_getelementptr">getelementptr</a></tt> instruction
1359 is associated with the addresses associated with the first operand
1360 of the <tt>getelementptr</tt>.</li>
1361 <li>An address of a global variable is associated with the address
Dan Gohman27b47012009-07-27 18:07:55 +0000[diff] [blame]1362 range of the variable's storage.</li>
1363 <li>The result value of an allocation instruction is associated with
1364 the address range of the allocated storage.</li>
1365 <li>A null pointer in the default address-space is associated with
Andreas Bolka11fbf432009-07-29 00:02:05 +0000[diff] [blame]1366 no address.</li>
1367 <li>A pointer value formed by an
1368 <tt><a href="#i_inttoptr">inttoptr</a></tt> is associated with all
1369 address ranges of all pointer values that contribute (directly or
1370 indirectly) to the computation of the pointer's value.</li>
1371 <li>The result value of a
1372 <tt><a href="#i_bitcast">bitcast</a></tt> is associated with all
Dan Gohman27b47012009-07-27 18:07:55 +0000[diff] [blame]1373 addresses associated with the operand of the <tt>bitcast</tt>.</li>
1374 <li>An integer constant other than zero or a pointer value returned
1375 from a function not defined within LLVM may be associated with address
1376 ranges allocated through mechanisms other than those provided by
Andreas Bolka11fbf432009-07-29 00:02:05 +0000[diff] [blame]1377 LLVM. Such ranges shall not overlap with any ranges of addresses
Dan Gohman27b47012009-07-27 18:07:55 +0000[diff] [blame]1378 allocated by mechanisms provided by LLVM.</li>
1379 </ul>
1380
1381<p>LLVM IR does not associate types with memory. The result type of a
Andreas Bolka11fbf432009-07-29 00:02:05 +0000[diff] [blame]1382<tt><a href="#i_load">load</a></tt> merely indicates the size and
1383alignment of the memory from which to load, as well as the
1384interpretation of the value. The first operand of a
1385<tt><a href="#i_store">store</a></tt> similarly only indicates the size
1386and alignment of the store.</p>
Dan Gohman27b47012009-07-27 18:07:55 +0000[diff] [blame]1387
1388<p>Consequently, type-based alias analysis, aka TBAA, aka
1389<tt>-fstrict-aliasing</tt>, is not applicable to general unadorned
1390LLVM IR. <a href="#metadata">Metadata</a> may be used to encode
1391additional information which specialized optimization passes may use
1392to implement type-based alias analysis.</p>
1393
1394</div>
1395
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1396<!-- *********************************************************************** -->
1397<div class="doc_section"> <a name="typesystem">Type System</a> </div>
1398<!-- *********************************************************************** -->
1399
1400<div class="doc_text">
1401
1402<p>The LLVM type system is one of the most important features of the
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1403 intermediate representation. Being typed enables a number of optimizations
1404 to be performed on the intermediate representation directly, without having
1405 to do extra analyses on the side before the transformation. A strong type
1406 system makes it easier to read the generated code and enables novel analyses
1407 and transformations that are not feasible to perform on normal three address
1408 code representations.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1409
1410</div>
1411
1412<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1413<div class="doc_subsection"> <a name="t_classifications">Type
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1414Classifications</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1415
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1416<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1417
1418<p>The types fall into a few useful classifications:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1419
1420<table border="1" cellspacing="0" cellpadding="4">
1421 <tbody>
1422 <tr><th>Classification</th><th>Types</th></tr>
1423 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1424 <td><a href="#t_integer">integer</a></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1425 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
1426 </tr>
1427 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1428 <td><a href="#t_floating">floating point</a></td>
1429 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1430 </tr>
1431 <tr>
1432 <td><a name="t_firstclass">first class</a></td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1433 <td><a href="#t_integer">integer</a>,
1434 <a href="#t_floating">floating point</a>,
1435 <a href="#t_pointer">pointer</a>,
Dan Gohmanf6237db2008-06-18 18:42:13 +0000[diff] [blame]1436 <a href="#t_vector">vector</a>,
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]1437 <a href="#t_struct">structure</a>,
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame]1438 <a href="#t_union">union</a>,
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]1439 <a href="#t_array">array</a>,
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]1440 <a href="#t_label">label</a>,
1441 <a href="#t_metadata">metadata</a>.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1442 </td>
1443 </tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1444 <tr>
1445 <td><a href="#t_primitive">primitive</a></td>
1446 <td><a href="#t_label">label</a>,
1447 <a href="#t_void">void</a>,
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]1448 <a href="#t_floating">floating point</a>,
1449 <a href="#t_metadata">metadata</a>.</td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1450 </tr>
1451 <tr>
1452 <td><a href="#t_derived">derived</a></td>
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame]1453 <td><a href="#t_array">array</a>,
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1454 <a href="#t_function">function</a>,
1455 <a href="#t_pointer">pointer</a>,
1456 <a href="#t_struct">structure</a>,
1457 <a href="#t_pstruct">packed structure</a>,
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame]1458 <a href="#t_union">union</a>,
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1459 <a href="#t_vector">vector</a>,
1460 <a href="#t_opaque">opaque</a>.
Dan Gohman032ba852008-10-14 16:32:04 +0000[diff] [blame]1461 </td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1462 </tr>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1463 </tbody>
1464</table>
1465
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1466<p>The <a href="#t_firstclass">first class</a> types are perhaps the most
1467 important. Values of these types are the only ones which can be produced by
Nick Lewycky244cf482009-09-27 00:45:11 +0000[diff] [blame]1468 instructions.</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1469
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1470</div>
1471
1472<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1473<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1474
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1475<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1476
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1477<p>The primitive types are the fundamental building blocks of the LLVM
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1478 system.</p>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1479
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1480</div>
1481
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1482<!-- _______________________________________________________________________ -->
Nick Lewycky244cf482009-09-27 00:45:11 +0000[diff] [blame]1483<div class="doc_subsubsection"> <a name="t_integer">Integer Type</a> </div>
1484
1485<div class="doc_text">
1486
1487<h5>Overview:</h5>
1488<p>The integer type is a very simple type that simply specifies an arbitrary
1489 bit width for the integer type desired. Any bit width from 1 bit to
1490 2<sup>23</sup>-1 (about 8 million) can be specified.</p>
1491
1492<h5>Syntax:</h5>
1493<pre>
1494 iN
1495</pre>
1496
1497<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1498 value.</p>
1499
1500<h5>Examples:</h5>
1501<table class="layout">
1502 <tr class="layout">
1503 <td class="left"><tt>i1</tt></td>
1504 <td class="left">a single-bit integer.</td>
1505 </tr>
1506 <tr class="layout">
1507 <td class="left"><tt>i32</tt></td>
1508 <td class="left">a 32-bit integer.</td>
1509 </tr>
1510 <tr class="layout">
1511 <td class="left"><tt>i1942652</tt></td>
1512 <td class="left">a really big integer of over 1 million bits.</td>
1513 </tr>
1514</table>
1515
Nick Lewycky244cf482009-09-27 00:45:11 +0000[diff] [blame]1516</div>
1517
1518<!-- _______________________________________________________________________ -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1519<div class="doc_subsubsection"> <a name="t_floating">Floating Point Types</a> </div>
1520
1521<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1522
1523<table>
1524 <tbody>
1525 <tr><th>Type</th><th>Description</th></tr>
1526 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1527 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1528 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1529 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1530 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1531 </tbody>
1532</table>
1533
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1534</div>
1535
1536<!-- _______________________________________________________________________ -->
1537<div class="doc_subsubsection"> <a name="t_void">Void Type</a> </div>
1538
1539<div class="doc_text">
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]1540
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1541<h5>Overview:</h5>
1542<p>The void type does not represent any value and has no size.</p>
1543
1544<h5>Syntax:</h5>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1545<pre>
1546 void
1547</pre>
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]1548
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1549</div>
1550
1551<!-- _______________________________________________________________________ -->
1552<div class="doc_subsubsection"> <a name="t_label">Label Type</a> </div>
1553
1554<div class="doc_text">
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]1555
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1556<h5>Overview:</h5>
1557<p>The label type represents code labels.</p>
1558
1559<h5>Syntax:</h5>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1560<pre>
1561 label
1562</pre>
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]1563
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1564</div>
1565
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]1566<!-- _______________________________________________________________________ -->
1567<div class="doc_subsubsection"> <a name="t_metadata">Metadata Type</a> </div>
1568
1569<div class="doc_text">
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]1570
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]1571<h5>Overview:</h5>
Nick Lewyckyeb021882009-09-27 23:27:42 +0000[diff] [blame]1572<p>The metadata type represents embedded metadata. No derived types may be
1573 created from metadata except for <a href="#t_function">function</a>
1574 arguments.
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]1575
1576<h5>Syntax:</h5>
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]1577<pre>
1578 metadata
1579</pre>
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]1580
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]1581</div>
1582
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1583
1584<!-- ======================================================================= -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1585<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
1586
1587<div class="doc_text">
1588
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1589<p>The real power in LLVM comes from the derived types in the system. This is
1590 what allows a programmer to represent arrays, functions, pointers, and other
Nick Lewycky244cf482009-09-27 00:45:11 +0000[diff] [blame]1591 useful types. Each of these types contain one or more element types which
1592 may be a primitive type, or another derived type. For example, it is
1593 possible to have a two dimensional array, using an array as the element type
1594 of another array.</p>
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1595
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame]1596
1597</div>
1598
1599<!-- _______________________________________________________________________ -->
1600<div class="doc_subsubsection"> <a name="t_aggregate">Aggregate Types</a> </div>
1601
1602<div class="doc_text">
1603
1604<p>Aggregate Types are a subset of derived types that can contain multiple
1605 member types. <a href="#t_array">Arrays</a>,
1606 <a href="#t_struct">structs</a>, <a href="#t_vector">vectors</a> and
1607 <a href="#t_union">unions</a> are aggregate types.</p>
1608
1609</div>
1610
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1611</div>
1612
1613<!-- _______________________________________________________________________ -->
1614<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
1615
1616<div class="doc_text">
1617
1618<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1619<p>The array type is a very simple derived type that arranges elements
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1620 sequentially in memory. The array type requires a size (number of elements)
1621 and an underlying data type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1622
1623<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1624<pre>
1625 [&lt;# elements&gt; x &lt;elementtype&gt;]
1626</pre>
1627
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1628<p>The number of elements is a constant integer value; <tt>elementtype</tt> may
1629 be any type with a size.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1630
1631<h5>Examples:</h5>
1632<table class="layout">
1633 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1634 <td class="left"><tt>[40 x i32]</tt></td>
1635 <td class="left">Array of 40 32-bit integer values.</td>
1636 </tr>
1637 <tr class="layout">
1638 <td class="left"><tt>[41 x i32]</tt></td>
1639 <td class="left">Array of 41 32-bit integer values.</td>
1640 </tr>
1641 <tr class="layout">
1642 <td class="left"><tt>[4 x i8]</tt></td>
1643 <td class="left">Array of 4 8-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1644 </tr>
1645</table>
1646<p>Here are some examples of multidimensional arrays:</p>
1647<table class="layout">
1648 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1649 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1650 <td class="left">3x4 array of 32-bit integer values.</td>
1651 </tr>
1652 <tr class="layout">
1653 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1654 <td class="left">12x10 array of single precision floating point values.</td>
1655 </tr>
1656 <tr class="layout">
1657 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1658 <td class="left">2x3x4 array of 16-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1659 </tr>
1660</table>
1661
Dan Gohman87ec5ad2009-11-09 19:01:53 +0000[diff] [blame]1662<p>There is no restriction on indexing beyond the end of the array implied by
1663 a static type (though there are restrictions on indexing beyond the bounds
1664 of an allocated object in some cases). This means that single-dimension
1665 'variable sized array' addressing can be implemented in LLVM with a zero
1666 length array type. An implementation of 'pascal style arrays' in LLVM could
1667 use the type "<tt>{ i32, [0 x float]}</tt>", for example.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1668
1669</div>
1670
1671<!-- _______________________________________________________________________ -->
1672<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1673
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1674<div class="doc_text">
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1675
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1676<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1677<p>The function type can be thought of as a function signature. It consists of
1678 a return type and a list of formal parameter types. The return type of a
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame]1679 function type is a scalar type, a void type, a struct type, or a union
1680 type. If the return type is a struct type then all struct elements must be
1681 of first class types, and the struct must have at least one element.</p>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1682
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1683<h5>Syntax:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1684<pre>
Nick Lewycky3b3b54d2009-09-27 07:55:32 +0000[diff] [blame]1685 &lt;returntype&gt; (&lt;parameter list&gt;)
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1686</pre>
1687
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1688<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1689 specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1690 which indicates that the function takes a variable number of arguments.
1691 Variable argument functions can access their arguments with
1692 the <a href="#int_varargs">variable argument handling intrinsic</a>
Chris Lattner553fb1e2010-03-02 06:36:51 +0000[diff] [blame]1693 functions. '<tt>&lt;returntype&gt;</tt>' is any type except
Nick Lewyckyeb021882009-09-27 23:27:42 +0000[diff] [blame]1694 <a href="#t_label">label</a>.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1695
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1696<h5>Examples:</h5>
1697<table class="layout">
1698 <tr class="layout">
1699 <td class="left"><tt>i32 (i32)</tt></td>
1700 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
1701 </td>
1702 </tr><tr class="layout">
Chris Lattner553fb1e2010-03-02 06:36:51 +0000[diff] [blame]1703 <td class="left"><tt>float&nbsp;(i16,&nbsp;i32&nbsp;*)&nbsp;*
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1704 </tt></td>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]1705 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
Chris Lattner553fb1e2010-03-02 06:36:51 +0000[diff] [blame]1706 an <tt>i16</tt> and a <a href="#t_pointer">pointer</a> to <tt>i32</tt>,
1707 returning <tt>float</tt>.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1708 </td>
1709 </tr><tr class="layout">
1710 <td class="left"><tt>i32 (i8*, ...)</tt></td>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]1711 <td class="left">A vararg function that takes at least one
1712 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1713 which returns an integer. This is the signature for <tt>printf</tt> in
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1714 LLVM.
1715 </td>
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1716 </tr><tr class="layout">
1717 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Nick Lewycky3b3b54d2009-09-27 07:55:32 +0000[diff] [blame]1718 <td class="left">A function taking an <tt>i32</tt>, returning a
1719 <a href="#t_struct">structure</a> containing two <tt>i32</tt> values
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1720 </td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1721 </tr>
1722</table>
1723
1724</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1725
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1726<!-- _______________________________________________________________________ -->
1727<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1728
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1729<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1730
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1731<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1732<p>The structure type is used to represent a collection of data members together
1733 in memory. The packing of the field types is defined to match the ABI of the
1734 underlying processor. The elements of a structure may be any type that has a
1735 size.</p>
1736
Jeffrey Yasskine05d7732010-01-11 19:19:26 +0000[diff] [blame]1737<p>Structures in memory are accessed using '<tt><a href="#i_load">load</a></tt>'
1738 and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field
1739 with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
1740 Structures in registers are accessed using the
1741 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' and
1742 '<tt><a href="#i_insertvalue">insertvalue</a></tt>' instructions.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1743<h5>Syntax:</h5>
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]1744<pre>
1745 { &lt;type list&gt; }
1746</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1747
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1748<h5>Examples:</h5>
1749<table class="layout">
1750 <tr class="layout">
1751 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1752 <td class="left">A triple of three <tt>i32</tt> values</td>
1753 </tr><tr class="layout">
1754 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1755 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1756 second element is a <a href="#t_pointer">pointer</a> to a
1757 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1758 an <tt>i32</tt>.</td>
1759 </tr>
1760</table>
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1761
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1762</div>
1763
1764<!-- _______________________________________________________________________ -->
1765<div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
1766</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1767
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1768<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1769
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1770<h5>Overview:</h5>
1771<p>The packed structure type is used to represent a collection of data members
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1772 together in memory. There is no padding between fields. Further, the
1773 alignment of a packed structure is 1 byte. The elements of a packed
1774 structure may be any type that has a size.</p>
1775
1776<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt> and
1777 '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field with
1778 the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
1779
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1780<h5>Syntax:</h5>
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]1781<pre>
1782 &lt; { &lt;type list&gt; } &gt;
1783</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1784
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1785<h5>Examples:</h5>
1786<table class="layout">
1787 <tr class="layout">
1788 <td class="left"><tt>&lt; { i32, i32, i32 } &gt;</tt></td>
1789 <td class="left">A triple of three <tt>i32</tt> values</td>
1790 </tr><tr class="layout">
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1791 <td class="left">
1792<tt>&lt;&nbsp;{&nbsp;float,&nbsp;i32&nbsp;(i32)*&nbsp;}&nbsp;&gt;</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1793 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1794 second element is a <a href="#t_pointer">pointer</a> to a
1795 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1796 an <tt>i32</tt>.</td>
1797 </tr>
1798</table>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1799
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1800</div>
1801
1802<!-- _______________________________________________________________________ -->
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame]1803<div class="doc_subsubsection"> <a name="t_union">Union Type</a> </div>
1804
1805<div class="doc_text">
1806
1807<h5>Overview:</h5>
1808<p>A union type describes an object with size and alignment suitable for
1809 an object of any one of a given set of types (also known as an "untagged"
1810 union). It is similar in concept and usage to a
1811 <a href="#t_struct">struct</a>, except that all members of the union
1812 have an offset of zero. The elements of a union may be any type that has a
1813 size. Unions must have at least one member - empty unions are not allowed.
1814 </p>
1815
1816<p>The size of the union as a whole will be the size of its largest member,
1817 and the alignment requirements of the union as a whole will be the largest
1818 alignment requirement of any member.</p>
1819
Dan Gohmanef8400c2010-02-25 16:51:31 +0000[diff] [blame]1820<p>Union members are accessed using '<tt><a href="#i_load">load</a></tt> and
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame]1821 '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field with
1822 the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
1823 Since all members are at offset zero, the getelementptr instruction does
1824 not affect the address, only the type of the resulting pointer.</p>
1825
1826<h5>Syntax:</h5>
1827<pre>
1828 union { &lt;type list&gt; }
1829</pre>
1830
1831<h5>Examples:</h5>
1832<table class="layout">
1833 <tr class="layout">
1834 <td class="left"><tt>union { i32, i32*, float }</tt></td>
1835 <td class="left">A union of three types: an <tt>i32</tt>, a pointer to
1836 an <tt>i32</tt>, and a <tt>float</tt>.</td>
1837 </tr><tr class="layout">
1838 <td class="left">
1839 <tt>union {&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1840 <td class="left">A union, where the first element is a <tt>float</tt> and the
1841 second element is a <a href="#t_pointer">pointer</a> to a
1842 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1843 an <tt>i32</tt>.</td>
1844 </tr>
1845</table>
1846
1847</div>
1848
1849<!-- _______________________________________________________________________ -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1850<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
Chris Lattner96edbd32009-02-08 19:53:29 +0000[diff] [blame]1851
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1852<div class="doc_text">
1853
1854<h5>Overview:</h5>
Dan Gohmanb2f72c82010-02-25 16:50:07 +0000[diff] [blame]1855<p>The pointer type is used to specify memory locations.
1856 Pointers are commonly used to reference objects in memory.</p>
1857
1858<p>Pointer types may have an optional address space attribute defining the
1859 numbered address space where the pointed-to object resides. The default
1860 address space is number zero. The semantics of non-zero address
1861 spaces are target-specific.</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1862
1863<p>Note that LLVM does not permit pointers to void (<tt>void*</tt>) nor does it
1864 permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</tt> instead.</p>
Chris Lattner96edbd32009-02-08 19:53:29 +0000[diff] [blame]1865
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1866<h5>Syntax:</h5>
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]1867<pre>
1868 &lt;type&gt; *
1869</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1870
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1871<h5>Examples:</h5>
1872<table class="layout">
1873 <tr class="layout">
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]1874 <td class="left"><tt>[4 x i32]*</tt></td>
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1875 <td class="left">A <a href="#t_pointer">pointer</a> to <a
1876 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
1877 </tr>
1878 <tr class="layout">
1879 <td class="left"><tt>i32 (i32 *) *</tt></td>
1880 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1881 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1882 <tt>i32</tt>.</td>
1883 </tr>
1884 <tr class="layout">
1885 <td class="left"><tt>i32 addrspace(5)*</tt></td>
1886 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
1887 that resides in address space #5.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1888 </tr>
1889</table>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1890
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1891</div>
1892
1893<!-- _______________________________________________________________________ -->
1894<div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1895
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1896<div class="doc_text">
1897
1898<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1899<p>A vector type is a simple derived type that represents a vector of elements.
1900 Vector types are used when multiple primitive data are operated in parallel
1901 using a single instruction (SIMD). A vector type requires a size (number of
Duncan Sands87598b02009-11-27 13:38:03 +0000[diff] [blame]1902 elements) and an underlying primitive data type. Vector types are considered
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1903 <a href="#t_firstclass">first class</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1904
1905<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1906<pre>
1907 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1908</pre>
1909
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1910<p>The number of elements is a constant integer value; elementtype may be any
1911 integer or floating point type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1912
1913<h5>Examples:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1914<table class="layout">
1915 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1916 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
1917 <td class="left">Vector of 4 32-bit integer values.</td>
1918 </tr>
1919 <tr class="layout">
1920 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
1921 <td class="left">Vector of 8 32-bit floating-point values.</td>
1922 </tr>
1923 <tr class="layout">
1924 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
1925 <td class="left">Vector of 2 64-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1926 </tr>
1927</table>
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1928
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1929</div>
1930
1931<!-- _______________________________________________________________________ -->
1932<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1933<div class="doc_text">
1934
1935<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1936<p>Opaque types are used to represent unknown types in the system. This
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1937 corresponds (for example) to the C notion of a forward declared structure
1938 type. In LLVM, opaque types can eventually be resolved to any type (not just
1939 a structure type).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1940
1941<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1942<pre>
1943 opaque
1944</pre>
1945
1946<h5>Examples:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1947<table class="layout">
1948 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1949 <td class="left"><tt>opaque</tt></td>
1950 <td class="left">An opaque type.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1951 </tr>
1952</table>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1953
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1954</div>
1955
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]1956<!-- ======================================================================= -->
1957<div class="doc_subsection">
1958 <a name="t_uprefs">Type Up-references</a>
1959</div>
1960
1961<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1962
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]1963<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1964<p>An "up reference" allows you to refer to a lexically enclosing type without
1965 requiring it to have a name. For instance, a structure declaration may
1966 contain a pointer to any of the types it is lexically a member of. Example
1967 of up references (with their equivalent as named type declarations)
1968 include:</p>
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]1969
1970<pre>
Chris Lattner5ad632d2009-02-09 10:00:56 +0000[diff] [blame]1971 { \2 * } %x = type { %x* }
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]1972 { \2 }* %y = type { %y }*
1973 \1* %z = type %z*
1974</pre>
1975
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1976<p>An up reference is needed by the asmprinter for printing out cyclic types
1977 when there is no declared name for a type in the cycle. Because the
1978 asmprinter does not want to print out an infinite type string, it needs a
1979 syntax to handle recursive types that have no names (all names are optional
1980 in llvm IR).</p>
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]1981
1982<h5>Syntax:</h5>
1983<pre>
1984 \&lt;level&gt;
1985</pre>
1986
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1987<p>The level is the count of the lexical type that is being referred to.</p>
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]1988
1989<h5>Examples:</h5>
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]1990<table class="layout">
1991 <tr class="layout">
1992 <td class="left"><tt>\1*</tt></td>
1993 <td class="left">Self-referential pointer.</td>
1994 </tr>
1995 <tr class="layout">
1996 <td class="left"><tt>{ { \3*, i8 }, i32 }</tt></td>
1997 <td class="left">Recursive structure where the upref refers to the out-most
1998 structure.</td>
1999 </tr>
2000</table>
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]2001
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2002</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2003
2004<!-- *********************************************************************** -->
2005<div class="doc_section"> <a name="constants">Constants</a> </div>
2006<!-- *********************************************************************** -->
2007
2008<div class="doc_text">
2009
2010<p>LLVM has several different basic types of constants. This section describes
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2011 them all and their syntax.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2012
2013</div>
2014
2015<!-- ======================================================================= -->
2016<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
2017
2018<div class="doc_text">
2019
2020<dl>
2021 <dt><b>Boolean constants</b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2022 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
Nick Lewycky244cf482009-09-27 00:45:11 +0000[diff] [blame]2023 constants of the <tt><a href="#t_integer">i1</a></tt> type.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2024
2025 <dt><b>Integer constants</b></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2026 <dd>Standard integers (such as '4') are constants of
2027 the <a href="#t_integer">integer</a> type. Negative numbers may be used
2028 with integer types.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2029
2030 <dt><b>Floating point constants</b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2031 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2032 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
2033 notation (see below). The assembler requires the exact decimal value of a
2034 floating-point constant. For example, the assembler accepts 1.25 but
2035 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
2036 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2037
2038 <dt><b>Null pointer constants</b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2039 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2040 and must be of <a href="#t_pointer">pointer type</a>.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2041</dl>
2042
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2043<p>The one non-intuitive notation for constants is the hexadecimal form of
2044 floating point constants. For example, the form '<tt>double
2045 0x432ff973cafa8000</tt>' is equivalent to (but harder to read than)
2046 '<tt>double 4.5e+15</tt>'. The only time hexadecimal floating point
2047 constants are required (and the only time that they are generated by the
2048 disassembler) is when a floating point constant must be emitted but it cannot
2049 be represented as a decimal floating point number in a reasonable number of
2050 digits. For example, NaN's, infinities, and other special values are
2051 represented in their IEEE hexadecimal format so that assembly and disassembly
2052 do not cause any bits to change in the constants.</p>
2053
Dale Johannesenf82a52f2009-02-11 22:14:51 +0000[diff] [blame]2054<p>When using the hexadecimal form, constants of types float and double are
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2055 represented using the 16-digit form shown above (which matches the IEEE754
2056 representation for double); float values must, however, be exactly
2057 representable as IEE754 single precision. Hexadecimal format is always used
2058 for long double, and there are three forms of long double. The 80-bit format
2059 used by x86 is represented as <tt>0xK</tt> followed by 20 hexadecimal digits.
2060 The 128-bit format used by PowerPC (two adjacent doubles) is represented
2061 by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit format
2062 is represented by <tt>0xL</tt> followed by 32 hexadecimal digits; no
2063 currently supported target uses this format. Long doubles will only work if
2064 they match the long double format on your target. All hexadecimal formats
2065 are big-endian (sign bit at the left).</p>
2066
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2067</div>
2068
2069<!-- ======================================================================= -->
Chris Lattner97063852009-02-28 18:32:25 +0000[diff] [blame]2070<div class="doc_subsection">
Bill Wendling1a2630a2009-07-20 02:32:41 +0000[diff] [blame]2071<a name="aggregateconstants"></a> <!-- old anchor -->
2072<a name="complexconstants">Complex Constants</a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2073</div>
2074
2075<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2076
Chris Lattner97063852009-02-28 18:32:25 +0000[diff] [blame]2077<p>Complex constants are a (potentially recursive) combination of simple
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2078 constants and smaller complex constants.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2079
2080<dl>
2081 <dt><b>Structure constants</b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2082 <dd>Structure constants are represented with notation similar to structure
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2083 type definitions (a comma separated list of elements, surrounded by braces
2084 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
2085 where "<tt>@G</tt>" is declared as "<tt>@G = external global i32</tt>".
2086 Structure constants must have <a href="#t_struct">structure type</a>, and
2087 the number and types of elements must match those specified by the
2088 type.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2089
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame]2090 <dt><b>Union constants</b></dt>
2091 <dd>Union constants are represented with notation similar to a structure with
2092 a single element - that is, a single typed element surrounded
2093 by braces (<tt>{}</tt>)). For example: "<tt>{ i32 4 }</tt>". The
2094 <a href="#t_union">union type</a> can be initialized with a single-element
2095 struct as long as the type of the struct element matches the type of
2096 one of the union members.</dd>
2097
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2098 <dt><b>Array constants</b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2099 <dd>Array constants are represented with notation similar to array type
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2100 definitions (a comma separated list of elements, surrounded by square
2101 brackets (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74
2102 ]</tt>". Array constants must have <a href="#t_array">array type</a>, and
2103 the number and types of elements must match those specified by the
2104 type.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2105
2106 <dt><b>Vector constants</b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2107 <dd>Vector constants are represented with notation similar to vector type
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2108 definitions (a comma separated list of elements, surrounded by
2109 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32
2110 42, i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must
2111 have <a href="#t_vector">vector type</a>, and the number and types of
2112 elements must match those specified by the type.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2113
2114 <dt><b>Zero initialization</b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2115 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame]2116 value to zero of <em>any</em> type, including scalar and
2117 <a href="#t_aggregate">aggregate</a> types.
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2118 This is often used to avoid having to print large zero initializers
2119 (e.g. for large arrays) and is always exactly equivalent to using explicit
2120 zero initializers.</dd>
Nick Lewycky4dcf8102009-04-04 07:22:01 +0000[diff] [blame]2121
2122 <dt><b>Metadata node</b></dt>
Nick Lewyckyf122c7e2009-05-30 16:08:30 +0000[diff] [blame]2123 <dd>A metadata node is a structure-like constant with
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2124 <a href="#t_metadata">metadata type</a>. For example: "<tt>metadata !{
2125 i32 0, metadata !"test" }</tt>". Unlike other constants that are meant to
2126 be interpreted as part of the instruction stream, metadata is a place to
2127 attach additional information such as debug info.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2128</dl>
2129
2130</div>
2131
2132<!-- ======================================================================= -->
2133<div class="doc_subsection">
2134 <a name="globalconstants">Global Variable and Function Addresses</a>
2135</div>
2136
2137<div class="doc_text">
2138
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2139<p>The addresses of <a href="#globalvars">global variables</a>
2140 and <a href="#functionstructure">functions</a> are always implicitly valid
2141 (link-time) constants. These constants are explicitly referenced when
2142 the <a href="#identifiers">identifier for the global</a> is used and always
2143 have <a href="#t_pointer">pointer</a> type. For example, the following is a
2144 legal LLVM file:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2145
2146<div class="doc_code">
2147<pre>
2148@X = global i32 17
2149@Y = global i32 42
2150@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
2151</pre>
2152</div>
2153
2154</div>
2155
2156<!-- ======================================================================= -->
2157<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
2158<div class="doc_text">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2159
Chris Lattner3d72cd82009-09-07 22:52:39 +0000[diff] [blame]2160<p>The string '<tt>undef</tt>' can be used anywhere a constant is expected, and
Benjamin Kramer5fb9d7e2009-10-12 14:46:08 +0000[diff] [blame]2161 indicates that the user of the value may receive an unspecified bit-pattern.
Chris Lattner3d72cd82009-09-07 22:52:39 +0000[diff] [blame]2162 Undefined values may be of any type (other than label or void) and be used
2163 anywhere a constant is permitted.</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2164
Chris Lattner9d1e4fa2009-09-11 01:49:31 +0000[diff] [blame]2165<p>Undefined values are useful because they indicate to the compiler that the
Chris Lattner3d72cd82009-09-07 22:52:39 +0000[diff] [blame]2166 program is well defined no matter what value is used. This gives the
2167 compiler more freedom to optimize. Here are some examples of (potentially
2168 surprising) transformations that are valid (in pseudo IR):</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2169
Chris Lattner3d72cd82009-09-07 22:52:39 +0000[diff] [blame]2170
2171<div class="doc_code">
2172<pre>
2173 %A = add %X, undef
2174 %B = sub %X, undef
2175 %C = xor %X, undef
2176Safe:
2177 %A = undef
2178 %B = undef
2179 %C = undef
2180</pre>
2181</div>
2182
2183<p>This is safe because all of the output bits are affected by the undef bits.
2184Any output bit can have a zero or one depending on the input bits.</p>
2185
2186<div class="doc_code">
2187<pre>
2188 %A = or %X, undef
2189 %B = and %X, undef
2190Safe:
2191 %A = -1
2192 %B = 0
2193Unsafe:
2194 %A = undef
2195 %B = undef
2196</pre>
2197</div>
2198
2199<p>These logical operations have bits that are not always affected by the input.
2200For example, if "%X" has a zero bit, then the output of the 'and' operation will
2201always be a zero, no matter what the corresponding bit from the undef is. As
Chris Lattner9d1e4fa2009-09-11 01:49:31 +0000[diff] [blame]2202such, it is unsafe to optimize or assume that the result of the and is undef.
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]2203However, it is safe to assume that all bits of the undef could be 0, and
2204optimize the and to 0. Likewise, it is safe to assume that all the bits of
2205the undef operand to the or could be set, allowing the or to be folded to
Chris Lattner9d1e4fa2009-09-11 01:49:31 +0000[diff] [blame]2206-1.</p>
Chris Lattner3d72cd82009-09-07 22:52:39 +0000[diff] [blame]2207
2208<div class="doc_code">
2209<pre>
2210 %A = select undef, %X, %Y
2211 %B = select undef, 42, %Y
2212 %C = select %X, %Y, undef
2213Safe:
2214 %A = %X (or %Y)
2215 %B = 42 (or %Y)
2216 %C = %Y
2217Unsafe:
2218 %A = undef
2219 %B = undef
2220 %C = undef
2221</pre>
2222</div>
2223
2224<p>This set of examples show that undefined select (and conditional branch)
2225conditions can go "either way" but they have to come from one of the two
2226operands. In the %A example, if %X and %Y were both known to have a clear low
2227bit, then %A would have to have a cleared low bit. However, in the %C example,
2228the optimizer is allowed to assume that the undef operand could be the same as
2229%Y, allowing the whole select to be eliminated.</p>
2230
2231
2232<div class="doc_code">
2233<pre>
2234 %A = xor undef, undef
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]2235
Chris Lattner3d72cd82009-09-07 22:52:39 +0000[diff] [blame]2236 %B = undef
2237 %C = xor %B, %B
2238
2239 %D = undef
2240 %E = icmp lt %D, 4
2241 %F = icmp gte %D, 4
2242
2243Safe:
2244 %A = undef
2245 %B = undef
2246 %C = undef
2247 %D = undef
2248 %E = undef
2249 %F = undef
2250</pre>
2251</div>
2252
2253<p>This example points out that two undef operands are not necessarily the same.
2254This can be surprising to people (and also matches C semantics) where they
2255assume that "X^X" is always zero, even if X is undef. This isn't true for a
2256number of reasons, but the short answer is that an undef "variable" can
2257arbitrarily change its value over its "live range". This is true because the
2258"variable" doesn't actually <em>have a live range</em>. Instead, the value is
2259logically read from arbitrary registers that happen to be around when needed,
Benjamin Kramer5fb9d7e2009-10-12 14:46:08 +0000[diff] [blame]2260so the value is not necessarily consistent over time. In fact, %A and %C need
Chris Lattnerc802e482009-09-08 15:13:16 +0000[diff] [blame]2261to have the same semantics or the core LLVM "replace all uses with" concept
Chris Lattner3d72cd82009-09-07 22:52:39 +0000[diff] [blame]2262would not hold.</p>
Chris Lattner466291f2009-09-07 23:33:52 +0000[diff] [blame]2263
2264<div class="doc_code">
2265<pre>
2266 %A = fdiv undef, %X
2267 %B = fdiv %X, undef
2268Safe:
2269 %A = undef
2270b: unreachable
2271</pre>
2272</div>
2273
2274<p>These examples show the crucial difference between an <em>undefined
2275value</em> and <em>undefined behavior</em>. An undefined value (like undef) is
2276allowed to have an arbitrary bit-pattern. This means that the %A operation
2277can be constant folded to undef because the undef could be an SNaN, and fdiv is
2278not (currently) defined on SNaN's. However, in the second example, we can make
2279a more aggressive assumption: because the undef is allowed to be an arbitrary
2280value, we are allowed to assume that it could be zero. Since a divide by zero
Chris Lattner83d45d72009-09-08 19:45:34 +0000[diff] [blame]2281has <em>undefined behavior</em>, we are allowed to assume that the operation
Chris Lattner466291f2009-09-07 23:33:52 +0000[diff] [blame]2282does not execute at all. This allows us to delete the divide and all code after
2283it: since the undefined operation "can't happen", the optimizer can assume that
2284it occurs in dead code.
2285</p>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]2286
Chris Lattner466291f2009-09-07 23:33:52 +0000[diff] [blame]2287<div class="doc_code">
2288<pre>
2289a: store undef -> %X
2290b: store %X -> undef
2291Safe:
2292a: &lt;deleted&gt;
2293b: unreachable
2294</pre>
2295</div>
2296
2297<p>These examples reiterate the fdiv example: a store "of" an undefined value
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]2298can be assumed to not have any effect: we can assume that the value is
Chris Lattner466291f2009-09-07 23:33:52 +0000[diff] [blame]2299overwritten with bits that happen to match what was already there. However, a
2300store "to" an undefined location could clobber arbitrary memory, therefore, it
2301has undefined behavior.</p>
2302
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2303</div>
2304
2305<!-- ======================================================================= -->
Chris Lattner29246b52009-10-27 21:19:13 +0000[diff] [blame]2306<div class="doc_subsection"><a name="blockaddress">Addresses of Basic
2307 Blocks</a></div>
Chris Lattnerd07c8372009-10-27 21:01:34 +0000[diff] [blame]2308<div class="doc_text">
2309
Chris Lattner620cead2009-11-01 01:27:45 +0000[diff] [blame]2310<p><b><tt>blockaddress(@function, %block)</tt></b></p>
Chris Lattnerd07c8372009-10-27 21:01:34 +0000[diff] [blame]2311
2312<p>The '<tt>blockaddress</tt>' constant computes the address of the specified
Chris Lattner24f2e852009-10-27 21:49:40 +0000[diff] [blame]2313 basic block in the specified function, and always has an i8* type. Taking
Chris Lattner620cead2009-11-01 01:27:45 +0000[diff] [blame]2314 the address of the entry block is illegal.</p>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]2315
Chris Lattnerd07c8372009-10-27 21:01:34 +0000[diff] [blame]2316<p>This value only has defined behavior when used as an operand to the
Chris Lattner4c3800f2009-10-28 00:19:10 +0000[diff] [blame]2317 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>' instruction or for comparisons
Chris Lattnerd07c8372009-10-27 21:01:34 +0000[diff] [blame]2318 against null. Pointer equality tests between labels addresses is undefined
2319 behavior - though, again, comparison against null is ok, and no label is
Chris Lattner29246b52009-10-27 21:19:13 +0000[diff] [blame]2320 equal to the null pointer. This may also be passed around as an opaque
2321 pointer sized value as long as the bits are not inspected. This allows
Chris Lattner0bae7b32009-10-27 21:44:20 +0000[diff] [blame]2322 <tt>ptrtoint</tt> and arithmetic to be performed on these values so long as
Chris Lattner4c3800f2009-10-28 00:19:10 +0000[diff] [blame]2323 the original value is reconstituted before the <tt>indirectbr</tt>.</p>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]2324
Chris Lattner29246b52009-10-27 21:19:13 +0000[diff] [blame]2325<p>Finally, some targets may provide defined semantics when
Chris Lattnerd07c8372009-10-27 21:01:34 +0000[diff] [blame]2326 using the value as the operand to an inline assembly, but that is target
2327 specific.
2328 </p>
2329
2330</div>
2331
2332
2333<!-- ======================================================================= -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2334<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
2335</div>
2336
2337<div class="doc_text">
2338
2339<p>Constant expressions are used to allow expressions involving other constants
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2340 to be used as constants. Constant expressions may be of
2341 any <a href="#t_firstclass">first class</a> type and may involve any LLVM
2342 operation that does not have side effects (e.g. load and call are not
2343 supported). The following is the syntax for constant expressions:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2344
2345<dl>
2346 <dt><b><tt>trunc ( CST to TYPE )</tt></b></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2347 <dd>Truncate a constant to another type. The bit size of CST must be larger
2348 than the bit size of TYPE. Both types must be integers.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2349
2350 <dt><b><tt>zext ( CST to TYPE )</tt></b></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2351 <dd>Zero extend a constant to another type. The bit size of CST must be
2352 smaller or equal to the bit size of TYPE. Both types must be
2353 integers.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2354
2355 <dt><b><tt>sext ( CST to TYPE )</tt></b></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2356 <dd>Sign extend a constant to another type. The bit size of CST must be
2357 smaller or equal to the bit size of TYPE. Both types must be
2358 integers.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2359
2360 <dt><b><tt>fptrunc ( CST to TYPE )</tt></b></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2361 <dd>Truncate a floating point constant to another floating point type. The
2362 size of CST must be larger than the size of TYPE. Both types must be
2363 floating point.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2364
2365 <dt><b><tt>fpext ( CST to TYPE )</tt></b></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2366 <dd>Floating point extend a constant to another type. The size of CST must be
2367 smaller or equal to the size of TYPE. Both types must be floating
2368 point.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2369
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]2370 <dt><b><tt>fptoui ( CST to TYPE )</tt></b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2371 <dd>Convert a floating point constant to the corresponding unsigned integer
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2372 constant. TYPE must be a scalar or vector integer type. CST must be of
2373 scalar or vector floating point type. Both CST and TYPE must be scalars,
2374 or vectors of the same number of elements. If the value won't fit in the
2375 integer type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2376
2377 <dt><b><tt>fptosi ( CST to TYPE )</tt></b></dt>
2378 <dd>Convert a floating point constant to the corresponding signed integer
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2379 constant. TYPE must be a scalar or vector integer type. CST must be of
2380 scalar or vector floating point type. Both CST and TYPE must be scalars,
2381 or vectors of the same number of elements. If the value won't fit in the
2382 integer type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2383
2384 <dt><b><tt>uitofp ( CST to TYPE )</tt></b></dt>
2385 <dd>Convert an unsigned integer constant to the corresponding floating point
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2386 constant. TYPE must be a scalar or vector floating point type. CST must be
2387 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2388 vectors of the same number of elements. If the value won't fit in the
2389 floating point type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2390
2391 <dt><b><tt>sitofp ( CST to TYPE )</tt></b></dt>
2392 <dd>Convert a signed integer constant to the corresponding floating point
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2393 constant. TYPE must be a scalar or vector floating point type. CST must be
2394 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2395 vectors of the same number of elements. If the value won't fit in the
2396 floating point type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2397
2398 <dt><b><tt>ptrtoint ( CST to TYPE )</tt></b></dt>
2399 <dd>Convert a pointer typed constant to the corresponding integer constant
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2400 <tt>TYPE</tt> must be an integer type. <tt>CST</tt> must be of pointer
2401 type. The <tt>CST</tt> value is zero extended, truncated, or unchanged to
2402 make it fit in <tt>TYPE</tt>.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2403
2404 <dt><b><tt>inttoptr ( CST to TYPE )</tt></b></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2405 <dd>Convert a integer constant to a pointer constant. TYPE must be a pointer
2406 type. CST must be of integer type. The CST value is zero extended,
2407 truncated, or unchanged to make it fit in a pointer size. This one is
2408 <i>really</i> dangerous!</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2409
2410 <dt><b><tt>bitcast ( CST to TYPE )</tt></b></dt>
Chris Lattner557bc5d2009-02-28 18:27:03 +0000[diff] [blame]2411 <dd>Convert a constant, CST, to another TYPE. The constraints of the operands
2412 are the same as those for the <a href="#i_bitcast">bitcast
2413 instruction</a>.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2414
2415 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
Dan Gohman106b2ae2009-07-27 21:53:46 +0000[diff] [blame]2416 <dt><b><tt>getelementptr inbounds ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2417 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2418 constants. As with the <a href="#i_getelementptr">getelementptr</a>
2419 instruction, the index list may have zero or more indexes, which are
2420 required to make sense for the type of "CSTPTR".</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2421
2422 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2423 <dd>Perform the <a href="#i_select">select operation</a> on constants.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2424
2425 <dt><b><tt>icmp COND ( VAL1, VAL2 )</tt></b></dt>
2426 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
2427
2428 <dt><b><tt>fcmp COND ( VAL1, VAL2 )</tt></b></dt>
2429 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
2430
2431 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2432 <dd>Perform the <a href="#i_extractelement">extractelement operation</a> on
2433 constants.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2434
2435 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2436 <dd>Perform the <a href="#i_insertelement">insertelement operation</a> on
2437 constants.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2438
2439 <dt><b><tt>shufflevector ( VEC1, VEC2, IDXMASK )</tt></b></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2440 <dd>Perform the <a href="#i_shufflevector">shufflevector operation</a> on
2441 constants.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2442
2443 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2444 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
2445 be any of the <a href="#binaryops">binary</a>
2446 or <a href="#bitwiseops">bitwise binary</a> operations. The constraints
2447 on operands are the same as those for the corresponding instruction
2448 (e.g. no bitwise operations on floating point values are allowed).</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2449</dl>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2450
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2451</div>
2452
2453<!-- *********************************************************************** -->
2454<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
2455<!-- *********************************************************************** -->
2456
2457<!-- ======================================================================= -->
2458<div class="doc_subsection">
2459<a name="inlineasm">Inline Assembler Expressions</a>
2460</div>
2461
2462<div class="doc_text">
2463
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2464<p>LLVM supports inline assembler expressions (as opposed
2465 to <a href="#moduleasm"> Module-Level Inline Assembly</a>) through the use of
2466 a special value. This value represents the inline assembler as a string
2467 (containing the instructions to emit), a list of operand constraints (stored
Dale Johannesen648950f2009-10-13 21:56:55 +0000[diff] [blame]2468 as a string), a flag that indicates whether or not the inline asm
Dale Johannesen5ee3e4b2009-10-21 23:28:00 +0000[diff] [blame]2469 expression has side effects, and a flag indicating whether the function
2470 containing the asm needs to align its stack conservatively. An example
2471 inline assembler expression is:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2472
2473<div class="doc_code">
2474<pre>
2475i32 (i32) asm "bswap $0", "=r,r"
2476</pre>
2477</div>
2478
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2479<p>Inline assembler expressions may <b>only</b> be used as the callee operand of
2480 a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we
2481 have:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2482
2483<div class="doc_code">
2484<pre>
2485%X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
2486</pre>
2487</div>
2488
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2489<p>Inline asms with side effects not visible in the constraint list must be
2490 marked as having side effects. This is done through the use of the
2491 '<tt>sideeffect</tt>' keyword, like so:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2492
2493<div class="doc_code">
2494<pre>
2495call void asm sideeffect "eieio", ""()
2496</pre>
2497</div>
2498
Dale Johannesen5ee3e4b2009-10-21 23:28:00 +0000[diff] [blame]2499<p>In some cases inline asms will contain code that will not work unless the
2500 stack is aligned in some way, such as calls or SSE instructions on x86,
2501 yet will not contain code that does that alignment within the asm.
2502 The compiler should make conservative assumptions about what the asm might
2503 contain and should generate its usual stack alignment code in the prologue
2504 if the '<tt>alignstack</tt>' keyword is present:</p>
Dale Johannesen648950f2009-10-13 21:56:55 +0000[diff] [blame]2505
2506<div class="doc_code">
2507<pre>
Dale Johannesen5ee3e4b2009-10-21 23:28:00 +0000[diff] [blame]2508call void asm alignstack "eieio", ""()
Dale Johannesen648950f2009-10-13 21:56:55 +0000[diff] [blame]2509</pre>
2510</div>
2511
2512<p>If both keywords appear the '<tt>sideeffect</tt>' keyword must come
2513 first.</p>
2514
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2515<p>TODO: The format of the asm and constraints string still need to be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2516 documented here. Constraints on what can be done (e.g. duplication, moving,
2517 etc need to be documented). This is probably best done by reference to
2518 another document that covers inline asm from a holistic perspective.</p>
Chris Lattnerbafc8372010-04-07 05:38:05 +0000[diff] [blame]2519</div>
2520
2521<div class="doc_subsubsection">
2522<a name="inlineasm_md">Inline Asm Metadata</a>
2523</div>
2524
2525<div class="doc_text">
2526
2527<p>The call instructions that wrap inline asm nodes may have a "!srcloc" MDNode
2528 attached to it that contains a constant integer. If present, the code
2529 generator will use the integer as the location cookie value when report
2530 errors through the LLVMContext error reporting mechanisms. This allows a
2531 front-end to corrolate backend errors that occur with inline asm back to the
2532 source code that produced it. For example:</p>
2533
2534<div class="doc_code">
2535<pre>
2536call void asm sideeffect "something bad", ""()<b>, !srcloc !42</b>
2537...
2538!42 = !{ i32 1234567 }
2539</pre>
2540</div>
2541
2542<p>It is up to the front-end to make sense of the magic numbers it places in the
2543 IR.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2544
2545</div>
2546
Chris Lattnerd0d96292010-01-15 21:50:19 +0000[diff] [blame]2547<!-- ======================================================================= -->
2548<div class="doc_subsection"><a name="metadata">Metadata Nodes and Metadata
2549 Strings</a>
2550</div>
2551
2552<div class="doc_text">
2553
2554<p>LLVM IR allows metadata to be attached to instructions in the program that
2555 can convey extra information about the code to the optimizers and code
2556 generator. One example application of metadata is source-level debug
2557 information. There are two metadata primitives: strings and nodes. All
2558 metadata has the <tt>metadata</tt> type and is identified in syntax by a
2559 preceding exclamation point ('<tt>!</tt>').</p>
2560
2561<p>A metadata string is a string surrounded by double quotes. It can contain
2562 any character by escaping non-printable characters with "\xx" where "xx" is
2563 the two digit hex code. For example: "<tt>!"test\00"</tt>".</p>
2564
2565<p>Metadata nodes are represented with notation similar to structure constants
2566 (a comma separated list of elements, surrounded by braces and preceded by an
2567 exclamation point). For example: "<tt>!{ metadata !"test\00", i32
2568 10}</tt>". Metadata nodes can have any values as their operand.</p>
2569
2570<p>A <a href="#namedmetadatastructure">named metadata</a> is a collection of
2571 metadata nodes, which can be looked up in the module symbol table. For
2572 example: "<tt>!foo = metadata !{!4, !3}</tt>".
2573
Devang Patelb1586922010-03-04 23:44:48 +0000[diff] [blame]2574<p>Metadata can be used as function arguments. Here <tt>llvm.dbg.value</tt>
2575 function is using two metadata arguments.
2576
2577 <div class="doc_code">
2578 <pre>
2579 call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2580 </pre>
2581 </div></p>
2582
2583<p>Metadata can be attached with an instruction. Here metadata <tt>!21</tt> is
2584 attached with <tt>add</tt> instruction using <tt>!dbg</tt> identifier.
2585
2586 <div class="doc_code">
2587 <pre>
2588 %indvar.next = add i64 %indvar, 1, !dbg !21
2589 </pre>
2590 </div></p>
Chris Lattnerd0d96292010-01-15 21:50:19 +0000[diff] [blame]2591</div>
2592
Chris Lattner75c24e02009-07-20 05:55:19 +0000[diff] [blame]2593
2594<!-- *********************************************************************** -->
2595<div class="doc_section">
2596 <a name="intrinsic_globals">Intrinsic Global Variables</a>
2597</div>
2598<!-- *********************************************************************** -->
2599
2600<p>LLVM has a number of "magic" global variables that contain data that affect
2601code generation or other IR semantics. These are documented here. All globals
Chris Lattner1e0e0d12009-07-20 06:14:25 +0000[diff] [blame]2602of this sort should have a section specified as "<tt>llvm.metadata</tt>". This
2603section and all globals that start with "<tt>llvm.</tt>" are reserved for use
2604by LLVM.</p>
Chris Lattner75c24e02009-07-20 05:55:19 +0000[diff] [blame]2605
2606<!-- ======================================================================= -->
2607<div class="doc_subsection">
2608<a name="intg_used">The '<tt>llvm.used</tt>' Global Variable</a>
2609</div>
2610
2611<div class="doc_text">
2612
2613<p>The <tt>@llvm.used</tt> global is an array with i8* element type which has <a
2614href="#linkage_appending">appending linkage</a>. This array contains a list of
2615pointers to global variables and functions which may optionally have a pointer
2616cast formed of bitcast or getelementptr. For example, a legal use of it is:</p>
2617
2618<pre>
2619 @X = global i8 4
2620 @Y = global i32 123
2621
2622 @llvm.used = appending global [2 x i8*] [
2623 i8* @X,
2624 i8* bitcast (i32* @Y to i8*)
2625 ], section "llvm.metadata"
2626</pre>
2627
2628<p>If a global variable appears in the <tt>@llvm.used</tt> list, then the
2629compiler, assembler, and linker are required to treat the symbol as if there is
2630a reference to the global that it cannot see. For example, if a variable has
2631internal linkage and no references other than that from the <tt>@llvm.used</tt>
2632list, it cannot be deleted. This is commonly used to represent references from
2633inline asms and other things the compiler cannot "see", and corresponds to
2634"attribute((used))" in GNU C.</p>
2635
2636<p>On some targets, the code generator must emit a directive to the assembler or
2637object file to prevent the assembler and linker from molesting the symbol.</p>
2638
2639</div>
2640
2641<!-- ======================================================================= -->
2642<div class="doc_subsection">
Chris Lattner1e0e0d12009-07-20 06:14:25 +0000[diff] [blame]2643<a name="intg_compiler_used">The '<tt>llvm.compiler.used</tt>' Global Variable</a>
2644</div>
2645
2646<div class="doc_text">
2647
2648<p>The <tt>@llvm.compiler.used</tt> directive is the same as the
2649<tt>@llvm.used</tt> directive, except that it only prevents the compiler from
2650touching the symbol. On targets that support it, this allows an intelligent
2651linker to optimize references to the symbol without being impeded as it would be
2652by <tt>@llvm.used</tt>.</p>
2653
2654<p>This is a rare construct that should only be used in rare circumstances, and
2655should not be exposed to source languages.</p>
2656
2657</div>
2658
2659<!-- ======================================================================= -->
2660<div class="doc_subsection">
Chris Lattner75c24e02009-07-20 05:55:19 +0000[diff] [blame]2661<a name="intg_global_ctors">The '<tt>llvm.global_ctors</tt>' Global Variable</a>
2662</div>
2663
2664<div class="doc_text">
2665
2666<p>TODO: Describe this.</p>
2667
2668</div>
2669
2670<!-- ======================================================================= -->
2671<div class="doc_subsection">
2672<a name="intg_global_dtors">The '<tt>llvm.global_dtors</tt>' Global Variable</a>
2673</div>
2674
2675<div class="doc_text">
2676
2677<p>TODO: Describe this.</p>
2678
2679</div>
2680
2681
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2682<!-- *********************************************************************** -->
2683<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
2684<!-- *********************************************************************** -->
2685
2686<div class="doc_text">
2687
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2688<p>The LLVM instruction set consists of several different classifications of
2689 instructions: <a href="#terminators">terminator
2690 instructions</a>, <a href="#binaryops">binary instructions</a>,
2691 <a href="#bitwiseops">bitwise binary instructions</a>,
2692 <a href="#memoryops">memory instructions</a>, and
2693 <a href="#otherops">other instructions</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2694
2695</div>
2696
2697<!-- ======================================================================= -->
2698<div class="doc_subsection"> <a name="terminators">Terminator
2699Instructions</a> </div>
2700
2701<div class="doc_text">
2702
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2703<p>As mentioned <a href="#functionstructure">previously</a>, every basic block
2704 in a program ends with a "Terminator" instruction, which indicates which
2705 block should be executed after the current block is finished. These
2706 terminator instructions typically yield a '<tt>void</tt>' value: they produce
2707 control flow, not values (the one exception being the
2708 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
2709
2710<p>There are six different terminator instructions: the
2711 '<a href="#i_ret"><tt>ret</tt></a>' instruction, the
2712 '<a href="#i_br"><tt>br</tt></a>' instruction, the
2713 '<a href="#i_switch"><tt>switch</tt></a>' instruction, the
Bill Wendlingb4ae2292009-11-02 00:25:26 +0000[diff] [blame]2714 '<a href="#i_indirectbr">'<tt>indirectbr</tt></a>' Instruction, the
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2715 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the
2716 '<a href="#i_unwind"><tt>unwind</tt></a>' instruction, and the
2717 '<a href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2718
2719</div>
2720
2721<!-- _______________________________________________________________________ -->
2722<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
2723Instruction</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2724
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2725<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2726
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2727<h5>Syntax:</h5>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2728<pre>
2729 ret &lt;type&gt; &lt;value&gt; <i>; Return a value from a non-void function</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2730 ret void <i>; Return from void function</i>
2731</pre>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2732
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2733<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2734<p>The '<tt>ret</tt>' instruction is used to return control flow (and optionally
2735 a value) from a function back to the caller.</p>
2736
2737<p>There are two forms of the '<tt>ret</tt>' instruction: one that returns a
2738 value and then causes control flow, and one that just causes control flow to
2739 occur.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2740
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2741<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2742<p>The '<tt>ret</tt>' instruction optionally accepts a single argument, the
2743 return value. The type of the return value must be a
2744 '<a href="#t_firstclass">first class</a>' type.</p>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2745
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2746<p>A function is not <a href="#wellformed">well formed</a> if it it has a
2747 non-void return type and contains a '<tt>ret</tt>' instruction with no return
2748 value or a return value with a type that does not match its type, or if it
2749 has a void return type and contains a '<tt>ret</tt>' instruction with a
2750 return value.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2751
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2752<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2753<p>When the '<tt>ret</tt>' instruction is executed, control flow returns back to
2754 the calling function's context. If the caller is a
2755 "<a href="#i_call"><tt>call</tt></a>" instruction, execution continues at the
2756 instruction after the call. If the caller was an
2757 "<a href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues at
2758 the beginning of the "normal" destination block. If the instruction returns
2759 a value, that value shall set the call or invoke instruction's return
2760 value.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2761
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2762<h5>Example:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2763<pre>
2764 ret i32 5 <i>; Return an integer value of 5</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2765 ret void <i>; Return from a void function</i>
Bill Wendlingd163e2d2009-02-28 22:12:54 +0000[diff] [blame]2766 ret { i32, i8 } { i32 4, i8 2 } <i>; Return a struct of values 4 and 2</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2767</pre>
Dan Gohman60967192009-01-12 23:12:39 +0000[diff] [blame]2768
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2769</div>
2770<!-- _______________________________________________________________________ -->
2771<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2772
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2773<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2774
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2775<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2776<pre>
2777 br i1 &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;<br> br label &lt;dest&gt; <i>; Unconditional branch</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2778</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2779
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2780<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2781<p>The '<tt>br</tt>' instruction is used to cause control flow to transfer to a
2782 different basic block in the current function. There are two forms of this
2783 instruction, corresponding to a conditional branch and an unconditional
2784 branch.</p>
2785
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2786<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2787<p>The conditional branch form of the '<tt>br</tt>' instruction takes a single
2788 '<tt>i1</tt>' value and two '<tt>label</tt>' values. The unconditional form
2789 of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>' value as a
2790 target.</p>
2791
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2792<h5>Semantics:</h5>
2793<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2794 argument is evaluated. If the value is <tt>true</tt>, control flows to the
2795 '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
2796 control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
2797
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2798<h5>Example:</h5>
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]2799<pre>
2800Test:
2801 %cond = <a href="#i_icmp">icmp</a> eq i32 %a, %b
2802 br i1 %cond, label %IfEqual, label %IfUnequal
2803IfEqual:
2804 <a href="#i_ret">ret</a> i32 1
2805IfUnequal:
2806 <a href="#i_ret">ret</a> i32 0
2807</pre>
2808
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2809</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2810
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2811<!-- _______________________________________________________________________ -->
2812<div class="doc_subsubsection">
2813 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
2814</div>
2815
2816<div class="doc_text">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2817
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2818<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2819<pre>
2820 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
2821</pre>
2822
2823<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2824<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2825 several different places. It is a generalization of the '<tt>br</tt>'
2826 instruction, allowing a branch to occur to one of many possible
2827 destinations.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2828
2829<h5>Arguments:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2830<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2831 comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination,
2832 and an array of pairs of comparison value constants and '<tt>label</tt>'s.
2833 The table is not allowed to contain duplicate constant entries.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2834
2835<h5>Semantics:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2836<p>The <tt>switch</tt> instruction specifies a table of values and
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2837 destinations. When the '<tt>switch</tt>' instruction is executed, this table
2838 is searched for the given value. If the value is found, control flow is
Benjamin Kramer5fb9d7e2009-10-12 14:46:08 +0000[diff] [blame]2839 transferred to the corresponding destination; otherwise, control flow is
2840 transferred to the default destination.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2841
2842<h5>Implementation:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2843<p>Depending on properties of the target machine and the particular
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2844 <tt>switch</tt> instruction, this instruction may be code generated in
2845 different ways. For example, it could be generated as a series of chained
2846 conditional branches or with a lookup table.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2847
2848<h5>Example:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2849<pre>
2850 <i>; Emulate a conditional br instruction</i>
2851 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]2852 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2853
2854 <i>; Emulate an unconditional br instruction</i>
2855 switch i32 0, label %dest [ ]
2856
2857 <i>; Implement a jump table:</i>
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]2858 switch i32 %val, label %otherwise [ i32 0, label %onzero
2859 i32 1, label %onone
2860 i32 2, label %ontwo ]
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2861</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2862
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2863</div>
2864
Chris Lattnere0787282009-10-27 19:13:16 +0000[diff] [blame]2865
2866<!-- _______________________________________________________________________ -->
2867<div class="doc_subsubsection">
Chris Lattner4c3800f2009-10-28 00:19:10 +0000[diff] [blame]2868 <a name="i_indirectbr">'<tt>indirectbr</tt>' Instruction</a>
Chris Lattnere0787282009-10-27 19:13:16 +0000[diff] [blame]2869</div>
2870
2871<div class="doc_text">
2872
2873<h5>Syntax:</h5>
2874<pre>
Chris Lattner4c3800f2009-10-28 00:19:10 +0000[diff] [blame]2875 indirectbr &lt;somety&gt;* &lt;address&gt;, [ label &lt;dest1&gt;, label &lt;dest2&gt;, ... ]
Chris Lattnere0787282009-10-27 19:13:16 +0000[diff] [blame]2876</pre>
2877
2878<h5>Overview:</h5>
2879
Chris Lattner4c3800f2009-10-28 00:19:10 +0000[diff] [blame]2880<p>The '<tt>indirectbr</tt>' instruction implements an indirect branch to a label
Chris Lattnere0787282009-10-27 19:13:16 +0000[diff] [blame]2881 within the current function, whose address is specified by
Chris Lattnerd07c8372009-10-27 21:01:34 +0000[diff] [blame]2882 "<tt>address</tt>". Address must be derived from a <a
2883 href="#blockaddress">blockaddress</a> constant.</p>
Chris Lattnere0787282009-10-27 19:13:16 +0000[diff] [blame]2884
2885<h5>Arguments:</h5>
2886
2887<p>The '<tt>address</tt>' argument is the address of the label to jump to. The
2888 rest of the arguments indicate the full set of possible destinations that the
2889 address may point to. Blocks are allowed to occur multiple times in the
2890 destination list, though this isn't particularly useful.</p>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]2891
Chris Lattnere0787282009-10-27 19:13:16 +0000[diff] [blame]2892<p>This destination list is required so that dataflow analysis has an accurate
2893 understanding of the CFG.</p>
2894
2895<h5>Semantics:</h5>
2896
2897<p>Control transfers to the block specified in the address argument. All
2898 possible destination blocks must be listed in the label list, otherwise this
2899 instruction has undefined behavior. This implies that jumps to labels
2900 defined in other functions have undefined behavior as well.</p>
2901
2902<h5>Implementation:</h5>
2903
2904<p>This is typically implemented with a jump through a register.</p>
2905
2906<h5>Example:</h5>
2907<pre>
Chris Lattner4c3800f2009-10-28 00:19:10 +0000[diff] [blame]2908 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
Chris Lattnere0787282009-10-27 19:13:16 +0000[diff] [blame]2909</pre>
2910
2911</div>
2912
2913
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2914<!-- _______________________________________________________________________ -->
2915<div class="doc_subsubsection">
2916 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
2917</div>
2918
2919<div class="doc_text">
2920
2921<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2922<pre>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]2923 &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>]
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2924 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
2925</pre>
2926
2927<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2928<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2929 function, with the possibility of control flow transfer to either the
2930 '<tt>normal</tt>' label or the '<tt>exception</tt>' label. If the callee
2931 function returns with the "<tt><a href="#i_ret">ret</a></tt>" instruction,
2932 control flow will return to the "normal" label. If the callee (or any
2933 indirect callees) returns with the "<a href="#i_unwind"><tt>unwind</tt></a>"
2934 instruction, control is interrupted and continued at the dynamically nearest
2935 "exception" label.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2936
2937<h5>Arguments:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2938<p>This instruction requires several arguments:</p>
2939
2940<ol>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2941 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
2942 convention</a> the call should use. If none is specified, the call
2943 defaults to using C calling conventions.</li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]2944
2945 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2946 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
2947 '<tt>inreg</tt>' attributes are valid here.</li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]2948
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2949 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2950 function value being invoked. In most cases, this is a direct function
2951 invocation, but indirect <tt>invoke</tt>s are just as possible, branching
2952 off an arbitrary pointer to function value.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2953
2954 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2955 function to be invoked. </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2956
2957 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner553fb1e2010-03-02 06:36:51 +0000[diff] [blame]2958 signature argument types and parameter attributes. All arguments must be
2959 of <a href="#t_firstclass">first class</a> type. If the function
2960 signature indicates the function accepts a variable number of arguments,
2961 the extra arguments can be specified.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2962
2963 <li>'<tt>normal label</tt>': the label reached when the called function
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2964 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2965
2966 <li>'<tt>exception label</tt>': the label reached when a callee returns with
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2967 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2968
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]2969 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2970 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
2971 '<tt>readnone</tt>' attributes are valid here.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2972</ol>
2973
2974<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2975<p>This instruction is designed to operate as a standard
2976 '<tt><a href="#i_call">call</a></tt>' instruction in most regards. The
2977 primary difference is that it establishes an association with a label, which
2978 is used by the runtime library to unwind the stack.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2979
2980<p>This instruction is used in languages with destructors to ensure that proper
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2981 cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
2982 exception. Additionally, this is important for implementation of
2983 '<tt>catch</tt>' clauses in high-level languages that support them.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2984
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2985<p>For the purposes of the SSA form, the definition of the value returned by the
2986 '<tt>invoke</tt>' instruction is deemed to occur on the edge from the current
2987 block to the "normal" label. If the callee unwinds then no return value is
2988 available.</p>
Dan Gohman140ba5d2009-05-22 21:47:08 +0000[diff] [blame]2989
Chris Lattner4a91ef42010-01-15 18:08:37 +0000[diff] [blame]2990<p>Note that the code generator does not yet completely support unwind, and
2991that the invoke/unwind semantics are likely to change in future versions.</p>
2992
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2993<h5>Example:</h5>
2994<pre>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2995 %retval = invoke i32 @Test(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2996 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2997 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2998 unwind label %TestCleanup <i>; {i32}:retval set</i>
2999</pre>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3000
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3001</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3002
3003<!-- _______________________________________________________________________ -->
3004
3005<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
3006Instruction</a> </div>
3007
3008<div class="doc_text">
3009
3010<h5>Syntax:</h5>
3011<pre>
3012 unwind
3013</pre>
3014
3015<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3016<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3017 at the first callee in the dynamic call stack which used
3018 an <a href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call.
3019 This is primarily used to implement exception handling.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3020
3021<h5>Semantics:</h5>
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]3022<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3023 immediately halt. The dynamic call stack is then searched for the
3024 first <a href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack.
3025 Once found, execution continues at the "exceptional" destination block
3026 specified by the <tt>invoke</tt> instruction. If there is no <tt>invoke</tt>
3027 instruction in the dynamic call chain, undefined behavior results.</p>
3028
Chris Lattner4a91ef42010-01-15 18:08:37 +0000[diff] [blame]3029<p>Note that the code generator does not yet completely support unwind, and
3030that the invoke/unwind semantics are likely to change in future versions.</p>
3031
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3032</div>
3033
3034<!-- _______________________________________________________________________ -->
3035
3036<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
3037Instruction</a> </div>
3038
3039<div class="doc_text">
3040
3041<h5>Syntax:</h5>
3042<pre>
3043 unreachable
3044</pre>
3045
3046<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3047<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3048 instruction is used to inform the optimizer that a particular portion of the
3049 code is not reachable. This can be used to indicate that the code after a
3050 no-return function cannot be reached, and other facts.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3051
3052<h5>Semantics:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3053<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3054
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3055</div>
3056
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3057<!-- ======================================================================= -->
3058<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3059
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3060<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3061
3062<p>Binary operators are used to do most of the computation in a program. They
3063 require two operands of the same type, execute an operation on them, and
3064 produce a single value. The operands might represent multiple data, as is
3065 the case with the <a href="#t_vector">vector</a> data type. The result value
3066 has the same type as its operands.</p>
3067
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3068<p>There are several different binary operators:</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3069
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3070</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3071
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3072<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3073<div class="doc_subsubsection">
3074 <a name="i_add">'<tt>add</tt>' Instruction</a>
3075</div>
3076
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3077<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3078
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3079<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3080<pre>
Dan Gohmaned5fcb22009-07-20 22:41:19 +0000[diff] [blame]3081 &lt;result&gt; = add &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanb3b03482009-09-02 17:31:42 +0000[diff] [blame]3082 &lt;result&gt; = add nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3083 &lt;result&gt; = add nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3084 &lt;result&gt; = add nuw nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3085</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3086
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3087<h5>Overview:</h5>
3088<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3089
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3090<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3091<p>The two arguments to the '<tt>add</tt>' instruction must
3092 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3093 integer values. Both arguments must have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3094
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3095<h5>Semantics:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3096<p>The value produced is the integer sum of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3097
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3098<p>If the sum has unsigned overflow, the result returned is the mathematical
3099 result modulo 2<sup>n</sup>, where n is the bit width of the result.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3100
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3101<p>Because LLVM integers use a two's complement representation, this instruction
3102 is appropriate for both signed and unsigned integers.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3103
Dan Gohman46e96012009-07-22 22:44:56 +0000[diff] [blame]3104<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3105 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3106 <tt>nsw</tt> keywords are present, the result value of the <tt>add</tt>
3107 is undefined if unsigned and/or signed overflow, respectively, occurs.</p>
Dan Gohmaned5fcb22009-07-20 22:41:19 +0000[diff] [blame]3108
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3109<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3110<pre>
3111 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3112</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3113
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3114</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3115
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3116<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3117<div class="doc_subsubsection">
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3118 <a name="i_fadd">'<tt>fadd</tt>' Instruction</a>
3119</div>
3120
3121<div class="doc_text">
3122
3123<h5>Syntax:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3124<pre>
3125 &lt;result&gt; = fadd &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3126</pre>
3127
3128<h5>Overview:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3129<p>The '<tt>fadd</tt>' instruction returns the sum of its two operands.</p>
3130
3131<h5>Arguments:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3132<p>The two arguments to the '<tt>fadd</tt>' instruction must be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3133 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3134 floating point values. Both arguments must have identical types.</p>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3135
3136<h5>Semantics:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3137<p>The value produced is the floating point sum of the two operands.</p>
3138
3139<h5>Example:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3140<pre>
3141 &lt;result&gt; = fadd float 4.0, %var <i>; yields {float}:result = 4.0 + %var</i>
3142</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3143
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3144</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3145
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3146<!-- _______________________________________________________________________ -->
3147<div class="doc_subsubsection">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3148 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
3149</div>
3150
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3151<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3152
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3153<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3154<pre>
Dan Gohman46e96012009-07-22 22:44:56 +0000[diff] [blame]3155 &lt;result&gt; = sub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanb3b03482009-09-02 17:31:42 +0000[diff] [blame]3156 &lt;result&gt; = sub nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3157 &lt;result&gt; = sub nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3158 &lt;result&gt; = sub nuw nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3159</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3160
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3161<h5>Overview:</h5>
3162<p>The '<tt>sub</tt>' instruction returns the difference of its two
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3163 operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3164
3165<p>Note that the '<tt>sub</tt>' instruction is used to represent the
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3166 '<tt>neg</tt>' instruction present in most other intermediate
3167 representations.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3168
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3169<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3170<p>The two arguments to the '<tt>sub</tt>' instruction must
3171 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3172 integer values. Both arguments must have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3173
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3174<h5>Semantics:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3175<p>The value produced is the integer difference of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3176
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3177<p>If the difference has unsigned overflow, the result returned is the
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3178 mathematical result modulo 2<sup>n</sup>, where n is the bit width of the
3179 result.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3180
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3181<p>Because LLVM integers use a two's complement representation, this instruction
3182 is appropriate for both signed and unsigned integers.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3183
Dan Gohman46e96012009-07-22 22:44:56 +0000[diff] [blame]3184<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3185 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3186 <tt>nsw</tt> keywords are present, the result value of the <tt>sub</tt>
3187 is undefined if unsigned and/or signed overflow, respectively, occurs.</p>
Dan Gohmaned5fcb22009-07-20 22:41:19 +0000[diff] [blame]3188
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3189<h5>Example:</h5>
3190<pre>
3191 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
3192 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
3193</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3194
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3195</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3196
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3197<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3198<div class="doc_subsubsection">
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3199 <a name="i_fsub">'<tt>fsub</tt>' Instruction</a>
3200</div>
3201
3202<div class="doc_text">
3203
3204<h5>Syntax:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3205<pre>
3206 &lt;result&gt; = fsub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3207</pre>
3208
3209<h5>Overview:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3210<p>The '<tt>fsub</tt>' instruction returns the difference of its two
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3211 operands.</p>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3212
3213<p>Note that the '<tt>fsub</tt>' instruction is used to represent the
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3214 '<tt>fneg</tt>' instruction present in most other intermediate
3215 representations.</p>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3216
3217<h5>Arguments:</h5>
Bill Wendling1a2630a2009-07-20 02:32:41 +0000[diff] [blame]3218<p>The two arguments to the '<tt>fsub</tt>' instruction must be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3219 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3220 floating point values. Both arguments must have identical types.</p>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3221
3222<h5>Semantics:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3223<p>The value produced is the floating point difference of the two operands.</p>
3224
3225<h5>Example:</h5>
3226<pre>
3227 &lt;result&gt; = fsub float 4.0, %var <i>; yields {float}:result = 4.0 - %var</i>
3228 &lt;result&gt; = fsub float -0.0, %val <i>; yields {float}:result = -%var</i>
3229</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3230
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3231</div>
3232
3233<!-- _______________________________________________________________________ -->
3234<div class="doc_subsubsection">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3235 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
3236</div>
3237
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3238<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3239
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3240<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3241<pre>
Dan Gohman46e96012009-07-22 22:44:56 +0000[diff] [blame]3242 &lt;result&gt; = mul &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanb3b03482009-09-02 17:31:42 +0000[diff] [blame]3243 &lt;result&gt; = mul nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3244 &lt;result&gt; = mul nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3245 &lt;result&gt; = mul nuw nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3246</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3247
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3248<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3249<p>The '<tt>mul</tt>' instruction returns the product of its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3250
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3251<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3252<p>The two arguments to the '<tt>mul</tt>' instruction must
3253 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3254 integer values. Both arguments must have identical types.</p>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3255
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3256<h5>Semantics:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3257<p>The value produced is the integer product of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3258
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3259<p>If the result of the multiplication has unsigned overflow, the result
3260 returned is the mathematical result modulo 2<sup>n</sup>, where n is the bit
3261 width of the result.</p>
3262
3263<p>Because LLVM integers use a two's complement representation, and the result
3264 is the same width as the operands, this instruction returns the correct
3265 result for both signed and unsigned integers. If a full product
3266 (e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands should
3267 be sign-extended or zero-extended as appropriate to the width of the full
3268 product.</p>
3269
Dan Gohman46e96012009-07-22 22:44:56 +0000[diff] [blame]3270<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3271 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3272 <tt>nsw</tt> keywords are present, the result value of the <tt>mul</tt>
3273 is undefined if unsigned and/or signed overflow, respectively, occurs.</p>
Dan Gohmaned5fcb22009-07-20 22:41:19 +0000[diff] [blame]3274
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3275<h5>Example:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3276<pre>
3277 &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3278</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3279
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3280</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3281
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3282<!-- _______________________________________________________________________ -->
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3283<div class="doc_subsubsection">
3284 <a name="i_fmul">'<tt>fmul</tt>' Instruction</a>
3285</div>
3286
3287<div class="doc_text">
3288
3289<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3290<pre>
3291 &lt;result&gt; = fmul &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3292</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3293
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3294<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3295<p>The '<tt>fmul</tt>' instruction returns the product of its two operands.</p>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3296
3297<h5>Arguments:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3298<p>The two arguments to the '<tt>fmul</tt>' instruction must be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3299 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3300 floating point values. Both arguments must have identical types.</p>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3301
3302<h5>Semantics:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3303<p>The value produced is the floating point product of the two operands.</p>
3304
3305<h5>Example:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3306<pre>
3307 &lt;result&gt; = fmul float 4.0, %var <i>; yields {float}:result = 4.0 * %var</i>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3308</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3309
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3310</div>
3311
3312<!-- _______________________________________________________________________ -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3313<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
3314</a></div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3315
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3316<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3317
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3318<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3319<pre>
3320 &lt;result&gt; = udiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3321</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3322
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3323<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3324<p>The '<tt>udiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3325
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3326<h5>Arguments:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3327<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3328 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3329 values. Both arguments must have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3330
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3331<h5>Semantics:</h5>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]3332<p>The value produced is the unsigned integer quotient of the two operands.</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3333
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]3334<p>Note that unsigned integer division and signed integer division are distinct
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3335 operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
3336
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]3337<p>Division by zero leads to undefined behavior.</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3338
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3339<h5>Example:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3340<pre>
3341 &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3342</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3343
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3344</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3345
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3346<!-- _______________________________________________________________________ -->
3347<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
3348</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3349
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3350<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3351
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3352<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3353<pre>
Dan Gohmaned5fcb22009-07-20 22:41:19 +0000[diff] [blame]3354 &lt;result&gt; = sdiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanb3b03482009-09-02 17:31:42 +0000[diff] [blame]3355 &lt;result&gt; = sdiv exact &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3356</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3357
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3358<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3359<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3360
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3361<h5>Arguments:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3362<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3363 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3364 values. Both arguments must have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3365
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3366<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3367<p>The value produced is the signed integer quotient of the two operands rounded
3368 towards zero.</p>
3369
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]3370<p>Note that signed integer division and unsigned integer division are distinct
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3371 operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
3372
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]3373<p>Division by zero leads to undefined behavior. Overflow also leads to
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3374 undefined behavior; this is a rare case, but can occur, for example, by doing
3375 a 32-bit division of -2147483648 by -1.</p>
3376
Dan Gohman67fa48e2009-07-22 00:04:19 +0000[diff] [blame]3377<p>If the <tt>exact</tt> keyword is present, the result value of the
3378 <tt>sdiv</tt> is undefined if the result would be rounded or if overflow
3379 would occur.</p>
Dan Gohmaned5fcb22009-07-20 22:41:19 +0000[diff] [blame]3380
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3381<h5>Example:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3382<pre>
3383 &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3384</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3385
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3386</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3387
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3388<!-- _______________________________________________________________________ -->
3389<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
3390Instruction</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3391
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3392<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3393
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3394<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3395<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3396 &lt;result&gt; = fdiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3397</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3398
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3399<h5>Overview:</h5>
3400<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3401
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3402<h5>Arguments:</h5>
3403<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3404 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3405 floating point values. Both arguments must have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3406
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3407<h5>Semantics:</h5>
3408<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3409
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3410<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3411<pre>
3412 &lt;result&gt; = fdiv float 4.0, %var <i>; yields {float}:result = 4.0 / %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3413</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3414
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3415</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3416
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3417<!-- _______________________________________________________________________ -->
3418<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
3419</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3420
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3421<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3422
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3423<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3424<pre>
3425 &lt;result&gt; = urem &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3426</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3427
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3428<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3429<p>The '<tt>urem</tt>' instruction returns the remainder from the unsigned
3430 division of its two arguments.</p>
3431
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3432<h5>Arguments:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3433<p>The two arguments to the '<tt>urem</tt>' instruction must be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3434 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3435 values. Both arguments must have identical types.</p>
3436
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3437<h5>Semantics:</h5>
3438<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3439 This instruction always performs an unsigned division to get the
3440 remainder.</p>
3441
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]3442<p>Note that unsigned integer remainder and signed integer remainder are
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3443 distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
3444
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]3445<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3446
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3447<h5>Example:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3448<pre>
3449 &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3450</pre>
3451
3452</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3453
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3454<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3455<div class="doc_subsubsection">
3456 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
3457</div>
3458
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3459<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3460
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3461<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3462<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3463 &lt;result&gt; = srem &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3464</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3465
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3466<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3467<p>The '<tt>srem</tt>' instruction returns the remainder from the signed
3468 division of its two operands. This instruction can also take
3469 <a href="#t_vector">vector</a> versions of the values in which case the
3470 elements must be integers.</p>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]3471
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3472<h5>Arguments:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3473<p>The two arguments to the '<tt>srem</tt>' instruction must be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3474 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3475 values. Both arguments must have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3476
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3477<h5>Semantics:</h5>
3478<p>This instruction returns the <i>remainder</i> of a division (where the result
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3479 has the same sign as the dividend, <tt>op1</tt>), not the <i>modulo</i>
3480 operator (where the result has the same sign as the divisor, <tt>op2</tt>) of
3481 a value. For more information about the difference,
3482 see <a href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
3483 Math Forum</a>. For a table of how this is implemented in various languages,
3484 please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
3485 Wikipedia: modulo operation</a>.</p>
3486
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]3487<p>Note that signed integer remainder and unsigned integer remainder are
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3488 distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
3489
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]3490<p>Taking the remainder of a division by zero leads to undefined behavior.
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3491 Overflow also leads to undefined behavior; this is a rare case, but can
3492 occur, for example, by taking the remainder of a 32-bit division of
3493 -2147483648 by -1. (The remainder doesn't actually overflow, but this rule
3494 lets srem be implemented using instructions that return both the result of
3495 the division and the remainder.)</p>
3496
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3497<h5>Example:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3498<pre>
3499 &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3500</pre>
3501
3502</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3503
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3504<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3505<div class="doc_subsubsection">
3506 <a name="i_frem">'<tt>frem</tt>' Instruction</a> </div>
3507
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3508<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3509
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3510<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3511<pre>
3512 &lt;result&gt; = frem &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3513</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3514
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3515<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3516<p>The '<tt>frem</tt>' instruction returns the remainder from the division of
3517 its two operands.</p>
3518
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3519<h5>Arguments:</h5>
3520<p>The two arguments to the '<tt>frem</tt>' instruction must be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3521 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3522 floating point values. Both arguments must have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3523
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3524<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3525<p>This instruction returns the <i>remainder</i> of a division. The remainder
3526 has the same sign as the dividend.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3527
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3528<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3529<pre>
3530 &lt;result&gt; = frem float 4.0, %var <i>; yields {float}:result = 4.0 % %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3531</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3532
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3533</div>
3534
3535<!-- ======================================================================= -->
3536<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
3537Operations</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3538
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3539<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3540
3541<p>Bitwise binary operators are used to do various forms of bit-twiddling in a
3542 program. They are generally very efficient instructions and can commonly be
3543 strength reduced from other instructions. They require two operands of the
3544 same type, execute an operation on them, and produce a single value. The
3545 resulting value is the same type as its operands.</p>
3546
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3547</div>
3548
3549<!-- _______________________________________________________________________ -->
3550<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
3551Instruction</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3552
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3553<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3554
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3555<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3556<pre>
3557 &lt;result&gt; = shl &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3558</pre>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]3559
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3560<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3561<p>The '<tt>shl</tt>' instruction returns the first operand shifted to the left
3562 a specified number of bits.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]3563
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3564<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3565<p>Both arguments to the '<tt>shl</tt>' instruction must be the
3566 same <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3567 integer type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3568
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3569<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3570<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod
3571 2<sup>n</sup>, where <tt>n</tt> is the width of the result. If <tt>op2</tt>
3572 is (statically or dynamically) negative or equal to or larger than the number
3573 of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3574 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3575 shift amount in <tt>op2</tt>.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]3576
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3577<h5>Example:</h5>
3578<pre>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3579 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
3580 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
3581 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]3582 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Mon P Wang9901e732008-12-09 05:46:39 +0000[diff] [blame]3583 &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>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3584</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3585
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3586</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3587
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3588<!-- _______________________________________________________________________ -->
3589<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
3590Instruction</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3591
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3592<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3593
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3594<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3595<pre>
3596 &lt;result&gt; = lshr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3597</pre>
3598
3599<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3600<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
3601 operand shifted to the right a specified number of bits with zero fill.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3602
3603<h5>Arguments:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3604<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3605 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3606 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3607
3608<h5>Semantics:</h5>
3609<p>This instruction always performs a logical shift right operation. The most
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3610 significant bits of the result will be filled with zero bits after the shift.
3611 If <tt>op2</tt> is (statically or dynamically) equal to or larger than the
3612 number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3613 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3614 shift amount in <tt>op2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3615
3616<h5>Example:</h5>
3617<pre>
3618 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
3619 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
3620 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
3621 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]3622 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Mon P Wang9901e732008-12-09 05:46:39 +0000[diff] [blame]3623 &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>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3624</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3625
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3626</div>
3627
3628<!-- _______________________________________________________________________ -->
3629<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
3630Instruction</a> </div>
3631<div class="doc_text">
3632
3633<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3634<pre>
3635 &lt;result&gt; = ashr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3636</pre>
3637
3638<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3639<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
3640 operand shifted to the right a specified number of bits with sign
3641 extension.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3642
3643<h5>Arguments:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3644<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3645 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3646 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3647
3648<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3649<p>This instruction always performs an arithmetic shift right operation, The
3650 most significant bits of the result will be filled with the sign bit
3651 of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
3652 larger than the number of bits in <tt>op1</tt>, the result is undefined. If
3653 the arguments are vectors, each vector element of <tt>op1</tt> is shifted by
3654 the corresponding shift amount in <tt>op2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3655
3656<h5>Example:</h5>
3657<pre>
3658 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
3659 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
3660 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
3661 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]3662 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Mon P Wang9901e732008-12-09 05:46:39 +0000[diff] [blame]3663 &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>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3664</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3665
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3666</div>
3667
3668<!-- _______________________________________________________________________ -->
3669<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
3670Instruction</a> </div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3671
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3672<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3673
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3674<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3675<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3676 &lt;result&gt; = and &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3677</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3678
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3679<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3680<p>The '<tt>and</tt>' instruction returns the bitwise logical and of its two
3681 operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3682
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3683<h5>Arguments:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3684<p>The two arguments to the '<tt>and</tt>' instruction must be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3685 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3686 values. Both arguments must have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3687
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3688<h5>Semantics:</h5>
3689<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3690
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3691<table border="1" cellspacing="0" cellpadding="4">
3692 <tbody>
3693 <tr>
3694 <td>In0</td>
3695 <td>In1</td>
3696 <td>Out</td>
3697 </tr>
3698 <tr>
3699 <td>0</td>
3700 <td>0</td>
3701 <td>0</td>
3702 </tr>
3703 <tr>
3704 <td>0</td>
3705 <td>1</td>
3706 <td>0</td>
3707 </tr>
3708 <tr>
3709 <td>1</td>
3710 <td>0</td>
3711 <td>0</td>
3712 </tr>
3713 <tr>
3714 <td>1</td>
3715 <td>1</td>
3716 <td>1</td>
3717 </tr>
3718 </tbody>
3719</table>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3720
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3721<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3722<pre>
3723 &lt;result&gt; = and i32 4, %var <i>; yields {i32}:result = 4 &amp; %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3724 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
3725 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
3726</pre>
3727</div>
3728<!-- _______________________________________________________________________ -->
3729<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3730
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3731<div class="doc_text">
3732
3733<h5>Syntax:</h5>
3734<pre>
3735 &lt;result&gt; = or &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3736</pre>
3737
3738<h5>Overview:</h5>
3739<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive or of its
3740 two operands.</p>
3741
3742<h5>Arguments:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3743<p>The two arguments to the '<tt>or</tt>' instruction must be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3744 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3745 values. Both arguments must have identical types.</p>
3746
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3747<h5>Semantics:</h5>
3748<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3749
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3750<table border="1" cellspacing="0" cellpadding="4">
3751 <tbody>
3752 <tr>
3753 <td>In0</td>
3754 <td>In1</td>
3755 <td>Out</td>
3756 </tr>
3757 <tr>
3758 <td>0</td>
3759 <td>0</td>
3760 <td>0</td>
3761 </tr>
3762 <tr>
3763 <td>0</td>
3764 <td>1</td>
3765 <td>1</td>
3766 </tr>
3767 <tr>
3768 <td>1</td>
3769 <td>0</td>
3770 <td>1</td>
3771 </tr>
3772 <tr>
3773 <td>1</td>
3774 <td>1</td>
3775 <td>1</td>
3776 </tr>
3777 </tbody>
3778</table>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3779
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3780<h5>Example:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3781<pre>
3782 &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3783 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
3784 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
3785</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3786
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3787</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3788
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3789<!-- _______________________________________________________________________ -->
3790<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
3791Instruction</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3792
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3793<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3794
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3795<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3796<pre>
3797 &lt;result&gt; = xor &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3798</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3799
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3800<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3801<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive or of
3802 its two operands. The <tt>xor</tt> is used to implement the "one's
3803 complement" operation, which is the "~" operator in C.</p>
3804
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3805<h5>Arguments:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3806<p>The two arguments to the '<tt>xor</tt>' instruction must be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3807 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3808 values. Both arguments must have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3809
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3810<h5>Semantics:</h5>
3811<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3812
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3813<table border="1" cellspacing="0" cellpadding="4">
3814 <tbody>
3815 <tr>
3816 <td>In0</td>
3817 <td>In1</td>
3818 <td>Out</td>
3819 </tr>
3820 <tr>
3821 <td>0</td>
3822 <td>0</td>
3823 <td>0</td>
3824 </tr>
3825 <tr>
3826 <td>0</td>
3827 <td>1</td>
3828 <td>1</td>
3829 </tr>
3830 <tr>
3831 <td>1</td>
3832 <td>0</td>
3833 <td>1</td>
3834 </tr>
3835 <tr>
3836 <td>1</td>
3837 <td>1</td>
3838 <td>0</td>
3839 </tr>
3840 </tbody>
3841</table>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3842
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3843<h5>Example:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3844<pre>
3845 &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3846 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
3847 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
3848 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
3849</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3850
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3851</div>
3852
3853<!-- ======================================================================= -->
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3854<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3855 <a name="vectorops">Vector Operations</a>
3856</div>
3857
3858<div class="doc_text">
3859
3860<p>LLVM supports several instructions to represent vector operations in a
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3861 target-independent manner. These instructions cover the element-access and
3862 vector-specific operations needed to process vectors effectively. While LLVM
3863 does directly support these vector operations, many sophisticated algorithms
3864 will want to use target-specific intrinsics to take full advantage of a
3865 specific target.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3866
3867</div>
3868
3869<!-- _______________________________________________________________________ -->
3870<div class="doc_subsubsection">
3871 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
3872</div>
3873
3874<div class="doc_text">
3875
3876<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3877<pre>
3878 &lt;result&gt; = extractelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, i32 &lt;idx&gt; <i>; yields &lt;ty&gt;</i>
3879</pre>
3880
3881<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3882<p>The '<tt>extractelement</tt>' instruction extracts a single scalar element
3883 from a vector at a specified index.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3884
3885
3886<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3887<p>The first operand of an '<tt>extractelement</tt>' instruction is a value
3888 of <a href="#t_vector">vector</a> type. The second operand is an index
3889 indicating the position from which to extract the element. The index may be
3890 a variable.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3891
3892<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3893<p>The result is a scalar of the same type as the element type of
3894 <tt>val</tt>. Its value is the value at position <tt>idx</tt> of
3895 <tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
3896 results are undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3897
3898<h5>Example:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3899<pre>
Gabor Greifb2c73142009-10-28 13:14:50 +0000[diff] [blame]3900 &lt;result&gt; = extractelement &lt;4 x i32&gt; %vec, i32 0 <i>; yields i32</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3901</pre>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3902
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3903</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3904
3905<!-- _______________________________________________________________________ -->
3906<div class="doc_subsubsection">
3907 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
3908</div>
3909
3910<div class="doc_text">
3911
3912<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3913<pre>
Dan Gohmanbcc3c502008-05-12 23:38:42 +0000[diff] [blame]3914 &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>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3915</pre>
3916
3917<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3918<p>The '<tt>insertelement</tt>' instruction inserts a scalar element into a
3919 vector at a specified index.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3920
3921<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3922<p>The first operand of an '<tt>insertelement</tt>' instruction is a value
3923 of <a href="#t_vector">vector</a> type. The second operand is a scalar value
3924 whose type must equal the element type of the first operand. The third
3925 operand is an index indicating the position at which to insert the value.
3926 The index may be a variable.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3927
3928<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3929<p>The result is a vector of the same type as <tt>val</tt>. Its element values
3930 are those of <tt>val</tt> except at position <tt>idx</tt>, where it gets the
3931 value <tt>elt</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
3932 results are undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3933
3934<h5>Example:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3935<pre>
Gabor Greifb2c73142009-10-28 13:14:50 +0000[diff] [blame]3936 &lt;result&gt; = insertelement &lt;4 x i32&gt; %vec, i32 1, i32 0 <i>; yields &lt;4 x i32&gt;</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3937</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3938
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3939</div>
3940
3941<!-- _______________________________________________________________________ -->
3942<div class="doc_subsubsection">
3943 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
3944</div>
3945
3946<div class="doc_text">
3947
3948<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3949<pre>
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3950 &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>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3951</pre>
3952
3953<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3954<p>The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
3955 from two input vectors, returning a vector with the same element type as the
3956 input and length that is the same as the shuffle mask.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3957
3958<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3959<p>The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
3960 with types that match each other. The third argument is a shuffle mask whose
3961 element type is always 'i32'. The result of the instruction is a vector
3962 whose length is the same as the shuffle mask and whose element type is the
3963 same as the element type of the first two operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3964
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3965<p>The shuffle mask operand is required to be a constant vector with either
3966 constant integer or undef values.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3967
3968<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3969<p>The elements of the two input vectors are numbered from left to right across
3970 both of the vectors. The shuffle mask operand specifies, for each element of
3971 the result vector, which element of the two input vectors the result element
3972 gets. The element selector may be undef (meaning "don't care") and the
3973 second operand may be undef if performing a shuffle from only one vector.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3974
3975<h5>Example:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3976<pre>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3977 &lt;result&gt; = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3978 &lt;4 x i32&gt; &lt;i32 0, i32 4, i32 1, i32 5&gt; <i>; yields &lt;4 x i32&gt;</i>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3979 &lt;result&gt; = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; undef,
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3980 &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 Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3981 &lt;result&gt; = shufflevector &lt;8 x i32&gt; %v1, &lt;8 x i32&gt; undef,
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3982 &lt;4 x i32&gt; &lt;i32 0, i32 1, i32 2, i32 3&gt; <i>; yields &lt;4 x i32&gt;</i>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3983 &lt;result&gt; = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3984 &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>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3985</pre>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3986
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3987</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3988
3989<!-- ======================================================================= -->
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3990<div class="doc_subsection">
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3991 <a name="aggregateops">Aggregate Operations</a>
3992</div>
3993
3994<div class="doc_text">
3995
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame]3996<p>LLVM supports several instructions for working with
3997 <a href="#t_aggregate">aggregate</a> values.</p>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3998
3999</div>
4000
4001<!-- _______________________________________________________________________ -->
4002<div class="doc_subsubsection">
4003 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
4004</div>
4005
4006<div class="doc_text">
4007
4008<h5>Syntax:</h5>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]4009<pre>
4010 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
4011</pre>
4012
4013<h5>Overview:</h5>
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame]4014<p>The '<tt>extractvalue</tt>' instruction extracts the value of a member field
4015 from an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]4016
4017<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4018<p>The first operand of an '<tt>extractvalue</tt>' instruction is a value
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame]4019 of <a href="#t_struct">struct</a>, <a href="#t_union">union</a> or
4020 <a href="#t_array">array</a> type. The operands are constant indices to
4021 specify which value to extract in a similar manner as indices in a
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4022 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]4023
4024<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4025<p>The result is the value at the position in the aggregate specified by the
4026 index operands.</p>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]4027
4028<h5>Example:</h5>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]4029<pre>
Gabor Greifb2c73142009-10-28 13:14:50 +0000[diff] [blame]4030 &lt;result&gt; = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]4031</pre>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]4032
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4033</div>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]4034
4035<!-- _______________________________________________________________________ -->
4036<div class="doc_subsubsection">
4037 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
4038</div>
4039
4040<div class="doc_text">
4041
4042<h5>Syntax:</h5>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]4043<pre>
Jeffrey Yasskine05d7732010-01-11 19:19:26 +0000[diff] [blame]4044 &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 Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]4045</pre>
4046
4047<h5>Overview:</h5>
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame]4048<p>The '<tt>insertvalue</tt>' instruction inserts a value into a member field
4049 in an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]4050
4051<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4052<p>The first operand of an '<tt>insertvalue</tt>' instruction is a value
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame]4053 of <a href="#t_struct">struct</a>, <a href="#t_union">union</a> or
4054 <a href="#t_array">array</a> type. The second operand is a first-class
4055 value to insert. The following operands are constant indices indicating
4056 the position at which to insert the value in a similar manner as indices in a
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4057 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction. The
4058 value to insert must have the same type as the value identified by the
4059 indices.</p>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]4060
4061<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4062<p>The result is an aggregate of the same type as <tt>val</tt>. Its value is
4063 that of <tt>val</tt> except that the value at the position specified by the
4064 indices is that of <tt>elt</tt>.</p>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]4065
4066<h5>Example:</h5>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]4067<pre>
Jeffrey Yasskine05d7732010-01-11 19:19:26 +0000[diff] [blame]4068 %agg1 = insertvalue {i32, float} undef, i32 1, 0 <i>; yields {i32 1, float undef}</i>
4069 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 <i>; yields {i32 1, float %val}</i>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]4070</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4071
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]4072</div>
4073
4074
4075<!-- ======================================================================= -->
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4076<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4077 <a name="memoryops">Memory Access and Addressing Operations</a>
4078</div>
4079
4080<div class="doc_text">
4081
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4082<p>A key design point of an SSA-based representation is how it represents
4083 memory. In LLVM, no memory locations are in SSA form, which makes things
Victor Hernandez6b054092009-10-26 23:44:29 +0000[diff] [blame]4084 very simple. This section describes how to read, write, and allocate
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4085 memory in LLVM.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4086
4087</div>
4088
4089<!-- _______________________________________________________________________ -->
4090<div class="doc_subsubsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4091 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
4092</div>
4093
4094<div class="doc_text">
4095
4096<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4097<pre>
4098 &lt;result&gt; = alloca &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
4099</pre>
4100
4101<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4102<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4103 currently executing function, to be automatically released when this function
4104 returns to its caller. The object is always allocated in the generic address
4105 space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4106
4107<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4108<p>The '<tt>alloca</tt>' instruction
4109 allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt> bytes of memory on the
4110 runtime stack, returning a pointer of the appropriate type to the program.
4111 If "NumElements" is specified, it is the number of elements allocated,
4112 otherwise "NumElements" is defaulted to be one. If a constant alignment is
4113 specified, the value result of the allocation is guaranteed to be aligned to
4114 at least that boundary. If not specified, or if zero, the target can choose
4115 to align the allocation on any convenient boundary compatible with the
4116 type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4117
4118<p>'<tt>type</tt>' may be any sized type.</p>
4119
4120<h5>Semantics:</h5>
Bill Wendling2a454572009-05-08 20:49:29 +0000[diff] [blame]4121<p>Memory is allocated; a pointer is returned. The operation is undefined if
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4122 there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
4123 memory is automatically released when the function returns. The
4124 '<tt>alloca</tt>' instruction is commonly used to represent automatic
4125 variables that must have an address available. When the function returns
4126 (either with the <tt><a href="#i_ret">ret</a></tt>
4127 or <tt><a href="#i_unwind">unwind</a></tt> instructions), the memory is
4128 reclaimed. Allocating zero bytes is legal, but the result is undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4129
4130<h5>Example:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4131<pre>
Dan Gohmanf54f50a2009-01-04 23:49:44 +0000[diff] [blame]4132 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
4133 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
4134 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
4135 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4136</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4137
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4138</div>
4139
4140<!-- _______________________________________________________________________ -->
4141<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
4142Instruction</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4143
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4144<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4145
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4146<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4147<pre>
Bill Wendling4197e452010-02-25 21:23:24 +0000[diff] [blame]4148 &lt;result&gt; = load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;]
4149 &lt;result&gt; = volatile load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;]
4150 !&lt;index&gt; = !{ i32 1 }
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4151</pre>
4152
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4153<h5>Overview:</h5>
4154<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4155
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4156<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4157<p>The argument to the '<tt>load</tt>' instruction specifies the memory address
4158 from which to load. The pointer must point to
4159 a <a href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
4160 marked as <tt>volatile</tt>, then the optimizer is not allowed to modify the
4161 number or order of execution of this <tt>load</tt> with other
4162 volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
David Greene02dfe202010-02-16 20:50:18 +0000[diff] [blame]4163 instructions.</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4164
Bill Wendling4197e452010-02-25 21:23:24 +0000[diff] [blame]4165<p>The optional constant <tt>align</tt> argument specifies the alignment of the
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4166 operation (that is, the alignment of the memory address). A value of 0 or an
Bill Wendling4197e452010-02-25 21:23:24 +0000[diff] [blame]4167 omitted <tt>align</tt> argument means that the operation has the preferential
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4168 alignment for the target. It is the responsibility of the code emitter to
4169 ensure that the alignment information is correct. Overestimating the
Bill Wendling4197e452010-02-25 21:23:24 +0000[diff] [blame]4170 alignment results in undefined behavior. Underestimating the alignment may
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4171 produce less efficient code. An alignment of 1 is always safe.</p>
4172
Bill Wendling4197e452010-02-25 21:23:24 +0000[diff] [blame]4173<p>The optional <tt>!nontemporal</tt> metadata must reference a single
4174 metatadata name &lt;index&gt; corresponding to a metadata node with
Dan Gohman22dc6682010-03-01 17:41:39 +0000[diff] [blame]4175 one <tt>i32</tt> entry of value 1. The existence of
Bill Wendling4197e452010-02-25 21:23:24 +0000[diff] [blame]4176 the <tt>!nontemporal</tt> metatadata on the instruction tells the optimizer
4177 and code generator that this load is not expected to be reused in the cache.
4178 The code generator may select special instructions to save cache bandwidth,
Dan Gohman22dc6682010-03-01 17:41:39 +0000[diff] [blame]4179 such as the <tt>MOVNT</tt> instruction on x86.</p>
David Greene02dfe202010-02-16 20:50:18 +0000[diff] [blame]4180
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4181<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4182<p>The location of memory pointed to is loaded. If the value being loaded is of
4183 scalar type then the number of bytes read does not exceed the minimum number
4184 of bytes needed to hold all bits of the type. For example, loading an
4185 <tt>i24</tt> reads at most three bytes. When loading a value of a type like
4186 <tt>i20</tt> with a size that is not an integral number of bytes, the result
4187 is undefined if the value was not originally written using a store of the
4188 same type.</p>
4189
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4190<h5>Examples:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4191<pre>
4192 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
4193 <a href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4194 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
4195</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4196
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4197</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4198
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4199<!-- _______________________________________________________________________ -->
4200<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
4201Instruction</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4202
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4203<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4204
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4205<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4206<pre>
David Greene02dfe202010-02-16 20:50:18 +0000[diff] [blame]4207 store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !<index>] <i>; yields {void}</i>
4208 volatile store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !<index>] <i>; yields {void}</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4209</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4210
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4211<h5>Overview:</h5>
4212<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4213
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4214<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4215<p>There are two arguments to the '<tt>store</tt>' instruction: a value to store
4216 and an address at which to store it. The type of the
4217 '<tt>&lt;pointer&gt;</tt>' operand must be a pointer to
4218 the <a href="#t_firstclass">first class</a> type of the
4219 '<tt>&lt;value&gt;</tt>' operand. If the <tt>store</tt> is marked
4220 as <tt>volatile</tt>, then the optimizer is not allowed to modify the number
4221 or order of execution of this <tt>store</tt> with other
4222 volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
4223 instructions.</p>
4224
4225<p>The optional constant "align" argument specifies the alignment of the
4226 operation (that is, the alignment of the memory address). A value of 0 or an
4227 omitted "align" argument means that the operation has the preferential
4228 alignment for the target. It is the responsibility of the code emitter to
4229 ensure that the alignment information is correct. Overestimating the
4230 alignment results in an undefined behavior. Underestimating the alignment may
4231 produce less efficient code. An alignment of 1 is always safe.</p>
4232
David Greene02dfe202010-02-16 20:50:18 +0000[diff] [blame]4233<p>The optional !nontemporal metadata must reference a single metatadata
4234 name <index> corresponding to a metadata node with one i32 entry of
Dan Gohman22dc6682010-03-01 17:41:39 +0000[diff] [blame]4235 value 1. The existence of the !nontemporal metatadata on the
David Greene02dfe202010-02-16 20:50:18 +0000[diff] [blame]4236 instruction tells the optimizer and code generator that this load is
4237 not expected to be reused in the cache. The code generator may
4238 select special instructions to save cache bandwidth, such as the
Dan Gohman22dc6682010-03-01 17:41:39 +0000[diff] [blame]4239 MOVNT instruction on x86.</p>
David Greene02dfe202010-02-16 20:50:18 +0000[diff] [blame]4240
4241
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4242<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4243<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>' at the
4244 location specified by the '<tt>&lt;pointer&gt;</tt>' operand. If
4245 '<tt>&lt;value&gt;</tt>' is of scalar type then the number of bytes written
4246 does not exceed the minimum number of bytes needed to hold all bits of the
4247 type. For example, storing an <tt>i24</tt> writes at most three bytes. When
4248 writing a value of a type like <tt>i20</tt> with a size that is not an
4249 integral number of bytes, it is unspecified what happens to the extra bits
4250 that do not belong to the type, but they will typically be overwritten.</p>
4251
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4252<h5>Example:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4253<pre>
4254 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling63ffa142007-10-22 05:10:05 +0000[diff] [blame]4255 store i32 3, i32* %ptr <i>; yields {void}</i>
4256 %val = <a href="#i_load">load</a> i32* %ptr <i>; yields {i32}:val = i32 3</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4257</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4258
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4259</div>
4260
4261<!-- _______________________________________________________________________ -->
4262<div class="doc_subsubsection">
4263 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
4264</div>
4265
4266<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4267
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4268<h5>Syntax:</h5>
4269<pre>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]4270 &lt;result&gt; = getelementptr &lt;pty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
Dan Gohman106b2ae2009-07-27 21:53:46 +0000[diff] [blame]4271 &lt;result&gt; = getelementptr inbounds &lt;pty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4272</pre>
4273
4274<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4275<p>The '<tt>getelementptr</tt>' instruction is used to get the address of a
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame]4276 subelement of an <a href="#t_aggregate">aggregate</a> data structure.
4277 It performs address calculation only and does not access memory.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4278
4279<h5>Arguments:</h5>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]4280<p>The first argument is always a pointer, and forms the basis of the
Chris Lattnere92fc832009-07-29 06:44:13 +0000[diff] [blame]4281 calculation. The remaining arguments are indices that indicate which of the
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4282 elements of the aggregate object are indexed. The interpretation of each
4283 index is dependent on the type being indexed into. The first index always
4284 indexes the pointer value given as the first argument, the second index
4285 indexes a value of the type pointed to (not necessarily the value directly
4286 pointed to, since the first index can be non-zero), etc. The first type
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame]4287 indexed into must be a pointer value, subsequent types can be arrays,
4288 vectors, structs and unions. Note that subsequent types being indexed into
4289 can never be pointers, since that would require loading the pointer before
4290 continuing calculation.</p>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]4291
4292<p>The type of each index argument depends on the type it is indexing into.
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame]4293 When indexing into a (optionally packed) structure or union, only <tt>i32</tt>
4294 integer <b>constants</b> are allowed. When indexing into an array, pointer
4295 or vector, integers of any width are allowed, and they are not required to be
Chris Lattnere92fc832009-07-29 06:44:13 +0000[diff] [blame]4296 constant.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4297
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4298<p>For example, let's consider a C code fragment and how it gets compiled to
4299 LLVM:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4300
4301<div class="doc_code">
4302<pre>
4303struct RT {
4304 char A;
4305 int B[10][20];
4306 char C;
4307};
4308struct ST {
4309 int X;
4310 double Y;
4311 struct RT Z;
4312};
4313
4314int *foo(struct ST *s) {
4315 return &amp;s[1].Z.B[5][13];
4316}
4317</pre>
4318</div>
4319
4320<p>The LLVM code generated by the GCC frontend is:</p>
4321
4322<div class="doc_code">
4323<pre>
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]4324%RT = <a href="#namedtypes">type</a> { i8 , [10 x [20 x i32]], i8 }
4325%ST = <a href="#namedtypes">type</a> { i32, double, %RT }
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4326
Dan Gohman47360842009-07-25 02:23:48 +0000[diff] [blame]4327define i32* @foo(%ST* %s) {
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4328entry:
4329 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
4330 ret i32* %reg
4331}
4332</pre>
4333</div>
4334
4335<h5>Semantics:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4336<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4337 type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
4338 }</tt>' type, a structure. The second index indexes into the third element
4339 of the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
4340 i8 }</tt>' type, another structure. The third index indexes into the second
4341 element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
4342 array. The two dimensions of the array are subscripted into, yielding an
4343 '<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a
4344 pointer to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4345
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4346<p>Note that it is perfectly legal to index partially through a structure,
4347 returning a pointer to an inner element. Because of this, the LLVM code for
4348 the given testcase is equivalent to:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4349
4350<pre>
Dan Gohman47360842009-07-25 02:23:48 +0000[diff] [blame]4351 define i32* @foo(%ST* %s) {
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4352 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
4353 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
4354 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
4355 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
4356 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
4357 ret i32* %t5
4358 }
4359</pre>
4360
Dan Gohman106b2ae2009-07-27 21:53:46 +0000[diff] [blame]4361<p>If the <tt>inbounds</tt> keyword is present, the result value of the
Dan Gohman4cba1562009-07-29 16:00:30 +0000[diff] [blame]4362 <tt>getelementptr</tt> is undefined if the base pointer is not an
4363 <i>in bounds</i> address of an allocated object, or if any of the addresses
Dan Gohmanbc4c87f2009-08-20 17:08:17 +0000[diff] [blame]4364 that would be formed by successive addition of the offsets implied by the
4365 indices to the base address with infinitely precise arithmetic are not an
4366 <i>in bounds</i> address of that allocated object.
Dan Gohman4cba1562009-07-29 16:00:30 +0000[diff] [blame]4367 The <i>in bounds</i> addresses for an allocated object are all the addresses
Dan Gohmanbc4c87f2009-08-20 17:08:17 +0000[diff] [blame]4368 that point into the object, plus the address one byte past the end.</p>
Dan Gohman106b2ae2009-07-27 21:53:46 +0000[diff] [blame]4369
4370<p>If the <tt>inbounds</tt> keyword is not present, the offsets are added to
4371 the base address with silently-wrapping two's complement arithmetic, and
4372 the result value of the <tt>getelementptr</tt> may be outside the object
4373 pointed to by the base pointer. The result value may not necessarily be
4374 used to access memory though, even if it happens to point into allocated
4375 storage. See the <a href="#pointeraliasing">Pointer Aliasing Rules</a>
4376 section for more information.</p>
4377
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4378<p>The getelementptr instruction is often confusing. For some more insight into
4379 how it works, see <a href="GetElementPtr.html">the getelementptr FAQ</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4380
4381<h5>Example:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4382<pre>
4383 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]4384 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
4385 <i>; yields i8*:vptr</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4386 %vptr = getelementptr {i32, &lt;2 x i8&gt;}* %svptr, i64 0, i32 1, i32 1
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]4387 <i>; yields i8*:eptr</i>
4388 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Sanjiv Gupta4f9a0dc2009-04-25 07:27:44 +0000[diff] [blame]4389 <i>; yields i32*:iptr</i>
Sanjiv Gupta1e46c582009-04-24 16:38:13 +0000[diff] [blame]4390 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4391</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4392
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4393</div>
4394
4395<!-- ======================================================================= -->
4396<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
4397</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4398
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4399<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4400
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4401<p>The instructions in this category are the conversion instructions (casting)
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4402 which all take a single operand and a type. They perform various bit
4403 conversions on the operand.</p>
4404
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4405</div>
4406
4407<!-- _______________________________________________________________________ -->
4408<div class="doc_subsubsection">
4409 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
4410</div>
4411<div class="doc_text">
4412
4413<h5>Syntax:</h5>
4414<pre>
4415 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4416</pre>
4417
4418<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4419<p>The '<tt>trunc</tt>' instruction truncates its operand to the
4420 type <tt>ty2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4421
4422<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4423<p>The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
4424 be an <a href="#t_integer">integer</a> type, and a type that specifies the
4425 size and type of the result, which must be
4426 an <a href="#t_integer">integer</a> type. The bit size of <tt>value</tt> must
4427 be larger than the bit size of <tt>ty2</tt>. Equal sized types are not
4428 allowed.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4429
4430<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4431<p>The '<tt>trunc</tt>' instruction truncates the high order bits
4432 in <tt>value</tt> and converts the remaining bits to <tt>ty2</tt>. Since the
4433 source size must be larger than the destination size, <tt>trunc</tt> cannot
4434 be a <i>no-op cast</i>. It will always truncate bits.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4435
4436<h5>Example:</h5>
4437<pre>
4438 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
4439 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
Gabor Greif463c9342009-10-28 09:21:30 +0000[diff] [blame]4440 %Z = trunc i32 122 to i1 <i>; yields i1:false</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4441</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4442
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4443</div>
4444
4445<!-- _______________________________________________________________________ -->
4446<div class="doc_subsubsection">
4447 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
4448</div>
4449<div class="doc_text">
4450
4451<h5>Syntax:</h5>
4452<pre>
4453 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4454</pre>
4455
4456<h5>Overview:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4457<p>The '<tt>zext</tt>' instruction zero extends its operand to type
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4458 <tt>ty2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4459
4460
4461<h5>Arguments:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4462<p>The '<tt>zext</tt>' instruction takes a value to cast, which must be of
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4463 <a href="#t_integer">integer</a> type, and a type to cast it to, which must
4464 also be of <a href="#t_integer">integer</a> type. The bit size of the
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4465 <tt>value</tt> must be smaller than the bit size of the destination type,
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4466 <tt>ty2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4467
4468<h5>Semantics:</h5>
4469<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4470 bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4471
4472<p>When zero extending from i1, the result will always be either 0 or 1.</p>
4473
4474<h5>Example:</h5>
4475<pre>
4476 %X = zext i32 257 to i64 <i>; yields i64:257</i>
4477 %Y = zext i1 true to i32 <i>; yields i32:1</i>
4478</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4479
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4480</div>
4481
4482<!-- _______________________________________________________________________ -->
4483<div class="doc_subsubsection">
4484 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
4485</div>
4486<div class="doc_text">
4487
4488<h5>Syntax:</h5>
4489<pre>
4490 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4491</pre>
4492
4493<h5>Overview:</h5>
4494<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
4495
4496<h5>Arguments:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4497<p>The '<tt>sext</tt>' instruction takes a value to cast, which must be of
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4498 <a href="#t_integer">integer</a> type, and a type to cast it to, which must
4499 also be of <a href="#t_integer">integer</a> type. The bit size of the
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4500 <tt>value</tt> must be smaller than the bit size of the destination type,
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4501 <tt>ty2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4502
4503<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4504<p>The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
4505 bit (highest order bit) of the <tt>value</tt> until it reaches the bit size
4506 of the type <tt>ty2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4507
4508<p>When sign extending from i1, the extension always results in -1 or 0.</p>
4509
4510<h5>Example:</h5>
4511<pre>
4512 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
4513 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
4514</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4515
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4516</div>
4517
4518<!-- _______________________________________________________________________ -->
4519<div class="doc_subsubsection">
4520 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
4521</div>
4522
4523<div class="doc_text">
4524
4525<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4526<pre>
4527 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4528</pre>
4529
4530<h5>Overview:</h5>
4531<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4532 <tt>ty2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4533
4534<h5>Arguments:</h5>
4535<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4536 point</a> value to cast and a <a href="#t_floating">floating point</a> type
4537 to cast it to. The size of <tt>value</tt> must be larger than the size of
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4538 <tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4539 <i>no-op cast</i>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4540
4541<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4542<p>The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4543 <a href="#t_floating">floating point</a> type to a smaller
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4544 <a href="#t_floating">floating point</a> type. If the value cannot fit
4545 within the destination type, <tt>ty2</tt>, then the results are
4546 undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4547
4548<h5>Example:</h5>
4549<pre>
4550 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
4551 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
4552</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4553
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4554</div>
4555
4556<!-- _______________________________________________________________________ -->
4557<div class="doc_subsubsection">
4558 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
4559</div>
4560<div class="doc_text">
4561
4562<h5>Syntax:</h5>
4563<pre>
4564 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4565</pre>
4566
4567<h5>Overview:</h5>
4568<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4569 floating point value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4570
4571<h5>Arguments:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4572<p>The '<tt>fpext</tt>' instruction takes a
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4573 <a href="#t_floating">floating point</a> <tt>value</tt> to cast, and
4574 a <a href="#t_floating">floating point</a> type to cast it to. The source
4575 type must be smaller than the destination type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4576
4577<h5>Semantics:</h5>
4578<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4579 <a href="#t_floating">floating point</a> type to a larger
4580 <a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
4581 used to make a <i>no-op cast</i> because it always changes bits. Use
4582 <tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4583
4584<h5>Example:</h5>
4585<pre>
4586 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
4587 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
4588</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4589
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4590</div>
4591
4592<!-- _______________________________________________________________________ -->
4593<div class="doc_subsubsection">
4594 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
4595</div>
4596<div class="doc_text">
4597
4598<h5>Syntax:</h5>
4599<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]4600 &lt;result&gt; = fptoui &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4601</pre>
4602
4603<h5>Overview:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]4604<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4605 unsigned integer equivalent of type <tt>ty2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4606
4607<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4608<p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
4609 scalar or vector <a href="#t_floating">floating point</a> value, and a type
4610 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4611 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4612 vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4613
4614<h5>Semantics:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4615<p>The '<tt>fptoui</tt>' instruction converts its
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4616 <a href="#t_floating">floating point</a> operand into the nearest (rounding
4617 towards zero) unsigned integer value. If the value cannot fit
4618 in <tt>ty2</tt>, the results are undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4619
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4620<h5>Example:</h5>
4621<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]4622 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]4623 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Gabor Greif463c9342009-10-28 09:21:30 +0000[diff] [blame]4624 %Z = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4625</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4626
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4627</div>
4628
4629<!-- _______________________________________________________________________ -->
4630<div class="doc_subsubsection">
4631 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
4632</div>
4633<div class="doc_text">
4634
4635<h5>Syntax:</h5>
4636<pre>
4637 &lt;result&gt; = fptosi &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4638</pre>
4639
4640<h5>Overview:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4641<p>The '<tt>fptosi</tt>' instruction converts
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4642 <a href="#t_floating">floating point</a> <tt>value</tt> to
4643 type <tt>ty2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4644
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4645<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4646<p>The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
4647 scalar or vector <a href="#t_floating">floating point</a> value, and a type
4648 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4649 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4650 vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4651
4652<h5>Semantics:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4653<p>The '<tt>fptosi</tt>' instruction converts its
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4654 <a href="#t_floating">floating point</a> operand into the nearest (rounding
4655 towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
4656 the results are undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4657
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4658<h5>Example:</h5>
4659<pre>
4660 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]4661 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Gabor Greif463c9342009-10-28 09:21:30 +0000[diff] [blame]4662 %Z = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4663</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4664
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4665</div>
4666
4667<!-- _______________________________________________________________________ -->
4668<div class="doc_subsubsection">
4669 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
4670</div>
4671<div class="doc_text">
4672
4673<h5>Syntax:</h5>
4674<pre>
4675 &lt;result&gt; = uitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4676</pre>
4677
4678<h5>Overview:</h5>
4679<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4680 integer and converts that value to the <tt>ty2</tt> type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4681
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4682<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]4683<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be a
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4684 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
4685 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4686 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4687 floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4688
4689<h5>Semantics:</h5>
4690<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4691 integer quantity and converts it to the corresponding floating point
4692 value. If the value cannot fit in the floating point value, the results are
4693 undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4694
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4695<h5>Example:</h5>
4696<pre>
4697 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4698 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4699</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4700
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4701</div>
4702
4703<!-- _______________________________________________________________________ -->
4704<div class="doc_subsubsection">
4705 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
4706</div>
4707<div class="doc_text">
4708
4709<h5>Syntax:</h5>
4710<pre>
4711 &lt;result&gt; = sitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4712</pre>
4713
4714<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4715<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed integer
4716 and converts that value to the <tt>ty2</tt> type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4717
4718<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]4719<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be a
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4720 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
4721 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4722 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4723 floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4724
4725<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4726<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed integer
4727 quantity and converts it to the corresponding floating point value. If the
4728 value cannot fit in the floating point value, the results are undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4729
4730<h5>Example:</h5>
4731<pre>
4732 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4733 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4734</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4735
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4736</div>
4737
4738<!-- _______________________________________________________________________ -->
4739<div class="doc_subsubsection">
4740 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
4741</div>
4742<div class="doc_text">
4743
4744<h5>Syntax:</h5>
4745<pre>
4746 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4747</pre>
4748
4749<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4750<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
4751 the integer type <tt>ty2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4752
4753<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4754<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
4755 must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
4756 <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4757
4758<h5>Semantics:</h5>
4759<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4760 <tt>ty2</tt> by interpreting the pointer value as an integer and either
4761 truncating or zero extending that value to the size of the integer type. If
4762 <tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
4763 <tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
4764 are the same size, then nothing is done (<i>no-op cast</i>) other than a type
4765 change.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4766
4767<h5>Example:</h5>
4768<pre>
4769 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
4770 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
4771</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4772
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4773</div>
4774
4775<!-- _______________________________________________________________________ -->
4776<div class="doc_subsubsection">
4777 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
4778</div>
4779<div class="doc_text">
4780
4781<h5>Syntax:</h5>
4782<pre>
4783 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4784</pre>
4785
4786<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4787<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to a
4788 pointer type, <tt>ty2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4789
4790<h5>Arguments:</h5>
4791<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4792 value to cast, and a type to cast it to, which must be a
4793 <a href="#t_pointer">pointer</a> type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4794
4795<h5>Semantics:</h5>
4796<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4797 <tt>ty2</tt> by applying either a zero extension or a truncation depending on
4798 the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
4799 size of a pointer then a truncation is done. If <tt>value</tt> is smaller
4800 than the size of a pointer then a zero extension is done. If they are the
4801 same size, nothing is done (<i>no-op cast</i>).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4802
4803<h5>Example:</h5>
4804<pre>
4805 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
Gabor Greif463c9342009-10-28 09:21:30 +0000[diff] [blame]4806 %Y = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
4807 %Z = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4808</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4809
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4810</div>
4811
4812<!-- _______________________________________________________________________ -->
4813<div class="doc_subsubsection">
4814 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
4815</div>
4816<div class="doc_text">
4817
4818<h5>Syntax:</h5>
4819<pre>
4820 &lt;result&gt; = bitcast &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4821</pre>
4822
4823<h5>Overview:</h5>
4824<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4825 <tt>ty2</tt> without changing any bits.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4826
4827<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4828<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be a
4829 non-aggregate first class value, and a type to cast it to, which must also be
4830 a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes
4831 of <tt>value</tt> and the destination type, <tt>ty2</tt>, must be
4832 identical. If the source type is a pointer, the destination type must also be
4833 a pointer. This instruction supports bitwise conversion of vectors to
4834 integers and to vectors of other types (as long as they have the same
4835 size).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4836
4837<h5>Semantics:</h5>
4838<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4839 <tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
4840 this conversion. The conversion is done as if the <tt>value</tt> had been
4841 stored to memory and read back as type <tt>ty2</tt>. Pointer types may only
4842 be converted to other pointer types with this instruction. To convert
4843 pointers to other types, use the <a href="#i_inttoptr">inttoptr</a> or
4844 <a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4845
4846<h5>Example:</h5>
4847<pre>
4848 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
4849 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4850 %Z = bitcast &lt;2 x int&gt; %V to i64; <i>; yields i64: %V</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4851</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4852
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4853</div>
4854
4855<!-- ======================================================================= -->
4856<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4857
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4858<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4859
4860<p>The instructions in this category are the "miscellaneous" instructions, which
4861 defy better classification.</p>
4862
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4863</div>
4864
4865<!-- _______________________________________________________________________ -->
4866<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
4867</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4868
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4869<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4870
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4871<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4872<pre>
4873 &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>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4874</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4875
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4876<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4877<p>The '<tt>icmp</tt>' instruction returns a boolean value or a vector of
4878 boolean values based on comparison of its two integer, integer vector, or
4879 pointer operands.</p>
4880
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4881<h5>Arguments:</h5>
4882<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4883 the condition code indicating the kind of comparison to perform. It is not a
4884 value, just a keyword. The possible condition code are:</p>
4885
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4886<ol>
4887 <li><tt>eq</tt>: equal</li>
4888 <li><tt>ne</tt>: not equal </li>
4889 <li><tt>ugt</tt>: unsigned greater than</li>
4890 <li><tt>uge</tt>: unsigned greater or equal</li>
4891 <li><tt>ult</tt>: unsigned less than</li>
4892 <li><tt>ule</tt>: unsigned less or equal</li>
4893 <li><tt>sgt</tt>: signed greater than</li>
4894 <li><tt>sge</tt>: signed greater or equal</li>
4895 <li><tt>slt</tt>: signed less than</li>
4896 <li><tt>sle</tt>: signed less or equal</li>
4897</ol>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4898
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4899<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4900 <a href="#t_pointer">pointer</a> or integer <a href="#t_vector">vector</a>
4901 typed. They must also be identical types.</p>
4902
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4903<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4904<p>The '<tt>icmp</tt>' compares <tt>op1</tt> and <tt>op2</tt> according to the
4905 condition code given as <tt>cond</tt>. The comparison performed always yields
Nick Lewycky244cf482009-09-27 00:45:11 +0000[diff] [blame]4906 either an <a href="#t_integer"><tt>i1</tt></a> or vector of <tt>i1</tt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4907 result, as follows:</p>
4908
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4909<ol>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4910 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4911 <tt>false</tt> otherwise. No sign interpretation is necessary or
4912 performed.</li>
4913
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4914 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4915 <tt>false</tt> otherwise. No sign interpretation is necessary or
4916 performed.</li>
4917
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4918 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4919 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
4920
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4921 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4922 <tt>true</tt> if <tt>op1</tt> is greater than or equal
4923 to <tt>op2</tt>.</li>
4924
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4925 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4926 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
4927
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4928 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4929 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
4930
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4931 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4932 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
4933
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4934 <li><tt>sge</tt>: interprets the operands as signed values and yields
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4935 <tt>true</tt> if <tt>op1</tt> is greater than or equal
4936 to <tt>op2</tt>.</li>
4937
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4938 <li><tt>slt</tt>: interprets the operands as signed values and yields
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4939 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
4940
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4941 <li><tt>sle</tt>: interprets the operands as signed values and yields
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4942 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4943</ol>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4944
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4945<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4946 values are compared as if they were integers.</p>
4947
4948<p>If the operands are integer vectors, then they are compared element by
4949 element. The result is an <tt>i1</tt> vector with the same number of elements
4950 as the values being compared. Otherwise, the result is an <tt>i1</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4951
4952<h5>Example:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4953<pre>
4954 &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4955 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
4956 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
4957 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
4958 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
4959 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
4960</pre>
Dan Gohmana53eb382009-01-22 01:39:38 +0000[diff] [blame]4961
4962<p>Note that the code generator does not yet support vector types with
4963 the <tt>icmp</tt> instruction.</p>
4964
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4965</div>
4966
4967<!-- _______________________________________________________________________ -->
4968<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
4969</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4970
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4971<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4972
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4973<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4974<pre>
4975 &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>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4976</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4977
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4978<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4979<p>The '<tt>fcmp</tt>' instruction returns a boolean value or vector of boolean
4980 values based on comparison of its operands.</p>
4981
4982<p>If the operands are floating point scalars, then the result type is a boolean
Nick Lewycky244cf482009-09-27 00:45:11 +0000[diff] [blame]4983(<a href="#t_integer"><tt>i1</tt></a>).</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4984
4985<p>If the operands are floating point vectors, then the result type is a vector
4986 of boolean with the same number of elements as the operands being
4987 compared.</p>
4988
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4989<h5>Arguments:</h5>
4990<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4991 the condition code indicating the kind of comparison to perform. It is not a
4992 value, just a keyword. The possible condition code are:</p>
4993
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4994<ol>
4995 <li><tt>false</tt>: no comparison, always returns false</li>
4996 <li><tt>oeq</tt>: ordered and equal</li>
4997 <li><tt>ogt</tt>: ordered and greater than </li>
4998 <li><tt>oge</tt>: ordered and greater than or equal</li>
4999 <li><tt>olt</tt>: ordered and less than </li>
5000 <li><tt>ole</tt>: ordered and less than or equal</li>
5001 <li><tt>one</tt>: ordered and not equal</li>
5002 <li><tt>ord</tt>: ordered (no nans)</li>
5003 <li><tt>ueq</tt>: unordered or equal</li>
5004 <li><tt>ugt</tt>: unordered or greater than </li>
5005 <li><tt>uge</tt>: unordered or greater than or equal</li>
5006 <li><tt>ult</tt>: unordered or less than </li>
5007 <li><tt>ule</tt>: unordered or less than or equal</li>
5008 <li><tt>une</tt>: unordered or not equal</li>
5009 <li><tt>uno</tt>: unordered (either nans)</li>
5010 <li><tt>true</tt>: no comparison, always returns true</li>
5011</ol>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5012
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5013<p><i>Ordered</i> means that neither operand is a QNAN while
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5014 <i>unordered</i> means that either operand may be a QNAN.</p>
5015
5016<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be either
5017 a <a href="#t_floating">floating point</a> type or
5018 a <a href="#t_vector">vector</a> of floating point type. They must have
5019 identical types.</p>
5020
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5021<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]5022<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5023 according to the condition code given as <tt>cond</tt>. If the operands are
5024 vectors, then the vectors are compared element by element. Each comparison
Nick Lewycky244cf482009-09-27 00:45:11 +0000[diff] [blame]5025 performed always yields an <a href="#t_integer">i1</a> result, as
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5026 follows:</p>
5027
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5028<ol>
5029 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5030
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]5031 <li><tt>oeq</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5032 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5033
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5034 <li><tt>ogt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Dan Gohman22dc6682010-03-01 17:41:39 +0000[diff] [blame]5035 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5036
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]5037 <li><tt>oge</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5038 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5039
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]5040 <li><tt>olt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5041 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5042
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]5043 <li><tt>ole</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5044 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5045
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]5046 <li><tt>one</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5047 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5048
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5049 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5050
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]5051 <li><tt>ueq</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5052 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5053
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]5054 <li><tt>ugt</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5055 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5056
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]5057 <li><tt>uge</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5058 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5059
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]5060 <li><tt>ult</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5061 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5062
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]5063 <li><tt>ule</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5064 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5065
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]5066 <li><tt>une</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5067 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5068
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5069 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5070
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5071 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
5072</ol>
5073
5074<h5>Example:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5075<pre>
5076 &lt;result&gt; = fcmp oeq float 4.0, 5.0 <i>; yields: result=false</i>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]5077 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
5078 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
5079 &lt;result&gt; = fcmp ueq double 1.0, 2.0 <i>; yields: result=false</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5080</pre>
Dan Gohmana53eb382009-01-22 01:39:38 +0000[diff] [blame]5081
5082<p>Note that the code generator does not yet support vector types with
5083 the <tt>fcmp</tt> instruction.</p>
5084
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5085</div>
5086
5087<!-- _______________________________________________________________________ -->
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]5088<div class="doc_subsubsection">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]5089 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
5090</div>
5091
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5092<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]5093
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5094<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5095<pre>
5096 &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...
5097</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]5098
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5099<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5100<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in the
5101 SSA graph representing the function.</p>
5102
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5103<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5104<p>The type of the incoming values is specified with the first type field. After
5105 this, the '<tt>phi</tt>' instruction takes a list of pairs as arguments, with
5106 one pair for each predecessor basic block of the current block. Only values
5107 of <a href="#t_firstclass">first class</a> type may be used as the value
5108 arguments to the PHI node. Only labels may be used as the label
5109 arguments.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]5110
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5111<p>There must be no non-phi instructions between the start of a basic block and
5112 the PHI instructions: i.e. PHI instructions must be first in a basic
5113 block.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]5114
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5115<p>For the purposes of the SSA form, the use of each incoming value is deemed to
5116 occur on the edge from the corresponding predecessor block to the current
5117 block (but after any definition of an '<tt>invoke</tt>' instruction's return
5118 value on the same edge).</p>
Jay Foad8e2fd2c2009-06-03 10:20:10 +0000[diff] [blame]5119
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5120<h5>Semantics:</h5>
5121<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5122 specified by the pair corresponding to the predecessor basic block that
5123 executed just prior to the current block.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]5124
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5125<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]5126<pre>
5127Loop: ; Infinite loop that counts from 0 on up...
5128 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
5129 %nextindvar = add i32 %indvar, 1
5130 br label %Loop
5131</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5132
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5133</div>
5134
5135<!-- _______________________________________________________________________ -->
5136<div class="doc_subsubsection">
5137 <a name="i_select">'<tt>select</tt>' Instruction</a>
5138</div>
5139
5140<div class="doc_text">
5141
5142<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5143<pre>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]5144 &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>
5145
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5146 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5147</pre>
5148
5149<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5150<p>The '<tt>select</tt>' instruction is used to choose one value based on a
5151 condition, without branching.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5152
5153
5154<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5155<p>The '<tt>select</tt>' instruction requires an 'i1' value or a vector of 'i1'
5156 values indicating the condition, and two values of the
5157 same <a href="#t_firstclass">first class</a> type. If the val1/val2 are
5158 vectors and the condition is a scalar, then entire vectors are selected, not
5159 individual elements.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5160
5161<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5162<p>If the condition is an i1 and it evaluates to 1, the instruction returns the
5163 first value argument; otherwise, it returns the second value argument.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5164
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5165<p>If the condition is a vector of i1, then the value arguments must be vectors
5166 of the same size, and the selection is done element by element.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5167
5168<h5>Example:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5169<pre>
5170 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
5171</pre>
Dan Gohmana53eb382009-01-22 01:39:38 +0000[diff] [blame]5172
5173<p>Note that the code generator does not yet support conditions
5174 with vector type.</p>
5175
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5176</div>
5177
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5178<!-- _______________________________________________________________________ -->
5179<div class="doc_subsubsection">
5180 <a name="i_call">'<tt>call</tt>' Instruction</a>
5181</div>
5182
5183<div class="doc_text">
5184
5185<h5>Syntax:</h5>
5186<pre>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]5187 &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>]
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5188</pre>
5189
5190<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5191<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
5192
5193<h5>Arguments:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5194<p>This instruction requires several arguments:</p>
5195
5196<ol>
Jeffrey Yasskin3e1a3a02010-01-09 19:44:16 +0000[diff] [blame]5197 <li>The optional "tail" marker indicates that the callee function does not
5198 access any allocas or varargs in the caller. Note that calls may be
5199 marked "tail" even if they do not occur before
5200 a <a href="#i_ret"><tt>ret</tt></a> instruction. If the "tail" marker is
5201 present, the function call is eligible for tail call optimization,
5202 but <a href="CodeGenerator.html#tailcallopt">might not in fact be
Evan Chengcc7495c2010-03-08 21:05:02 +0000[diff] [blame]5203 optimized into a jump</a>. The code generator may optimize calls marked
5204 "tail" with either 1) automatic <a href="CodeGenerator.html#sibcallopt">
5205 sibling call optimization</a> when the caller and callee have
5206 matching signatures, or 2) forced tail call optimization when the
5207 following extra requirements are met:
Jeffrey Yasskin3e1a3a02010-01-09 19:44:16 +0000[diff] [blame]5208 <ul>
5209 <li>Caller and callee both have the calling
5210 convention <tt>fastcc</tt>.</li>
5211 <li>The call is in tail position (ret immediately follows call and ret
5212 uses value of call or is void).</li>
5213 <li>Option <tt>-tailcallopt</tt> is enabled,
Dan Gohman1be84f02010-03-02 01:08:11 +0000[diff] [blame]5214 or <code>llvm::GuaranteedTailCallOpt</code> is <code>true</code>.</li>
Jeffrey Yasskin3e1a3a02010-01-09 19:44:16 +0000[diff] [blame]5215 <li><a href="CodeGenerator.html#tailcallopt">Platform specific
5216 constraints are met.</a></li>
5217 </ul>
5218 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]5219
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5220 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
5221 convention</a> the call should use. If none is specified, the call
Jeffrey Yasskin3e1a3a02010-01-09 19:44:16 +0000[diff] [blame]5222 defaults to using C calling conventions. The calling convention of the
5223 call must match the calling convention of the target function, or else the
5224 behavior is undefined.</li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]5225
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5226 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
5227 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
5228 '<tt>inreg</tt>' attributes are valid here.</li>
5229
5230 <li>'<tt>ty</tt>': the type of the call instruction itself which is also the
5231 type of the return value. Functions that return no value are marked
5232 <tt><a href="#t_void">void</a></tt>.</li>
5233
5234 <li>'<tt>fnty</tt>': shall be the signature of the pointer to function value
5235 being invoked. The argument types must match the types implied by this
5236 signature. This type can be omitted if the function is not varargs and if
5237 the function type does not return a pointer to a function.</li>
5238
5239 <li>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
5240 be invoked. In most cases, this is a direct function invocation, but
5241 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
5242 to function value.</li>
5243
5244 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner553fb1e2010-03-02 06:36:51 +0000[diff] [blame]5245 signature argument types and parameter attributes. All arguments must be
5246 of <a href="#t_firstclass">first class</a> type. If the function
5247 signature indicates the function accepts a variable number of arguments,
5248 the extra arguments can be specified.</li>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5249
5250 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
5251 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
5252 '<tt>readnone</tt>' attributes are valid here.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5253</ol>
5254
5255<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5256<p>The '<tt>call</tt>' instruction is used to cause control flow to transfer to
5257 a specified function, with its incoming arguments bound to the specified
5258 values. Upon a '<tt><a href="#i_ret">ret</a></tt>' instruction in the called
5259 function, control flow continues with the instruction after the function
5260 call, and the return value of the function is bound to the result
5261 argument.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5262
5263<h5>Example:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5264<pre>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]5265 %retval = call i32 @test(i32 %argc)
Chris Lattner5e893ef2008-03-21 17:24:17 +0000[diff] [blame]5266 call i32 (i8 *, ...)* @printf(i8 * %msg, i32 12, i8 42) <i>; yields i32</i>
5267 %X = tail call i32 @foo() <i>; yields i32</i>
5268 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
5269 call void %foo(i8 97 signext)
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]5270
5271 %struct.A = type { i32, i8 }
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]5272 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]5273 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
5274 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattnerac454b32008-10-08 06:26:11 +0000[diff] [blame]5275 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijman2c4e05a2008-10-07 10:03:45 +0000[diff] [blame]5276 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5277</pre>
5278
Dale Johannesen9fcf4a92009-09-24 18:38:21 +0000[diff] [blame]5279<p>llvm treats calls to some functions with names and arguments that match the
Dale Johannesen4730cd32009-09-25 17:04:42 +0000[diff] [blame]5280standard C99 library as being the C99 library functions, and may perform
5281optimizations or generate code for them under that assumption. This is
5282something we'd like to change in the future to provide better support for
Dan Gohman22dc6682010-03-01 17:41:39 +0000[diff] [blame]5283freestanding environments and non-C-based languages.</p>
Dale Johannesen9fcf4a92009-09-24 18:38:21 +0000[diff] [blame]5284
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5285</div>
5286
5287<!-- _______________________________________________________________________ -->
5288<div class="doc_subsubsection">
5289 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
5290</div>
5291
5292<div class="doc_text">
5293
5294<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5295<pre>
5296 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
5297</pre>
5298
5299<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5300<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5301 the "variable argument" area of a function call. It is used to implement the
5302 <tt>va_arg</tt> macro in C.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5303
5304<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5305<p>This instruction takes a <tt>va_list*</tt> value and the type of the
5306 argument. It returns a value of the specified argument type and increments
5307 the <tt>va_list</tt> to point to the next argument. The actual type
5308 of <tt>va_list</tt> is target specific.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5309
5310<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5311<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified type
5312 from the specified <tt>va_list</tt> and causes the <tt>va_list</tt> to point
5313 to the next argument. For more information, see the variable argument
5314 handling <a href="#int_varargs">Intrinsic Functions</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5315
5316<p>It is legal for this instruction to be called in a function which does not
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5317 take a variable number of arguments, for example, the <tt>vfprintf</tt>
5318 function.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5319
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5320<p><tt>va_arg</tt> is an LLVM instruction instead of
5321 an <a href="#intrinsics">intrinsic function</a> because it takes a type as an
5322 argument.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5323
5324<h5>Example:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5325<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
5326
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5327<p>Note that the code generator does not yet fully support va_arg on many
5328 targets. Also, it does not currently support va_arg with aggregate types on
5329 any target.</p>
Dan Gohman60967192009-01-12 23:12:39 +0000[diff] [blame]5330
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5331</div>
5332
5333<!-- *********************************************************************** -->
5334<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
5335<!-- *********************************************************************** -->
5336
5337<div class="doc_text">
5338
5339<p>LLVM supports the notion of an "intrinsic function". These functions have
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5340 well known names and semantics and are required to follow certain
5341 restrictions. Overall, these intrinsics represent an extension mechanism for
5342 the LLVM language that does not require changing all of the transformations
5343 in LLVM when adding to the language (or the bitcode reader/writer, the
5344 parser, etc...).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5345
5346<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5347 prefix is reserved in LLVM for intrinsic names; thus, function names may not
5348 begin with this prefix. Intrinsic functions must always be external
5349 functions: you cannot define the body of intrinsic functions. Intrinsic
5350 functions may only be used in call or invoke instructions: it is illegal to
5351 take the address of an intrinsic function. Additionally, because intrinsic
5352 functions are part of the LLVM language, it is required if any are added that
5353 they be documented here.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5354
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5355<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents a
5356 family of functions that perform the same operation but on different data
5357 types. Because LLVM can represent over 8 million different integer types,
5358 overloading is used commonly to allow an intrinsic function to operate on any
5359 integer type. One or more of the argument types or the result type can be
5360 overloaded to accept any integer type. Argument types may also be defined as
5361 exactly matching a previous argument's type or the result type. This allows
5362 an intrinsic function which accepts multiple arguments, but needs all of them
5363 to be of the same type, to only be overloaded with respect to a single
5364 argument or the result.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5365
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5366<p>Overloaded intrinsics will have the names of its overloaded argument types
5367 encoded into its function name, each preceded by a period. Only those types
5368 which are overloaded result in a name suffix. Arguments whose type is matched
5369 against another type do not. For example, the <tt>llvm.ctpop</tt> function
5370 can take an integer of any width and returns an integer of exactly the same
5371 integer width. This leads to a family of functions such as
5372 <tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29
5373 %val)</tt>. Only one type, the return type, is overloaded, and only one type
5374 suffix is required. Because the argument's type is matched against the return
5375 type, it does not require its own name suffix.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5376
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]5377<p>To learn how to add an intrinsic function, please see the
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5378 <a href="ExtendingLLVM.html">Extending LLVM Guide</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5379
5380</div>
5381
5382<!-- ======================================================================= -->
5383<div class="doc_subsection">
5384 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
5385</div>
5386
5387<div class="doc_text">
5388
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5389<p>Variable argument support is defined in LLVM with
5390 the <a href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
5391 intrinsic functions. These functions are related to the similarly named
5392 macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5393
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5394<p>All of these functions operate on arguments that use a target-specific value
5395 type "<tt>va_list</tt>". The LLVM assembly language reference manual does
5396 not define what this type is, so all transformations should be prepared to
5397 handle these functions regardless of the type used.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5398
5399<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5400 instruction and the variable argument handling intrinsic functions are
5401 used.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5402
5403<div class="doc_code">
5404<pre>
5405define i32 @test(i32 %X, ...) {
5406 ; Initialize variable argument processing
5407 %ap = alloca i8*
5408 %ap2 = bitcast i8** %ap to i8*
5409 call void @llvm.va_start(i8* %ap2)
5410
5411 ; Read a single integer argument
5412 %tmp = va_arg i8** %ap, i32
5413
5414 ; Demonstrate usage of llvm.va_copy and llvm.va_end
5415 %aq = alloca i8*
5416 %aq2 = bitcast i8** %aq to i8*
5417 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
5418 call void @llvm.va_end(i8* %aq2)
5419
5420 ; Stop processing of arguments.
5421 call void @llvm.va_end(i8* %ap2)
5422 ret i32 %tmp
5423}
5424
5425declare void @llvm.va_start(i8*)
5426declare void @llvm.va_copy(i8*, i8*)
5427declare void @llvm.va_end(i8*)
5428</pre>
5429</div>
5430
5431</div>
5432
5433<!-- _______________________________________________________________________ -->
5434<div class="doc_subsubsection">
5435 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
5436</div>
5437
5438
5439<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5440
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5441<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5442<pre>
5443 declare void %llvm.va_start(i8* &lt;arglist&gt;)
5444</pre>
5445
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5446<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5447<p>The '<tt>llvm.va_start</tt>' intrinsic initializes <tt>*&lt;arglist&gt;</tt>
5448 for subsequent use by <tt><a href="#i_va_arg">va_arg</a></tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5449
5450<h5>Arguments:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5451<p>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5452
5453<h5>Semantics:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5454<p>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5455 macro available in C. In a target-dependent way, it initializes
5456 the <tt>va_list</tt> element to which the argument points, so that the next
5457 call to <tt>va_arg</tt> will produce the first variable argument passed to
5458 the function. Unlike the C <tt>va_start</tt> macro, this intrinsic does not
5459 need to know the last argument of the function as the compiler can figure
5460 that out.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5461
5462</div>
5463
5464<!-- _______________________________________________________________________ -->
5465<div class="doc_subsubsection">
5466 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
5467</div>
5468
5469<div class="doc_text">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5470
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5471<h5>Syntax:</h5>
5472<pre>
5473 declare void @llvm.va_end(i8* &lt;arglist&gt;)
5474</pre>
5475
5476<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5477<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5478 which has been initialized previously
5479 with <tt><a href="#int_va_start">llvm.va_start</a></tt>
5480 or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5481
5482<h5>Arguments:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5483<p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
5484
5485<h5>Semantics:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5486<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5487 macro available in C. In a target-dependent way, it destroys
5488 the <tt>va_list</tt> element to which the argument points. Calls
5489 to <a href="#int_va_start"><tt>llvm.va_start</tt></a>
5490 and <a href="#int_va_copy"> <tt>llvm.va_copy</tt></a> must be matched exactly
5491 with calls to <tt>llvm.va_end</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5492
5493</div>
5494
5495<!-- _______________________________________________________________________ -->
5496<div class="doc_subsubsection">
5497 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
5498</div>
5499
5500<div class="doc_text">
5501
5502<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5503<pre>
5504 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
5505</pre>
5506
5507<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5508<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5509 from the source argument list to the destination argument list.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5510
5511<h5>Arguments:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5512<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5513 The second argument is a pointer to a <tt>va_list</tt> element to copy
5514 from.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5515
5516<h5>Semantics:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5517<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5518 macro available in C. In a target-dependent way, it copies the
5519 source <tt>va_list</tt> element into the destination <tt>va_list</tt>
5520 element. This intrinsic is necessary because
5521 the <tt><a href="#int_va_start"> llvm.va_start</a></tt> intrinsic may be
5522 arbitrarily complex and require, for example, memory allocation.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5523
5524</div>
5525
5526<!-- ======================================================================= -->
5527<div class="doc_subsection">
5528 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
5529</div>
5530
5531<div class="doc_text">
5532
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5533<p>LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]5534Collection</a> (GC) requires the implementation and generation of these
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5535intrinsics. These intrinsics allow identification of <a href="#int_gcroot">GC
5536roots on the stack</a>, as well as garbage collector implementations that
5537require <a href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a>
5538barriers. Front-ends for type-safe garbage collected languages should generate
5539these intrinsics to make use of the LLVM garbage collectors. For more details,
5540see <a href="GarbageCollection.html">Accurate Garbage Collection with
5541LLVM</a>.</p>
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]5542
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5543<p>The garbage collection intrinsics only operate on objects in the generic
5544 address space (address space zero).</p>
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]5545
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5546</div>
5547
5548<!-- _______________________________________________________________________ -->
5549<div class="doc_subsubsection">
5550 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
5551</div>
5552
5553<div class="doc_text">
5554
5555<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5556<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5557 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5558</pre>
5559
5560<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5561<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5562 the code generator, and allows some metadata to be associated with it.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5563
5564<h5>Arguments:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5565<p>The first argument specifies the address of a stack object that contains the
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5566 root pointer. The second pointer (which must be either a constant or a
5567 global value address) contains the meta-data to be associated with the
5568 root.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5569
5570<h5>Semantics:</h5>
Chris Lattnera7d94ba2008-04-24 05:59:56 +0000[diff] [blame]5571<p>At runtime, a call to this intrinsic stores a null pointer into the "ptrloc"
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5572 location. At compile-time, the code generator generates information to allow
5573 the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
5574 intrinsic may only be used in a function which <a href="#gc">specifies a GC
5575 algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5576
5577</div>
5578
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5579<!-- _______________________________________________________________________ -->
5580<div class="doc_subsubsection">
5581 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
5582</div>
5583
5584<div class="doc_text">
5585
5586<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5587<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5588 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5589</pre>
5590
5591<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5592<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5593 locations, allowing garbage collector implementations that require read
5594 barriers.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5595
5596<h5>Arguments:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5597<p>The second argument is the address to read from, which should be an address
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5598 allocated from the garbage collector. The first object is a pointer to the
5599 start of the referenced object, if needed by the language runtime (otherwise
5600 null).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5601
5602<h5>Semantics:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5603<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5604 instruction, but may be replaced with substantially more complex code by the
5605 garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
5606 may only be used in a function which <a href="#gc">specifies a GC
5607 algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5608
5609</div>
5610
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5611<!-- _______________________________________________________________________ -->
5612<div class="doc_subsubsection">
5613 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
5614</div>
5615
5616<div class="doc_text">
5617
5618<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5619<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5620 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5621</pre>
5622
5623<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5624<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5625 locations, allowing garbage collector implementations that require write
5626 barriers (such as generational or reference counting collectors).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5627
5628<h5>Arguments:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5629<p>The first argument is the reference to store, the second is the start of the
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5630 object to store it to, and the third is the address of the field of Obj to
5631 store to. If the runtime does not require a pointer to the object, Obj may
5632 be null.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5633
5634<h5>Semantics:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5635<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5636 instruction, but may be replaced with substantially more complex code by the
5637 garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
5638 may only be used in a function which <a href="#gc">specifies a GC
5639 algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5640
5641</div>
5642
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5643<!-- ======================================================================= -->
5644<div class="doc_subsection">
5645 <a name="int_codegen">Code Generator Intrinsics</a>
5646</div>
5647
5648<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5649
5650<p>These intrinsics are provided by LLVM to expose special features that may
5651 only be implemented with code generator support.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5652
5653</div>
5654
5655<!-- _______________________________________________________________________ -->
5656<div class="doc_subsubsection">
5657 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
5658</div>
5659
5660<div class="doc_text">
5661
5662<h5>Syntax:</h5>
5663<pre>
5664 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
5665</pre>
5666
5667<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5668<p>The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
5669 target-specific value indicating the return address of the current function
5670 or one of its callers.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5671
5672<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5673<p>The argument to this intrinsic indicates which function to return the address
5674 for. Zero indicates the calling function, one indicates its caller, etc.
5675 The argument is <b>required</b> to be a constant integer value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5676
5677<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5678<p>The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer
5679 indicating the return address of the specified call frame, or zero if it
5680 cannot be identified. The value returned by this intrinsic is likely to be
5681 incorrect or 0 for arguments other than zero, so it should only be used for
5682 debugging purposes.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5683
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5684<p>Note that calling this intrinsic does not prevent function inlining or other
5685 aggressive transformations, so the value returned may not be that of the
5686 obvious source-language caller.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5687
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5688</div>
5689
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5690<!-- _______________________________________________________________________ -->
5691<div class="doc_subsubsection">
5692 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
5693</div>
5694
5695<div class="doc_text">
5696
5697<h5>Syntax:</h5>
5698<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5699 declare i8 *@llvm.frameaddress(i32 &lt;level&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5700</pre>
5701
5702<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5703<p>The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
5704 target-specific frame pointer value for the specified stack frame.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5705
5706<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5707<p>The argument to this intrinsic indicates which function to return the frame
5708 pointer for. Zero indicates the calling function, one indicates its caller,
5709 etc. The argument is <b>required</b> to be a constant integer value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5710
5711<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5712<p>The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer
5713 indicating the frame address of the specified call frame, or zero if it
5714 cannot be identified. The value returned by this intrinsic is likely to be
5715 incorrect or 0 for arguments other than zero, so it should only be used for
5716 debugging purposes.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5717
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5718<p>Note that calling this intrinsic does not prevent function inlining or other
5719 aggressive transformations, so the value returned may not be that of the
5720 obvious source-language caller.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5721
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5722</div>
5723
5724<!-- _______________________________________________________________________ -->
5725<div class="doc_subsubsection">
5726 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
5727</div>
5728
5729<div class="doc_text">
5730
5731<h5>Syntax:</h5>
5732<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5733 declare i8 *@llvm.stacksave()
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5734</pre>
5735
5736<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5737<p>The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state
5738 of the function stack, for use
5739 with <a href="#int_stackrestore"> <tt>llvm.stackrestore</tt></a>. This is
5740 useful for implementing language features like scoped automatic variable
5741 sized arrays in C99.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5742
5743<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5744<p>This intrinsic returns a opaque pointer value that can be passed
5745 to <a href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When
5746 an <tt>llvm.stackrestore</tt> intrinsic is executed with a value saved
5747 from <tt>llvm.stacksave</tt>, it effectively restores the state of the stack
5748 to the state it was in when the <tt>llvm.stacksave</tt> intrinsic executed.
5749 In practice, this pops any <a href="#i_alloca">alloca</a> blocks from the
5750 stack that were allocated after the <tt>llvm.stacksave</tt> was executed.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5751
5752</div>
5753
5754<!-- _______________________________________________________________________ -->
5755<div class="doc_subsubsection">
5756 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
5757</div>
5758
5759<div class="doc_text">
5760
5761<h5>Syntax:</h5>
5762<pre>
5763 declare void @llvm.stackrestore(i8 * %ptr)
5764</pre>
5765
5766<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5767<p>The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
5768 the function stack to the state it was in when the
5769 corresponding <a href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic
5770 executed. This is useful for implementing language features like scoped
5771 automatic variable sized arrays in C99.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5772
5773<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5774<p>See the description
5775 for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5776
5777</div>
5778
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5779<!-- _______________________________________________________________________ -->
5780<div class="doc_subsubsection">
5781 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
5782</div>
5783
5784<div class="doc_text">
5785
5786<h5>Syntax:</h5>
5787<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5788 declare void @llvm.prefetch(i8* &lt;address&gt;, i32 &lt;rw&gt;, i32 &lt;locality&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5789</pre>
5790
5791<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5792<p>The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to
5793 insert a prefetch instruction if supported; otherwise, it is a noop.
5794 Prefetches have no effect on the behavior of the program but can change its
5795 performance characteristics.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5796
5797<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5798<p><tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the
5799 specifier determining if the fetch should be for a read (0) or write (1),
5800 and <tt>locality</tt> is a temporal locality specifier ranging from (0) - no
5801 locality, to (3) - extremely local keep in cache. The <tt>rw</tt>
5802 and <tt>locality</tt> arguments must be constant integers.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5803
5804<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5805<p>This intrinsic does not modify the behavior of the program. In particular,
5806 prefetches cannot trap and do not produce a value. On targets that support
5807 this intrinsic, the prefetch can provide hints to the processor cache for
5808 better performance.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5809
5810</div>
5811
5812<!-- _______________________________________________________________________ -->
5813<div class="doc_subsubsection">
5814 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
5815</div>
5816
5817<div class="doc_text">
5818
5819<h5>Syntax:</h5>
5820<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5821 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5822</pre>
5823
5824<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5825<p>The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program
5826 Counter (PC) in a region of code to simulators and other tools. The method
5827 is target specific, but it is expected that the marker will use exported
5828 symbols to transmit the PC of the marker. The marker makes no guarantees
5829 that it will remain with any specific instruction after optimizations. It is
5830 possible that the presence of a marker will inhibit optimizations. The
5831 intended use is to be inserted after optimizations to allow correlations of
5832 simulation runs.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5833
5834<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5835<p><tt>id</tt> is a numerical id identifying the marker.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5836
5837<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5838<p>This intrinsic does not modify the behavior of the program. Backends that do
Dan Gohman22dc6682010-03-01 17:41:39 +0000[diff] [blame]5839 not support this intrinsic may ignore it.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5840
5841</div>
5842
5843<!-- _______________________________________________________________________ -->
5844<div class="doc_subsubsection">
5845 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
5846</div>
5847
5848<div class="doc_text">
5849
5850<h5>Syntax:</h5>
5851<pre>
5852 declare i64 @llvm.readcyclecounter( )
5853</pre>
5854
5855<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5856<p>The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
5857 counter register (or similar low latency, high accuracy clocks) on those
5858 targets that support it. On X86, it should map to RDTSC. On Alpha, it
5859 should map to RPCC. As the backing counters overflow quickly (on the order
5860 of 9 seconds on alpha), this should only be used for small timings.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5861
5862<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5863<p>When directly supported, reading the cycle counter should not modify any
5864 memory. Implementations are allowed to either return a application specific
5865 value or a system wide value. On backends without support, this is lowered
5866 to a constant 0.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5867
5868</div>
5869
5870<!-- ======================================================================= -->
5871<div class="doc_subsection">
5872 <a name="int_libc">Standard C Library Intrinsics</a>
5873</div>
5874
5875<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5876
5877<p>LLVM provides intrinsics for a few important standard C library functions.
5878 These intrinsics allow source-language front-ends to pass information about
5879 the alignment of the pointer arguments to the code generator, providing
5880 opportunity for more efficient code generation.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5881
5882</div>
5883
5884<!-- _______________________________________________________________________ -->
5885<div class="doc_subsubsection">
5886 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
5887</div>
5888
5889<div class="doc_text">
5890
5891<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5892<p>This is an overloaded intrinsic. You can use <tt>llvm.memcpy</tt> on any
Mon P Wang238462c2010-04-07 06:35:53 +0000[diff] [blame]5893 integer bit width and for different address spaces. Not all targets support
5894 all bit widths however.</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5895
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5896<pre>
Chris Lattner5aa5aab2010-04-08 00:53:57 +0000[diff] [blame^]5897 declare void @llvm.memcpy.p0i8.p0i8.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5898 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
5899 declare void @llvm.memcpy.p0i8.p0i8.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5900 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5901</pre>
5902
5903<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5904<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
5905 source location to the destination location.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5906
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5907<p>Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
Chris Lattner5aa5aab2010-04-08 00:53:57 +0000[diff] [blame^]5908 intrinsics do not return a value, takes extra alignment/isvolatile arguments
5909 and the pointers can be in specified address spaces.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5910
5911<h5>Arguments:</h5>
Chris Lattner5aa5aab2010-04-08 00:53:57 +0000[diff] [blame^]5912
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5913<p>The first argument is a pointer to the destination, the second is a pointer
5914 to the source. The third argument is an integer argument specifying the
Chris Lattner5aa5aab2010-04-08 00:53:57 +0000[diff] [blame^]5915 number of bytes to copy, the fourth argument is the alignment of the
5916 source and destination locations, and the fifth is a boolean indicating a
5917 volatile access.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5918
Dan Gohman22dc6682010-03-01 17:41:39 +0000[diff] [blame]5919<p>If the call to this intrinsic has an alignment value that is not 0 or 1,
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5920 then the caller guarantees that both the source and destination pointers are
5921 aligned to that boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5922
Chris Lattner5aa5aab2010-04-08 00:53:57 +0000[diff] [blame^]5923<p>Volatile accesses should not be deleted if dead, but the access behavior is
5924 not very cleanly specified and it is unwise to depend on it.</p>
5925
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5926<h5>Semantics:</h5>
Chris Lattner5aa5aab2010-04-08 00:53:57 +0000[diff] [blame^]5927
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5928<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
5929 source location to the destination location, which are not allowed to
5930 overlap. It copies "len" bytes of memory over. If the argument is known to
5931 be aligned to some boundary, this can be specified as the fourth argument,
5932 otherwise it should be set to 0 or 1.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5933
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5934</div>
5935
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5936<!-- _______________________________________________________________________ -->
5937<div class="doc_subsubsection">
5938 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
5939</div>
5940
5941<div class="doc_text">
5942
5943<h5>Syntax:</h5>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5944<p>This is an overloaded intrinsic. You can use llvm.memmove on any integer bit
Mon P Wang238462c2010-04-07 06:35:53 +0000[diff] [blame]5945 width and for different address space. Not all targets support all bit
5946 widths however.</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5947
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5948<pre>
Chris Lattner5aa5aab2010-04-08 00:53:57 +0000[diff] [blame^]5949 declare void @llvm.memmove.p0i8.p0i8.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5950 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
5951 declare void @llvm.memmove.p0i8.p0i8.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5952 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5953</pre>
5954
5955<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5956<p>The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the
5957 source location to the destination location. It is similar to the
5958 '<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to
5959 overlap.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5960
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5961<p>Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
Chris Lattner5aa5aab2010-04-08 00:53:57 +0000[diff] [blame^]5962 intrinsics do not return a value, takes extra alignment/isvolatile arguments
5963 and the pointers can be in specified address spaces.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5964
5965<h5>Arguments:</h5>
Chris Lattner5aa5aab2010-04-08 00:53:57 +0000[diff] [blame^]5966
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5967<p>The first argument is a pointer to the destination, the second is a pointer
5968 to the source. The third argument is an integer argument specifying the
Chris Lattner5aa5aab2010-04-08 00:53:57 +0000[diff] [blame^]5969 number of bytes to copy, the fourth argument is the alignment of the
5970 source and destination locations, and the fifth is a boolean indicating a
5971 volatile access.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5972
Dan Gohman22dc6682010-03-01 17:41:39 +0000[diff] [blame]5973<p>If the call to this intrinsic has an alignment value that is not 0 or 1,
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5974 then the caller guarantees that the source and destination pointers are
5975 aligned to that boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5976
Chris Lattner5aa5aab2010-04-08 00:53:57 +0000[diff] [blame^]5977<p>Volatile accesses should not be deleted if dead, but the access behavior is
5978 not very cleanly specified and it is unwise to depend on it.</p>
5979
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5980<h5>Semantics:</h5>
Chris Lattner5aa5aab2010-04-08 00:53:57 +0000[diff] [blame^]5981
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5982<p>The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the
5983 source location to the destination location, which may overlap. It copies
5984 "len" bytes of memory over. If the argument is known to be aligned to some
5985 boundary, this can be specified as the fourth argument, otherwise it should
5986 be set to 0 or 1.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5987
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5988</div>
5989
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5990<!-- _______________________________________________________________________ -->
5991<div class="doc_subsubsection">
5992 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
5993</div>
5994
5995<div class="doc_text">
5996
5997<h5>Syntax:</h5>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5998<p>This is an overloaded intrinsic. You can use llvm.memset on any integer bit
Mon P Wang238462c2010-04-07 06:35:53 +0000[diff] [blame]5999 width and for different address spaces. Not all targets support all bit
6000 widths however.</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6001
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6002<pre>
Chris Lattner5aa5aab2010-04-08 00:53:57 +0000[diff] [blame^]6003 declare void @llvm.memset.p0i8.i32(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
6004 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lgt;isvolatile&gt;)
6005 declare void @llvm.memset.p0i8.i64(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
6006 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lgt;isvolatile&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6007</pre>
6008
6009<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6010<p>The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a
6011 particular byte value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6012
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6013<p>Note that, unlike the standard libc function, the <tt>llvm.memset</tt>
Chris Lattner5aa5aab2010-04-08 00:53:57 +0000[diff] [blame^]6014 intrinsic does not return a value, takes extra alignment/volatile arguments,
6015 and the destination can be in an arbitrary address space.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6016
6017<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6018<p>The first argument is a pointer to the destination to fill, the second is the
6019 byte value to fill it with, the third argument is an integer argument
6020 specifying the number of bytes to fill, and the fourth argument is the known
6021 alignment of destination location.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6022
Dan Gohman22dc6682010-03-01 17:41:39 +0000[diff] [blame]6023<p>If the call to this intrinsic has an alignment value that is not 0 or 1,
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6024 then the caller guarantees that the destination pointer is aligned to that
6025 boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6026
Chris Lattner5aa5aab2010-04-08 00:53:57 +0000[diff] [blame^]6027<p>Volatile accesses should not be deleted if dead, but the access behavior is
6028 not very cleanly specified and it is unwise to depend on it.</p>
6029
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6030<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6031<p>The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting
6032 at the destination location. If the argument is known to be aligned to some
6033 boundary, this can be specified as the fourth argument, otherwise it should
6034 be set to 0 or 1.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6035
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6036</div>
6037
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6038<!-- _______________________________________________________________________ -->
6039<div class="doc_subsubsection">
6040 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
6041</div>
6042
6043<div class="doc_text">
6044
6045<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6046<p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
6047 floating point or vector of floating point type. Not all targets support all
6048 types however.</p>
6049
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6050<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]6051 declare float @llvm.sqrt.f32(float %Val)
6052 declare double @llvm.sqrt.f64(double %Val)
6053 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
6054 declare fp128 @llvm.sqrt.f128(fp128 %Val)
6055 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6056</pre>
6057
6058<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6059<p>The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
6060 returning the same value as the libm '<tt>sqrt</tt>' functions would.
6061 Unlike <tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined
6062 behavior for negative numbers other than -0.0 (which allows for better
6063 optimization, because there is no need to worry about errno being
6064 set). <tt>llvm.sqrt(-0.0)</tt> is defined to return -0.0 like IEEE sqrt.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6065
6066<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6067<p>The argument and return value are floating point numbers of the same
6068 type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6069
6070<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6071<p>This function returns the sqrt of the specified operand if it is a
6072 nonnegative floating point number.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6073
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6074</div>
6075
6076<!-- _______________________________________________________________________ -->
6077<div class="doc_subsubsection">
6078 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
6079</div>
6080
6081<div class="doc_text">
6082
6083<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6084<p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
6085 floating point or vector of floating point type. Not all targets support all
6086 types however.</p>
6087
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6088<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]6089 declare float @llvm.powi.f32(float %Val, i32 %power)
6090 declare double @llvm.powi.f64(double %Val, i32 %power)
6091 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
6092 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
6093 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6094</pre>
6095
6096<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6097<p>The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
6098 specified (positive or negative) power. The order of evaluation of
6099 multiplications is not defined. When a vector of floating point type is
6100 used, the second argument remains a scalar integer value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6101
6102<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6103<p>The second argument is an integer power, and the first is a value to raise to
6104 that power.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6105
6106<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6107<p>This function returns the first value raised to the second power with an
6108 unspecified sequence of rounding operations.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6109
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6110</div>
6111
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6112<!-- _______________________________________________________________________ -->
6113<div class="doc_subsubsection">
6114 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
6115</div>
6116
6117<div class="doc_text">
6118
6119<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6120<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
6121 floating point or vector of floating point type. Not all targets support all
6122 types however.</p>
6123
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6124<pre>
6125 declare float @llvm.sin.f32(float %Val)
6126 declare double @llvm.sin.f64(double %Val)
6127 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
6128 declare fp128 @llvm.sin.f128(fp128 %Val)
6129 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
6130</pre>
6131
6132<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6133<p>The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6134
6135<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6136<p>The argument and return value are floating point numbers of the same
6137 type.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6138
6139<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6140<p>This function returns the sine of the specified operand, returning the same
6141 values as the libm <tt>sin</tt> functions would, and handles error conditions
6142 in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6143
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6144</div>
6145
6146<!-- _______________________________________________________________________ -->
6147<div class="doc_subsubsection">
6148 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
6149</div>
6150
6151<div class="doc_text">
6152
6153<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6154<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
6155 floating point or vector of floating point type. Not all targets support all
6156 types however.</p>
6157
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6158<pre>
6159 declare float @llvm.cos.f32(float %Val)
6160 declare double @llvm.cos.f64(double %Val)
6161 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
6162 declare fp128 @llvm.cos.f128(fp128 %Val)
6163 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
6164</pre>
6165
6166<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6167<p>The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6168
6169<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6170<p>The argument and return value are floating point numbers of the same
6171 type.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6172
6173<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6174<p>This function returns the cosine of the specified operand, returning the same
6175 values as the libm <tt>cos</tt> functions would, and handles error conditions
6176 in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6177
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6178</div>
6179
6180<!-- _______________________________________________________________________ -->
6181<div class="doc_subsubsection">
6182 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
6183</div>
6184
6185<div class="doc_text">
6186
6187<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6188<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
6189 floating point or vector of floating point type. Not all targets support all
6190 types however.</p>
6191
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6192<pre>
6193 declare float @llvm.pow.f32(float %Val, float %Power)
6194 declare double @llvm.pow.f64(double %Val, double %Power)
6195 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
6196 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
6197 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
6198</pre>
6199
6200<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6201<p>The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
6202 specified (positive or negative) power.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6203
6204<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6205<p>The second argument is a floating point power, and the first is a value to
6206 raise to that power.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6207
6208<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6209<p>This function returns the first value raised to the second power, returning
6210 the same values as the libm <tt>pow</tt> functions would, and handles error
6211 conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6212
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6213</div>
6214
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6215<!-- ======================================================================= -->
6216<div class="doc_subsection">
6217 <a name="int_manip">Bit Manipulation Intrinsics</a>
6218</div>
6219
6220<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6221
6222<p>LLVM provides intrinsics for a few important bit manipulation operations.
6223 These allow efficient code generation for some algorithms.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6224
6225</div>
6226
6227<!-- _______________________________________________________________________ -->
6228<div class="doc_subsubsection">
6229 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
6230</div>
6231
6232<div class="doc_text">
6233
6234<h5>Syntax:</h5>
6235<p>This is an overloaded intrinsic function. You can use bswap on any integer
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6236 type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
6237
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6238<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]6239 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
6240 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
6241 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6242</pre>
6243
6244<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6245<p>The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
6246 values with an even number of bytes (positive multiple of 16 bits). These
6247 are useful for performing operations on data that is not in the target's
6248 native byte order.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6249
6250<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6251<p>The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
6252 and low byte of the input i16 swapped. Similarly,
6253 the <tt>llvm.bswap.i32</tt> intrinsic returns an i32 value that has the four
6254 bytes of the input i32 swapped, so that if the input bytes are numbered 0, 1,
6255 2, 3 then the returned i32 will have its bytes in 3, 2, 1, 0 order.
6256 The <tt>llvm.bswap.i48</tt>, <tt>llvm.bswap.i64</tt> and other intrinsics
6257 extend this concept to additional even-byte lengths (6 bytes, 8 bytes and
6258 more, respectively).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6259
6260</div>
6261
6262<!-- _______________________________________________________________________ -->
6263<div class="doc_subsubsection">
6264 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
6265</div>
6266
6267<div class="doc_text">
6268
6269<h5>Syntax:</h5>
6270<p>This is an overloaded intrinsic. You can use llvm.ctpop on any integer bit
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6271 width. Not all targets support all bit widths however.</p>
6272
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6273<pre>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6274 declare i8 @llvm.ctpop.i8(i8 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]6275 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6276 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]6277 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
6278 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6279</pre>
6280
6281<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6282<p>The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set
6283 in a value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6284
6285<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6286<p>The only argument is the value to be counted. The argument may be of any
6287 integer type. The return type must match the argument type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6288
6289<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6290<p>The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6291
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6292</div>
6293
6294<!-- _______________________________________________________________________ -->
6295<div class="doc_subsubsection">
6296 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
6297</div>
6298
6299<div class="doc_text">
6300
6301<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6302<p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
6303 integer bit width. Not all targets support all bit widths however.</p>
6304
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6305<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]6306 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
6307 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6308 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]6309 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
6310 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6311</pre>
6312
6313<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6314<p>The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
6315 leading zeros in a variable.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6316
6317<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6318<p>The only argument is the value to be counted. The argument may be of any
6319 integer type. The return type must match the argument type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6320
6321<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6322<p>The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant)
6323 zeros in a variable. If the src == 0 then the result is the size in bits of
6324 the type of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6325
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6326</div>
6327
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6328<!-- _______________________________________________________________________ -->
6329<div class="doc_subsubsection">
6330 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
6331</div>
6332
6333<div class="doc_text">
6334
6335<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6336<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
6337 integer bit width. Not all targets support all bit widths however.</p>
6338
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6339<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]6340 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
6341 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6342 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]6343 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
6344 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6345</pre>
6346
6347<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6348<p>The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
6349 trailing zeros.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6350
6351<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6352<p>The only argument is the value to be counted. The argument may be of any
6353 integer type. The return type must match the argument type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6354
6355<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6356<p>The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant)
6357 zeros in a variable. If the src == 0 then the result is the size in bits of
6358 the type of src. For example, <tt>llvm.cttz(2) = 1</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6359
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6360</div>
6361
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6362<!-- ======================================================================= -->
6363<div class="doc_subsection">
6364 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
6365</div>
6366
6367<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6368
6369<p>LLVM provides intrinsics for some arithmetic with overflow operations.</p>
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6370
6371</div>
6372
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6373<!-- _______________________________________________________________________ -->
6374<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6375 <a name="int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6376</div>
6377
6378<div class="doc_text">
6379
6380<h5>Syntax:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6381<p>This is an overloaded intrinsic. You can use <tt>llvm.sadd.with.overflow</tt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6382 on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6383
6384<pre>
6385 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
6386 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6387 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
6388</pre>
6389
6390<h5>Overview:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6391<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6392 a signed addition of the two arguments, and indicate whether an overflow
6393 occurred during the signed summation.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6394
6395<h5>Arguments:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6396<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6397 be of integer types of any bit width, but they must have the same bit
6398 width. The second element of the result structure must be of
6399 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6400 undergo signed addition.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6401
6402<h5>Semantics:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6403<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6404 a signed addition of the two variables. They return a structure &mdash; the
6405 first element of which is the signed summation, and the second element of
6406 which is a bit specifying if the signed summation resulted in an
6407 overflow.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6408
6409<h5>Examples:</h5>
6410<pre>
6411 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6412 %sum = extractvalue {i32, i1} %res, 0
6413 %obit = extractvalue {i32, i1} %res, 1
6414 br i1 %obit, label %overflow, label %normal
6415</pre>
6416
6417</div>
6418
6419<!-- _______________________________________________________________________ -->
6420<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6421 <a name="int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6422</div>
6423
6424<div class="doc_text">
6425
6426<h5>Syntax:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6427<p>This is an overloaded intrinsic. You can use <tt>llvm.uadd.with.overflow</tt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6428 on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6429
6430<pre>
6431 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
6432 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6433 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
6434</pre>
6435
6436<h5>Overview:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6437<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6438 an unsigned addition of the two arguments, and indicate whether a carry
6439 occurred during the unsigned summation.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6440
6441<h5>Arguments:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6442<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6443 be of integer types of any bit width, but they must have the same bit
6444 width. The second element of the result structure must be of
6445 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6446 undergo unsigned addition.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6447
6448<h5>Semantics:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6449<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6450 an unsigned addition of the two arguments. They return a structure &mdash;
6451 the first element of which is the sum, and the second element of which is a
6452 bit specifying if the unsigned summation resulted in a carry.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6453
6454<h5>Examples:</h5>
6455<pre>
6456 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6457 %sum = extractvalue {i32, i1} %res, 0
6458 %obit = extractvalue {i32, i1} %res, 1
6459 br i1 %obit, label %carry, label %normal
6460</pre>
6461
6462</div>
6463
6464<!-- _______________________________________________________________________ -->
6465<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6466 <a name="int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6467</div>
6468
6469<div class="doc_text">
6470
6471<h5>Syntax:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6472<p>This is an overloaded intrinsic. You can use <tt>llvm.ssub.with.overflow</tt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6473 on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6474
6475<pre>
6476 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
6477 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6478 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
6479</pre>
6480
6481<h5>Overview:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6482<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6483 a signed subtraction of the two arguments, and indicate whether an overflow
6484 occurred during the signed subtraction.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6485
6486<h5>Arguments:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6487<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6488 be of integer types of any bit width, but they must have the same bit
6489 width. The second element of the result structure must be of
6490 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6491 undergo signed subtraction.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6492
6493<h5>Semantics:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6494<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6495 a signed subtraction of the two arguments. They return a structure &mdash;
6496 the first element of which is the subtraction, and the second element of
6497 which is a bit specifying if the signed subtraction resulted in an
6498 overflow.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6499
6500<h5>Examples:</h5>
6501<pre>
6502 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6503 %sum = extractvalue {i32, i1} %res, 0
6504 %obit = extractvalue {i32, i1} %res, 1
6505 br i1 %obit, label %overflow, label %normal
6506</pre>
6507
6508</div>
6509
6510<!-- _______________________________________________________________________ -->
6511<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6512 <a name="int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6513</div>
6514
6515<div class="doc_text">
6516
6517<h5>Syntax:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6518<p>This is an overloaded intrinsic. You can use <tt>llvm.usub.with.overflow</tt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6519 on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6520
6521<pre>
6522 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
6523 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6524 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
6525</pre>
6526
6527<h5>Overview:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6528<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6529 an unsigned subtraction of the two arguments, and indicate whether an
6530 overflow occurred during the unsigned subtraction.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6531
6532<h5>Arguments:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6533<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6534 be of integer types of any bit width, but they must have the same bit
6535 width. The second element of the result structure must be of
6536 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6537 undergo unsigned subtraction.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6538
6539<h5>Semantics:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6540<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6541 an unsigned subtraction of the two arguments. They return a structure &mdash;
6542 the first element of which is the subtraction, and the second element of
6543 which is a bit specifying if the unsigned subtraction resulted in an
6544 overflow.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6545
6546<h5>Examples:</h5>
6547<pre>
6548 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6549 %sum = extractvalue {i32, i1} %res, 0
6550 %obit = extractvalue {i32, i1} %res, 1
6551 br i1 %obit, label %overflow, label %normal
6552</pre>
6553
6554</div>
6555
6556<!-- _______________________________________________________________________ -->
6557<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6558 <a name="int_smul_overflow">'<tt>llvm.smul.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6559</div>
6560
6561<div class="doc_text">
6562
6563<h5>Syntax:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6564<p>This is an overloaded intrinsic. You can use <tt>llvm.smul.with.overflow</tt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6565 on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6566
6567<pre>
6568 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
6569 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6570 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
6571</pre>
6572
6573<h5>Overview:</h5>
6574
6575<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6576 a signed multiplication of the two arguments, and indicate whether an
6577 overflow occurred during the signed multiplication.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6578
6579<h5>Arguments:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6580<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6581 be of integer types of any bit width, but they must have the same bit
6582 width. The second element of the result structure must be of
6583 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6584 undergo signed multiplication.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6585
6586<h5>Semantics:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6587<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6588 a signed multiplication of the two arguments. They return a structure &mdash;
6589 the first element of which is the multiplication, and the second element of
6590 which is a bit specifying if the signed multiplication resulted in an
6591 overflow.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6592
6593<h5>Examples:</h5>
6594<pre>
6595 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6596 %sum = extractvalue {i32, i1} %res, 0
6597 %obit = extractvalue {i32, i1} %res, 1
6598 br i1 %obit, label %overflow, label %normal
6599</pre>
6600
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6601</div>
6602
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]6603<!-- _______________________________________________________________________ -->
6604<div class="doc_subsubsection">
6605 <a name="int_umul_overflow">'<tt>llvm.umul.with.overflow.*</tt>' Intrinsics</a>
6606</div>
6607
6608<div class="doc_text">
6609
6610<h5>Syntax:</h5>
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]6611<p>This is an overloaded intrinsic. You can use <tt>llvm.umul.with.overflow</tt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6612 on any integer bit width.</p>
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]6613
6614<pre>
6615 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
6616 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6617 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
6618</pre>
6619
6620<h5>Overview:</h5>
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]6621<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6622 a unsigned multiplication of the two arguments, and indicate whether an
6623 overflow occurred during the unsigned multiplication.</p>
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]6624
6625<h5>Arguments:</h5>
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]6626<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6627 be of integer types of any bit width, but they must have the same bit
6628 width. The second element of the result structure must be of
6629 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6630 undergo unsigned multiplication.</p>
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]6631
6632<h5>Semantics:</h5>
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]6633<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6634 an unsigned multiplication of the two arguments. They return a structure
6635 &mdash; the first element of which is the multiplication, and the second
6636 element of which is a bit specifying if the unsigned multiplication resulted
6637 in an overflow.</p>
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]6638
6639<h5>Examples:</h5>
6640<pre>
6641 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6642 %sum = extractvalue {i32, i1} %res, 0
6643 %obit = extractvalue {i32, i1} %res, 1
6644 br i1 %obit, label %overflow, label %normal
6645</pre>
6646
6647</div>
6648
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6649<!-- ======================================================================= -->
6650<div class="doc_subsection">
Anton Korobeynikov8c98a892010-03-14 18:42:47 +0000[diff] [blame]6651 <a name="int_fp16">Half Precision Floating Point Intrinsics</a>
6652</div>
6653
6654<div class="doc_text">
6655
Chris Lattnere5969c62010-03-15 04:12:21 +0000[diff] [blame]6656<p>Half precision floating point is a storage-only format. This means that it is
6657 a dense encoding (in memory) but does not support computation in the
6658 format.</p>
Chris Lattnerebc48e52010-03-14 23:03:31 +0000[diff] [blame]6659
Chris Lattnere5969c62010-03-15 04:12:21 +0000[diff] [blame]6660<p>This means that code must first load the half-precision floating point
Chris Lattnerebc48e52010-03-14 23:03:31 +0000[diff] [blame]6661 value as an i16, then convert it to float with <a
6662 href="#int_convert_from_fp16"><tt>llvm.convert.from.fp16</tt></a>.
6663 Computation can then be performed on the float value (including extending to
Chris Lattnere5969c62010-03-15 04:12:21 +0000[diff] [blame]6664 double etc). To store the value back to memory, it is first converted to
6665 float if needed, then converted to i16 with
Chris Lattnerebc48e52010-03-14 23:03:31 +0000[diff] [blame]6666 <a href="#int_convert_to_fp16"><tt>llvm.convert.to.fp16</tt></a>, then
6667 storing as an i16 value.</p>
Anton Korobeynikov8c98a892010-03-14 18:42:47 +0000[diff] [blame]6668</div>
6669
6670<!-- _______________________________________________________________________ -->
6671<div class="doc_subsubsection">
Chris Lattnerebc48e52010-03-14 23:03:31 +0000[diff] [blame]6672 <a name="int_convert_to_fp16">'<tt>llvm.convert.to.fp16</tt>' Intrinsic</a>
Anton Korobeynikov8c98a892010-03-14 18:42:47 +0000[diff] [blame]6673</div>
6674
6675<div class="doc_text">
6676
6677<h5>Syntax:</h5>
6678<pre>
6679 declare i16 @llvm.convert.to.fp16(f32 %a)
6680</pre>
6681
6682<h5>Overview:</h5>
6683<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
6684 a conversion from single precision floating point format to half precision
6685 floating point format.</p>
6686
6687<h5>Arguments:</h5>
6688<p>The intrinsic function contains single argument - the value to be
6689 converted.</p>
6690
6691<h5>Semantics:</h5>
6692<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
6693 a conversion from single precision floating point format to half precision
Chris Lattnere5969c62010-03-15 04:12:21 +0000[diff] [blame]6694 floating point format. The return value is an <tt>i16</tt> which
Chris Lattnerebc48e52010-03-14 23:03:31 +0000[diff] [blame]6695 contains the converted number.</p>
Anton Korobeynikov8c98a892010-03-14 18:42:47 +0000[diff] [blame]6696
6697<h5>Examples:</h5>
6698<pre>
6699 %res = call i16 @llvm.convert.to.fp16(f32 %a)
6700 store i16 %res, i16* @x, align 2
6701</pre>
6702
6703</div>
6704
6705<!-- _______________________________________________________________________ -->
6706<div class="doc_subsubsection">
Chris Lattnerebc48e52010-03-14 23:03:31 +0000[diff] [blame]6707 <a name="int_convert_from_fp16">'<tt>llvm.convert.from.fp16</tt>' Intrinsic</a>
Anton Korobeynikov8c98a892010-03-14 18:42:47 +0000[diff] [blame]6708</div>
6709
6710<div class="doc_text">
6711
6712<h5>Syntax:</h5>
6713<pre>
6714 declare f32 @llvm.convert.from.fp16(i16 %a)
6715</pre>
6716
6717<h5>Overview:</h5>
6718<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs
6719 a conversion from half precision floating point format to single precision
6720 floating point format.</p>
6721
6722<h5>Arguments:</h5>
6723<p>The intrinsic function contains single argument - the value to be
6724 converted.</p>
6725
6726<h5>Semantics:</h5>
6727<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs a
Chris Lattnere5969c62010-03-15 04:12:21 +0000[diff] [blame]6728 conversion from half single precision floating point format to single
Chris Lattnerebc48e52010-03-14 23:03:31 +0000[diff] [blame]6729 precision floating point format. The input half-float value is represented by
6730 an <tt>i16</tt> value.</p>
Anton Korobeynikov8c98a892010-03-14 18:42:47 +0000[diff] [blame]6731
6732<h5>Examples:</h5>
6733<pre>
6734 %a = load i16* @x, align 2
6735 %res = call f32 @llvm.convert.from.fp16(i16 %a)
6736</pre>
6737
6738</div>
6739
6740<!-- ======================================================================= -->
6741<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6742 <a name="int_debugger">Debugger Intrinsics</a>
6743</div>
6744
6745<div class="doc_text">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6746
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6747<p>The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt>
6748 prefix), are described in
6749 the <a href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source
6750 Level Debugging</a> document.</p>
6751
6752</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6753
6754<!-- ======================================================================= -->
6755<div class="doc_subsection">
6756 <a name="int_eh">Exception Handling Intrinsics</a>
6757</div>
6758
6759<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6760
6761<p>The LLVM exception handling intrinsics (which all start with
6762 <tt>llvm.eh.</tt> prefix), are described in
6763 the <a href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
6764 Handling</a> document.</p>
6765
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6766</div>
6767
6768<!-- ======================================================================= -->
6769<div class="doc_subsection">
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6770 <a name="int_trampoline">Trampoline Intrinsic</a>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6771</div>
6772
6773<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6774
6775<p>This intrinsic makes it possible to excise one parameter, marked with
6776 the <tt>nest</tt> attribute, from a function. The result is a callable
6777 function pointer lacking the nest parameter - the caller does not need to
6778 provide a value for it. Instead, the value to use is stored in advance in a
6779 "trampoline", a block of memory usually allocated on the stack, which also
6780 contains code to splice the nest value into the argument list. This is used
6781 to implement the GCC nested function address extension.</p>
6782
6783<p>For example, if the function is
6784 <tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
6785 pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as
6786 follows:</p>
6787
6788<div class="doc_code">
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6789<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6790 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
6791 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
6792 %p = call i8* @llvm.init.trampoline( i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8*), i8* %nval )
6793 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6794</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6795</div>
6796
6797<p>The call <tt>%val = call i32 %fp( i32 %x, i32 %y )</tt> is then equivalent
6798 to <tt>%val = call i32 %f( i8* %nval, i32 %x, i32 %y )</tt>.</p>
6799
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6800</div>
6801
6802<!-- _______________________________________________________________________ -->
6803<div class="doc_subsubsection">
6804 <a name="int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a>
6805</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6806
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6807<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6808
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6809<h5>Syntax:</h5>
6810<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6811 declare i8* @llvm.init.trampoline(i8* &lt;tramp&gt;, i8* &lt;func&gt;, i8* &lt;nval&gt;)
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6812</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6813
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6814<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6815<p>This fills the memory pointed to by <tt>tramp</tt> with code and returns a
6816 function pointer suitable for executing it.</p>
6817
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6818<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6819<p>The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
6820 pointers. The <tt>tramp</tt> argument must point to a sufficiently large and
6821 sufficiently aligned block of memory; this memory is written to by the
6822 intrinsic. Note that the size and the alignment are target-specific - LLVM
6823 currently provides no portable way of determining them, so a front-end that
6824 generates this intrinsic needs to have some target-specific knowledge.
6825 The <tt>func</tt> argument must hold a function bitcast to
6826 an <tt>i8*</tt>.</p>
6827
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6828<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6829<p>The block of memory pointed to by <tt>tramp</tt> is filled with target
6830 dependent code, turning it into a function. A pointer to this function is
6831 returned, but needs to be bitcast to an <a href="#int_trampoline">appropriate
6832 function pointer type</a> before being called. The new function's signature
6833 is the same as that of <tt>func</tt> with any arguments marked with
6834 the <tt>nest</tt> attribute removed. At most one such <tt>nest</tt> argument
6835 is allowed, and it must be of pointer type. Calling the new function is
6836 equivalent to calling <tt>func</tt> with the same argument list, but
6837 with <tt>nval</tt> used for the missing <tt>nest</tt> argument. If, after
6838 calling <tt>llvm.init.trampoline</tt>, the memory pointed to
6839 by <tt>tramp</tt> is modified, then the effect of any later call to the
6840 returned function pointer is undefined.</p>
6841
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6842</div>
6843
6844<!-- ======================================================================= -->
6845<div class="doc_subsection">
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6846 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
6847</div>
6848
6849<div class="doc_text">
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6850
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6851<p>These intrinsic functions expand the "universal IR" of LLVM to represent
6852 hardware constructs for atomic operations and memory synchronization. This
6853 provides an interface to the hardware, not an interface to the programmer. It
6854 is aimed at a low enough level to allow any programming models or APIs
6855 (Application Programming Interfaces) which need atomic behaviors to map
6856 cleanly onto it. It is also modeled primarily on hardware behavior. Just as
6857 hardware provides a "universal IR" for source languages, it also provides a
6858 starting point for developing a "universal" atomic operation and
6859 synchronization IR.</p>
6860
6861<p>These do <em>not</em> form an API such as high-level threading libraries,
6862 software transaction memory systems, atomic primitives, and intrinsic
6863 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
6864 application libraries. The hardware interface provided by LLVM should allow
6865 a clean implementation of all of these APIs and parallel programming models.
6866 No one model or paradigm should be selected above others unless the hardware
6867 itself ubiquitously does so.</p>
6868
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6869</div>
6870
6871<!-- _______________________________________________________________________ -->
6872<div class="doc_subsubsection">
6873 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
6874</div>
6875<div class="doc_text">
6876<h5>Syntax:</h5>
6877<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6878 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 Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6879</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6880
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6881<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6882<p>The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
6883 specific pairs of memory access types.</p>
6884
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6885<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6886<p>The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
6887 The first four arguments enables a specific barrier as listed below. The
Dan Gohman22dc6682010-03-01 17:41:39 +0000[diff] [blame]6888 fifth argument specifies that the barrier applies to io or device or uncached
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6889 memory.</p>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6890
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6891<ul>
6892 <li><tt>ll</tt>: load-load barrier</li>
6893 <li><tt>ls</tt>: load-store barrier</li>
6894 <li><tt>sl</tt>: store-load barrier</li>
6895 <li><tt>ss</tt>: store-store barrier</li>
6896 <li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
6897</ul>
6898
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6899<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6900<p>This intrinsic causes the system to enforce some ordering constraints upon
6901 the loads and stores of the program. This barrier does not
6902 indicate <em>when</em> any events will occur, it only enforces
6903 an <em>order</em> in which they occur. For any of the specified pairs of load
6904 and store operations (f.ex. load-load, or store-load), all of the first
6905 operations preceding the barrier will complete before any of the second
6906 operations succeeding the barrier begin. Specifically the semantics for each
6907 pairing is as follows:</p>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6908
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6909<ul>
6910 <li><tt>ll</tt>: All loads before the barrier must complete before any load
6911 after the barrier begins.</li>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]6912 <li><tt>ls</tt>: All loads before the barrier must complete before any
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6913 store after the barrier begins.</li>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]6914 <li><tt>ss</tt>: All stores before the barrier must complete before any
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6915 store after the barrier begins.</li>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]6916 <li><tt>sl</tt>: All stores before the barrier must complete before any
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6917 load after the barrier begins.</li>
6918</ul>
6919
6920<p>These semantics are applied with a logical "and" behavior when more than one
6921 is enabled in a single memory barrier intrinsic.</p>
6922
6923<p>Backends may implement stronger barriers than those requested when they do
6924 not support as fine grained a barrier as requested. Some architectures do
6925 not need all types of barriers and on such architectures, these become
6926 noops.</p>
6927
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6928<h5>Example:</h5>
6929<pre>
Victor Hernandez6b054092009-10-26 23:44:29 +0000[diff] [blame]6930%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
6931%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6932 store i32 4, %ptr
6933
6934%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
6935 call void @llvm.memory.barrier( i1 false, i1 true, i1 false, i1 false )
6936 <i>; guarantee the above finishes</i>
6937 store i32 8, %ptr <i>; before this begins</i>
6938</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6939
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6940</div>
6941
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6942<!-- _______________________________________________________________________ -->
6943<div class="doc_subsubsection">
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6944 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6945</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6946
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6947<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6948
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6949<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6950<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
6951 any integer bit width and for different address spaces. Not all targets
6952 support all bit widths however.</p>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6953
6954<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6955 declare i8 @llvm.atomic.cmp.swap.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt; )
6956 declare i16 @llvm.atomic.cmp.swap.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt; )
6957 declare i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt; )
6958 declare i64 @llvm.atomic.cmp.swap.i64.p0i64( i64* &lt;ptr&gt;, i64 &lt;cmp&gt;, i64 &lt;val&gt; )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6959</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6960
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6961<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6962<p>This loads a value in memory and compares it to a given value. If they are
6963 equal, it stores a new value into the memory.</p>
6964
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6965<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6966<p>The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result
6967 as well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
6968 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
6969 this integer type. While any bit width integer may be used, targets may only
6970 lower representations they support in hardware.</p>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6971
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6972<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6973<p>This entire intrinsic must be executed atomically. It first loads the value
6974 in memory pointed to by <tt>ptr</tt> and compares it with the
6975 value <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the
6976 memory. The loaded value is yielded in all cases. This provides the
6977 equivalent of an atomic compare-and-swap operation within the SSA
6978 framework.</p>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6979
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6980<h5>Examples:</h5>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6981<pre>
Victor Hernandez6b054092009-10-26 23:44:29 +0000[diff] [blame]6982%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
6983%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6984 store i32 4, %ptr
6985
6986%val1 = add i32 4, 4
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6987%result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 4, %val1 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6988 <i>; yields {i32}:result1 = 4</i>
6989%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
6990%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
6991
6992%val2 = add i32 1, 1
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6993%result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 5, %val2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6994 <i>; yields {i32}:result2 = 8</i>
6995%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
6996
6997%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
6998</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6999
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7000</div>
7001
7002<!-- _______________________________________________________________________ -->
7003<div class="doc_subsubsection">
7004 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
7005</div>
7006<div class="doc_text">
7007<h5>Syntax:</h5>
7008
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7009<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
7010 integer bit width. Not all targets support all bit widths however.</p>
7011
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7012<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7013 declare i8 @llvm.atomic.swap.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;val&gt; )
7014 declare i16 @llvm.atomic.swap.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;val&gt; )
7015 declare i32 @llvm.atomic.swap.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;val&gt; )
7016 declare i64 @llvm.atomic.swap.i64.p0i64( i64* &lt;ptr&gt;, i64 &lt;val&gt; )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7017</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7018
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7019<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7020<p>This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
7021 the value from memory. It then stores the value in <tt>val</tt> in the memory
7022 at <tt>ptr</tt>.</p>
7023
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7024<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7025<p>The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both
7026 the <tt>val</tt> argument and the result must be integers of the same bit
7027 width. The first argument, <tt>ptr</tt>, must be a pointer to a value of this
7028 integer type. The targets may only lower integer representations they
7029 support.</p>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7030
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7031<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7032<p>This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
7033 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
7034 equivalent of an atomic swap operation within the SSA framework.</p>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7035
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7036<h5>Examples:</h5>
7037<pre>
Victor Hernandez6b054092009-10-26 23:44:29 +0000[diff] [blame]7038%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7039%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7040 store i32 4, %ptr
7041
7042%val1 = add i32 4, 4
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7043%result1 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val1 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7044 <i>; yields {i32}:result1 = 4</i>
7045%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
7046%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
7047
7048%val2 = add i32 1, 1
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7049%result2 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7050 <i>; yields {i32}:result2 = 8</i>
7051
7052%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
7053%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
7054</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7055
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7056</div>
7057
7058<!-- _______________________________________________________________________ -->
7059<div class="doc_subsubsection">
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7060 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7061
7062</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7063
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7064<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7065
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7066<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7067<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.add</tt> on
7068 any integer bit width. Not all targets support all bit widths however.</p>
7069
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7070<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7071 declare i8 @llvm.atomic.load.add.i8..p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
7072 declare i16 @llvm.atomic.load.add.i16..p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
7073 declare i32 @llvm.atomic.load.add.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
7074 declare i64 @llvm.atomic.load.add.i64..p0i64( i64* &lt;ptr&gt;, i64 &lt;delta&gt; )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7075</pre>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7076
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7077<h5>Overview:</h5>
7078<p>This intrinsic adds <tt>delta</tt> to the value stored in memory
7079 at <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
7080
7081<h5>Arguments:</h5>
7082<p>The intrinsic takes two arguments, the first a pointer to an integer value
7083 and the second an integer value. The result is also an integer value. These
7084 integer types can have any bit width, but they must all have the same bit
7085 width. The targets may only lower integer representations they support.</p>
7086
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7087<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7088<p>This intrinsic does a series of operations atomically. It first loads the
7089 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
7090 to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.</p>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7091
7092<h5>Examples:</h5>
7093<pre>
Victor Hernandez6b054092009-10-26 23:44:29 +0000[diff] [blame]7094%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7095%ptr = bitcast i8* %mallocP to i32*
7096 store i32 4, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7097%result1 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 4 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7098 <i>; yields {i32}:result1 = 4</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7099%result2 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7100 <i>; yields {i32}:result2 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7101%result3 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 5 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7102 <i>; yields {i32}:result3 = 10</i>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7103%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7104</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7105
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]7106</div>
7107
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7108<!-- _______________________________________________________________________ -->
7109<div class="doc_subsubsection">
7110 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
7111
7112</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7113
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7114<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7115
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7116<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7117<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.sub</tt> on
7118 any integer bit width and for different address spaces. Not all targets
7119 support all bit widths however.</p>
7120
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7121<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7122 declare i8 @llvm.atomic.load.sub.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
7123 declare i16 @llvm.atomic.load.sub.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
7124 declare i32 @llvm.atomic.load.sub.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
7125 declare i64 @llvm.atomic.load.sub.i64.p0i32( i64* &lt;ptr&gt;, i64 &lt;delta&gt; )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7126</pre>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7127
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7128<h5>Overview:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]7129<p>This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7130 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
7131
7132<h5>Arguments:</h5>
7133<p>The intrinsic takes two arguments, the first a pointer to an integer value
7134 and the second an integer value. The result is also an integer value. These
7135 integer types can have any bit width, but they must all have the same bit
7136 width. The targets may only lower integer representations they support.</p>
7137
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7138<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7139<p>This intrinsic does a series of operations atomically. It first loads the
7140 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
7141 result to <tt>ptr</tt>. It yields the original value stored
7142 at <tt>ptr</tt>.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7143
7144<h5>Examples:</h5>
7145<pre>
Victor Hernandez6b054092009-10-26 23:44:29 +0000[diff] [blame]7146%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7147%ptr = bitcast i8* %mallocP to i32*
7148 store i32 8, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7149%result1 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 4 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7150 <i>; yields {i32}:result1 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7151%result2 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 2 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7152 <i>; yields {i32}:result2 = 4</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7153%result3 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 5 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7154 <i>; yields {i32}:result3 = 2</i>
7155%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
7156</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7157
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7158</div>
7159
7160<!-- _______________________________________________________________________ -->
7161<div class="doc_subsubsection">
7162 <a name="int_atomic_load_and">'<tt>llvm.atomic.load.and.*</tt>' Intrinsic</a><br>
7163 <a name="int_atomic_load_nand">'<tt>llvm.atomic.load.nand.*</tt>' Intrinsic</a><br>
7164 <a name="int_atomic_load_or">'<tt>llvm.atomic.load.or.*</tt>' Intrinsic</a><br>
7165 <a name="int_atomic_load_xor">'<tt>llvm.atomic.load.xor.*</tt>' Intrinsic</a><br>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7166</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7167
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7168<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7169
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7170<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7171<p>These are overloaded intrinsics. You can
7172 use <tt>llvm.atomic.load_and</tt>, <tt>llvm.atomic.load_nand</tt>,
7173 <tt>llvm.atomic.load_or</tt>, and <tt>llvm.atomic.load_xor</tt> on any integer
7174 bit width and for different address spaces. Not all targets support all bit
7175 widths however.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7176
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7177<pre>
7178 declare i8 @llvm.atomic.load.and.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
7179 declare i16 @llvm.atomic.load.and.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
7180 declare i32 @llvm.atomic.load.and.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
7181 declare i64 @llvm.atomic.load.and.i64.p0i64( i64* &lt;ptr&gt;, i64 &lt;delta&gt; )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7182</pre>
7183
7184<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7185 declare i8 @llvm.atomic.load.or.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
7186 declare i16 @llvm.atomic.load.or.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
7187 declare i32 @llvm.atomic.load.or.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
7188 declare i64 @llvm.atomic.load.or.i64.p0i64( i64* &lt;ptr&gt;, i64 &lt;delta&gt; )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7189</pre>
7190
7191<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7192 declare i8 @llvm.atomic.load.nand.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
7193 declare i16 @llvm.atomic.load.nand.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
7194 declare i32 @llvm.atomic.load.nand.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
7195 declare i64 @llvm.atomic.load.nand.i64.p0i32( i64* &lt;ptr&gt;, i64 &lt;delta&gt; )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7196</pre>
7197
7198<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7199 declare i8 @llvm.atomic.load.xor.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
7200 declare i16 @llvm.atomic.load.xor.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
7201 declare i32 @llvm.atomic.load.xor.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
7202 declare i64 @llvm.atomic.load.xor.i64.p0i32( i64* &lt;ptr&gt;, i64 &lt;delta&gt; )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7203</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7204
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7205<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7206<p>These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
7207 the value stored in memory at <tt>ptr</tt>. It yields the original value
7208 at <tt>ptr</tt>.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7209
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7210<h5>Arguments:</h5>
7211<p>These intrinsics take two arguments, the first a pointer to an integer value
7212 and the second an integer value. The result is also an integer value. These
7213 integer types can have any bit width, but they must all have the same bit
7214 width. The targets may only lower integer representations they support.</p>
7215
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7216<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7217<p>These intrinsics does a series of operations atomically. They first load the
7218 value stored at <tt>ptr</tt>. They then do the bitwise
7219 operation <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the
7220 original value stored at <tt>ptr</tt>.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7221
7222<h5>Examples:</h5>
7223<pre>
Victor Hernandez6b054092009-10-26 23:44:29 +0000[diff] [blame]7224%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7225%ptr = bitcast i8* %mallocP to i32*
7226 store i32 0x0F0F, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7227%result0 = call i32 @llvm.atomic.load.nand.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7228 <i>; yields {i32}:result0 = 0x0F0F</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7229%result1 = call i32 @llvm.atomic.load.and.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7230 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7231%result2 = call i32 @llvm.atomic.load.or.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7232 <i>; yields {i32}:result2 = 0xF0</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7233%result3 = call i32 @llvm.atomic.load.xor.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7234 <i>; yields {i32}:result3 = FF</i>
7235%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
7236</pre>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7237
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7238</div>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7239
7240<!-- _______________________________________________________________________ -->
7241<div class="doc_subsubsection">
7242 <a name="int_atomic_load_max">'<tt>llvm.atomic.load.max.*</tt>' Intrinsic</a><br>
7243 <a name="int_atomic_load_min">'<tt>llvm.atomic.load.min.*</tt>' Intrinsic</a><br>
7244 <a name="int_atomic_load_umax">'<tt>llvm.atomic.load.umax.*</tt>' Intrinsic</a><br>
7245 <a name="int_atomic_load_umin">'<tt>llvm.atomic.load.umin.*</tt>' Intrinsic</a><br>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7246</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7247
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7248<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7249
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7250<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7251<p>These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
7252 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
7253 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
7254 address spaces. Not all targets support all bit widths however.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7255
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7256<pre>
7257 declare i8 @llvm.atomic.load.max.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
7258 declare i16 @llvm.atomic.load.max.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
7259 declare i32 @llvm.atomic.load.max.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
7260 declare i64 @llvm.atomic.load.max.i64.p0i64( i64* &lt;ptr&gt;, i64 &lt;delta&gt; )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7261</pre>
7262
7263<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7264 declare i8 @llvm.atomic.load.min.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
7265 declare i16 @llvm.atomic.load.min.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
7266 declare i32 @llvm.atomic.load.min.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
7267 declare i64 @llvm.atomic.load.min.i64..p0i64( i64* &lt;ptr&gt;, i64 &lt;delta&gt; )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7268</pre>
7269
7270<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7271 declare i8 @llvm.atomic.load.umax.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
7272 declare i16 @llvm.atomic.load.umax.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
7273 declare i32 @llvm.atomic.load.umax.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
7274 declare i64 @llvm.atomic.load.umax.i64.p0i64( i64* &lt;ptr&gt;, i64 &lt;delta&gt; )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7275</pre>
7276
7277<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7278 declare i8 @llvm.atomic.load.umin.i8..p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
7279 declare i16 @llvm.atomic.load.umin.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
7280 declare i32 @llvm.atomic.load.umin.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
7281 declare i64 @llvm.atomic.load.umin.i64..p0i64( i64* &lt;ptr&gt;, i64 &lt;delta&gt; )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7282</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7283
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7284<h5>Overview:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]7285<p>These intrinsics takes the signed or unsigned minimum or maximum of
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7286 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
7287 original value at <tt>ptr</tt>.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7288
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7289<h5>Arguments:</h5>
7290<p>These intrinsics take two arguments, the first a pointer to an integer value
7291 and the second an integer value. The result is also an integer value. These
7292 integer types can have any bit width, but they must all have the same bit
7293 width. The targets may only lower integer representations they support.</p>
7294
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7295<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7296<p>These intrinsics does a series of operations atomically. They first load the
7297 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or
7298 max <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They
7299 yield the original value stored at <tt>ptr</tt>.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7300
7301<h5>Examples:</h5>
7302<pre>
Victor Hernandez6b054092009-10-26 23:44:29 +0000[diff] [blame]7303%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7304%ptr = bitcast i8* %mallocP to i32*
7305 store i32 7, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7306%result0 = call i32 @llvm.atomic.load.min.i32.p0i32( i32* %ptr, i32 -2 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7307 <i>; yields {i32}:result0 = 7</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7308%result1 = call i32 @llvm.atomic.load.max.i32.p0i32( i32* %ptr, i32 8 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7309 <i>; yields {i32}:result1 = -2</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7310%result2 = call i32 @llvm.atomic.load.umin.i32.p0i32( i32* %ptr, i32 10 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7311 <i>; yields {i32}:result2 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7312%result3 = call i32 @llvm.atomic.load.umax.i32.p0i32( i32* %ptr, i32 30 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7313 <i>; yields {i32}:result3 = 8</i>
7314%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
7315</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7316
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7317</div>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]7318
Nick Lewyckyc888d352009-10-13 07:03:23 +0000[diff] [blame]7319
7320<!-- ======================================================================= -->
7321<div class="doc_subsection">
7322 <a name="int_memorymarkers">Memory Use Markers</a>
7323</div>
7324
7325<div class="doc_text">
7326
7327<p>This class of intrinsics exists to information about the lifetime of memory
7328 objects and ranges where variables are immutable.</p>
7329
7330</div>
7331
7332<!-- _______________________________________________________________________ -->
7333<div class="doc_subsubsection">
7334 <a name="int_lifetime_start">'<tt>llvm.lifetime.start</tt>' Intrinsic</a>
7335</div>
7336
7337<div class="doc_text">
7338
7339<h5>Syntax:</h5>
7340<pre>
7341 declare void @llvm.lifetime.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7342</pre>
7343
7344<h5>Overview:</h5>
7345<p>The '<tt>llvm.lifetime.start</tt>' intrinsic specifies the start of a memory
7346 object's lifetime.</p>
7347
7348<h5>Arguments:</h5>
Nick Lewycky23f49982009-10-13 07:57:33 +0000[diff] [blame]7349<p>The first argument is a constant integer representing the size of the
7350 object, or -1 if it is variable sized. The second argument is a pointer to
7351 the object.</p>
Nick Lewyckyc888d352009-10-13 07:03:23 +0000[diff] [blame]7352
7353<h5>Semantics:</h5>
7354<p>This intrinsic indicates that before this point in the code, the value of the
7355 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
Nick Lewyckyf8a5f302009-10-27 16:56:58 +0000[diff] [blame]7356 never be used and has an undefined value. A load from the pointer that
7357 precedes this intrinsic can be replaced with
Nick Lewyckyc888d352009-10-13 07:03:23 +0000[diff] [blame]7358 <tt>'<a href="#undefvalues">undef</a>'</tt>.</p>
7359
7360</div>
7361
7362<!-- _______________________________________________________________________ -->
7363<div class="doc_subsubsection">
7364 <a name="int_lifetime_end">'<tt>llvm.lifetime.end</tt>' Intrinsic</a>
7365</div>
7366
7367<div class="doc_text">
7368
7369<h5>Syntax:</h5>
7370<pre>
7371 declare void @llvm.lifetime.end(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7372</pre>
7373
7374<h5>Overview:</h5>
7375<p>The '<tt>llvm.lifetime.end</tt>' intrinsic specifies the end of a memory
7376 object's lifetime.</p>
7377
7378<h5>Arguments:</h5>
Nick Lewycky23f49982009-10-13 07:57:33 +0000[diff] [blame]7379<p>The first argument is a constant integer representing the size of the
7380 object, or -1 if it is variable sized. The second argument is a pointer to
7381 the object.</p>
Nick Lewyckyc888d352009-10-13 07:03:23 +0000[diff] [blame]7382
7383<h5>Semantics:</h5>
7384<p>This intrinsic indicates that after this point in the code, the value of the
7385 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
7386 never be used and has an undefined value. Any stores into the memory object
7387 following this intrinsic may be removed as dead.
7388
7389</div>
7390
7391<!-- _______________________________________________________________________ -->
7392<div class="doc_subsubsection">
7393 <a name="int_invariant_start">'<tt>llvm.invariant.start</tt>' Intrinsic</a>
7394</div>
7395
7396<div class="doc_text">
7397
7398<h5>Syntax:</h5>
7399<pre>
7400 declare {}* @llvm.invariant.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;) readonly
7401</pre>
7402
7403<h5>Overview:</h5>
7404<p>The '<tt>llvm.invariant.start</tt>' intrinsic specifies that the contents of
7405 a memory object will not change.</p>
7406
7407<h5>Arguments:</h5>
Nick Lewycky23f49982009-10-13 07:57:33 +0000[diff] [blame]7408<p>The first argument is a constant integer representing the size of the
7409 object, or -1 if it is variable sized. The second argument is a pointer to
7410 the object.</p>
Nick Lewyckyc888d352009-10-13 07:03:23 +0000[diff] [blame]7411
7412<h5>Semantics:</h5>
7413<p>This intrinsic indicates that until an <tt>llvm.invariant.end</tt> that uses
7414 the return value, the referenced memory location is constant and
7415 unchanging.</p>
7416
7417</div>
7418
7419<!-- _______________________________________________________________________ -->
7420<div class="doc_subsubsection">
7421 <a name="int_invariant_end">'<tt>llvm.invariant.end</tt>' Intrinsic</a>
7422</div>
7423
7424<div class="doc_text">
7425
7426<h5>Syntax:</h5>
7427<pre>
7428 declare void @llvm.invariant.end({}* &lt;start&gt;, i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7429</pre>
7430
7431<h5>Overview:</h5>
7432<p>The '<tt>llvm.invariant.end</tt>' intrinsic specifies that the contents of
7433 a memory object are mutable.</p>
7434
7435<h5>Arguments:</h5>
7436<p>The first argument is the matching <tt>llvm.invariant.start</tt> intrinsic.
Nick Lewycky23f49982009-10-13 07:57:33 +0000[diff] [blame]7437 The second argument is a constant integer representing the size of the
7438 object, or -1 if it is variable sized and the third argument is a pointer
7439 to the object.</p>
Nick Lewyckyc888d352009-10-13 07:03:23 +0000[diff] [blame]7440
7441<h5>Semantics:</h5>
7442<p>This intrinsic indicates that the memory is mutable again.</p>
7443
7444</div>
7445
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]7446<!-- ======================================================================= -->
7447<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7448 <a name="int_general">General Intrinsics</a>
7449</div>
7450
7451<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7452
7453<p>This class of intrinsics is designed to be generic and has no specific
7454 purpose.</p>
7455
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7456</div>
7457
7458<!-- _______________________________________________________________________ -->
7459<div class="doc_subsubsection">
7460 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
7461</div>
7462
7463<div class="doc_text">
7464
7465<h5>Syntax:</h5>
7466<pre>
7467 declare void @llvm.var.annotation(i8* &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
7468</pre>
7469
7470<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7471<p>The '<tt>llvm.var.annotation</tt>' intrinsic.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7472
7473<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7474<p>The first argument is a pointer to a value, the second is a pointer to a
7475 global string, the third is a pointer to a global string which is the source
7476 file name, and the last argument is the line number.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7477
7478<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7479<p>This intrinsic allows annotation of local variables with arbitrary strings.
7480 This can be useful for special purpose optimizations that want to look for
7481 these annotations. These have no other defined use, they are ignored by code
7482 generation and optimization.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7483
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7484</div>
7485
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7486<!-- _______________________________________________________________________ -->
7487<div class="doc_subsubsection">
Tanya Lattnerc9869b12007-09-21 23:57:59 +0000[diff] [blame]7488 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7489</div>
7490
7491<div class="doc_text">
7492
7493<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7494<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
7495 any integer bit width.</p>
7496
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7497<pre>
Tanya Lattner09161fe2007-09-22 00:03:01 +0000[diff] [blame]7498 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
7499 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
7500 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
7501 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
7502 declare i256 @llvm.annotation.i256(i256 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7503</pre>
7504
7505<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7506<p>The '<tt>llvm.annotation</tt>' intrinsic.</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7507
7508<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7509<p>The first argument is an integer value (result of some expression), the
7510 second is a pointer to a global string, the third is a pointer to a global
7511 string which is the source file name, and the last argument is the line
7512 number. It returns the value of the first argument.</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7513
7514<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7515<p>This intrinsic allows annotations to be put on arbitrary expressions with
7516 arbitrary strings. This can be useful for special purpose optimizations that
7517 want to look for these annotations. These have no other defined use, they
7518 are ignored by code generation and optimization.</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7519
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7520</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7521
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]7522<!-- _______________________________________________________________________ -->
7523<div class="doc_subsubsection">
7524 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
7525</div>
7526
7527<div class="doc_text">
7528
7529<h5>Syntax:</h5>
7530<pre>
7531 declare void @llvm.trap()
7532</pre>
7533
7534<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7535<p>The '<tt>llvm.trap</tt>' intrinsic.</p>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]7536
7537<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7538<p>None.</p>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]7539
7540<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7541<p>This intrinsics is lowered to the target dependent trap instruction. If the
7542 target does not have a trap instruction, this intrinsic will be lowered to
7543 the call of the <tt>abort()</tt> function.</p>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]7544
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]7545</div>
7546
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]7547<!-- _______________________________________________________________________ -->
7548<div class="doc_subsubsection">
Misha Brukman5dd7f4d2008-11-22 23:55:29 +0000[diff] [blame]7549 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]7550</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7551
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]7552<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7553
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]7554<h5>Syntax:</h5>
7555<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7556 declare void @llvm.stackprotector( i8* &lt;guard&gt;, i8** &lt;slot&gt; )
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]7557</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7558
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]7559<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7560<p>The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and
7561 stores it onto the stack at <tt>slot</tt>. The stack slot is adjusted to
7562 ensure that it is placed on the stack before local variables.</p>
7563
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]7564<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7565<p>The <tt>llvm.stackprotector</tt> intrinsic requires two pointer
7566 arguments. The first argument is the value loaded from the stack
7567 guard <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt>
7568 that has enough space to hold the value of the guard.</p>
7569
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]7570<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7571<p>This intrinsic causes the prologue/epilogue inserter to force the position of
7572 the <tt>AllocaInst</tt> stack slot to be before local variables on the
7573 stack. This is to ensure that if a local variable on the stack is
7574 overwritten, it will destroy the value of the guard. When the function exits,
7575 the guard on the stack is checked against the original guard. If they're
7576 different, then the program aborts by calling the <tt>__stack_chk_fail()</tt>
7577 function.</p>
7578
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]7579</div>
7580
Eric Christopher767a3722009-11-30 08:03:53 +0000[diff] [blame]7581<!-- _______________________________________________________________________ -->
7582<div class="doc_subsubsection">
7583 <a name="int_objectsize">'<tt>llvm.objectsize</tt>' Intrinsic</a>
7584</div>
7585
7586<div class="doc_text">
7587
7588<h5>Syntax:</h5>
7589<pre>
Eric Christopher0101f9d2009-12-23 00:29:49 +0000[diff] [blame]7590 declare i32 @llvm.objectsize.i32( i8* &lt;object&gt;, i1 &lt;type&gt; )
7591 declare i64 @llvm.objectsize.i64( i8* &lt;object&gt;, i1 &lt;type&gt; )
Eric Christopher767a3722009-11-30 08:03:53 +0000[diff] [blame]7592</pre>
7593
7594<h5>Overview:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]7595<p>The <tt>llvm.objectsize</tt> intrinsic is designed to provide information
Eric Christopher184f44d2010-01-08 21:42:39 +0000[diff] [blame]7596 to the optimizers to discover at compile time either a) when an
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]7597 operation like memcpy will either overflow a buffer that corresponds to
7598 an object, or b) to determine that a runtime check for overflow isn't
7599 necessary. An object in this context means an allocation of a
Eric Christopher0101f9d2009-12-23 00:29:49 +0000[diff] [blame]7600 specific class, structure, array, or other object.</p>
Eric Christopher767a3722009-11-30 08:03:53 +0000[diff] [blame]7601
7602<h5>Arguments:</h5>
7603<p>The <tt>llvm.objectsize</tt> intrinsic takes two arguments. The first
Eric Christopher0101f9d2009-12-23 00:29:49 +0000[diff] [blame]7604 argument is a pointer to or into the <tt>object</tt>. The second argument
7605 is a boolean 0 or 1. This argument determines whether you want the
7606 maximum (0) or minimum (1) bytes remaining. This needs to be a literal 0 or
7607 1, variables are not allowed.</p>
7608
Eric Christopher767a3722009-11-30 08:03:53 +0000[diff] [blame]7609<h5>Semantics:</h5>
7610<p>The <tt>llvm.objectsize</tt> intrinsic is lowered to either a constant
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]7611 representing the size of the object concerned or <tt>i32/i64 -1 or 0</tt>
7612 (depending on the <tt>type</tt> argument if the size cannot be determined
7613 at compile time.</p>
Eric Christopher767a3722009-11-30 08:03:53 +0000[diff] [blame]7614
7615</div>
7616
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7617<!-- *********************************************************************** -->
7618<hr>
7619<address>
7620 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman947321d2008-12-11 17:34:48 +0000[diff] [blame]7621 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7622 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman947321d2008-12-11 17:34:48 +0000[diff] [blame]7623 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7624
7625 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
7626 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
7627 Last modified: $Date$
7628</address>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]7629
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7630</body>
7631</html>