Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / d5d5172c3c572f2f6ce33254511747c58398bf14 / . / docs / LangRef.html
blob: b337b6a8571f857634156ed510a48d798d259f17 [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>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]258 <li><a href="#int_debugger">Debugger intrinsics</a></li>
259 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]260 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]261 <ol>
262 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]263 </ol>
264 </li>
Bill Wendling9127adb2008-11-18 22:10:53 +0000[diff] [blame]265 <li><a href="#int_atomics">Atomic intrinsics</a>
266 <ol>
267 <li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
268 <li><a href="#int_atomic_cmp_swap"><tt>llvm.atomic.cmp.swap</tt></a></li>
269 <li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
270 <li><a href="#int_atomic_load_add"><tt>llvm.atomic.load.add</tt></a></li>
271 <li><a href="#int_atomic_load_sub"><tt>llvm.atomic.load.sub</tt></a></li>
272 <li><a href="#int_atomic_load_and"><tt>llvm.atomic.load.and</tt></a></li>
273 <li><a href="#int_atomic_load_nand"><tt>llvm.atomic.load.nand</tt></a></li>
274 <li><a href="#int_atomic_load_or"><tt>llvm.atomic.load.or</tt></a></li>
275 <li><a href="#int_atomic_load_xor"><tt>llvm.atomic.load.xor</tt></a></li>
276 <li><a href="#int_atomic_load_max"><tt>llvm.atomic.load.max</tt></a></li>
277 <li><a href="#int_atomic_load_min"><tt>llvm.atomic.load.min</tt></a></li>
278 <li><a href="#int_atomic_load_umax"><tt>llvm.atomic.load.umax</tt></a></li>
279 <li><a href="#int_atomic_load_umin"><tt>llvm.atomic.load.umin</tt></a></li>
280 </ol>
281 </li>
Nick Lewyckyc888d352009-10-13 07:03:23 +0000[diff] [blame]282 <li><a href="#int_memorymarkers">Memory Use Markers</a>
283 <ol>
284 <li><a href="#int_lifetime_start"><tt>llvm.lifetime.start</tt></a></li>
285 <li><a href="#int_lifetime_end"><tt>llvm.lifetime.end</tt></a></li>
286 <li><a href="#int_invariant_start"><tt>llvm.invariant.start</tt></a></li>
287 <li><a href="#int_invariant_end"><tt>llvm.invariant.end</tt></a></li>
288 </ol>
289 </li>
Reid Spencerb043f672007-07-20 19:59:11 +0000[diff] [blame]290 <li><a href="#int_general">General intrinsics</a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]291 <ol>
Reid Spencerb043f672007-07-20 19:59:11 +0000[diff] [blame]292 <li><a href="#int_var_annotation">
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]293 '<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]294 <li><a href="#int_annotation">
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]295 '<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]296 <li><a href="#int_trap">
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]297 '<tt>llvm.trap</tt>' Intrinsic</a></li>
298 <li><a href="#int_stackprotector">
299 '<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
Eric Christopher767a3722009-11-30 08:03:53 +0000[diff] [blame]300 <li><a href="#int_objectsize">
301 '<tt>llvm.objectsize</tt>' Intrinsic</a></li>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]302 </ol>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]303 </li>
304 </ol>
305 </li>
306</ol>
307
308<div class="doc_author">
309 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
310 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
311</div>
312
313<!-- *********************************************************************** -->
314<div class="doc_section"> <a name="abstract">Abstract </a></div>
315<!-- *********************************************************************** -->
316
317<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]318
319<p>This document is a reference manual for the LLVM assembly language. LLVM is
320 a Static Single Assignment (SSA) based representation that provides type
321 safety, low-level operations, flexibility, and the capability of representing
322 'all' high-level languages cleanly. It is the common code representation
323 used throughout all phases of the LLVM compilation strategy.</p>
324
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]325</div>
326
327<!-- *********************************************************************** -->
328<div class="doc_section"> <a name="introduction">Introduction</a> </div>
329<!-- *********************************************************************** -->
330
331<div class="doc_text">
332
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]333<p>The LLVM code representation is designed to be used in three different forms:
334 as an in-memory compiler IR, as an on-disk bitcode representation (suitable
335 for fast loading by a Just-In-Time compiler), and as a human readable
336 assembly language representation. This allows LLVM to provide a powerful
337 intermediate representation for efficient compiler transformations and
338 analysis, while providing a natural means to debug and visualize the
339 transformations. The three different forms of LLVM are all equivalent. This
340 document describes the human readable representation and notation.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]341
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]342<p>The LLVM representation aims to be light-weight and low-level while being
343 expressive, typed, and extensible at the same time. It aims to be a
344 "universal IR" of sorts, by being at a low enough level that high-level ideas
345 may be cleanly mapped to it (similar to how microprocessors are "universal
346 IR's", allowing many source languages to be mapped to them). By providing
347 type information, LLVM can be used as the target of optimizations: for
348 example, through pointer analysis, it can be proven that a C automatic
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]349 variable is never accessed outside of the current function, allowing it to
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]350 be promoted to a simple SSA value instead of a memory location.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]351
352</div>
353
354<!-- _______________________________________________________________________ -->
355<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
356
357<div class="doc_text">
358
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]359<p>It is important to note that this document describes 'well formed' LLVM
360 assembly language. There is a difference between what the parser accepts and
361 what is considered 'well formed'. For example, the following instruction is
362 syntactically okay, but not well formed:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]363
364<div class="doc_code">
365<pre>
366%x = <a href="#i_add">add</a> i32 1, %x
367</pre>
368</div>
369
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]370<p>because the definition of <tt>%x</tt> does not dominate all of its uses. The
371 LLVM infrastructure provides a verification pass that may be used to verify
372 that an LLVM module is well formed. This pass is automatically run by the
373 parser after parsing input assembly and by the optimizer before it outputs
374 bitcode. The violations pointed out by the verifier pass indicate bugs in
375 transformation passes or input to the parser.</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]376
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]377</div>
378
Chris Lattnera83fdc02007-10-03 17:34:29 +0000[diff] [blame]379<!-- Describe the typesetting conventions here. -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]380
381<!-- *********************************************************************** -->
382<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
383<!-- *********************************************************************** -->
384
385<div class="doc_text">
386
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]387<p>LLVM identifiers come in two basic types: global and local. Global
388 identifiers (functions, global variables) begin with the <tt>'@'</tt>
389 character. Local identifiers (register names, types) begin with
390 the <tt>'%'</tt> character. Additionally, there are three different formats
391 for identifiers, for different purposes:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]392
393<ol>
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]394 <li>Named values are represented as a string of characters with their prefix.
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]395 For example, <tt>%foo</tt>, <tt>@DivisionByZero</tt>,
396 <tt>%a.really.long.identifier</tt>. The actual regular expression used is
397 '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'. Identifiers which require
398 other characters in their names can be surrounded with quotes. Special
399 characters may be escaped using <tt>"\xx"</tt> where <tt>xx</tt> is the
400 ASCII code for the character in hexadecimal. In this way, any character
401 can be used in a name value, even quotes themselves.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]402
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]403 <li>Unnamed values are represented as an unsigned numeric value with their
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]404 prefix. For example, <tt>%12</tt>, <tt>@2</tt>, <tt>%44</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]405
406 <li>Constants, which are described in a <a href="#constants">section about
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]407 constants</a>, below.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]408</ol>
409
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]410<p>LLVM requires that values start with a prefix for two reasons: Compilers
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]411 don't need to worry about name clashes with reserved words, and the set of
412 reserved words may be expanded in the future without penalty. Additionally,
413 unnamed identifiers allow a compiler to quickly come up with a temporary
414 variable without having to avoid symbol table conflicts.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]415
416<p>Reserved words in LLVM are very similar to reserved words in other
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]417 languages. There are keywords for different opcodes
418 ('<tt><a href="#i_add">add</a></tt>',
419 '<tt><a href="#i_bitcast">bitcast</a></tt>',
420 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names
421 ('<tt><a href="#t_void">void</a></tt>',
422 '<tt><a href="#t_primitive">i32</a></tt>', etc...), and others. These
423 reserved words cannot conflict with variable names, because none of them
424 start with a prefix character (<tt>'%'</tt> or <tt>'@'</tt>).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]425
426<p>Here is an example of LLVM code to multiply the integer variable
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]427 '<tt>%X</tt>' by 8:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]428
429<p>The easy way:</p>
430
431<div class="doc_code">
432<pre>
433%result = <a href="#i_mul">mul</a> i32 %X, 8
434</pre>
435</div>
436
437<p>After strength reduction:</p>
438
439<div class="doc_code">
440<pre>
441%result = <a href="#i_shl">shl</a> i32 %X, i8 3
442</pre>
443</div>
444
445<p>And the hard way:</p>
446
447<div class="doc_code">
448<pre>
Gabor Greifc0ea7672009-10-28 13:05:07 +0000[diff] [blame]449%0 = <a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
450%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]451%result = <a href="#i_add">add</a> i32 %1, %1
452</pre>
453</div>
454
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]455<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several important
456 lexical features of LLVM:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]457
458<ol>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]459 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]460 line.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]461
462 <li>Unnamed temporaries are created when the result of a computation is not
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]463 assigned to a named value.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]464
465 <li>Unnamed temporaries are numbered sequentially</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]466</ol>
467
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]468<p>It also shows a convention that we follow in this document. When
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]469 demonstrating instructions, we will follow an instruction with a comment that
470 defines the type and name of value produced. Comments are shown in italic
471 text.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]472
473</div>
474
475<!-- *********************************************************************** -->
476<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
477<!-- *********************************************************************** -->
478
479<!-- ======================================================================= -->
480<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
481</div>
482
483<div class="doc_text">
484
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]485<p>LLVM programs are composed of "Module"s, each of which is a translation unit
486 of the input programs. Each module consists of functions, global variables,
487 and symbol table entries. Modules may be combined together with the LLVM
488 linker, which merges function (and global variable) definitions, resolves
489 forward declarations, and merges symbol table entries. Here is an example of
490 the "hello world" module:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]491
492<div class="doc_code">
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]493<pre>
494<i>; Declare the string constant as a global constant.</i>
495<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]496
497<i>; External declaration of the puts function</i>
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]498<a href="#functionstructure">declare</a> i32 @puts(i8 *) <i>; i32(i8 *)* </i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]499
500<i>; Definition of main function</i>
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]501define i32 @main() { <i>; i32()* </i>
502 <i>; Convert [13 x i8]* to i8 *...</i>
503 %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]504
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]505 <i>; Call puts function to write out the string to stdout.</i>
506 <a href="#i_call">call</a> i32 @puts(i8 * %cast210) <i>; i32</i>
Devang Patela4bb6792010-01-11 19:35:55 +0000[diff] [blame]507 <a href="#i_ret">ret</a> i32 0<br>}
508
509<i>; Named metadata</i>
510!1 = metadata !{i32 41}
511!foo = !{!1, null}
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]512</pre>
513</div>
514
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]515<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]516 "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>" function,
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]517 a <a href="#functionstructure">function definition</a> for
Devang Patela4bb6792010-01-11 19:35:55 +0000[diff] [blame]518 "<tt>main</tt>" and <a href="#namedmetadatastructure">named metadata</a>
519 "<tt>foo"</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]520
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]521<p>In general, a module is made up of a list of global values, where both
522 functions and global variables are global values. Global values are
523 represented by a pointer to a memory location (in this case, a pointer to an
524 array of char, and a pointer to a function), and have one of the
525 following <a href="#linkage">linkage types</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]526
527</div>
528
529<!-- ======================================================================= -->
530<div class="doc_subsection">
531 <a name="linkage">Linkage Types</a>
532</div>
533
534<div class="doc_text">
535
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]536<p>All Global Variables and Functions have one of the following types of
537 linkage:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]538
539<dl>
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]540 <dt><tt><b><a name="linkage_private">private</a></b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]541 <dd>Global values with private linkage are only directly accessible by objects
542 in the current module. In particular, linking code into a module with an
543 private global value may cause the private to be renamed as necessary to
544 avoid collisions. Because the symbol is private to the module, all
545 references can be updated. This doesn't show up in any symbol table in the
546 object file.</dd>
Rafael Espindolaa168fc92009-01-15 20:18:42 +0000[diff] [blame]547
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]548 <dt><tt><b><a name="linkage_linker_private">linker_private</a></b></tt></dt>
Bill Wendling41a07852009-07-20 01:03:30 +0000[diff] [blame]549 <dd>Similar to private, but the symbol is passed through the assembler and
Chris Lattner06c1ecc2009-08-24 04:32:16 +0000[diff] [blame]550 removed by the linker after evaluation. Note that (unlike private
551 symbols) linker_private symbols are subject to coalescing by the linker:
552 weak symbols get merged and redefinitions are rejected. However, unlike
553 normal strong symbols, they are removed by the linker from the final
554 linked image (executable or dynamic library).</dd>
Bill Wendling41a07852009-07-20 01:03:30 +0000[diff] [blame]555
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]556 <dt><tt><b><a name="linkage_internal">internal</a></b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]557 <dd>Similar to private, but the value shows as a local symbol
558 (<tt>STB_LOCAL</tt> in the case of ELF) in the object file. This
559 corresponds to the notion of the '<tt>static</tt>' keyword in C.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]560
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]561 <dt><tt><b><a name="linkage_available_externally">available_externally</a></b></tt></dt>
Chris Lattner68433442009-04-13 05:44:34 +0000[diff] [blame]562 <dd>Globals with "<tt>available_externally</tt>" linkage are never emitted
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]563 into the object file corresponding to the LLVM module. They exist to
564 allow inlining and other optimizations to take place given knowledge of
565 the definition of the global, which is known to be somewhere outside the
566 module. Globals with <tt>available_externally</tt> linkage are allowed to
567 be discarded at will, and are otherwise the same as <tt>linkonce_odr</tt>.
568 This linkage type is only allowed on definitions, not declarations.</dd>
Chris Lattner68433442009-04-13 05:44:34 +0000[diff] [blame]569
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]570 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]571 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
Chris Lattnerd16c5512010-01-09 19:15:14 +0000[diff] [blame]572 the same name when linkage occurs. This can be used to implement
573 some forms of inline functions, templates, or other code which must be
574 generated in each translation unit that uses it, but where the body may
575 be overridden with a more definitive definition later. Unreferenced
576 <tt>linkonce</tt> globals are allowed to be discarded. Note that
577 <tt>linkonce</tt> linkage does not actually allow the optimizer to
578 inline the body of this function into callers because it doesn't know if
579 this definition of the function is the definitive definition within the
580 program or whether it will be overridden by a stronger definition.
581 To enable inlining and other optimizations, use "<tt>linkonce_odr</tt>"
582 linkage.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]583
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]584 <dt><tt><b><a name="linkage_weak">weak</a></b></tt></dt>
Chris Lattner7215c7f2009-08-05 05:21:07 +0000[diff] [blame]585 <dd>"<tt>weak</tt>" linkage has the same merging semantics as
586 <tt>linkonce</tt> linkage, except that unreferenced globals with
587 <tt>weak</tt> linkage may not be discarded. This is used for globals that
588 are declared "weak" in C source code.</dd>
589
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]590 <dt><tt><b><a name="linkage_common">common</a></b></tt></dt>
Chris Lattner7215c7f2009-08-05 05:21:07 +0000[diff] [blame]591 <dd>"<tt>common</tt>" linkage is most similar to "<tt>weak</tt>" linkage, but
592 they are used for tentative definitions in C, such as "<tt>int X;</tt>" at
593 global scope.
594 Symbols with "<tt>common</tt>" linkage are merged in the same way as
595 <tt>weak symbols</tt>, and they may not be deleted if unreferenced.
Chris Lattner83c4a712009-08-05 05:41:44 +0000[diff] [blame]596 <tt>common</tt> symbols may not have an explicit section,
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]597 must have a zero initializer, and may not be marked '<a
Chris Lattner83c4a712009-08-05 05:41:44 +0000[diff] [blame]598 href="#globalvars"><tt>constant</tt></a>'. Functions and aliases may not
599 have common linkage.</dd>
Chris Lattner7215c7f2009-08-05 05:21:07 +0000[diff] [blame]600
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]601
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]602 <dt><tt><b><a name="linkage_appending">appending</a></b></tt></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]603 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]604 pointer to array type. When two global variables with appending linkage
605 are linked together, the two global arrays are appended together. This is
606 the LLVM, typesafe, equivalent of having the system linker append together
607 "sections" with identical names when .o files are linked.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]608
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]609 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]610 <dd>The semantics of this linkage follow the ELF object file model: the symbol
611 is weak until linked, if not linked, the symbol becomes null instead of
612 being an undefined reference.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]613
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]614 <dt><tt><b><a name="linkage_linkonce_odr">linkonce_odr</a></b></tt></dt>
615 <dt><tt><b><a name="linkage_weak_odr">weak_odr</a></b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]616 <dd>Some languages allow differing globals to be merged, such as two functions
617 with different semantics. Other languages, such as <tt>C++</tt>, ensure
618 that only equivalent globals are ever merged (the "one definition rule" -
619 "ODR"). Such languages can use the <tt>linkonce_odr</tt>
620 and <tt>weak_odr</tt> linkage types to indicate that the global will only
621 be merged with equivalent globals. These linkage types are otherwise the
622 same as their non-<tt>odr</tt> versions.</dd>
Duncan Sands19d161f2009-03-07 15:45:40 +0000[diff] [blame]623
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]624 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]625 <dd>If none of the above identifiers are used, the global is externally
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]626 visible, meaning that it participates in linkage and can be used to
627 resolve external symbol references.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]628</dl>
629
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]630<p>The next two types of linkage are targeted for Microsoft Windows platform
631 only. They are designed to support importing (exporting) symbols from (to)
632 DLLs (Dynamic Link Libraries).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]633
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]634<dl>
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]635 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]636 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]637 or variable via a global pointer to a pointer that is set up by the DLL
638 exporting the symbol. On Microsoft Windows targets, the pointer name is
639 formed by combining <code>__imp_</code> and the function or variable
640 name.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]641
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]642 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]643 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]644 pointer to a pointer in a DLL, so that it can be referenced with the
645 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
646 name is formed by combining <code>__imp_</code> and the function or
647 variable name.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]648</dl>
649
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]650<p>For example, since the "<tt>.LC0</tt>" variable is defined to be internal, if
651 another module defined a "<tt>.LC0</tt>" variable and was linked with this
652 one, one of the two would be renamed, preventing a collision. Since
653 "<tt>main</tt>" and "<tt>puts</tt>" are external (i.e., lacking any linkage
654 declarations), they are accessible outside of the current module.</p>
655
656<p>It is illegal for a function <i>declaration</i> to have any linkage type
657 other than "externally visible", <tt>dllimport</tt>
658 or <tt>extern_weak</tt>.</p>
659
Duncan Sands19d161f2009-03-07 15:45:40 +0000[diff] [blame]660<p>Aliases can have only <tt>external</tt>, <tt>internal</tt>, <tt>weak</tt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]661 or <tt>weak_odr</tt> linkages.</p>
662
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]663</div>
664
665<!-- ======================================================================= -->
666<div class="doc_subsection">
667 <a name="callingconv">Calling Conventions</a>
668</div>
669
670<div class="doc_text">
671
672<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]673 and <a href="#i_invoke">invokes</a> can all have an optional calling
674 convention specified for the call. The calling convention of any pair of
675 dynamic caller/callee must match, or the behavior of the program is
676 undefined. The following calling conventions are supported by LLVM, and more
677 may be added in the future:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]678
679<dl>
680 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]681 <dd>This calling convention (the default if no other calling convention is
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]682 specified) matches the target C calling conventions. This calling
683 convention supports varargs function calls and tolerates some mismatch in
684 the declared prototype and implemented declaration of the function (as
685 does normal C).</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]686
687 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]688 <dd>This calling convention attempts to make calls as fast as possible
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]689 (e.g. by passing things in registers). This calling convention allows the
690 target to use whatever tricks it wants to produce fast code for the
691 target, without having to conform to an externally specified ABI
Jeffrey Yasskin3e1a3a02010-01-09 19:44:16 +0000[diff] [blame]692 (Application Binary Interface).
693 <a href="CodeGenerator.html#tailcallopt">Tail calls can only be optimized
694 when this convention is used.</a> This calling convention does not
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]695 support varargs and requires the prototype of all callees to exactly match
696 the prototype of the function definition.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]697
698 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]699 <dd>This calling convention attempts to make code in the caller as efficient
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]700 as possible under the assumption that the call is not commonly executed.
701 As such, these calls often preserve all registers so that the call does
702 not break any live ranges in the caller side. This calling convention
703 does not support varargs and requires the prototype of all callees to
704 exactly match the prototype of the function definition.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]705
706 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]707 <dd>Any calling convention may be specified by number, allowing
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]708 target-specific calling conventions to be used. Target specific calling
709 conventions start at 64.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]710</dl>
711
712<p>More calling conventions can be added/defined on an as-needed basis, to
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]713 support Pascal conventions or any other well-known target-independent
714 convention.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]715
716</div>
717
718<!-- ======================================================================= -->
719<div class="doc_subsection">
720 <a name="visibility">Visibility Styles</a>
721</div>
722
723<div class="doc_text">
724
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]725<p>All Global Variables and Functions have one of the following visibility
726 styles:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]727
728<dl>
729 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]730 <dd>On targets that use the ELF object file format, default visibility means
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]731 that the declaration is visible to other modules and, in shared libraries,
732 means that the declared entity may be overridden. On Darwin, default
733 visibility means that the declaration is visible to other modules. Default
734 visibility corresponds to "external linkage" in the language.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]735
736 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]737 <dd>Two declarations of an object with hidden visibility refer to the same
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]738 object if they are in the same shared object. Usually, hidden visibility
739 indicates that the symbol will not be placed into the dynamic symbol
740 table, so no other module (executable or shared library) can reference it
741 directly.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]742
743 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]744 <dd>On ELF, protected visibility indicates that the symbol will be placed in
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]745 the dynamic symbol table, but that references within the defining module
746 will bind to the local symbol. That is, the symbol cannot be overridden by
747 another module.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]748</dl>
749
750</div>
751
752<!-- ======================================================================= -->
753<div class="doc_subsection">
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]754 <a name="namedtypes">Named Types</a>
755</div>
756
757<div class="doc_text">
758
759<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]760 it easier to read the IR and make the IR more condensed (particularly when
761 recursive types are involved). An example of a name specification is:</p>
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]762
763<div class="doc_code">
764<pre>
765%mytype = type { %mytype*, i32 }
766</pre>
767</div>
768
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]769<p>You may give a name to any <a href="#typesystem">type</a> except
770 "<a href="t_void">void</a>". Type name aliases may be used anywhere a type
771 is expected with the syntax "%mytype".</p>
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]772
773<p>Note that type names are aliases for the structural type that they indicate,
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]774 and that you can therefore specify multiple names for the same type. This
775 often leads to confusing behavior when dumping out a .ll file. Since LLVM IR
776 uses structural typing, the name is not part of the type. When printing out
777 LLVM IR, the printer will pick <em>one name</em> to render all types of a
778 particular shape. This means that if you have code where two different
779 source types end up having the same LLVM type, that the dumper will sometimes
780 print the "wrong" or unexpected type. This is an important design point and
781 isn't going to change.</p>
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]782
783</div>
784
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]785<!-- ======================================================================= -->
786<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]787 <a name="globalvars">Global Variables</a>
788</div>
789
790<div class="doc_text">
791
792<p>Global variables define regions of memory allocated at compilation time
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]793 instead of run-time. Global variables may optionally be initialized, may
794 have an explicit section to be placed in, and may have an optional explicit
795 alignment specified. A variable may be defined as "thread_local", which
796 means that it will not be shared by threads (each thread will have a
797 separated copy of the variable). A variable may be defined as a global
798 "constant," which indicates that the contents of the variable
799 will <b>never</b> be modified (enabling better optimization, allowing the
800 global data to be placed in the read-only section of an executable, etc).
801 Note that variables that need runtime initialization cannot be marked
802 "constant" as there is a store to the variable.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]803
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]804<p>LLVM explicitly allows <em>declarations</em> of global variables to be marked
805 constant, even if the final definition of the global is not. This capability
806 can be used to enable slightly better optimization of the program, but
807 requires the language definition to guarantee that optimizations based on the
808 'constantness' are valid for the translation units that do not include the
809 definition.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]810
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]811<p>As SSA values, global variables define pointer values that are in scope
812 (i.e. they dominate) all basic blocks in the program. Global variables
813 always define a pointer to their "content" type because they describe a
814 region of memory, and all memory objects in LLVM are accessed through
815 pointers.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]816
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]817<p>A global variable may be declared to reside in a target-specific numbered
818 address space. For targets that support them, address spaces may affect how
819 optimizations are performed and/or what target instructions are used to
820 access the variable. The default address space is zero. The address space
821 qualifier must precede any other attributes.</p>
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]822
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]823<p>LLVM allows an explicit section to be specified for globals. If the target
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]824 supports it, it will emit globals to the section specified.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]825
826<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]827 the alignment is set to zero, the alignment of the global is set by the
828 target to whatever it feels convenient. If an explicit alignment is
829 specified, the global is forced to have at least that much alignment. All
830 alignments must be a power of 2.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]831
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]832<p>For example, the following defines a global in a numbered address space with
833 an initializer, section, and alignment:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]834
835<div class="doc_code">
836<pre>
Dan Gohman21ef02c2009-01-11 00:40:00 +0000[diff] [blame]837@G = addrspace(5) constant float 1.0, section "foo", align 4
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]838</pre>
839</div>
840
841</div>
842
843
844<!-- ======================================================================= -->
845<div class="doc_subsection">
846 <a name="functionstructure">Functions</a>
847</div>
848
849<div class="doc_text">
850
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]851<p>LLVM function definitions consist of the "<tt>define</tt>" keyord, an
852 optional <a href="#linkage">linkage type</a>, an optional
853 <a href="#visibility">visibility style</a>, an optional
854 <a href="#callingconv">calling convention</a>, a return type, an optional
855 <a href="#paramattrs">parameter attribute</a> for the return type, a function
856 name, a (possibly empty) argument list (each with optional
857 <a href="#paramattrs">parameter attributes</a>), optional
858 <a href="#fnattrs">function attributes</a>, an optional section, an optional
859 alignment, an optional <a href="#gc">garbage collector name</a>, an opening
860 curly brace, a list of basic blocks, and a closing curly brace.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]861
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]862<p>LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
863 optional <a href="#linkage">linkage type</a>, an optional
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]864 <a href="#visibility">visibility style</a>, an optional
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]865 <a href="#callingconv">calling convention</a>, a return type, an optional
866 <a href="#paramattrs">parameter attribute</a> for the return type, a function
867 name, a possibly empty list of arguments, an optional alignment, and an
868 optional <a href="#gc">garbage collector name</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]869
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]870<p>A function definition contains a list of basic blocks, forming the CFG
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]871 (Control Flow Graph) for the function. Each basic block may optionally start
872 with a label (giving the basic block a symbol table entry), contains a list
873 of instructions, and ends with a <a href="#terminators">terminator</a>
874 instruction (such as a branch or function return).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]875
876<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]877 executed on entrance to the function, and it is not allowed to have
878 predecessor basic blocks (i.e. there can not be any branches to the entry
879 block of a function). Because the block can have no predecessors, it also
880 cannot have any <a href="#i_phi">PHI nodes</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]881
882<p>LLVM allows an explicit section to be specified for functions. If the target
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]883 supports it, it will emit functions to the section specified.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]884
885<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]886 the alignment is set to zero, the alignment of the function is set by the
887 target to whatever it feels convenient. If an explicit alignment is
888 specified, the function is forced to have at least that much alignment. All
889 alignments must be a power of 2.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]890
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]891<h5>Syntax:</h5>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]892<div class="doc_code">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]893<pre>
Chris Lattner1e5c5cd02008-10-13 16:55:18 +0000[diff] [blame]894define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]895 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
896 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
897 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
898 [<a href="#gc">gc</a>] { ... }
899</pre>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]900</div>
901
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]902</div>
903
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]904<!-- ======================================================================= -->
905<div class="doc_subsection">
906 <a name="aliasstructure">Aliases</a>
907</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]908
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]909<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]910
911<p>Aliases act as "second name" for the aliasee value (which can be either
912 function, global variable, another alias or bitcast of global value). Aliases
913 may have an optional <a href="#linkage">linkage type</a>, and an
914 optional <a href="#visibility">visibility style</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]915
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]916<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]917<div class="doc_code">
918<pre>
Duncan Sandsd7bfabf2008-09-12 20:48:21 +0000[diff] [blame]919@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]920</pre>
921</div>
922
923</div>
924
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]925<!-- ======================================================================= -->
Devang Patela4bb6792010-01-11 19:35:55 +0000[diff] [blame]926<div class="doc_subsection">
927 <a name="namedmetadatastructure">Named Metadata</a>
928</div>
929
930<div class="doc_text">
931
Chris Lattnerd0d96292010-01-15 21:50:19 +0000[diff] [blame]932<p>Named metadata is a collection of metadata. <a href="#metadata">Metadata
933 nodes</a> (but not metadata strings) and null are the only valid operands for
934 a named metadata.</p>
Devang Patela4bb6792010-01-11 19:35:55 +0000[diff] [blame]935
936<h5>Syntax:</h5>
937<div class="doc_code">
938<pre>
939!1 = metadata !{metadata !"one"}
940!name = !{null, !1}
941</pre>
942</div>
943
944</div>
945
946<!-- ======================================================================= -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]947<div class="doc_subsection"><a name="paramattrs">Parameter Attributes</a></div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]948
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]949<div class="doc_text">
950
951<p>The return type and each parameter of a function type may have a set of
952 <i>parameter attributes</i> associated with them. Parameter attributes are
953 used to communicate additional information about the result or parameters of
954 a function. Parameter attributes are considered to be part of the function,
955 not of the function type, so functions with different parameter attributes
956 can have the same function type.</p>
957
958<p>Parameter attributes are simple keywords that follow the type specified. If
959 multiple parameter attributes are needed, they are space separated. For
960 example:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]961
962<div class="doc_code">
963<pre>
Nick Lewycky3022a742009-02-15 23:06:14 +0000[diff] [blame]964declare i32 @printf(i8* noalias nocapture, ...)
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]965declare i32 @atoi(i8 zeroext)
966declare signext i8 @returns_signed_char()
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]967</pre>
968</div>
969
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]970<p>Note that any attributes for the function result (<tt>nounwind</tt>,
971 <tt>readonly</tt>) come immediately after the argument list.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]972
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]973<p>Currently, only the following parameter attributes are defined:</p>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]974
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]975<dl>
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]976 <dt><tt><b>zeroext</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]977 <dd>This indicates to the code generator that the parameter or return value
978 should be zero-extended to a 32-bit value by the caller (for a parameter)
979 or the callee (for a return value).</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]980
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]981 <dt><tt><b>signext</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]982 <dd>This indicates to the code generator that the parameter or return value
983 should be sign-extended to a 32-bit value by the caller (for a parameter)
984 or the callee (for a return value).</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]985
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]986 <dt><tt><b>inreg</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]987 <dd>This indicates that this parameter or return value should be treated in a
988 special target-dependent fashion during while emitting code for a function
989 call or return (usually, by putting it in a register as opposed to memory,
990 though some targets use it to distinguish between two different kinds of
991 registers). Use of this attribute is target-specific.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]992
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]993 <dt><tt><b><a name="byval">byval</a></b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]994 <dd>This indicates that the pointer parameter should really be passed by value
995 to the function. The attribute implies that a hidden copy of the pointee
996 is made between the caller and the callee, so the callee is unable to
997 modify the value in the callee. This attribute is only valid on LLVM
998 pointer arguments. It is generally used to pass structs and arrays by
999 value, but is also valid on pointers to scalars. The copy is considered
1000 to belong to the caller not the callee (for example,
1001 <tt><a href="#readonly">readonly</a></tt> functions should not write to
1002 <tt>byval</tt> parameters). This is not a valid attribute for return
1003 values. The byval attribute also supports specifying an alignment with
1004 the align attribute. This has a target-specific effect on the code
1005 generator that usually indicates a desired alignment for the synthesized
1006 stack slot.</dd>
1007
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1008 <dt><tt><b>sret</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1009 <dd>This indicates that the pointer parameter specifies the address of a
1010 structure that is the return value of the function in the source program.
1011 This pointer must be guaranteed by the caller to be valid: loads and
1012 stores to the structure may be assumed by the callee to not to trap. This
1013 may only be applied to the first parameter. This is not a valid attribute
1014 for return values. </dd>
1015
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1016 <dt><tt><b>noalias</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1017 <dd>This indicates that the pointer does not alias any global or any other
1018 parameter. The caller is responsible for ensuring that this is the
1019 case. On a function return value, <tt>noalias</tt> additionally indicates
1020 that the pointer does not alias any other pointers visible to the
1021 caller. For further details, please see the discussion of the NoAlias
1022 response in
1023 <a href="http://llvm.org/docs/AliasAnalysis.html#MustMayNo">alias
1024 analysis</a>.</dd>
1025
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1026 <dt><tt><b>nocapture</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1027 <dd>This indicates that the callee does not make any copies of the pointer
1028 that outlive the callee itself. This is not a valid attribute for return
1029 values.</dd>
1030
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1031 <dt><tt><b>nest</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1032 <dd>This indicates that the pointer parameter can be excised using the
1033 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
1034 attribute for return values.</dd>
1035</dl>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1036
1037</div>
1038
1039<!-- ======================================================================= -->
1040<div class="doc_subsection">
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]1041 <a name="gc">Garbage Collector Names</a>
1042</div>
1043
1044<div class="doc_text">
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]1045
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1046<p>Each function may specify a garbage collector name, which is simply a
1047 string:</p>
1048
1049<div class="doc_code">
1050<pre>
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1051define void @f() gc "name" { ... }
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1052</pre>
1053</div>
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]1054
1055<p>The compiler declares the supported values of <i>name</i>. Specifying a
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1056 collector which will cause the compiler to alter its output in order to
1057 support the named garbage collection algorithm.</p>
1058
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]1059</div>
1060
1061<!-- ======================================================================= -->
1062<div class="doc_subsection">
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1063 <a name="fnattrs">Function Attributes</a>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1064</div>
1065
1066<div class="doc_text">
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1067
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1068<p>Function attributes are set to communicate additional information about a
1069 function. Function attributes are considered to be part of the function, not
1070 of the function type, so functions with different parameter attributes can
1071 have the same function type.</p>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1072
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1073<p>Function attributes are simple keywords that follow the type specified. If
1074 multiple attributes are needed, they are space separated. For example:</p>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1075
1076<div class="doc_code">
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1077<pre>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1078define void @f() noinline { ... }
1079define void @f() alwaysinline { ... }
1080define void @f() alwaysinline optsize { ... }
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1081define void @f() optsize { ... }
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1082</pre>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1083</div>
1084
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1085<dl>
Charles Davisfaa8f752010-02-12 00:31:15 +0000[diff] [blame]1086 <dt><tt><b>alignstack(&lt;<em>n</em>&gt;)</b></tt></dt>
1087 <dd>This attribute indicates that, when emitting the prologue and epilogue,
1088 the backend should forcibly align the stack pointer. Specify the
1089 desired alignment, which must be a power of two, in parentheses.
1090
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1091 <dt><tt><b>alwaysinline</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1092 <dd>This attribute indicates that the inliner should attempt to inline this
1093 function into callers whenever possible, ignoring any active inlining size
1094 threshold for this caller.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1095
Jakob Stoklund Olesen77180732010-02-06 01:16:28 +0000[diff] [blame]1096 <dt><tt><b>inlinehint</b></tt></dt>
1097 <dd>This attribute indicates that the source code contained a hint that inlining
1098 this function is desirable (such as the "inline" keyword in C/C++). It
1099 is just a hint; it imposes no requirements on the inliner.</dd>
1100
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1101 <dt><tt><b>noinline</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1102 <dd>This attribute indicates that the inliner should never inline this
1103 function in any situation. This attribute may not be used together with
1104 the <tt>alwaysinline</tt> attribute.</dd>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1105
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1106 <dt><tt><b>optsize</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1107 <dd>This attribute suggests that optimization passes and code generator passes
1108 make choices that keep the code size of this function low, and otherwise
1109 do optimizations specifically to reduce code size.</dd>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1110
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1111 <dt><tt><b>noreturn</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1112 <dd>This function attribute indicates that the function never returns
1113 normally. This produces undefined behavior at runtime if the function
1114 ever does dynamically return.</dd>
Bill Wendlingdfaabba2008-11-13 01:02:51 +0000[diff] [blame]1115
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1116 <dt><tt><b>nounwind</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1117 <dd>This function attribute indicates that the function never returns with an
1118 unwind or exceptional control flow. If the function does unwind, its
1119 runtime behavior is undefined.</dd>
Bill Wendlingbe9ec3f2008-11-26 19:07:40 +0000[diff] [blame]1120
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1121 <dt><tt><b>readnone</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1122 <dd>This attribute indicates that the function computes its result (or decides
1123 to unwind an exception) based strictly on its arguments, without
1124 dereferencing any pointer arguments or otherwise accessing any mutable
1125 state (e.g. memory, control registers, etc) visible to caller functions.
1126 It does not write through any pointer arguments
1127 (including <tt><a href="#byval">byval</a></tt> arguments) and never
1128 changes any state visible to callers. This means that it cannot unwind
1129 exceptions by calling the <tt>C++</tt> exception throwing methods, but
1130 could use the <tt>unwind</tt> instruction.</dd>
Devang Patela2f9f412009-06-12 19:45:19 +0000[diff] [blame]1131
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1132 <dt><tt><b><a name="readonly">readonly</a></b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1133 <dd>This attribute indicates that the function does not write through any
1134 pointer arguments (including <tt><a href="#byval">byval</a></tt>
1135 arguments) or otherwise modify any state (e.g. memory, control registers,
1136 etc) visible to caller functions. It may dereference pointer arguments
1137 and read state that may be set in the caller. A readonly function always
1138 returns the same value (or unwinds an exception identically) when called
1139 with the same set of arguments and global state. It cannot unwind an
1140 exception by calling the <tt>C++</tt> exception throwing methods, but may
1141 use the <tt>unwind</tt> instruction.</dd>
Anton Korobeynikovedd7d112009-07-17 18:07:26 +0000[diff] [blame]1142
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1143 <dt><tt><b><a name="ssp">ssp</a></b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1144 <dd>This attribute indicates that the function should emit a stack smashing
1145 protector. It is in the form of a "canary"&mdash;a random value placed on
1146 the stack before the local variables that's checked upon return from the
1147 function to see if it has been overwritten. A heuristic is used to
1148 determine if a function needs stack protectors or not.<br>
1149<br>
1150 If a function that has an <tt>ssp</tt> attribute is inlined into a
1151 function that doesn't have an <tt>ssp</tt> attribute, then the resulting
1152 function will have an <tt>ssp</tt> attribute.</dd>
1153
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1154 <dt><tt><b>sspreq</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1155 <dd>This attribute indicates that the function should <em>always</em> emit a
1156 stack smashing protector. This overrides
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]1157 the <tt><a href="#ssp">ssp</a></tt> function attribute.<br>
1158<br>
1159 If a function that has an <tt>sspreq</tt> attribute is inlined into a
1160 function that doesn't have an <tt>sspreq</tt> attribute or which has
1161 an <tt>ssp</tt> attribute, then the resulting function will have
1162 an <tt>sspreq</tt> attribute.</dd>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1163
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1164 <dt><tt><b>noredzone</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1165 <dd>This attribute indicates that the code generator should not use a red
1166 zone, even if the target-specific ABI normally permits it.</dd>
1167
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1168 <dt><tt><b>noimplicitfloat</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1169 <dd>This attributes disables implicit floating point instructions.</dd>
1170
Bill Wendling614b32b2009-11-02 00:24:16 +0000[diff] [blame]1171 <dt><tt><b>naked</b></tt></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1172 <dd>This attribute disables prologue / epilogue emission for the function.
1173 This can have very system-specific consequences.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1174</dl>
1175
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1176</div>
1177
1178<!-- ======================================================================= -->
1179<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1180 <a name="moduleasm">Module-Level Inline Assembly</a>
1181</div>
1182
1183<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1184
1185<p>Modules may contain "module-level inline asm" blocks, which corresponds to
1186 the GCC "file scope inline asm" blocks. These blocks are internally
1187 concatenated by LLVM and treated as a single unit, but may be separated in
1188 the <tt>.ll</tt> file if desired. The syntax is very simple:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1189
1190<div class="doc_code">
1191<pre>
1192module asm "inline asm code goes here"
1193module asm "more can go here"
1194</pre>
1195</div>
1196
1197<p>The strings can contain any character by escaping non-printable characters.
1198 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]1199 for the number.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1200
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1201<p>The inline asm code is simply printed to the machine code .s file when
1202 assembly code is generated.</p>
1203
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1204</div>
1205
1206<!-- ======================================================================= -->
1207<div class="doc_subsection">
1208 <a name="datalayout">Data Layout</a>
1209</div>
1210
1211<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1212
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1213<p>A module may specify a target specific data layout string that specifies how
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1214 data is to be laid out in memory. The syntax for the data layout is
1215 simply:</p>
1216
1217<div class="doc_code">
1218<pre>
1219target datalayout = "<i>layout specification</i>"
1220</pre>
1221</div>
1222
1223<p>The <i>layout specification</i> consists of a list of specifications
1224 separated by the minus sign character ('-'). Each specification starts with
1225 a letter and may include other information after the letter to define some
1226 aspect of the data layout. The specifications accepted are as follows:</p>
1227
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1228<dl>
1229 <dt><tt>E</tt></dt>
1230 <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]1231 bits with the most significance have the lowest address location.</dd>
1232
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1233 <dt><tt>e</tt></dt>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1234 <dd>Specifies that the target lays out data in little-endian form. That is,
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1235 the bits with the least significance have the lowest address
1236 location.</dd>
1237
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1238 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]1239 <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]1240 <i>preferred</i> alignments. All sizes are in bits. Specifying
1241 the <i>pref</i> alignment is optional. If omitted, the
1242 preceding <tt>:</tt> should be omitted too.</dd>
1243
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1244 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1245 <dd>This specifies the alignment for an integer type of a given bit
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1246 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1247
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1248 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]1249 <dd>This specifies the alignment for a vector type of a given bit
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1250 <i>size</i>.</dd>
1251
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1252 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]1253 <dd>This specifies the alignment for a floating point type of a given bit
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1254 <i>size</i>. The value of <i>size</i> must be either 32 (float) or 64
1255 (double).</dd>
1256
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1257 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1258 <dd>This specifies the alignment for an aggregate type of a given bit
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1259 <i>size</i>.</dd>
1260
Daniel Dunbard88a97b2009-06-08 22:17:53 +0000[diff] [blame]1261 <dt><tt>s<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1262 <dd>This specifies the alignment for a stack object of a given bit
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1263 <i>size</i>.</dd>
Chris Lattnerece41802009-11-07 09:35:34 +0000[diff] [blame]1264
1265 <dt><tt>n<i>size1</i>:<i>size2</i>:<i>size3</i>...</tt></dt>
1266 <dd>This specifies a set of native integer widths for the target CPU
1267 in bits. For example, it might contain "n32" for 32-bit PowerPC,
1268 "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]1269 this set are considered to support most general arithmetic
Chris Lattnerece41802009-11-07 09:35:34 +0000[diff] [blame]1270 operations efficiently.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1271</dl>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1272
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1273<p>When constructing the data layout for a given target, LLVM starts with a
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1274 default set of specifications which are then (possibly) overriden by the
1275 specifications in the <tt>datalayout</tt> keyword. The default specifications
1276 are given in this list:</p>
1277
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1278<ul>
1279 <li><tt>E</tt> - big endian</li>
1280 <li><tt>p:32:64:64</tt> - 32-bit pointers with 64-bit alignment</li>
1281 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1282 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1283 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1284 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1285 <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]1286 alignment of 64-bits</li>
1287 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1288 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1289 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1290 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1291 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
Daniel Dunbard88a97b2009-06-08 22:17:53 +0000[diff] [blame]1292 <li><tt>s0:64:64</tt> - stack objects are 64-bit aligned</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1293</ul>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1294
1295<p>When LLVM is determining the alignment for a given type, it uses the
1296 following rules:</p>
1297
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1298<ol>
1299 <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]1300 specification is used.</li>
1301
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1302 <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]1303 smallest integer type that is larger than the bitwidth of the sought type
1304 is used. If none of the specifications are larger than the bitwidth then
1305 the the largest integer type is used. For example, given the default
1306 specifications above, the i7 type will use the alignment of i8 (next
1307 largest) while both i65 and i256 will use the alignment of i64 (largest
1308 specified).</li>
1309
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1310 <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]1311 largest vector type that is smaller than the sought vector type will be
1312 used as a fall back. This happens because &lt;128 x double&gt; can be
1313 implemented in terms of 64 &lt;2 x double&gt;, for example.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1314</ol>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1315
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1316</div>
1317
Dan Gohman27b47012009-07-27 18:07:55 +0000[diff] [blame]1318<!-- ======================================================================= -->
1319<div class="doc_subsection">
1320 <a name="pointeraliasing">Pointer Aliasing Rules</a>
1321</div>
1322
1323<div class="doc_text">
1324
Andreas Bolka11fbf432009-07-29 00:02:05 +0000[diff] [blame]1325<p>Any memory access must be done through a pointer value associated
Andreas Bolka23bece42009-07-27 20:37:10 +0000[diff] [blame]1326with an address range of the memory access, otherwise the behavior
Dan Gohman27b47012009-07-27 18:07:55 +0000[diff] [blame]1327is undefined. Pointer values are associated with address ranges
1328according to the following rules:</p>
1329
1330<ul>
Andreas Bolka11fbf432009-07-29 00:02:05 +0000[diff] [blame]1331 <li>A pointer value formed from a
1332 <tt><a href="#i_getelementptr">getelementptr</a></tt> instruction
1333 is associated with the addresses associated with the first operand
1334 of the <tt>getelementptr</tt>.</li>
1335 <li>An address of a global variable is associated with the address
Dan Gohman27b47012009-07-27 18:07:55 +0000[diff] [blame]1336 range of the variable's storage.</li>
1337 <li>The result value of an allocation instruction is associated with
1338 the address range of the allocated storage.</li>
1339 <li>A null pointer in the default address-space is associated with
Andreas Bolka11fbf432009-07-29 00:02:05 +0000[diff] [blame]1340 no address.</li>
1341 <li>A pointer value formed by an
1342 <tt><a href="#i_inttoptr">inttoptr</a></tt> is associated with all
1343 address ranges of all pointer values that contribute (directly or
1344 indirectly) to the computation of the pointer's value.</li>
1345 <li>The result value of a
1346 <tt><a href="#i_bitcast">bitcast</a></tt> is associated with all
Dan Gohman27b47012009-07-27 18:07:55 +0000[diff] [blame]1347 addresses associated with the operand of the <tt>bitcast</tt>.</li>
1348 <li>An integer constant other than zero or a pointer value returned
1349 from a function not defined within LLVM may be associated with address
1350 ranges allocated through mechanisms other than those provided by
Andreas Bolka11fbf432009-07-29 00:02:05 +0000[diff] [blame]1351 LLVM. Such ranges shall not overlap with any ranges of addresses
Dan Gohman27b47012009-07-27 18:07:55 +0000[diff] [blame]1352 allocated by mechanisms provided by LLVM.</li>
1353 </ul>
1354
1355<p>LLVM IR does not associate types with memory. The result type of a
Andreas Bolka11fbf432009-07-29 00:02:05 +0000[diff] [blame]1356<tt><a href="#i_load">load</a></tt> merely indicates the size and
1357alignment of the memory from which to load, as well as the
1358interpretation of the value. The first operand of a
1359<tt><a href="#i_store">store</a></tt> similarly only indicates the size
1360and alignment of the store.</p>
Dan Gohman27b47012009-07-27 18:07:55 +0000[diff] [blame]1361
1362<p>Consequently, type-based alias analysis, aka TBAA, aka
1363<tt>-fstrict-aliasing</tt>, is not applicable to general unadorned
1364LLVM IR. <a href="#metadata">Metadata</a> may be used to encode
1365additional information which specialized optimization passes may use
1366to implement type-based alias analysis.</p>
1367
1368</div>
1369
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1370<!-- *********************************************************************** -->
1371<div class="doc_section"> <a name="typesystem">Type System</a> </div>
1372<!-- *********************************************************************** -->
1373
1374<div class="doc_text">
1375
1376<p>The LLVM type system is one of the most important features of the
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1377 intermediate representation. Being typed enables a number of optimizations
1378 to be performed on the intermediate representation directly, without having
1379 to do extra analyses on the side before the transformation. A strong type
1380 system makes it easier to read the generated code and enables novel analyses
1381 and transformations that are not feasible to perform on normal three address
1382 code representations.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1383
1384</div>
1385
1386<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1387<div class="doc_subsection"> <a name="t_classifications">Type
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1388Classifications</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1389
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1390<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1391
1392<p>The types fall into a few useful classifications:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1393
1394<table border="1" cellspacing="0" cellpadding="4">
1395 <tbody>
1396 <tr><th>Classification</th><th>Types</th></tr>
1397 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1398 <td><a href="#t_integer">integer</a></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1399 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
1400 </tr>
1401 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1402 <td><a href="#t_floating">floating point</a></td>
1403 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1404 </tr>
1405 <tr>
1406 <td><a name="t_firstclass">first class</a></td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1407 <td><a href="#t_integer">integer</a>,
1408 <a href="#t_floating">floating point</a>,
1409 <a href="#t_pointer">pointer</a>,
Dan Gohmanf6237db2008-06-18 18:42:13 +0000[diff] [blame]1410 <a href="#t_vector">vector</a>,
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]1411 <a href="#t_struct">structure</a>,
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame^]1412 <a href="#t_union">union</a>,
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]1413 <a href="#t_array">array</a>,
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]1414 <a href="#t_label">label</a>,
1415 <a href="#t_metadata">metadata</a>.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1416 </td>
1417 </tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1418 <tr>
1419 <td><a href="#t_primitive">primitive</a></td>
1420 <td><a href="#t_label">label</a>,
1421 <a href="#t_void">void</a>,
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]1422 <a href="#t_floating">floating point</a>,
1423 <a href="#t_metadata">metadata</a>.</td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1424 </tr>
1425 <tr>
1426 <td><a href="#t_derived">derived</a></td>
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame^]1427 <td><a href="#t_array">array</a>,
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1428 <a href="#t_function">function</a>,
1429 <a href="#t_pointer">pointer</a>,
1430 <a href="#t_struct">structure</a>,
1431 <a href="#t_pstruct">packed structure</a>,
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame^]1432 <a href="#t_union">union</a>,
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1433 <a href="#t_vector">vector</a>,
1434 <a href="#t_opaque">opaque</a>.
Dan Gohman032ba852008-10-14 16:32:04 +0000[diff] [blame]1435 </td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1436 </tr>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1437 </tbody>
1438</table>
1439
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1440<p>The <a href="#t_firstclass">first class</a> types are perhaps the most
1441 important. Values of these types are the only ones which can be produced by
Nick Lewycky244cf482009-09-27 00:45:11 +0000[diff] [blame]1442 instructions.</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1443
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1444</div>
1445
1446<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1447<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1448
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1449<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1450
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1451<p>The primitive types are the fundamental building blocks of the LLVM
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1452 system.</p>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1453
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1454</div>
1455
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1456<!-- _______________________________________________________________________ -->
Nick Lewycky244cf482009-09-27 00:45:11 +0000[diff] [blame]1457<div class="doc_subsubsection"> <a name="t_integer">Integer Type</a> </div>
1458
1459<div class="doc_text">
1460
1461<h5>Overview:</h5>
1462<p>The integer type is a very simple type that simply specifies an arbitrary
1463 bit width for the integer type desired. Any bit width from 1 bit to
1464 2<sup>23</sup>-1 (about 8 million) can be specified.</p>
1465
1466<h5>Syntax:</h5>
1467<pre>
1468 iN
1469</pre>
1470
1471<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1472 value.</p>
1473
1474<h5>Examples:</h5>
1475<table class="layout">
1476 <tr class="layout">
1477 <td class="left"><tt>i1</tt></td>
1478 <td class="left">a single-bit integer.</td>
1479 </tr>
1480 <tr class="layout">
1481 <td class="left"><tt>i32</tt></td>
1482 <td class="left">a 32-bit integer.</td>
1483 </tr>
1484 <tr class="layout">
1485 <td class="left"><tt>i1942652</tt></td>
1486 <td class="left">a really big integer of over 1 million bits.</td>
1487 </tr>
1488</table>
1489
Nick Lewycky244cf482009-09-27 00:45:11 +0000[diff] [blame]1490</div>
1491
1492<!-- _______________________________________________________________________ -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1493<div class="doc_subsubsection"> <a name="t_floating">Floating Point Types</a> </div>
1494
1495<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1496
1497<table>
1498 <tbody>
1499 <tr><th>Type</th><th>Description</th></tr>
1500 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1501 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1502 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1503 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1504 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1505 </tbody>
1506</table>
1507
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1508</div>
1509
1510<!-- _______________________________________________________________________ -->
1511<div class="doc_subsubsection"> <a name="t_void">Void Type</a> </div>
1512
1513<div class="doc_text">
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]1514
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1515<h5>Overview:</h5>
1516<p>The void type does not represent any value and has no size.</p>
1517
1518<h5>Syntax:</h5>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1519<pre>
1520 void
1521</pre>
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]1522
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1523</div>
1524
1525<!-- _______________________________________________________________________ -->
1526<div class="doc_subsubsection"> <a name="t_label">Label Type</a> </div>
1527
1528<div class="doc_text">
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]1529
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1530<h5>Overview:</h5>
1531<p>The label type represents code labels.</p>
1532
1533<h5>Syntax:</h5>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1534<pre>
1535 label
1536</pre>
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]1537
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1538</div>
1539
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]1540<!-- _______________________________________________________________________ -->
1541<div class="doc_subsubsection"> <a name="t_metadata">Metadata Type</a> </div>
1542
1543<div class="doc_text">
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]1544
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]1545<h5>Overview:</h5>
Nick Lewyckyeb021882009-09-27 23:27:42 +0000[diff] [blame]1546<p>The metadata type represents embedded metadata. No derived types may be
1547 created from metadata except for <a href="#t_function">function</a>
1548 arguments.
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]1549
1550<h5>Syntax:</h5>
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]1551<pre>
1552 metadata
1553</pre>
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]1554
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]1555</div>
1556
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1557
1558<!-- ======================================================================= -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1559<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
1560
1561<div class="doc_text">
1562
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1563<p>The real power in LLVM comes from the derived types in the system. This is
1564 what allows a programmer to represent arrays, functions, pointers, and other
Nick Lewycky244cf482009-09-27 00:45:11 +0000[diff] [blame]1565 useful types. Each of these types contain one or more element types which
1566 may be a primitive type, or another derived type. For example, it is
1567 possible to have a two dimensional array, using an array as the element type
1568 of another array.</p>
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1569
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame^]1570
1571</div>
1572
1573<!-- _______________________________________________________________________ -->
1574<div class="doc_subsubsection"> <a name="t_aggregate">Aggregate Types</a> </div>
1575
1576<div class="doc_text">
1577
1578<p>Aggregate Types are a subset of derived types that can contain multiple
1579 member types. <a href="#t_array">Arrays</a>,
1580 <a href="#t_struct">structs</a>, <a href="#t_vector">vectors</a> and
1581 <a href="#t_union">unions</a> are aggregate types.</p>
1582
1583</div>
1584
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1585</div>
1586
1587<!-- _______________________________________________________________________ -->
1588<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
1589
1590<div class="doc_text">
1591
1592<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1593<p>The array type is a very simple derived type that arranges elements
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1594 sequentially in memory. The array type requires a size (number of elements)
1595 and an underlying data type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1596
1597<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1598<pre>
1599 [&lt;# elements&gt; x &lt;elementtype&gt;]
1600</pre>
1601
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1602<p>The number of elements is a constant integer value; <tt>elementtype</tt> may
1603 be any type with a size.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1604
1605<h5>Examples:</h5>
1606<table class="layout">
1607 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1608 <td class="left"><tt>[40 x i32]</tt></td>
1609 <td class="left">Array of 40 32-bit integer values.</td>
1610 </tr>
1611 <tr class="layout">
1612 <td class="left"><tt>[41 x i32]</tt></td>
1613 <td class="left">Array of 41 32-bit integer values.</td>
1614 </tr>
1615 <tr class="layout">
1616 <td class="left"><tt>[4 x i8]</tt></td>
1617 <td class="left">Array of 4 8-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1618 </tr>
1619</table>
1620<p>Here are some examples of multidimensional arrays:</p>
1621<table class="layout">
1622 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1623 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1624 <td class="left">3x4 array of 32-bit integer values.</td>
1625 </tr>
1626 <tr class="layout">
1627 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1628 <td class="left">12x10 array of single precision floating point values.</td>
1629 </tr>
1630 <tr class="layout">
1631 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1632 <td class="left">2x3x4 array of 16-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1633 </tr>
1634</table>
1635
Dan Gohman87ec5ad2009-11-09 19:01:53 +0000[diff] [blame]1636<p>There is no restriction on indexing beyond the end of the array implied by
1637 a static type (though there are restrictions on indexing beyond the bounds
1638 of an allocated object in some cases). This means that single-dimension
1639 'variable sized array' addressing can be implemented in LLVM with a zero
1640 length array type. An implementation of 'pascal style arrays' in LLVM could
1641 use the type "<tt>{ i32, [0 x float]}</tt>", for example.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1642
1643</div>
1644
1645<!-- _______________________________________________________________________ -->
1646<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1647
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1648<div class="doc_text">
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1649
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1650<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1651<p>The function type can be thought of as a function signature. It consists of
1652 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^]1653 function type is a scalar type, a void type, a struct type, or a union
1654 type. If the return type is a struct type then all struct elements must be
1655 of first class types, and the struct must have at least one element.</p>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1656
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1657<h5>Syntax:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1658<pre>
Nick Lewycky3b3b54d2009-09-27 07:55:32 +0000[diff] [blame]1659 &lt;returntype&gt; (&lt;parameter list&gt;)
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1660</pre>
1661
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1662<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]1663 specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1664 which indicates that the function takes a variable number of arguments.
1665 Variable argument functions can access their arguments with
1666 the <a href="#int_varargs">variable argument handling intrinsic</a>
Nick Lewycky3b3b54d2009-09-27 07:55:32 +0000[diff] [blame]1667 functions. '<tt>&lt;returntype&gt;</tt>' is a any type except
Nick Lewyckyeb021882009-09-27 23:27:42 +0000[diff] [blame]1668 <a href="#t_label">label</a>.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1669
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1670<h5>Examples:</h5>
1671<table class="layout">
1672 <tr class="layout">
1673 <td class="left"><tt>i32 (i32)</tt></td>
1674 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
1675 </td>
1676 </tr><tr class="layout">
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]1677 <td class="left"><tt>float&nbsp;(i16&nbsp;signext,&nbsp;i32&nbsp;*)&nbsp;*
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1678 </tt></td>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]1679 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
1680 an <tt>i16</tt> that should be sign extended and a
1681 <a href="#t_pointer">pointer</a> to <tt>i32</tt>, returning
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1682 <tt>float</tt>.
1683 </td>
1684 </tr><tr class="layout">
1685 <td class="left"><tt>i32 (i8*, ...)</tt></td>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]1686 <td class="left">A vararg function that takes at least one
1687 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1688 which returns an integer. This is the signature for <tt>printf</tt> in
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1689 LLVM.
1690 </td>
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1691 </tr><tr class="layout">
1692 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Nick Lewycky3b3b54d2009-09-27 07:55:32 +0000[diff] [blame]1693 <td class="left">A function taking an <tt>i32</tt>, returning a
1694 <a href="#t_struct">structure</a> containing two <tt>i32</tt> values
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1695 </td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1696 </tr>
1697</table>
1698
1699</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1700
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1701<!-- _______________________________________________________________________ -->
1702<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1703
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1704<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1705
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1706<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1707<p>The structure type is used to represent a collection of data members together
1708 in memory. The packing of the field types is defined to match the ABI of the
1709 underlying processor. The elements of a structure may be any type that has a
1710 size.</p>
1711
Jeffrey Yasskine05d7732010-01-11 19:19:26 +0000[diff] [blame]1712<p>Structures in memory are accessed using '<tt><a href="#i_load">load</a></tt>'
1713 and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field
1714 with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
1715 Structures in registers are accessed using the
1716 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' and
1717 '<tt><a href="#i_insertvalue">insertvalue</a></tt>' instructions.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1718<h5>Syntax:</h5>
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]1719<pre>
1720 { &lt;type list&gt; }
1721</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1722
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1723<h5>Examples:</h5>
1724<table class="layout">
1725 <tr class="layout">
1726 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1727 <td class="left">A triple of three <tt>i32</tt> values</td>
1728 </tr><tr class="layout">
1729 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1730 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1731 second element is a <a href="#t_pointer">pointer</a> to a
1732 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1733 an <tt>i32</tt>.</td>
1734 </tr>
1735</table>
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1736
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1737</div>
1738
1739<!-- _______________________________________________________________________ -->
1740<div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
1741</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1742
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1743<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1744
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1745<h5>Overview:</h5>
1746<p>The packed structure type is used to represent a collection of data members
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1747 together in memory. There is no padding between fields. Further, the
1748 alignment of a packed structure is 1 byte. The elements of a packed
1749 structure may be any type that has a size.</p>
1750
1751<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt> and
1752 '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field with
1753 the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
1754
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1755<h5>Syntax:</h5>
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]1756<pre>
1757 &lt; { &lt;type list&gt; } &gt;
1758</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1759
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1760<h5>Examples:</h5>
1761<table class="layout">
1762 <tr class="layout">
1763 <td class="left"><tt>&lt; { i32, i32, i32 } &gt;</tt></td>
1764 <td class="left">A triple of three <tt>i32</tt> values</td>
1765 </tr><tr class="layout">
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1766 <td class="left">
1767<tt>&lt;&nbsp;{&nbsp;float,&nbsp;i32&nbsp;(i32)*&nbsp;}&nbsp;&gt;</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1768 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1769 second element is a <a href="#t_pointer">pointer</a> to a
1770 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1771 an <tt>i32</tt>.</td>
1772 </tr>
1773</table>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1774
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1775</div>
1776
1777<!-- _______________________________________________________________________ -->
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame^]1778<div class="doc_subsubsection"> <a name="t_union">Union Type</a> </div>
1779
1780<div class="doc_text">
1781
1782<h5>Overview:</h5>
1783<p>A union type describes an object with size and alignment suitable for
1784 an object of any one of a given set of types (also known as an "untagged"
1785 union). It is similar in concept and usage to a
1786 <a href="#t_struct">struct</a>, except that all members of the union
1787 have an offset of zero. The elements of a union may be any type that has a
1788 size. Unions must have at least one member - empty unions are not allowed.
1789 </p>
1790
1791<p>The size of the union as a whole will be the size of its largest member,
1792 and the alignment requirements of the union as a whole will be the largest
1793 alignment requirement of any member.</p>
1794
1795<p>Unions members are accessed using '<tt><a href="#i_load">load</a></tt> and
1796 '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field with
1797 the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
1798 Since all members are at offset zero, the getelementptr instruction does
1799 not affect the address, only the type of the resulting pointer.</p>
1800
1801<h5>Syntax:</h5>
1802<pre>
1803 union { &lt;type list&gt; }
1804</pre>
1805
1806<h5>Examples:</h5>
1807<table class="layout">
1808 <tr class="layout">
1809 <td class="left"><tt>union { i32, i32*, float }</tt></td>
1810 <td class="left">A union of three types: an <tt>i32</tt>, a pointer to
1811 an <tt>i32</tt>, and a <tt>float</tt>.</td>
1812 </tr><tr class="layout">
1813 <td class="left">
1814 <tt>union {&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1815 <td class="left">A union, where the first element is a <tt>float</tt> and the
1816 second element is a <a href="#t_pointer">pointer</a> to a
1817 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1818 an <tt>i32</tt>.</td>
1819 </tr>
1820</table>
1821
1822</div>
1823
1824<!-- _______________________________________________________________________ -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1825<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
Chris Lattner96edbd32009-02-08 19:53:29 +0000[diff] [blame]1826
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1827<div class="doc_text">
1828
1829<h5>Overview:</h5>
1830<p>As in many languages, the pointer type represents a pointer or reference to
1831 another object, which must live in memory. Pointer types may have an optional
1832 address space attribute defining the target-specific numbered address space
1833 where the pointed-to object resides. The default address space is zero.</p>
1834
1835<p>Note that LLVM does not permit pointers to void (<tt>void*</tt>) nor does it
1836 permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</tt> instead.</p>
Chris Lattner96edbd32009-02-08 19:53:29 +0000[diff] [blame]1837
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1838<h5>Syntax:</h5>
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]1839<pre>
1840 &lt;type&gt; *
1841</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1842
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1843<h5>Examples:</h5>
1844<table class="layout">
1845 <tr class="layout">
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]1846 <td class="left"><tt>[4 x i32]*</tt></td>
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1847 <td class="left">A <a href="#t_pointer">pointer</a> to <a
1848 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
1849 </tr>
1850 <tr class="layout">
1851 <td class="left"><tt>i32 (i32 *) *</tt></td>
1852 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1853 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1854 <tt>i32</tt>.</td>
1855 </tr>
1856 <tr class="layout">
1857 <td class="left"><tt>i32 addrspace(5)*</tt></td>
1858 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
1859 that resides in address space #5.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1860 </tr>
1861</table>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1862
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1863</div>
1864
1865<!-- _______________________________________________________________________ -->
1866<div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1867
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1868<div class="doc_text">
1869
1870<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1871<p>A vector type is a simple derived type that represents a vector of elements.
1872 Vector types are used when multiple primitive data are operated in parallel
1873 using a single instruction (SIMD). A vector type requires a size (number of
Duncan Sands87598b02009-11-27 13:38:03 +0000[diff] [blame]1874 elements) and an underlying primitive data type. Vector types are considered
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1875 <a href="#t_firstclass">first class</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1876
1877<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1878<pre>
1879 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1880</pre>
1881
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1882<p>The number of elements is a constant integer value; elementtype may be any
1883 integer or floating point type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1884
1885<h5>Examples:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1886<table class="layout">
1887 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1888 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
1889 <td class="left">Vector of 4 32-bit integer values.</td>
1890 </tr>
1891 <tr class="layout">
1892 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
1893 <td class="left">Vector of 8 32-bit floating-point values.</td>
1894 </tr>
1895 <tr class="layout">
1896 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
1897 <td class="left">Vector of 2 64-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1898 </tr>
1899</table>
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1900
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1901</div>
1902
1903<!-- _______________________________________________________________________ -->
1904<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1905<div class="doc_text">
1906
1907<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1908<p>Opaque types are used to represent unknown types in the system. This
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1909 corresponds (for example) to the C notion of a forward declared structure
1910 type. In LLVM, opaque types can eventually be resolved to any type (not just
1911 a structure type).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1912
1913<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1914<pre>
1915 opaque
1916</pre>
1917
1918<h5>Examples:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1919<table class="layout">
1920 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1921 <td class="left"><tt>opaque</tt></td>
1922 <td class="left">An opaque type.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1923 </tr>
1924</table>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1925
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1926</div>
1927
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]1928<!-- ======================================================================= -->
1929<div class="doc_subsection">
1930 <a name="t_uprefs">Type Up-references</a>
1931</div>
1932
1933<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1934
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]1935<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1936<p>An "up reference" allows you to refer to a lexically enclosing type without
1937 requiring it to have a name. For instance, a structure declaration may
1938 contain a pointer to any of the types it is lexically a member of. Example
1939 of up references (with their equivalent as named type declarations)
1940 include:</p>
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]1941
1942<pre>
Chris Lattner5ad632d2009-02-09 10:00:56 +0000[diff] [blame]1943 { \2 * } %x = type { %x* }
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]1944 { \2 }* %y = type { %y }*
1945 \1* %z = type %z*
1946</pre>
1947
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1948<p>An up reference is needed by the asmprinter for printing out cyclic types
1949 when there is no declared name for a type in the cycle. Because the
1950 asmprinter does not want to print out an infinite type string, it needs a
1951 syntax to handle recursive types that have no names (all names are optional
1952 in llvm IR).</p>
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]1953
1954<h5>Syntax:</h5>
1955<pre>
1956 \&lt;level&gt;
1957</pre>
1958
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1959<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]1960
1961<h5>Examples:</h5>
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]1962<table class="layout">
1963 <tr class="layout">
1964 <td class="left"><tt>\1*</tt></td>
1965 <td class="left">Self-referential pointer.</td>
1966 </tr>
1967 <tr class="layout">
1968 <td class="left"><tt>{ { \3*, i8 }, i32 }</tt></td>
1969 <td class="left">Recursive structure where the upref refers to the out-most
1970 structure.</td>
1971 </tr>
1972</table>
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]1973
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1974</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1975
1976<!-- *********************************************************************** -->
1977<div class="doc_section"> <a name="constants">Constants</a> </div>
1978<!-- *********************************************************************** -->
1979
1980<div class="doc_text">
1981
1982<p>LLVM has several different basic types of constants. This section describes
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1983 them all and their syntax.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1984
1985</div>
1986
1987<!-- ======================================================================= -->
1988<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
1989
1990<div class="doc_text">
1991
1992<dl>
1993 <dt><b>Boolean constants</b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1994 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
Nick Lewycky244cf482009-09-27 00:45:11 +0000[diff] [blame]1995 constants of the <tt><a href="#t_integer">i1</a></tt> type.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1996
1997 <dt><b>Integer constants</b></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]1998 <dd>Standard integers (such as '4') are constants of
1999 the <a href="#t_integer">integer</a> type. Negative numbers may be used
2000 with integer types.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2001
2002 <dt><b>Floating point constants</b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2003 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2004 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
2005 notation (see below). The assembler requires the exact decimal value of a
2006 floating-point constant. For example, the assembler accepts 1.25 but
2007 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
2008 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2009
2010 <dt><b>Null pointer constants</b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2011 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2012 and must be of <a href="#t_pointer">pointer type</a>.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2013</dl>
2014
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2015<p>The one non-intuitive notation for constants is the hexadecimal form of
2016 floating point constants. For example, the form '<tt>double
2017 0x432ff973cafa8000</tt>' is equivalent to (but harder to read than)
2018 '<tt>double 4.5e+15</tt>'. The only time hexadecimal floating point
2019 constants are required (and the only time that they are generated by the
2020 disassembler) is when a floating point constant must be emitted but it cannot
2021 be represented as a decimal floating point number in a reasonable number of
2022 digits. For example, NaN's, infinities, and other special values are
2023 represented in their IEEE hexadecimal format so that assembly and disassembly
2024 do not cause any bits to change in the constants.</p>
2025
Dale Johannesenf82a52f2009-02-11 22:14:51 +0000[diff] [blame]2026<p>When using the hexadecimal form, constants of types float and double are
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2027 represented using the 16-digit form shown above (which matches the IEEE754
2028 representation for double); float values must, however, be exactly
2029 representable as IEE754 single precision. Hexadecimal format is always used
2030 for long double, and there are three forms of long double. The 80-bit format
2031 used by x86 is represented as <tt>0xK</tt> followed by 20 hexadecimal digits.
2032 The 128-bit format used by PowerPC (two adjacent doubles) is represented
2033 by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit format
2034 is represented by <tt>0xL</tt> followed by 32 hexadecimal digits; no
2035 currently supported target uses this format. Long doubles will only work if
2036 they match the long double format on your target. All hexadecimal formats
2037 are big-endian (sign bit at the left).</p>
2038
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2039</div>
2040
2041<!-- ======================================================================= -->
Chris Lattner97063852009-02-28 18:32:25 +0000[diff] [blame]2042<div class="doc_subsection">
Bill Wendling1a2630a2009-07-20 02:32:41 +0000[diff] [blame]2043<a name="aggregateconstants"></a> <!-- old anchor -->
2044<a name="complexconstants">Complex Constants</a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2045</div>
2046
2047<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2048
Chris Lattner97063852009-02-28 18:32:25 +0000[diff] [blame]2049<p>Complex constants are a (potentially recursive) combination of simple
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2050 constants and smaller complex constants.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2051
2052<dl>
2053 <dt><b>Structure constants</b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2054 <dd>Structure constants are represented with notation similar to structure
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2055 type definitions (a comma separated list of elements, surrounded by braces
2056 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
2057 where "<tt>@G</tt>" is declared as "<tt>@G = external global i32</tt>".
2058 Structure constants must have <a href="#t_struct">structure type</a>, and
2059 the number and types of elements must match those specified by the
2060 type.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2061
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame^]2062 <dt><b>Union constants</b></dt>
2063 <dd>Union constants are represented with notation similar to a structure with
2064 a single element - that is, a single typed element surrounded
2065 by braces (<tt>{}</tt>)). For example: "<tt>{ i32 4 }</tt>". The
2066 <a href="#t_union">union type</a> can be initialized with a single-element
2067 struct as long as the type of the struct element matches the type of
2068 one of the union members.</dd>
2069
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2070 <dt><b>Array constants</b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2071 <dd>Array constants are represented with notation similar to array type
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2072 definitions (a comma separated list of elements, surrounded by square
2073 brackets (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74
2074 ]</tt>". Array constants must have <a href="#t_array">array type</a>, and
2075 the number and types of elements must match those specified by the
2076 type.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2077
2078 <dt><b>Vector constants</b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2079 <dd>Vector constants are represented with notation similar to vector type
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2080 definitions (a comma separated list of elements, surrounded by
2081 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32
2082 42, i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must
2083 have <a href="#t_vector">vector type</a>, and the number and types of
2084 elements must match those specified by the type.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2085
2086 <dt><b>Zero initialization</b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2087 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame^]2088 value to zero of <em>any</em> type, including scalar and
2089 <a href="#t_aggregate">aggregate</a> types.
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2090 This is often used to avoid having to print large zero initializers
2091 (e.g. for large arrays) and is always exactly equivalent to using explicit
2092 zero initializers.</dd>
Nick Lewycky4dcf8102009-04-04 07:22:01 +0000[diff] [blame]2093
2094 <dt><b>Metadata node</b></dt>
Nick Lewyckyf122c7e2009-05-30 16:08:30 +0000[diff] [blame]2095 <dd>A metadata node is a structure-like constant with
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2096 <a href="#t_metadata">metadata type</a>. For example: "<tt>metadata !{
2097 i32 0, metadata !"test" }</tt>". Unlike other constants that are meant to
2098 be interpreted as part of the instruction stream, metadata is a place to
2099 attach additional information such as debug info.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2100</dl>
2101
2102</div>
2103
2104<!-- ======================================================================= -->
2105<div class="doc_subsection">
2106 <a name="globalconstants">Global Variable and Function Addresses</a>
2107</div>
2108
2109<div class="doc_text">
2110
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2111<p>The addresses of <a href="#globalvars">global variables</a>
2112 and <a href="#functionstructure">functions</a> are always implicitly valid
2113 (link-time) constants. These constants are explicitly referenced when
2114 the <a href="#identifiers">identifier for the global</a> is used and always
2115 have <a href="#t_pointer">pointer</a> type. For example, the following is a
2116 legal LLVM file:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2117
2118<div class="doc_code">
2119<pre>
2120@X = global i32 17
2121@Y = global i32 42
2122@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
2123</pre>
2124</div>
2125
2126</div>
2127
2128<!-- ======================================================================= -->
2129<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
2130<div class="doc_text">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2131
Chris Lattner3d72cd82009-09-07 22:52:39 +0000[diff] [blame]2132<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]2133 indicates that the user of the value may receive an unspecified bit-pattern.
Chris Lattner3d72cd82009-09-07 22:52:39 +0000[diff] [blame]2134 Undefined values may be of any type (other than label or void) and be used
2135 anywhere a constant is permitted.</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2136
Chris Lattner9d1e4fa2009-09-11 01:49:31 +0000[diff] [blame]2137<p>Undefined values are useful because they indicate to the compiler that the
Chris Lattner3d72cd82009-09-07 22:52:39 +0000[diff] [blame]2138 program is well defined no matter what value is used. This gives the
2139 compiler more freedom to optimize. Here are some examples of (potentially
2140 surprising) transformations that are valid (in pseudo IR):</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2141
Chris Lattner3d72cd82009-09-07 22:52:39 +0000[diff] [blame]2142
2143<div class="doc_code">
2144<pre>
2145 %A = add %X, undef
2146 %B = sub %X, undef
2147 %C = xor %X, undef
2148Safe:
2149 %A = undef
2150 %B = undef
2151 %C = undef
2152</pre>
2153</div>
2154
2155<p>This is safe because all of the output bits are affected by the undef bits.
2156Any output bit can have a zero or one depending on the input bits.</p>
2157
2158<div class="doc_code">
2159<pre>
2160 %A = or %X, undef
2161 %B = and %X, undef
2162Safe:
2163 %A = -1
2164 %B = 0
2165Unsafe:
2166 %A = undef
2167 %B = undef
2168</pre>
2169</div>
2170
2171<p>These logical operations have bits that are not always affected by the input.
2172For example, if "%X" has a zero bit, then the output of the 'and' operation will
2173always be a zero, no matter what the corresponding bit from the undef is. As
Chris Lattner9d1e4fa2009-09-11 01:49:31 +0000[diff] [blame]2174such, 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]2175However, it is safe to assume that all bits of the undef could be 0, and
2176optimize the and to 0. Likewise, it is safe to assume that all the bits of
2177the 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]2178-1.</p>
Chris Lattner3d72cd82009-09-07 22:52:39 +0000[diff] [blame]2179
2180<div class="doc_code">
2181<pre>
2182 %A = select undef, %X, %Y
2183 %B = select undef, 42, %Y
2184 %C = select %X, %Y, undef
2185Safe:
2186 %A = %X (or %Y)
2187 %B = 42 (or %Y)
2188 %C = %Y
2189Unsafe:
2190 %A = undef
2191 %B = undef
2192 %C = undef
2193</pre>
2194</div>
2195
2196<p>This set of examples show that undefined select (and conditional branch)
2197conditions can go "either way" but they have to come from one of the two
2198operands. In the %A example, if %X and %Y were both known to have a clear low
2199bit, then %A would have to have a cleared low bit. However, in the %C example,
2200the optimizer is allowed to assume that the undef operand could be the same as
2201%Y, allowing the whole select to be eliminated.</p>
2202
2203
2204<div class="doc_code">
2205<pre>
2206 %A = xor undef, undef
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]2207
Chris Lattner3d72cd82009-09-07 22:52:39 +0000[diff] [blame]2208 %B = undef
2209 %C = xor %B, %B
2210
2211 %D = undef
2212 %E = icmp lt %D, 4
2213 %F = icmp gte %D, 4
2214
2215Safe:
2216 %A = undef
2217 %B = undef
2218 %C = undef
2219 %D = undef
2220 %E = undef
2221 %F = undef
2222</pre>
2223</div>
2224
2225<p>This example points out that two undef operands are not necessarily the same.
2226This can be surprising to people (and also matches C semantics) where they
2227assume that "X^X" is always zero, even if X is undef. This isn't true for a
2228number of reasons, but the short answer is that an undef "variable" can
2229arbitrarily change its value over its "live range". This is true because the
2230"variable" doesn't actually <em>have a live range</em>. Instead, the value is
2231logically read from arbitrary registers that happen to be around when needed,
Benjamin Kramer5fb9d7e2009-10-12 14:46:08 +0000[diff] [blame]2232so the value is not necessarily consistent over time. In fact, %A and %C need
Chris Lattnerc802e482009-09-08 15:13:16 +0000[diff] [blame]2233to have the same semantics or the core LLVM "replace all uses with" concept
Chris Lattner3d72cd82009-09-07 22:52:39 +0000[diff] [blame]2234would not hold.</p>
Chris Lattner466291f2009-09-07 23:33:52 +0000[diff] [blame]2235
2236<div class="doc_code">
2237<pre>
2238 %A = fdiv undef, %X
2239 %B = fdiv %X, undef
2240Safe:
2241 %A = undef
2242b: unreachable
2243</pre>
2244</div>
2245
2246<p>These examples show the crucial difference between an <em>undefined
2247value</em> and <em>undefined behavior</em>. An undefined value (like undef) is
2248allowed to have an arbitrary bit-pattern. This means that the %A operation
2249can be constant folded to undef because the undef could be an SNaN, and fdiv is
2250not (currently) defined on SNaN's. However, in the second example, we can make
2251a more aggressive assumption: because the undef is allowed to be an arbitrary
2252value, 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]2253has <em>undefined behavior</em>, we are allowed to assume that the operation
Chris Lattner466291f2009-09-07 23:33:52 +0000[diff] [blame]2254does not execute at all. This allows us to delete the divide and all code after
2255it: since the undefined operation "can't happen", the optimizer can assume that
2256it occurs in dead code.
2257</p>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]2258
Chris Lattner466291f2009-09-07 23:33:52 +0000[diff] [blame]2259<div class="doc_code">
2260<pre>
2261a: store undef -> %X
2262b: store %X -> undef
2263Safe:
2264a: &lt;deleted&gt;
2265b: unreachable
2266</pre>
2267</div>
2268
2269<p>These examples reiterate the fdiv example: a store "of" an undefined value
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]2270can be assumed to not have any effect: we can assume that the value is
Chris Lattner466291f2009-09-07 23:33:52 +0000[diff] [blame]2271overwritten with bits that happen to match what was already there. However, a
2272store "to" an undefined location could clobber arbitrary memory, therefore, it
2273has undefined behavior.</p>
2274
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2275</div>
2276
2277<!-- ======================================================================= -->
Chris Lattner29246b52009-10-27 21:19:13 +0000[diff] [blame]2278<div class="doc_subsection"><a name="blockaddress">Addresses of Basic
2279 Blocks</a></div>
Chris Lattnerd07c8372009-10-27 21:01:34 +0000[diff] [blame]2280<div class="doc_text">
2281
Chris Lattner620cead2009-11-01 01:27:45 +0000[diff] [blame]2282<p><b><tt>blockaddress(@function, %block)</tt></b></p>
Chris Lattnerd07c8372009-10-27 21:01:34 +0000[diff] [blame]2283
2284<p>The '<tt>blockaddress</tt>' constant computes the address of the specified
Chris Lattner24f2e852009-10-27 21:49:40 +0000[diff] [blame]2285 basic block in the specified function, and always has an i8* type. Taking
Chris Lattner620cead2009-11-01 01:27:45 +0000[diff] [blame]2286 the address of the entry block is illegal.</p>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]2287
Chris Lattnerd07c8372009-10-27 21:01:34 +0000[diff] [blame]2288<p>This value only has defined behavior when used as an operand to the
Chris Lattner4c3800f2009-10-28 00:19:10 +0000[diff] [blame]2289 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>' instruction or for comparisons
Chris Lattnerd07c8372009-10-27 21:01:34 +0000[diff] [blame]2290 against null. Pointer equality tests between labels addresses is undefined
2291 behavior - though, again, comparison against null is ok, and no label is
Chris Lattner29246b52009-10-27 21:19:13 +0000[diff] [blame]2292 equal to the null pointer. This may also be passed around as an opaque
2293 pointer sized value as long as the bits are not inspected. This allows
Chris Lattner0bae7b32009-10-27 21:44:20 +0000[diff] [blame]2294 <tt>ptrtoint</tt> and arithmetic to be performed on these values so long as
Chris Lattner4c3800f2009-10-28 00:19:10 +0000[diff] [blame]2295 the original value is reconstituted before the <tt>indirectbr</tt>.</p>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]2296
Chris Lattner29246b52009-10-27 21:19:13 +0000[diff] [blame]2297<p>Finally, some targets may provide defined semantics when
Chris Lattnerd07c8372009-10-27 21:01:34 +0000[diff] [blame]2298 using the value as the operand to an inline assembly, but that is target
2299 specific.
2300 </p>
2301
2302</div>
2303
2304
2305<!-- ======================================================================= -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2306<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
2307</div>
2308
2309<div class="doc_text">
2310
2311<p>Constant expressions are used to allow expressions involving other constants
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2312 to be used as constants. Constant expressions may be of
2313 any <a href="#t_firstclass">first class</a> type and may involve any LLVM
2314 operation that does not have side effects (e.g. load and call are not
2315 supported). The following is the syntax for constant expressions:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2316
2317<dl>
2318 <dt><b><tt>trunc ( CST to TYPE )</tt></b></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2319 <dd>Truncate a constant to another type. The bit size of CST must be larger
2320 than the bit size of TYPE. Both types must be integers.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2321
2322 <dt><b><tt>zext ( CST to TYPE )</tt></b></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2323 <dd>Zero extend a constant to another type. The bit size of CST must be
2324 smaller or equal to the bit size of TYPE. Both types must be
2325 integers.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2326
2327 <dt><b><tt>sext ( CST to TYPE )</tt></b></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2328 <dd>Sign extend a constant to another type. The bit size of CST must be
2329 smaller or equal to the bit size of TYPE. Both types must be
2330 integers.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2331
2332 <dt><b><tt>fptrunc ( CST to TYPE )</tt></b></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2333 <dd>Truncate a floating point constant to another floating point type. The
2334 size of CST must be larger than the size of TYPE. Both types must be
2335 floating point.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2336
2337 <dt><b><tt>fpext ( CST to TYPE )</tt></b></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2338 <dd>Floating point extend a constant to another type. The size of CST must be
2339 smaller or equal to the size of TYPE. Both types must be floating
2340 point.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2341
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]2342 <dt><b><tt>fptoui ( CST to TYPE )</tt></b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2343 <dd>Convert a floating point constant to the corresponding unsigned integer
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2344 constant. TYPE must be a scalar or vector integer type. CST must be of
2345 scalar or vector floating point type. Both CST and TYPE must be scalars,
2346 or vectors of the same number of elements. If the value won't fit in the
2347 integer type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2348
2349 <dt><b><tt>fptosi ( CST to TYPE )</tt></b></dt>
2350 <dd>Convert a floating point constant to the corresponding signed integer
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2351 constant. TYPE must be a scalar or vector integer type. CST must be of
2352 scalar or vector floating point type. Both CST and TYPE must be scalars,
2353 or vectors of the same number of elements. If the value won't fit in the
2354 integer type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2355
2356 <dt><b><tt>uitofp ( CST to TYPE )</tt></b></dt>
2357 <dd>Convert an unsigned integer constant to the corresponding floating point
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2358 constant. TYPE must be a scalar or vector floating point type. CST must be
2359 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2360 vectors of the same number of elements. If the value won't fit in the
2361 floating point type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2362
2363 <dt><b><tt>sitofp ( CST to TYPE )</tt></b></dt>
2364 <dd>Convert a signed integer constant to the corresponding floating point
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2365 constant. TYPE must be a scalar or vector floating point type. CST must be
2366 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2367 vectors of the same number of elements. If the value won't fit in the
2368 floating point type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2369
2370 <dt><b><tt>ptrtoint ( CST to TYPE )</tt></b></dt>
2371 <dd>Convert a pointer typed constant to the corresponding integer constant
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2372 <tt>TYPE</tt> must be an integer type. <tt>CST</tt> must be of pointer
2373 type. The <tt>CST</tt> value is zero extended, truncated, or unchanged to
2374 make it fit in <tt>TYPE</tt>.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2375
2376 <dt><b><tt>inttoptr ( CST to TYPE )</tt></b></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2377 <dd>Convert a integer constant to a pointer constant. TYPE must be a pointer
2378 type. CST must be of integer type. The CST value is zero extended,
2379 truncated, or unchanged to make it fit in a pointer size. This one is
2380 <i>really</i> dangerous!</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2381
2382 <dt><b><tt>bitcast ( CST to TYPE )</tt></b></dt>
Chris Lattner557bc5d2009-02-28 18:27:03 +0000[diff] [blame]2383 <dd>Convert a constant, CST, to another TYPE. The constraints of the operands
2384 are the same as those for the <a href="#i_bitcast">bitcast
2385 instruction</a>.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2386
2387 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
Dan Gohman106b2ae2009-07-27 21:53:46 +0000[diff] [blame]2388 <dt><b><tt>getelementptr inbounds ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2389 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2390 constants. As with the <a href="#i_getelementptr">getelementptr</a>
2391 instruction, the index list may have zero or more indexes, which are
2392 required to make sense for the type of "CSTPTR".</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2393
2394 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2395 <dd>Perform the <a href="#i_select">select operation</a> on constants.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2396
2397 <dt><b><tt>icmp COND ( VAL1, VAL2 )</tt></b></dt>
2398 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
2399
2400 <dt><b><tt>fcmp COND ( VAL1, VAL2 )</tt></b></dt>
2401 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
2402
2403 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2404 <dd>Perform the <a href="#i_extractelement">extractelement operation</a> on
2405 constants.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2406
2407 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2408 <dd>Perform the <a href="#i_insertelement">insertelement operation</a> on
2409 constants.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2410
2411 <dt><b><tt>shufflevector ( VEC1, VEC2, IDXMASK )</tt></b></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2412 <dd>Perform the <a href="#i_shufflevector">shufflevector operation</a> on
2413 constants.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2414
2415 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2416 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
2417 be any of the <a href="#binaryops">binary</a>
2418 or <a href="#bitwiseops">bitwise binary</a> operations. The constraints
2419 on operands are the same as those for the corresponding instruction
2420 (e.g. no bitwise operations on floating point values are allowed).</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2421</dl>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2422
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2423</div>
2424
2425<!-- *********************************************************************** -->
2426<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
2427<!-- *********************************************************************** -->
2428
2429<!-- ======================================================================= -->
2430<div class="doc_subsection">
2431<a name="inlineasm">Inline Assembler Expressions</a>
2432</div>
2433
2434<div class="doc_text">
2435
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2436<p>LLVM supports inline assembler expressions (as opposed
2437 to <a href="#moduleasm"> Module-Level Inline Assembly</a>) through the use of
2438 a special value. This value represents the inline assembler as a string
2439 (containing the instructions to emit), a list of operand constraints (stored
Dale Johannesen648950f2009-10-13 21:56:55 +0000[diff] [blame]2440 as a string), a flag that indicates whether or not the inline asm
Dale Johannesen5ee3e4b2009-10-21 23:28:00 +0000[diff] [blame]2441 expression has side effects, and a flag indicating whether the function
2442 containing the asm needs to align its stack conservatively. An example
2443 inline assembler expression is:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2444
2445<div class="doc_code">
2446<pre>
2447i32 (i32) asm "bswap $0", "=r,r"
2448</pre>
2449</div>
2450
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2451<p>Inline assembler expressions may <b>only</b> be used as the callee operand of
2452 a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we
2453 have:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2454
2455<div class="doc_code">
2456<pre>
2457%X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
2458</pre>
2459</div>
2460
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2461<p>Inline asms with side effects not visible in the constraint list must be
2462 marked as having side effects. This is done through the use of the
2463 '<tt>sideeffect</tt>' keyword, like so:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2464
2465<div class="doc_code">
2466<pre>
2467call void asm sideeffect "eieio", ""()
2468</pre>
2469</div>
2470
Dale Johannesen5ee3e4b2009-10-21 23:28:00 +0000[diff] [blame]2471<p>In some cases inline asms will contain code that will not work unless the
2472 stack is aligned in some way, such as calls or SSE instructions on x86,
2473 yet will not contain code that does that alignment within the asm.
2474 The compiler should make conservative assumptions about what the asm might
2475 contain and should generate its usual stack alignment code in the prologue
2476 if the '<tt>alignstack</tt>' keyword is present:</p>
Dale Johannesen648950f2009-10-13 21:56:55 +0000[diff] [blame]2477
2478<div class="doc_code">
2479<pre>
Dale Johannesen5ee3e4b2009-10-21 23:28:00 +0000[diff] [blame]2480call void asm alignstack "eieio", ""()
Dale Johannesen648950f2009-10-13 21:56:55 +0000[diff] [blame]2481</pre>
2482</div>
2483
2484<p>If both keywords appear the '<tt>sideeffect</tt>' keyword must come
2485 first.</p>
2486
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2487<p>TODO: The format of the asm and constraints string still need to be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2488 documented here. Constraints on what can be done (e.g. duplication, moving,
2489 etc need to be documented). This is probably best done by reference to
2490 another document that covers inline asm from a holistic perspective.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2491
2492</div>
2493
Chris Lattnerd0d96292010-01-15 21:50:19 +0000[diff] [blame]2494<!-- ======================================================================= -->
2495<div class="doc_subsection"><a name="metadata">Metadata Nodes and Metadata
2496 Strings</a>
2497</div>
2498
2499<div class="doc_text">
2500
2501<p>LLVM IR allows metadata to be attached to instructions in the program that
2502 can convey extra information about the code to the optimizers and code
2503 generator. One example application of metadata is source-level debug
2504 information. There are two metadata primitives: strings and nodes. All
2505 metadata has the <tt>metadata</tt> type and is identified in syntax by a
2506 preceding exclamation point ('<tt>!</tt>').</p>
2507
2508<p>A metadata string is a string surrounded by double quotes. It can contain
2509 any character by escaping non-printable characters with "\xx" where "xx" is
2510 the two digit hex code. For example: "<tt>!"test\00"</tt>".</p>
2511
2512<p>Metadata nodes are represented with notation similar to structure constants
2513 (a comma separated list of elements, surrounded by braces and preceded by an
2514 exclamation point). For example: "<tt>!{ metadata !"test\00", i32
2515 10}</tt>". Metadata nodes can have any values as their operand.</p>
2516
2517<p>A <a href="#namedmetadatastructure">named metadata</a> is a collection of
2518 metadata nodes, which can be looked up in the module symbol table. For
2519 example: "<tt>!foo = metadata !{!4, !3}</tt>".
2520
2521</div>
2522
Chris Lattner75c24e02009-07-20 05:55:19 +0000[diff] [blame]2523
2524<!-- *********************************************************************** -->
2525<div class="doc_section">
2526 <a name="intrinsic_globals">Intrinsic Global Variables</a>
2527</div>
2528<!-- *********************************************************************** -->
2529
2530<p>LLVM has a number of "magic" global variables that contain data that affect
2531code generation or other IR semantics. These are documented here. All globals
Chris Lattner1e0e0d12009-07-20 06:14:25 +0000[diff] [blame]2532of this sort should have a section specified as "<tt>llvm.metadata</tt>". This
2533section and all globals that start with "<tt>llvm.</tt>" are reserved for use
2534by LLVM.</p>
Chris Lattner75c24e02009-07-20 05:55:19 +0000[diff] [blame]2535
2536<!-- ======================================================================= -->
2537<div class="doc_subsection">
2538<a name="intg_used">The '<tt>llvm.used</tt>' Global Variable</a>
2539</div>
2540
2541<div class="doc_text">
2542
2543<p>The <tt>@llvm.used</tt> global is an array with i8* element type which has <a
2544href="#linkage_appending">appending linkage</a>. This array contains a list of
2545pointers to global variables and functions which may optionally have a pointer
2546cast formed of bitcast or getelementptr. For example, a legal use of it is:</p>
2547
2548<pre>
2549 @X = global i8 4
2550 @Y = global i32 123
2551
2552 @llvm.used = appending global [2 x i8*] [
2553 i8* @X,
2554 i8* bitcast (i32* @Y to i8*)
2555 ], section "llvm.metadata"
2556</pre>
2557
2558<p>If a global variable appears in the <tt>@llvm.used</tt> list, then the
2559compiler, assembler, and linker are required to treat the symbol as if there is
2560a reference to the global that it cannot see. For example, if a variable has
2561internal linkage and no references other than that from the <tt>@llvm.used</tt>
2562list, it cannot be deleted. This is commonly used to represent references from
2563inline asms and other things the compiler cannot "see", and corresponds to
2564"attribute((used))" in GNU C.</p>
2565
2566<p>On some targets, the code generator must emit a directive to the assembler or
2567object file to prevent the assembler and linker from molesting the symbol.</p>
2568
2569</div>
2570
2571<!-- ======================================================================= -->
2572<div class="doc_subsection">
Chris Lattner1e0e0d12009-07-20 06:14:25 +0000[diff] [blame]2573<a name="intg_compiler_used">The '<tt>llvm.compiler.used</tt>' Global Variable</a>
2574</div>
2575
2576<div class="doc_text">
2577
2578<p>The <tt>@llvm.compiler.used</tt> directive is the same as the
2579<tt>@llvm.used</tt> directive, except that it only prevents the compiler from
2580touching the symbol. On targets that support it, this allows an intelligent
2581linker to optimize references to the symbol without being impeded as it would be
2582by <tt>@llvm.used</tt>.</p>
2583
2584<p>This is a rare construct that should only be used in rare circumstances, and
2585should not be exposed to source languages.</p>
2586
2587</div>
2588
2589<!-- ======================================================================= -->
2590<div class="doc_subsection">
Chris Lattner75c24e02009-07-20 05:55:19 +0000[diff] [blame]2591<a name="intg_global_ctors">The '<tt>llvm.global_ctors</tt>' Global Variable</a>
2592</div>
2593
2594<div class="doc_text">
2595
2596<p>TODO: Describe this.</p>
2597
2598</div>
2599
2600<!-- ======================================================================= -->
2601<div class="doc_subsection">
2602<a name="intg_global_dtors">The '<tt>llvm.global_dtors</tt>' Global Variable</a>
2603</div>
2604
2605<div class="doc_text">
2606
2607<p>TODO: Describe this.</p>
2608
2609</div>
2610
2611
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2612<!-- *********************************************************************** -->
2613<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
2614<!-- *********************************************************************** -->
2615
2616<div class="doc_text">
2617
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2618<p>The LLVM instruction set consists of several different classifications of
2619 instructions: <a href="#terminators">terminator
2620 instructions</a>, <a href="#binaryops">binary instructions</a>,
2621 <a href="#bitwiseops">bitwise binary instructions</a>,
2622 <a href="#memoryops">memory instructions</a>, and
2623 <a href="#otherops">other instructions</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2624
2625</div>
2626
2627<!-- ======================================================================= -->
2628<div class="doc_subsection"> <a name="terminators">Terminator
2629Instructions</a> </div>
2630
2631<div class="doc_text">
2632
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2633<p>As mentioned <a href="#functionstructure">previously</a>, every basic block
2634 in a program ends with a "Terminator" instruction, which indicates which
2635 block should be executed after the current block is finished. These
2636 terminator instructions typically yield a '<tt>void</tt>' value: they produce
2637 control flow, not values (the one exception being the
2638 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
2639
2640<p>There are six different terminator instructions: the
2641 '<a href="#i_ret"><tt>ret</tt></a>' instruction, the
2642 '<a href="#i_br"><tt>br</tt></a>' instruction, the
2643 '<a href="#i_switch"><tt>switch</tt></a>' instruction, the
Bill Wendlingb4ae2292009-11-02 00:25:26 +0000[diff] [blame]2644 '<a href="#i_indirectbr">'<tt>indirectbr</tt></a>' Instruction, the
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2645 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the
2646 '<a href="#i_unwind"><tt>unwind</tt></a>' instruction, and the
2647 '<a href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2648
2649</div>
2650
2651<!-- _______________________________________________________________________ -->
2652<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
2653Instruction</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2654
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2655<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2656
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2657<h5>Syntax:</h5>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2658<pre>
2659 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]2660 ret void <i>; Return from void function</i>
2661</pre>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2662
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2663<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2664<p>The '<tt>ret</tt>' instruction is used to return control flow (and optionally
2665 a value) from a function back to the caller.</p>
2666
2667<p>There are two forms of the '<tt>ret</tt>' instruction: one that returns a
2668 value and then causes control flow, and one that just causes control flow to
2669 occur.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2670
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2671<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2672<p>The '<tt>ret</tt>' instruction optionally accepts a single argument, the
2673 return value. The type of the return value must be a
2674 '<a href="#t_firstclass">first class</a>' type.</p>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2675
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2676<p>A function is not <a href="#wellformed">well formed</a> if it it has a
2677 non-void return type and contains a '<tt>ret</tt>' instruction with no return
2678 value or a return value with a type that does not match its type, or if it
2679 has a void return type and contains a '<tt>ret</tt>' instruction with a
2680 return value.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2681
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2682<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2683<p>When the '<tt>ret</tt>' instruction is executed, control flow returns back to
2684 the calling function's context. If the caller is a
2685 "<a href="#i_call"><tt>call</tt></a>" instruction, execution continues at the
2686 instruction after the call. If the caller was an
2687 "<a href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues at
2688 the beginning of the "normal" destination block. If the instruction returns
2689 a value, that value shall set the call or invoke instruction's return
2690 value.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2691
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2692<h5>Example:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2693<pre>
2694 ret i32 5 <i>; Return an integer value of 5</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2695 ret void <i>; Return from a void function</i>
Bill Wendlingd163e2d2009-02-28 22:12:54 +0000[diff] [blame]2696 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]2697</pre>
Dan Gohman60967192009-01-12 23:12:39 +0000[diff] [blame]2698
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2699</div>
2700<!-- _______________________________________________________________________ -->
2701<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2702
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2703<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2704
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2705<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2706<pre>
2707 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]2708</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2709
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2710<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2711<p>The '<tt>br</tt>' instruction is used to cause control flow to transfer to a
2712 different basic block in the current function. There are two forms of this
2713 instruction, corresponding to a conditional branch and an unconditional
2714 branch.</p>
2715
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2716<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2717<p>The conditional branch form of the '<tt>br</tt>' instruction takes a single
2718 '<tt>i1</tt>' value and two '<tt>label</tt>' values. The unconditional form
2719 of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>' value as a
2720 target.</p>
2721
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2722<h5>Semantics:</h5>
2723<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2724 argument is evaluated. If the value is <tt>true</tt>, control flows to the
2725 '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
2726 control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
2727
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2728<h5>Example:</h5>
Bill Wendling6ec40612009-07-20 02:39:26 +0000[diff] [blame]2729<pre>
2730Test:
2731 %cond = <a href="#i_icmp">icmp</a> eq i32 %a, %b
2732 br i1 %cond, label %IfEqual, label %IfUnequal
2733IfEqual:
2734 <a href="#i_ret">ret</a> i32 1
2735IfUnequal:
2736 <a href="#i_ret">ret</a> i32 0
2737</pre>
2738
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2739</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2740
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2741<!-- _______________________________________________________________________ -->
2742<div class="doc_subsubsection">
2743 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
2744</div>
2745
2746<div class="doc_text">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2747
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2748<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2749<pre>
2750 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
2751</pre>
2752
2753<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2754<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]2755 several different places. It is a generalization of the '<tt>br</tt>'
2756 instruction, allowing a branch to occur to one of many possible
2757 destinations.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2758
2759<h5>Arguments:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2760<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2761 comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination,
2762 and an array of pairs of comparison value constants and '<tt>label</tt>'s.
2763 The table is not allowed to contain duplicate constant entries.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2764
2765<h5>Semantics:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2766<p>The <tt>switch</tt> instruction specifies a table of values and
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2767 destinations. When the '<tt>switch</tt>' instruction is executed, this table
2768 is searched for the given value. If the value is found, control flow is
Benjamin Kramer5fb9d7e2009-10-12 14:46:08 +0000[diff] [blame]2769 transferred to the corresponding destination; otherwise, control flow is
2770 transferred to the default destination.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2771
2772<h5>Implementation:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2773<p>Depending on properties of the target machine and the particular
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2774 <tt>switch</tt> instruction, this instruction may be code generated in
2775 different ways. For example, it could be generated as a series of chained
2776 conditional branches or with a lookup table.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2777
2778<h5>Example:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2779<pre>
2780 <i>; Emulate a conditional br instruction</i>
2781 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]2782 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2783
2784 <i>; Emulate an unconditional br instruction</i>
2785 switch i32 0, label %dest [ ]
2786
2787 <i>; Implement a jump table:</i>
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]2788 switch i32 %val, label %otherwise [ i32 0, label %onzero
2789 i32 1, label %onone
2790 i32 2, label %ontwo ]
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2791</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2792
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2793</div>
2794
Chris Lattnere0787282009-10-27 19:13:16 +0000[diff] [blame]2795
2796<!-- _______________________________________________________________________ -->
2797<div class="doc_subsubsection">
Chris Lattner4c3800f2009-10-28 00:19:10 +0000[diff] [blame]2798 <a name="i_indirectbr">'<tt>indirectbr</tt>' Instruction</a>
Chris Lattnere0787282009-10-27 19:13:16 +0000[diff] [blame]2799</div>
2800
2801<div class="doc_text">
2802
2803<h5>Syntax:</h5>
2804<pre>
Chris Lattner4c3800f2009-10-28 00:19:10 +0000[diff] [blame]2805 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]2806</pre>
2807
2808<h5>Overview:</h5>
2809
Chris Lattner4c3800f2009-10-28 00:19:10 +0000[diff] [blame]2810<p>The '<tt>indirectbr</tt>' instruction implements an indirect branch to a label
Chris Lattnere0787282009-10-27 19:13:16 +0000[diff] [blame]2811 within the current function, whose address is specified by
Chris Lattnerd07c8372009-10-27 21:01:34 +0000[diff] [blame]2812 "<tt>address</tt>". Address must be derived from a <a
2813 href="#blockaddress">blockaddress</a> constant.</p>
Chris Lattnere0787282009-10-27 19:13:16 +0000[diff] [blame]2814
2815<h5>Arguments:</h5>
2816
2817<p>The '<tt>address</tt>' argument is the address of the label to jump to. The
2818 rest of the arguments indicate the full set of possible destinations that the
2819 address may point to. Blocks are allowed to occur multiple times in the
2820 destination list, though this isn't particularly useful.</p>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]2821
Chris Lattnere0787282009-10-27 19:13:16 +0000[diff] [blame]2822<p>This destination list is required so that dataflow analysis has an accurate
2823 understanding of the CFG.</p>
2824
2825<h5>Semantics:</h5>
2826
2827<p>Control transfers to the block specified in the address argument. All
2828 possible destination blocks must be listed in the label list, otherwise this
2829 instruction has undefined behavior. This implies that jumps to labels
2830 defined in other functions have undefined behavior as well.</p>
2831
2832<h5>Implementation:</h5>
2833
2834<p>This is typically implemented with a jump through a register.</p>
2835
2836<h5>Example:</h5>
2837<pre>
Chris Lattner4c3800f2009-10-28 00:19:10 +0000[diff] [blame]2838 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
Chris Lattnere0787282009-10-27 19:13:16 +0000[diff] [blame]2839</pre>
2840
2841</div>
2842
2843
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2844<!-- _______________________________________________________________________ -->
2845<div class="doc_subsubsection">
2846 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
2847</div>
2848
2849<div class="doc_text">
2850
2851<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2852<pre>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]2853 &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]2854 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
2855</pre>
2856
2857<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2858<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2859 function, with the possibility of control flow transfer to either the
2860 '<tt>normal</tt>' label or the '<tt>exception</tt>' label. If the callee
2861 function returns with the "<tt><a href="#i_ret">ret</a></tt>" instruction,
2862 control flow will return to the "normal" label. If the callee (or any
2863 indirect callees) returns with the "<a href="#i_unwind"><tt>unwind</tt></a>"
2864 instruction, control is interrupted and continued at the dynamically nearest
2865 "exception" label.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2866
2867<h5>Arguments:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2868<p>This instruction requires several arguments:</p>
2869
2870<ol>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2871 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
2872 convention</a> the call should use. If none is specified, the call
2873 defaults to using C calling conventions.</li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]2874
2875 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2876 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
2877 '<tt>inreg</tt>' attributes are valid here.</li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]2878
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2879 <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]2880 function value being invoked. In most cases, this is a direct function
2881 invocation, but indirect <tt>invoke</tt>s are just as possible, branching
2882 off an arbitrary pointer to function value.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2883
2884 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2885 function to be invoked. </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2886
2887 <li>'<tt>function args</tt>': argument list whose types match the function
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2888 signature argument types. If the function signature indicates the
2889 function accepts a variable number of arguments, the extra arguments can
2890 be specified.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2891
2892 <li>'<tt>normal label</tt>': the label reached when the called function
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2893 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2894
2895 <li>'<tt>exception label</tt>': the label reached when a callee returns with
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2896 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2897
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]2898 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2899 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
2900 '<tt>readnone</tt>' attributes are valid here.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2901</ol>
2902
2903<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2904<p>This instruction is designed to operate as a standard
2905 '<tt><a href="#i_call">call</a></tt>' instruction in most regards. The
2906 primary difference is that it establishes an association with a label, which
2907 is used by the runtime library to unwind the stack.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2908
2909<p>This instruction is used in languages with destructors to ensure that proper
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2910 cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
2911 exception. Additionally, this is important for implementation of
2912 '<tt>catch</tt>' clauses in high-level languages that support them.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2913
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2914<p>For the purposes of the SSA form, the definition of the value returned by the
2915 '<tt>invoke</tt>' instruction is deemed to occur on the edge from the current
2916 block to the "normal" label. If the callee unwinds then no return value is
2917 available.</p>
Dan Gohman140ba5d2009-05-22 21:47:08 +0000[diff] [blame]2918
Chris Lattner4a91ef42010-01-15 18:08:37 +0000[diff] [blame]2919<p>Note that the code generator does not yet completely support unwind, and
2920that the invoke/unwind semantics are likely to change in future versions.</p>
2921
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2922<h5>Example:</h5>
2923<pre>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2924 %retval = invoke i32 @Test(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2925 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2926 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2927 unwind label %TestCleanup <i>; {i32}:retval set</i>
2928</pre>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2929
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2930</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2931
2932<!-- _______________________________________________________________________ -->
2933
2934<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
2935Instruction</a> </div>
2936
2937<div class="doc_text">
2938
2939<h5>Syntax:</h5>
2940<pre>
2941 unwind
2942</pre>
2943
2944<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2945<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2946 at the first callee in the dynamic call stack which used
2947 an <a href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call.
2948 This is primarily used to implement exception handling.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2949
2950<h5>Semantics:</h5>
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]2951<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2952 immediately halt. The dynamic call stack is then searched for the
2953 first <a href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack.
2954 Once found, execution continues at the "exceptional" destination block
2955 specified by the <tt>invoke</tt> instruction. If there is no <tt>invoke</tt>
2956 instruction in the dynamic call chain, undefined behavior results.</p>
2957
Chris Lattner4a91ef42010-01-15 18:08:37 +0000[diff] [blame]2958<p>Note that the code generator does not yet completely support unwind, and
2959that the invoke/unwind semantics are likely to change in future versions.</p>
2960
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2961</div>
2962
2963<!-- _______________________________________________________________________ -->
2964
2965<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
2966Instruction</a> </div>
2967
2968<div class="doc_text">
2969
2970<h5>Syntax:</h5>
2971<pre>
2972 unreachable
2973</pre>
2974
2975<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2976<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2977 instruction is used to inform the optimizer that a particular portion of the
2978 code is not reachable. This can be used to indicate that the code after a
2979 no-return function cannot be reached, and other facts.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2980
2981<h5>Semantics:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2982<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2983
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2984</div>
2985
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2986<!-- ======================================================================= -->
2987<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2988
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2989<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2990
2991<p>Binary operators are used to do most of the computation in a program. They
2992 require two operands of the same type, execute an operation on them, and
2993 produce a single value. The operands might represent multiple data, as is
2994 the case with the <a href="#t_vector">vector</a> data type. The result value
2995 has the same type as its operands.</p>
2996
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2997<p>There are several different binary operators:</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]2998
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2999</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3000
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3001<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3002<div class="doc_subsubsection">
3003 <a name="i_add">'<tt>add</tt>' Instruction</a>
3004</div>
3005
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3006<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3007
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3008<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3009<pre>
Dan Gohmaned5fcb22009-07-20 22:41:19 +0000[diff] [blame]3010 &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]3011 &lt;result&gt; = add nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3012 &lt;result&gt; = add nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3013 &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]3014</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3015
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3016<h5>Overview:</h5>
3017<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3018
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3019<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3020<p>The two arguments to the '<tt>add</tt>' instruction must
3021 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3022 integer values. Both arguments must have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3023
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3024<h5>Semantics:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3025<p>The value produced is the integer sum of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3026
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3027<p>If the sum has unsigned overflow, the result returned is the mathematical
3028 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]3029
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3030<p>Because LLVM integers use a two's complement representation, this instruction
3031 is appropriate for both signed and unsigned integers.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3032
Dan Gohman46e96012009-07-22 22:44:56 +0000[diff] [blame]3033<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3034 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3035 <tt>nsw</tt> keywords are present, the result value of the <tt>add</tt>
3036 is undefined if unsigned and/or signed overflow, respectively, occurs.</p>
Dan Gohmaned5fcb22009-07-20 22:41:19 +0000[diff] [blame]3037
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3038<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3039<pre>
3040 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3041</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3042
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3043</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3044
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3045<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3046<div class="doc_subsubsection">
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3047 <a name="i_fadd">'<tt>fadd</tt>' Instruction</a>
3048</div>
3049
3050<div class="doc_text">
3051
3052<h5>Syntax:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3053<pre>
3054 &lt;result&gt; = fadd &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3055</pre>
3056
3057<h5>Overview:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3058<p>The '<tt>fadd</tt>' instruction returns the sum of its two operands.</p>
3059
3060<h5>Arguments:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3061<p>The two arguments to the '<tt>fadd</tt>' instruction must be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3062 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3063 floating point values. Both arguments must have identical types.</p>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3064
3065<h5>Semantics:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3066<p>The value produced is the floating point sum of the two operands.</p>
3067
3068<h5>Example:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3069<pre>
3070 &lt;result&gt; = fadd float 4.0, %var <i>; yields {float}:result = 4.0 + %var</i>
3071</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3072
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3073</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3074
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3075<!-- _______________________________________________________________________ -->
3076<div class="doc_subsubsection">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3077 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
3078</div>
3079
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3080<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3081
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3082<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3083<pre>
Dan Gohman46e96012009-07-22 22:44:56 +0000[diff] [blame]3084 &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]3085 &lt;result&gt; = sub nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3086 &lt;result&gt; = sub nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3087 &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]3088</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3089
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3090<h5>Overview:</h5>
3091<p>The '<tt>sub</tt>' instruction returns the difference of its two
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3092 operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3093
3094<p>Note that the '<tt>sub</tt>' instruction is used to represent the
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3095 '<tt>neg</tt>' instruction present in most other intermediate
3096 representations.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3097
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3098<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3099<p>The two arguments to the '<tt>sub</tt>' instruction must
3100 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3101 integer values. Both arguments must have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3102
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3103<h5>Semantics:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3104<p>The value produced is the integer difference of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3105
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3106<p>If the difference has unsigned overflow, the result returned is the
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3107 mathematical result modulo 2<sup>n</sup>, where n is the bit width of the
3108 result.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3109
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3110<p>Because LLVM integers use a two's complement representation, this instruction
3111 is appropriate for both signed and unsigned integers.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3112
Dan Gohman46e96012009-07-22 22:44:56 +0000[diff] [blame]3113<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3114 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3115 <tt>nsw</tt> keywords are present, the result value of the <tt>sub</tt>
3116 is undefined if unsigned and/or signed overflow, respectively, occurs.</p>
Dan Gohmaned5fcb22009-07-20 22:41:19 +0000[diff] [blame]3117
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3118<h5>Example:</h5>
3119<pre>
3120 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
3121 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
3122</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3123
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3124</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3125
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3126<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3127<div class="doc_subsubsection">
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3128 <a name="i_fsub">'<tt>fsub</tt>' Instruction</a>
3129</div>
3130
3131<div class="doc_text">
3132
3133<h5>Syntax:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3134<pre>
3135 &lt;result&gt; = fsub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3136</pre>
3137
3138<h5>Overview:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3139<p>The '<tt>fsub</tt>' instruction returns the difference of its two
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3140 operands.</p>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3141
3142<p>Note that the '<tt>fsub</tt>' instruction is used to represent the
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3143 '<tt>fneg</tt>' instruction present in most other intermediate
3144 representations.</p>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3145
3146<h5>Arguments:</h5>
Bill Wendling1a2630a2009-07-20 02:32:41 +0000[diff] [blame]3147<p>The two arguments to the '<tt>fsub</tt>' instruction must be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3148 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3149 floating point values. Both arguments must have identical types.</p>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3150
3151<h5>Semantics:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3152<p>The value produced is the floating point difference of the two operands.</p>
3153
3154<h5>Example:</h5>
3155<pre>
3156 &lt;result&gt; = fsub float 4.0, %var <i>; yields {float}:result = 4.0 - %var</i>
3157 &lt;result&gt; = fsub float -0.0, %val <i>; yields {float}:result = -%var</i>
3158</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3159
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3160</div>
3161
3162<!-- _______________________________________________________________________ -->
3163<div class="doc_subsubsection">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3164 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
3165</div>
3166
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3167<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3168
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3169<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3170<pre>
Dan Gohman46e96012009-07-22 22:44:56 +0000[diff] [blame]3171 &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]3172 &lt;result&gt; = mul nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3173 &lt;result&gt; = mul nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3174 &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]3175</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3176
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3177<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3178<p>The '<tt>mul</tt>' instruction returns the product of its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3179
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3180<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3181<p>The two arguments to the '<tt>mul</tt>' instruction must
3182 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3183 integer values. Both arguments must have identical types.</p>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3184
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3185<h5>Semantics:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3186<p>The value produced is the integer product of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3187
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3188<p>If the result of the multiplication has unsigned overflow, the result
3189 returned is the mathematical result modulo 2<sup>n</sup>, where n is the bit
3190 width of the result.</p>
3191
3192<p>Because LLVM integers use a two's complement representation, and the result
3193 is the same width as the operands, this instruction returns the correct
3194 result for both signed and unsigned integers. If a full product
3195 (e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands should
3196 be sign-extended or zero-extended as appropriate to the width of the full
3197 product.</p>
3198
Dan Gohman46e96012009-07-22 22:44:56 +0000[diff] [blame]3199<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3200 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3201 <tt>nsw</tt> keywords are present, the result value of the <tt>mul</tt>
3202 is undefined if unsigned and/or signed overflow, respectively, occurs.</p>
Dan Gohmaned5fcb22009-07-20 22:41:19 +0000[diff] [blame]3203
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3204<h5>Example:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3205<pre>
3206 &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3207</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3208
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3209</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3210
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3211<!-- _______________________________________________________________________ -->
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3212<div class="doc_subsubsection">
3213 <a name="i_fmul">'<tt>fmul</tt>' Instruction</a>
3214</div>
3215
3216<div class="doc_text">
3217
3218<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3219<pre>
3220 &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]3221</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3222
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3223<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3224<p>The '<tt>fmul</tt>' instruction returns the product of its two operands.</p>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3225
3226<h5>Arguments:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3227<p>The two arguments to the '<tt>fmul</tt>' instruction must be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3228 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3229 floating point values. Both arguments must have identical types.</p>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3230
3231<h5>Semantics:</h5>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3232<p>The value produced is the floating point product of the two operands.</p>
3233
3234<h5>Example:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3235<pre>
3236 &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]3237</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3238
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]3239</div>
3240
3241<!-- _______________________________________________________________________ -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3242<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
3243</a></div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3244
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3245<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3246
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3247<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3248<pre>
3249 &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]3250</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3251
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3252<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3253<p>The '<tt>udiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3254
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3255<h5>Arguments:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3256<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3257 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3258 values. Both arguments must have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3259
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3260<h5>Semantics:</h5>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]3261<p>The value produced is the unsigned integer quotient of the two operands.</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3262
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]3263<p>Note that unsigned integer division and signed integer division are distinct
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3264 operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
3265
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]3266<p>Division by zero leads to undefined behavior.</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3267
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3268<h5>Example:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3269<pre>
3270 &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3271</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3272
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3273</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3274
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3275<!-- _______________________________________________________________________ -->
3276<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
3277</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3278
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3279<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3280
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3281<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3282<pre>
Dan Gohmaned5fcb22009-07-20 22:41:19 +0000[diff] [blame]3283 &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]3284 &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]3285</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3286
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3287<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3288<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3289
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3290<h5>Arguments:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3291<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3292 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3293 values. Both arguments must have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3294
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3295<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3296<p>The value produced is the signed integer quotient of the two operands rounded
3297 towards zero.</p>
3298
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]3299<p>Note that signed integer division and unsigned integer division are distinct
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3300 operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
3301
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]3302<p>Division by zero leads to undefined behavior. Overflow also leads to
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3303 undefined behavior; this is a rare case, but can occur, for example, by doing
3304 a 32-bit division of -2147483648 by -1.</p>
3305
Dan Gohman67fa48e2009-07-22 00:04:19 +0000[diff] [blame]3306<p>If the <tt>exact</tt> keyword is present, the result value of the
3307 <tt>sdiv</tt> is undefined if the result would be rounded or if overflow
3308 would occur.</p>
Dan Gohmaned5fcb22009-07-20 22:41:19 +0000[diff] [blame]3309
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3310<h5>Example:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3311<pre>
3312 &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3313</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3314
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3315</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3316
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3317<!-- _______________________________________________________________________ -->
3318<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
3319Instruction</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3320
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3321<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3322
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3323<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3324<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3325 &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]3326</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3327
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3328<h5>Overview:</h5>
3329<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3330
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3331<h5>Arguments:</h5>
3332<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3333 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3334 floating point values. Both arguments must have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3335
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3336<h5>Semantics:</h5>
3337<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3338
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3339<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3340<pre>
3341 &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]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>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3345
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3346<!-- _______________________________________________________________________ -->
3347<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
3348</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>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3353<pre>
3354 &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]3355</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3356
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3357<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3358<p>The '<tt>urem</tt>' instruction returns the remainder from the unsigned
3359 division of its two arguments.</p>
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>urem</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>
3365
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3366<h5>Semantics:</h5>
3367<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3368 This instruction always performs an unsigned division to get the
3369 remainder.</p>
3370
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]3371<p>Note that unsigned integer remainder and signed integer remainder are
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3372 distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
3373
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]3374<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3375
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3376<h5>Example:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3377<pre>
3378 &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3379</pre>
3380
3381</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3382
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3383<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3384<div class="doc_subsubsection">
3385 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
3386</div>
3387
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3388<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3389
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3390<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3391<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3392 &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]3393</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3394
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3395<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3396<p>The '<tt>srem</tt>' instruction returns the remainder from the signed
3397 division of its two operands. This instruction can also take
3398 <a href="#t_vector">vector</a> versions of the values in which case the
3399 elements must be integers.</p>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]3400
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3401<h5>Arguments:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3402<p>The two arguments to the '<tt>srem</tt>' instruction must be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3403 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3404 values. Both arguments must have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3405
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3406<h5>Semantics:</h5>
3407<p>This instruction returns the <i>remainder</i> of a division (where the result
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3408 has the same sign as the dividend, <tt>op1</tt>), not the <i>modulo</i>
3409 operator (where the result has the same sign as the divisor, <tt>op2</tt>) of
3410 a value. For more information about the difference,
3411 see <a href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
3412 Math Forum</a>. For a table of how this is implemented in various languages,
3413 please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
3414 Wikipedia: modulo operation</a>.</p>
3415
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]3416<p>Note that signed integer remainder and unsigned integer remainder are
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3417 distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
3418
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]3419<p>Taking the remainder of a division by zero leads to undefined behavior.
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3420 Overflow also leads to undefined behavior; this is a rare case, but can
3421 occur, for example, by taking the remainder of a 32-bit division of
3422 -2147483648 by -1. (The remainder doesn't actually overflow, but this rule
3423 lets srem be implemented using instructions that return both the result of
3424 the division and the remainder.)</p>
3425
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3426<h5>Example:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3427<pre>
3428 &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3429</pre>
3430
3431</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3432
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3433<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3434<div class="doc_subsubsection">
3435 <a name="i_frem">'<tt>frem</tt>' Instruction</a> </div>
3436
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3437<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3438
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3439<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3440<pre>
3441 &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]3442</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3443
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3444<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3445<p>The '<tt>frem</tt>' instruction returns the remainder from the division of
3446 its two operands.</p>
3447
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3448<h5>Arguments:</h5>
3449<p>The two arguments to the '<tt>frem</tt>' instruction must be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3450 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3451 floating point values. Both arguments must have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3452
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3453<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3454<p>This instruction returns the <i>remainder</i> of a division. The remainder
3455 has the same sign as the dividend.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3456
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3457<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3458<pre>
3459 &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]3460</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3461
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3462</div>
3463
3464<!-- ======================================================================= -->
3465<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
3466Operations</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3467
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3468<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3469
3470<p>Bitwise binary operators are used to do various forms of bit-twiddling in a
3471 program. They are generally very efficient instructions and can commonly be
3472 strength reduced from other instructions. They require two operands of the
3473 same type, execute an operation on them, and produce a single value. The
3474 resulting value is the same type as its operands.</p>
3475
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3476</div>
3477
3478<!-- _______________________________________________________________________ -->
3479<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
3480Instruction</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3481
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3482<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3483
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3484<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3485<pre>
3486 &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]3487</pre>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]3488
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3489<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3490<p>The '<tt>shl</tt>' instruction returns the first operand shifted to the left
3491 a specified number of bits.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]3492
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3493<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3494<p>Both arguments to the '<tt>shl</tt>' instruction must be the
3495 same <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3496 integer type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3497
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3498<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3499<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod
3500 2<sup>n</sup>, where <tt>n</tt> is the width of the result. If <tt>op2</tt>
3501 is (statically or dynamically) negative or equal to or larger than the number
3502 of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3503 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3504 shift amount in <tt>op2</tt>.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]3505
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3506<h5>Example:</h5>
3507<pre>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3508 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
3509 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
3510 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]3511 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Mon P Wang9901e732008-12-09 05:46:39 +0000[diff] [blame]3512 &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]3513</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3514
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3515</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3516
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3517<!-- _______________________________________________________________________ -->
3518<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
3519Instruction</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3520
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3521<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3522
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3523<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3524<pre>
3525 &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]3526</pre>
3527
3528<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3529<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
3530 operand shifted to the right a specified number of bits with zero fill.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3531
3532<h5>Arguments:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3533<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3534 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3535 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3536
3537<h5>Semantics:</h5>
3538<p>This instruction always performs a logical shift right operation. The most
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3539 significant bits of the result will be filled with zero bits after the shift.
3540 If <tt>op2</tt> is (statically or dynamically) equal to or larger than the
3541 number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3542 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3543 shift amount in <tt>op2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3544
3545<h5>Example:</h5>
3546<pre>
3547 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
3548 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
3549 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
3550 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]3551 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Mon P Wang9901e732008-12-09 05:46:39 +0000[diff] [blame]3552 &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]3553</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3554
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3555</div>
3556
3557<!-- _______________________________________________________________________ -->
3558<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
3559Instruction</a> </div>
3560<div class="doc_text">
3561
3562<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3563<pre>
3564 &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]3565</pre>
3566
3567<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3568<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
3569 operand shifted to the right a specified number of bits with sign
3570 extension.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3571
3572<h5>Arguments:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3573<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3574 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3575 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3576
3577<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3578<p>This instruction always performs an arithmetic shift right operation, The
3579 most significant bits of the result will be filled with the sign bit
3580 of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
3581 larger than the number of bits in <tt>op1</tt>, the result is undefined. If
3582 the arguments are vectors, each vector element of <tt>op1</tt> is shifted by
3583 the corresponding shift amount in <tt>op2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3584
3585<h5>Example:</h5>
3586<pre>
3587 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
3588 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
3589 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
3590 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]3591 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Mon P Wang9901e732008-12-09 05:46:39 +0000[diff] [blame]3592 &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]3593</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3594
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3595</div>
3596
3597<!-- _______________________________________________________________________ -->
3598<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
3599Instruction</a> </div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3600
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3601<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3602
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3603<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3604<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3605 &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]3606</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3607
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3608<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3609<p>The '<tt>and</tt>' instruction returns the bitwise logical and of its two
3610 operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3611
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3612<h5>Arguments:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3613<p>The two arguments to the '<tt>and</tt>' instruction must be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3614 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3615 values. Both arguments must have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3616
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3617<h5>Semantics:</h5>
3618<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3619
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3620<table border="1" cellspacing="0" cellpadding="4">
3621 <tbody>
3622 <tr>
3623 <td>In0</td>
3624 <td>In1</td>
3625 <td>Out</td>
3626 </tr>
3627 <tr>
3628 <td>0</td>
3629 <td>0</td>
3630 <td>0</td>
3631 </tr>
3632 <tr>
3633 <td>0</td>
3634 <td>1</td>
3635 <td>0</td>
3636 </tr>
3637 <tr>
3638 <td>1</td>
3639 <td>0</td>
3640 <td>0</td>
3641 </tr>
3642 <tr>
3643 <td>1</td>
3644 <td>1</td>
3645 <td>1</td>
3646 </tr>
3647 </tbody>
3648</table>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3649
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3650<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3651<pre>
3652 &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]3653 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
3654 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
3655</pre>
3656</div>
3657<!-- _______________________________________________________________________ -->
3658<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3659
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3660<div class="doc_text">
3661
3662<h5>Syntax:</h5>
3663<pre>
3664 &lt;result&gt; = or &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3665</pre>
3666
3667<h5>Overview:</h5>
3668<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive or of its
3669 two operands.</p>
3670
3671<h5>Arguments:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3672<p>The two arguments to the '<tt>or</tt>' instruction must be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3673 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3674 values. Both arguments must have identical types.</p>
3675
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3676<h5>Semantics:</h5>
3677<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3678
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3679<table border="1" cellspacing="0" cellpadding="4">
3680 <tbody>
3681 <tr>
3682 <td>In0</td>
3683 <td>In1</td>
3684 <td>Out</td>
3685 </tr>
3686 <tr>
3687 <td>0</td>
3688 <td>0</td>
3689 <td>0</td>
3690 </tr>
3691 <tr>
3692 <td>0</td>
3693 <td>1</td>
3694 <td>1</td>
3695 </tr>
3696 <tr>
3697 <td>1</td>
3698 <td>0</td>
3699 <td>1</td>
3700 </tr>
3701 <tr>
3702 <td>1</td>
3703 <td>1</td>
3704 <td>1</td>
3705 </tr>
3706 </tbody>
3707</table>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3708
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3709<h5>Example:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3710<pre>
3711 &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3712 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
3713 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
3714</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3715
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3716</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3717
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3718<!-- _______________________________________________________________________ -->
3719<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
3720Instruction</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3721
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3722<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3723
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3724<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3725<pre>
3726 &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]3727</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3728
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3729<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3730<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive or of
3731 its two operands. The <tt>xor</tt> is used to implement the "one's
3732 complement" operation, which is the "~" operator in C.</p>
3733
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3734<h5>Arguments:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3735<p>The two arguments to the '<tt>xor</tt>' instruction must be
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3736 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3737 values. Both arguments must have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3738
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3739<h5>Semantics:</h5>
3740<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3741
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3742<table border="1" cellspacing="0" cellpadding="4">
3743 <tbody>
3744 <tr>
3745 <td>In0</td>
3746 <td>In1</td>
3747 <td>Out</td>
3748 </tr>
3749 <tr>
3750 <td>0</td>
3751 <td>0</td>
3752 <td>0</td>
3753 </tr>
3754 <tr>
3755 <td>0</td>
3756 <td>1</td>
3757 <td>1</td>
3758 </tr>
3759 <tr>
3760 <td>1</td>
3761 <td>0</td>
3762 <td>1</td>
3763 </tr>
3764 <tr>
3765 <td>1</td>
3766 <td>1</td>
3767 <td>0</td>
3768 </tr>
3769 </tbody>
3770</table>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3771
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3772<h5>Example:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3773<pre>
3774 &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3775 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
3776 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
3777 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
3778</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3779
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3780</div>
3781
3782<!-- ======================================================================= -->
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3783<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3784 <a name="vectorops">Vector Operations</a>
3785</div>
3786
3787<div class="doc_text">
3788
3789<p>LLVM supports several instructions to represent vector operations in a
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3790 target-independent manner. These instructions cover the element-access and
3791 vector-specific operations needed to process vectors effectively. While LLVM
3792 does directly support these vector operations, many sophisticated algorithms
3793 will want to use target-specific intrinsics to take full advantage of a
3794 specific target.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3795
3796</div>
3797
3798<!-- _______________________________________________________________________ -->
3799<div class="doc_subsubsection">
3800 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
3801</div>
3802
3803<div class="doc_text">
3804
3805<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3806<pre>
3807 &lt;result&gt; = extractelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, i32 &lt;idx&gt; <i>; yields &lt;ty&gt;</i>
3808</pre>
3809
3810<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3811<p>The '<tt>extractelement</tt>' instruction extracts a single scalar element
3812 from a vector at a specified index.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3813
3814
3815<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3816<p>The first operand of an '<tt>extractelement</tt>' instruction is a value
3817 of <a href="#t_vector">vector</a> type. The second operand is an index
3818 indicating the position from which to extract the element. The index may be
3819 a variable.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3820
3821<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3822<p>The result is a scalar of the same type as the element type of
3823 <tt>val</tt>. Its value is the value at position <tt>idx</tt> of
3824 <tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
3825 results are undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3826
3827<h5>Example:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3828<pre>
Gabor Greifb2c73142009-10-28 13:14:50 +0000[diff] [blame]3829 &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]3830</pre>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3831
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3832</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3833
3834<!-- _______________________________________________________________________ -->
3835<div class="doc_subsubsection">
3836 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
3837</div>
3838
3839<div class="doc_text">
3840
3841<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3842<pre>
Dan Gohmanbcc3c502008-05-12 23:38:42 +0000[diff] [blame]3843 &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]3844</pre>
3845
3846<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3847<p>The '<tt>insertelement</tt>' instruction inserts a scalar element into a
3848 vector at a specified index.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3849
3850<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3851<p>The first operand of an '<tt>insertelement</tt>' instruction is a value
3852 of <a href="#t_vector">vector</a> type. The second operand is a scalar value
3853 whose type must equal the element type of the first operand. The third
3854 operand is an index indicating the position at which to insert the value.
3855 The index may be a variable.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3856
3857<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3858<p>The result is a vector of the same type as <tt>val</tt>. Its element values
3859 are those of <tt>val</tt> except at position <tt>idx</tt>, where it gets the
3860 value <tt>elt</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
3861 results are undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3862
3863<h5>Example:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3864<pre>
Gabor Greifb2c73142009-10-28 13:14:50 +0000[diff] [blame]3865 &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]3866</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3867
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3868</div>
3869
3870<!-- _______________________________________________________________________ -->
3871<div class="doc_subsubsection">
3872 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
3873</div>
3874
3875<div class="doc_text">
3876
3877<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3878<pre>
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3879 &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]3880</pre>
3881
3882<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3883<p>The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
3884 from two input vectors, returning a vector with the same element type as the
3885 input and length that is the same as the shuffle mask.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3886
3887<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3888<p>The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
3889 with types that match each other. The third argument is a shuffle mask whose
3890 element type is always 'i32'. The result of the instruction is a vector
3891 whose length is the same as the shuffle mask and whose element type is the
3892 same as the element type of the first two operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3893
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3894<p>The shuffle mask operand is required to be a constant vector with either
3895 constant integer or undef values.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3896
3897<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3898<p>The elements of the two input vectors are numbered from left to right across
3899 both of the vectors. The shuffle mask operand specifies, for each element of
3900 the result vector, which element of the two input vectors the result element
3901 gets. The element selector may be undef (meaning "don't care") and the
3902 second operand may be undef if performing a shuffle from only one vector.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3903
3904<h5>Example:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3905<pre>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3906 &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]3907 &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]3908 &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]3909 &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]3910 &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]3911 &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]3912 &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]3913 &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]3914</pre>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3915
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3916</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3917
3918<!-- ======================================================================= -->
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]3919<div class="doc_subsection">
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3920 <a name="aggregateops">Aggregate Operations</a>
3921</div>
3922
3923<div class="doc_text">
3924
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame^]3925<p>LLVM supports several instructions for working with
3926 <a href="#t_aggregate">aggregate</a> values.</p>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3927
3928</div>
3929
3930<!-- _______________________________________________________________________ -->
3931<div class="doc_subsubsection">
3932 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
3933</div>
3934
3935<div class="doc_text">
3936
3937<h5>Syntax:</h5>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3938<pre>
3939 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
3940</pre>
3941
3942<h5>Overview:</h5>
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame^]3943<p>The '<tt>extractvalue</tt>' instruction extracts the value of a member field
3944 from an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3945
3946<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3947<p>The first operand of an '<tt>extractvalue</tt>' instruction is a value
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame^]3948 of <a href="#t_struct">struct</a>, <a href="#t_union">union</a> or
3949 <a href="#t_array">array</a> type. The operands are constant indices to
3950 specify which value to extract in a similar manner as indices in a
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3951 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3952
3953<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3954<p>The result is the value at the position in the aggregate specified by the
3955 index operands.</p>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3956
3957<h5>Example:</h5>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3958<pre>
Gabor Greifb2c73142009-10-28 13:14:50 +0000[diff] [blame]3959 &lt;result&gt; = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3960</pre>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3961
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3962</div>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3963
3964<!-- _______________________________________________________________________ -->
3965<div class="doc_subsubsection">
3966 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
3967</div>
3968
3969<div class="doc_text">
3970
3971<h5>Syntax:</h5>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3972<pre>
Jeffrey Yasskine05d7732010-01-11 19:19:26 +0000[diff] [blame]3973 &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]3974</pre>
3975
3976<h5>Overview:</h5>
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame^]3977<p>The '<tt>insertvalue</tt>' instruction inserts a value into a member field
3978 in an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3979
3980<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3981<p>The first operand of an '<tt>insertvalue</tt>' instruction is a value
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame^]3982 of <a href="#t_struct">struct</a>, <a href="#t_union">union</a> or
3983 <a href="#t_array">array</a> type. The second operand is a first-class
3984 value to insert. The following operands are constant indices indicating
3985 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]3986 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction. The
3987 value to insert must have the same type as the value identified by the
3988 indices.</p>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3989
3990<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]3991<p>The result is an aggregate of the same type as <tt>val</tt>. Its value is
3992 that of <tt>val</tt> except that the value at the position specified by the
3993 indices is that of <tt>elt</tt>.</p>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3994
3995<h5>Example:</h5>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3996<pre>
Jeffrey Yasskine05d7732010-01-11 19:19:26 +0000[diff] [blame]3997 %agg1 = insertvalue {i32, float} undef, i32 1, 0 <i>; yields {i32 1, float undef}</i>
3998 %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]3999</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4000
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]4001</div>
4002
4003
4004<!-- ======================================================================= -->
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4005<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4006 <a name="memoryops">Memory Access and Addressing Operations</a>
4007</div>
4008
4009<div class="doc_text">
4010
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4011<p>A key design point of an SSA-based representation is how it represents
4012 memory. In LLVM, no memory locations are in SSA form, which makes things
Victor Hernandez6b054092009-10-26 23:44:29 +0000[diff] [blame]4013 very simple. This section describes how to read, write, and allocate
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4014 memory in LLVM.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4015
4016</div>
4017
4018<!-- _______________________________________________________________________ -->
4019<div class="doc_subsubsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4020 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
4021</div>
4022
4023<div class="doc_text">
4024
4025<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4026<pre>
4027 &lt;result&gt; = alloca &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
4028</pre>
4029
4030<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4031<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4032 currently executing function, to be automatically released when this function
4033 returns to its caller. The object is always allocated in the generic address
4034 space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4035
4036<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4037<p>The '<tt>alloca</tt>' instruction
4038 allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt> bytes of memory on the
4039 runtime stack, returning a pointer of the appropriate type to the program.
4040 If "NumElements" is specified, it is the number of elements allocated,
4041 otherwise "NumElements" is defaulted to be one. If a constant alignment is
4042 specified, the value result of the allocation is guaranteed to be aligned to
4043 at least that boundary. If not specified, or if zero, the target can choose
4044 to align the allocation on any convenient boundary compatible with the
4045 type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4046
4047<p>'<tt>type</tt>' may be any sized type.</p>
4048
4049<h5>Semantics:</h5>
Bill Wendling2a454572009-05-08 20:49:29 +0000[diff] [blame]4050<p>Memory is allocated; a pointer is returned. The operation is undefined if
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4051 there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
4052 memory is automatically released when the function returns. The
4053 '<tt>alloca</tt>' instruction is commonly used to represent automatic
4054 variables that must have an address available. When the function returns
4055 (either with the <tt><a href="#i_ret">ret</a></tt>
4056 or <tt><a href="#i_unwind">unwind</a></tt> instructions), the memory is
4057 reclaimed. Allocating zero bytes is legal, but the result is undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4058
4059<h5>Example:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4060<pre>
Dan Gohmanf54f50a2009-01-04 23:49:44 +0000[diff] [blame]4061 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
4062 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
4063 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
4064 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4065</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4066
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4067</div>
4068
4069<!-- _______________________________________________________________________ -->
4070<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
4071Instruction</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4072
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4073<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4074
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4075<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4076<pre>
4077 &lt;result&gt; = load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;]
4078 &lt;result&gt; = volatile load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;]
4079</pre>
4080
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4081<h5>Overview:</h5>
4082<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4083
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4084<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4085<p>The argument to the '<tt>load</tt>' instruction specifies the memory address
4086 from which to load. The pointer must point to
4087 a <a href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
4088 marked as <tt>volatile</tt>, then the optimizer is not allowed to modify the
4089 number or order of execution of this <tt>load</tt> with other
4090 volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
4091 instructions. </p>
4092
4093<p>The optional constant "align" argument specifies the alignment of the
4094 operation (that is, the alignment of the memory address). A value of 0 or an
4095 omitted "align" argument means that the operation has the preferential
4096 alignment for the target. It is the responsibility of the code emitter to
4097 ensure that the alignment information is correct. Overestimating the
4098 alignment results in an undefined behavior. Underestimating the alignment may
4099 produce less efficient code. An alignment of 1 is always safe.</p>
4100
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4101<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4102<p>The location of memory pointed to is loaded. If the value being loaded is of
4103 scalar type then the number of bytes read does not exceed the minimum number
4104 of bytes needed to hold all bits of the type. For example, loading an
4105 <tt>i24</tt> reads at most three bytes. When loading a value of a type like
4106 <tt>i20</tt> with a size that is not an integral number of bytes, the result
4107 is undefined if the value was not originally written using a store of the
4108 same type.</p>
4109
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4110<h5>Examples:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4111<pre>
4112 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
4113 <a href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4114 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
4115</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4116
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4117</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4118
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4119<!-- _______________________________________________________________________ -->
4120<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
4121Instruction</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4122
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4123<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4124
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4125<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4126<pre>
4127 store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4128 volatile store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
4129</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4130
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4131<h5>Overview:</h5>
4132<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4133
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4134<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4135<p>There are two arguments to the '<tt>store</tt>' instruction: a value to store
4136 and an address at which to store it. The type of the
4137 '<tt>&lt;pointer&gt;</tt>' operand must be a pointer to
4138 the <a href="#t_firstclass">first class</a> type of the
4139 '<tt>&lt;value&gt;</tt>' operand. If the <tt>store</tt> is marked
4140 as <tt>volatile</tt>, then the optimizer is not allowed to modify the number
4141 or order of execution of this <tt>store</tt> with other
4142 volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
4143 instructions.</p>
4144
4145<p>The optional constant "align" argument specifies the alignment of the
4146 operation (that is, the alignment of the memory address). A value of 0 or an
4147 omitted "align" argument means that the operation has the preferential
4148 alignment for the target. It is the responsibility of the code emitter to
4149 ensure that the alignment information is correct. Overestimating the
4150 alignment results in an undefined behavior. Underestimating the alignment may
4151 produce less efficient code. An alignment of 1 is always safe.</p>
4152
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4153<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4154<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>' at the
4155 location specified by the '<tt>&lt;pointer&gt;</tt>' operand. If
4156 '<tt>&lt;value&gt;</tt>' is of scalar type then the number of bytes written
4157 does not exceed the minimum number of bytes needed to hold all bits of the
4158 type. For example, storing an <tt>i24</tt> writes at most three bytes. When
4159 writing a value of a type like <tt>i20</tt> with a size that is not an
4160 integral number of bytes, it is unspecified what happens to the extra bits
4161 that do not belong to the type, but they will typically be overwritten.</p>
4162
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4163<h5>Example:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4164<pre>
4165 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling63ffa142007-10-22 05:10:05 +0000[diff] [blame]4166 store i32 3, i32* %ptr <i>; yields {void}</i>
4167 %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]4168</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4169
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4170</div>
4171
4172<!-- _______________________________________________________________________ -->
4173<div class="doc_subsubsection">
4174 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
4175</div>
4176
4177<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4178
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4179<h5>Syntax:</h5>
4180<pre>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]4181 &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]4182 &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]4183</pre>
4184
4185<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4186<p>The '<tt>getelementptr</tt>' instruction is used to get the address of a
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame^]4187 subelement of an <a href="#t_aggregate">aggregate</a> data structure.
4188 It performs address calculation only and does not access memory.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4189
4190<h5>Arguments:</h5>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]4191<p>The first argument is always a pointer, and forms the basis of the
Chris Lattnere92fc832009-07-29 06:44:13 +0000[diff] [blame]4192 calculation. The remaining arguments are indices that indicate which of the
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4193 elements of the aggregate object are indexed. The interpretation of each
4194 index is dependent on the type being indexed into. The first index always
4195 indexes the pointer value given as the first argument, the second index
4196 indexes a value of the type pointed to (not necessarily the value directly
4197 pointed to, since the first index can be non-zero), etc. The first type
Chris Lattnerd5d51722010-02-12 20:49:41 +0000[diff] [blame^]4198 indexed into must be a pointer value, subsequent types can be arrays,
4199 vectors, structs and unions. Note that subsequent types being indexed into
4200 can never be pointers, since that would require loading the pointer before
4201 continuing calculation.</p>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]4202
4203<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^]4204 When indexing into a (optionally packed) structure or union, only <tt>i32</tt>
4205 integer <b>constants</b> are allowed. When indexing into an array, pointer
4206 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]4207 constant.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4208
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4209<p>For example, let's consider a C code fragment and how it gets compiled to
4210 LLVM:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4211
4212<div class="doc_code">
4213<pre>
4214struct RT {
4215 char A;
4216 int B[10][20];
4217 char C;
4218};
4219struct ST {
4220 int X;
4221 double Y;
4222 struct RT Z;
4223};
4224
4225int *foo(struct ST *s) {
4226 return &amp;s[1].Z.B[5][13];
4227}
4228</pre>
4229</div>
4230
4231<p>The LLVM code generated by the GCC frontend is:</p>
4232
4233<div class="doc_code">
4234<pre>
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]4235%RT = <a href="#namedtypes">type</a> { i8 , [10 x [20 x i32]], i8 }
4236%ST = <a href="#namedtypes">type</a> { i32, double, %RT }
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4237
Dan Gohman47360842009-07-25 02:23:48 +0000[diff] [blame]4238define i32* @foo(%ST* %s) {
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4239entry:
4240 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
4241 ret i32* %reg
4242}
4243</pre>
4244</div>
4245
4246<h5>Semantics:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4247<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]4248 type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
4249 }</tt>' type, a structure. The second index indexes into the third element
4250 of the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
4251 i8 }</tt>' type, another structure. The third index indexes into the second
4252 element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
4253 array. The two dimensions of the array are subscripted into, yielding an
4254 '<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a
4255 pointer to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4256
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4257<p>Note that it is perfectly legal to index partially through a structure,
4258 returning a pointer to an inner element. Because of this, the LLVM code for
4259 the given testcase is equivalent to:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4260
4261<pre>
Dan Gohman47360842009-07-25 02:23:48 +0000[diff] [blame]4262 define i32* @foo(%ST* %s) {
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4263 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
4264 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
4265 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
4266 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
4267 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
4268 ret i32* %t5
4269 }
4270</pre>
4271
Dan Gohman106b2ae2009-07-27 21:53:46 +0000[diff] [blame]4272<p>If the <tt>inbounds</tt> keyword is present, the result value of the
Dan Gohman4cba1562009-07-29 16:00:30 +0000[diff] [blame]4273 <tt>getelementptr</tt> is undefined if the base pointer is not an
4274 <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]4275 that would be formed by successive addition of the offsets implied by the
4276 indices to the base address with infinitely precise arithmetic are not an
4277 <i>in bounds</i> address of that allocated object.
Dan Gohman4cba1562009-07-29 16:00:30 +0000[diff] [blame]4278 The <i>in bounds</i> addresses for an allocated object are all the addresses
Dan Gohmanbc4c87f2009-08-20 17:08:17 +0000[diff] [blame]4279 that point into the object, plus the address one byte past the end.</p>
Dan Gohman106b2ae2009-07-27 21:53:46 +0000[diff] [blame]4280
4281<p>If the <tt>inbounds</tt> keyword is not present, the offsets are added to
4282 the base address with silently-wrapping two's complement arithmetic, and
4283 the result value of the <tt>getelementptr</tt> may be outside the object
4284 pointed to by the base pointer. The result value may not necessarily be
4285 used to access memory though, even if it happens to point into allocated
4286 storage. See the <a href="#pointeraliasing">Pointer Aliasing Rules</a>
4287 section for more information.</p>
4288
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4289<p>The getelementptr instruction is often confusing. For some more insight into
4290 how it works, see <a href="GetElementPtr.html">the getelementptr FAQ</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4291
4292<h5>Example:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4293<pre>
4294 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]4295 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
4296 <i>; yields i8*:vptr</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4297 %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]4298 <i>; yields i8*:eptr</i>
4299 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Sanjiv Gupta4f9a0dc2009-04-25 07:27:44 +0000[diff] [blame]4300 <i>; yields i32*:iptr</i>
Sanjiv Gupta1e46c582009-04-24 16:38:13 +0000[diff] [blame]4301 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4302</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4303
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4304</div>
4305
4306<!-- ======================================================================= -->
4307<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
4308</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4309
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4310<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4311
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4312<p>The instructions in this category are the conversion instructions (casting)
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4313 which all take a single operand and a type. They perform various bit
4314 conversions on the operand.</p>
4315
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4316</div>
4317
4318<!-- _______________________________________________________________________ -->
4319<div class="doc_subsubsection">
4320 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
4321</div>
4322<div class="doc_text">
4323
4324<h5>Syntax:</h5>
4325<pre>
4326 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4327</pre>
4328
4329<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4330<p>The '<tt>trunc</tt>' instruction truncates its operand to the
4331 type <tt>ty2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4332
4333<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4334<p>The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
4335 be an <a href="#t_integer">integer</a> type, and a type that specifies the
4336 size and type of the result, which must be
4337 an <a href="#t_integer">integer</a> type. The bit size of <tt>value</tt> must
4338 be larger than the bit size of <tt>ty2</tt>. Equal sized types are not
4339 allowed.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4340
4341<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4342<p>The '<tt>trunc</tt>' instruction truncates the high order bits
4343 in <tt>value</tt> and converts the remaining bits to <tt>ty2</tt>. Since the
4344 source size must be larger than the destination size, <tt>trunc</tt> cannot
4345 be a <i>no-op cast</i>. It will always truncate bits.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4346
4347<h5>Example:</h5>
4348<pre>
4349 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
4350 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
Gabor Greif463c9342009-10-28 09:21:30 +0000[diff] [blame]4351 %Z = trunc i32 122 to i1 <i>; yields i1:false</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4352</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4353
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4354</div>
4355
4356<!-- _______________________________________________________________________ -->
4357<div class="doc_subsubsection">
4358 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
4359</div>
4360<div class="doc_text">
4361
4362<h5>Syntax:</h5>
4363<pre>
4364 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4365</pre>
4366
4367<h5>Overview:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4368<p>The '<tt>zext</tt>' instruction zero extends its operand to type
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4369 <tt>ty2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4370
4371
4372<h5>Arguments:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4373<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]4374 <a href="#t_integer">integer</a> type, and a type to cast it to, which must
4375 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]4376 <tt>value</tt> must be smaller than the bit size of the destination type,
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4377 <tt>ty2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4378
4379<h5>Semantics:</h5>
4380<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]4381 bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4382
4383<p>When zero extending from i1, the result will always be either 0 or 1.</p>
4384
4385<h5>Example:</h5>
4386<pre>
4387 %X = zext i32 257 to i64 <i>; yields i64:257</i>
4388 %Y = zext i1 true to i32 <i>; yields i32:1</i>
4389</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4390
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4391</div>
4392
4393<!-- _______________________________________________________________________ -->
4394<div class="doc_subsubsection">
4395 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
4396</div>
4397<div class="doc_text">
4398
4399<h5>Syntax:</h5>
4400<pre>
4401 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4402</pre>
4403
4404<h5>Overview:</h5>
4405<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
4406
4407<h5>Arguments:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4408<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]4409 <a href="#t_integer">integer</a> type, and a type to cast it to, which must
4410 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]4411 <tt>value</tt> must be smaller than the bit size of the destination type,
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4412 <tt>ty2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4413
4414<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4415<p>The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
4416 bit (highest order bit) of the <tt>value</tt> until it reaches the bit size
4417 of the type <tt>ty2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4418
4419<p>When sign extending from i1, the extension always results in -1 or 0.</p>
4420
4421<h5>Example:</h5>
4422<pre>
4423 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
4424 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
4425</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4426
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4427</div>
4428
4429<!-- _______________________________________________________________________ -->
4430<div class="doc_subsubsection">
4431 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
4432</div>
4433
4434<div class="doc_text">
4435
4436<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4437<pre>
4438 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4439</pre>
4440
4441<h5>Overview:</h5>
4442<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4443 <tt>ty2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4444
4445<h5>Arguments:</h5>
4446<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4447 point</a> value to cast and a <a href="#t_floating">floating point</a> type
4448 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]4449 <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]4450 <i>no-op cast</i>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4451
4452<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4453<p>The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4454 <a href="#t_floating">floating point</a> type to a smaller
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4455 <a href="#t_floating">floating point</a> type. If the value cannot fit
4456 within the destination type, <tt>ty2</tt>, then the results are
4457 undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4458
4459<h5>Example:</h5>
4460<pre>
4461 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
4462 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
4463</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4464
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4465</div>
4466
4467<!-- _______________________________________________________________________ -->
4468<div class="doc_subsubsection">
4469 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
4470</div>
4471<div class="doc_text">
4472
4473<h5>Syntax:</h5>
4474<pre>
4475 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4476</pre>
4477
4478<h5>Overview:</h5>
4479<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]4480 floating point value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4481
4482<h5>Arguments:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4483<p>The '<tt>fpext</tt>' instruction takes a
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4484 <a href="#t_floating">floating point</a> <tt>value</tt> to cast, and
4485 a <a href="#t_floating">floating point</a> type to cast it to. The source
4486 type must be smaller than the destination type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4487
4488<h5>Semantics:</h5>
4489<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4490 <a href="#t_floating">floating point</a> type to a larger
4491 <a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
4492 used to make a <i>no-op cast</i> because it always changes bits. Use
4493 <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]4494
4495<h5>Example:</h5>
4496<pre>
4497 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
4498 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
4499</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4500
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4501</div>
4502
4503<!-- _______________________________________________________________________ -->
4504<div class="doc_subsubsection">
4505 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
4506</div>
4507<div class="doc_text">
4508
4509<h5>Syntax:</h5>
4510<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]4511 &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]4512</pre>
4513
4514<h5>Overview:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]4515<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4516 unsigned integer equivalent of type <tt>ty2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4517
4518<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4519<p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
4520 scalar or vector <a href="#t_floating">floating point</a> value, and a type
4521 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4522 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4523 vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4524
4525<h5>Semantics:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4526<p>The '<tt>fptoui</tt>' instruction converts its
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4527 <a href="#t_floating">floating point</a> operand into the nearest (rounding
4528 towards zero) unsigned integer value. If the value cannot fit
4529 in <tt>ty2</tt>, the results are undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4530
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4531<h5>Example:</h5>
4532<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]4533 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]4534 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Gabor Greif463c9342009-10-28 09:21:30 +0000[diff] [blame]4535 %Z = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4536</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4537
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4538</div>
4539
4540<!-- _______________________________________________________________________ -->
4541<div class="doc_subsubsection">
4542 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
4543</div>
4544<div class="doc_text">
4545
4546<h5>Syntax:</h5>
4547<pre>
4548 &lt;result&gt; = fptosi &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4549</pre>
4550
4551<h5>Overview:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4552<p>The '<tt>fptosi</tt>' instruction converts
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4553 <a href="#t_floating">floating point</a> <tt>value</tt> to
4554 type <tt>ty2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4555
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4556<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4557<p>The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
4558 scalar or vector <a href="#t_floating">floating point</a> value, and a type
4559 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4560 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4561 vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4562
4563<h5>Semantics:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4564<p>The '<tt>fptosi</tt>' instruction converts its
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4565 <a href="#t_floating">floating point</a> operand into the nearest (rounding
4566 towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
4567 the results are undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4568
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4569<h5>Example:</h5>
4570<pre>
4571 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]4572 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Gabor Greif463c9342009-10-28 09:21:30 +0000[diff] [blame]4573 %Z = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4574</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4575
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4576</div>
4577
4578<!-- _______________________________________________________________________ -->
4579<div class="doc_subsubsection">
4580 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
4581</div>
4582<div class="doc_text">
4583
4584<h5>Syntax:</h5>
4585<pre>
4586 &lt;result&gt; = uitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4587</pre>
4588
4589<h5>Overview:</h5>
4590<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4591 integer and converts that value to the <tt>ty2</tt> type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4592
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4593<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]4594<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]4595 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
4596 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4597 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4598 floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4599
4600<h5>Semantics:</h5>
4601<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4602 integer quantity and converts it to the corresponding floating point
4603 value. If the value cannot fit in the floating point value, the results are
4604 undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4605
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4606<h5>Example:</h5>
4607<pre>
4608 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4609 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4610</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4611
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4612</div>
4613
4614<!-- _______________________________________________________________________ -->
4615<div class="doc_subsubsection">
4616 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
4617</div>
4618<div class="doc_text">
4619
4620<h5>Syntax:</h5>
4621<pre>
4622 &lt;result&gt; = sitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4623</pre>
4624
4625<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4626<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed integer
4627 and converts that value to the <tt>ty2</tt> type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4628
4629<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]4630<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]4631 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
4632 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4633 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4634 floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4635
4636<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4637<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed integer
4638 quantity and converts it to the corresponding floating point value. If the
4639 value cannot fit in the floating point value, the results are undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4640
4641<h5>Example:</h5>
4642<pre>
4643 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4644 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4645</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4646
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4647</div>
4648
4649<!-- _______________________________________________________________________ -->
4650<div class="doc_subsubsection">
4651 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
4652</div>
4653<div class="doc_text">
4654
4655<h5>Syntax:</h5>
4656<pre>
4657 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4658</pre>
4659
4660<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4661<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
4662 the integer type <tt>ty2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4663
4664<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4665<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
4666 must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
4667 <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]4668
4669<h5>Semantics:</h5>
4670<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4671 <tt>ty2</tt> by interpreting the pointer value as an integer and either
4672 truncating or zero extending that value to the size of the integer type. If
4673 <tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
4674 <tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
4675 are the same size, then nothing is done (<i>no-op cast</i>) other than a type
4676 change.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4677
4678<h5>Example:</h5>
4679<pre>
4680 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
4681 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
4682</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4683
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4684</div>
4685
4686<!-- _______________________________________________________________________ -->
4687<div class="doc_subsubsection">
4688 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
4689</div>
4690<div class="doc_text">
4691
4692<h5>Syntax:</h5>
4693<pre>
4694 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4695</pre>
4696
4697<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4698<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to a
4699 pointer type, <tt>ty2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4700
4701<h5>Arguments:</h5>
4702<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4703 value to cast, and a type to cast it to, which must be a
4704 <a href="#t_pointer">pointer</a> type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4705
4706<h5>Semantics:</h5>
4707<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4708 <tt>ty2</tt> by applying either a zero extension or a truncation depending on
4709 the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
4710 size of a pointer then a truncation is done. If <tt>value</tt> is smaller
4711 than the size of a pointer then a zero extension is done. If they are the
4712 same size, nothing is done (<i>no-op cast</i>).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4713
4714<h5>Example:</h5>
4715<pre>
4716 %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]4717 %Y = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
4718 %Z = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4719</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4720
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4721</div>
4722
4723<!-- _______________________________________________________________________ -->
4724<div class="doc_subsubsection">
4725 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
4726</div>
4727<div class="doc_text">
4728
4729<h5>Syntax:</h5>
4730<pre>
4731 &lt;result&gt; = bitcast &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4732</pre>
4733
4734<h5>Overview:</h5>
4735<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4736 <tt>ty2</tt> without changing any bits.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4737
4738<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4739<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be a
4740 non-aggregate first class value, and a type to cast it to, which must also be
4741 a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes
4742 of <tt>value</tt> and the destination type, <tt>ty2</tt>, must be
4743 identical. If the source type is a pointer, the destination type must also be
4744 a pointer. This instruction supports bitwise conversion of vectors to
4745 integers and to vectors of other types (as long as they have the same
4746 size).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4747
4748<h5>Semantics:</h5>
4749<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4750 <tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
4751 this conversion. The conversion is done as if the <tt>value</tt> had been
4752 stored to memory and read back as type <tt>ty2</tt>. Pointer types may only
4753 be converted to other pointer types with this instruction. To convert
4754 pointers to other types, use the <a href="#i_inttoptr">inttoptr</a> or
4755 <a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4756
4757<h5>Example:</h5>
4758<pre>
4759 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
4760 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4761 %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]4762</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4763
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4764</div>
4765
4766<!-- ======================================================================= -->
4767<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4768
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4769<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4770
4771<p>The instructions in this category are the "miscellaneous" instructions, which
4772 defy better classification.</p>
4773
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4774</div>
4775
4776<!-- _______________________________________________________________________ -->
4777<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
4778</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4779
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4780<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4781
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4782<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4783<pre>
4784 &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]4785</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4786
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4787<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4788<p>The '<tt>icmp</tt>' instruction returns a boolean value or a vector of
4789 boolean values based on comparison of its two integer, integer vector, or
4790 pointer operands.</p>
4791
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4792<h5>Arguments:</h5>
4793<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4794 the condition code indicating the kind of comparison to perform. It is not a
4795 value, just a keyword. The possible condition code are:</p>
4796
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4797<ol>
4798 <li><tt>eq</tt>: equal</li>
4799 <li><tt>ne</tt>: not equal </li>
4800 <li><tt>ugt</tt>: unsigned greater than</li>
4801 <li><tt>uge</tt>: unsigned greater or equal</li>
4802 <li><tt>ult</tt>: unsigned less than</li>
4803 <li><tt>ule</tt>: unsigned less or equal</li>
4804 <li><tt>sgt</tt>: signed greater than</li>
4805 <li><tt>sge</tt>: signed greater or equal</li>
4806 <li><tt>slt</tt>: signed less than</li>
4807 <li><tt>sle</tt>: signed less or equal</li>
4808</ol>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4809
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4810<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4811 <a href="#t_pointer">pointer</a> or integer <a href="#t_vector">vector</a>
4812 typed. They must also be identical types.</p>
4813
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4814<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4815<p>The '<tt>icmp</tt>' compares <tt>op1</tt> and <tt>op2</tt> according to the
4816 condition code given as <tt>cond</tt>. The comparison performed always yields
Nick Lewycky244cf482009-09-27 00:45:11 +0000[diff] [blame]4817 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]4818 result, as follows:</p>
4819
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4820<ol>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4821 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4822 <tt>false</tt> otherwise. No sign interpretation is necessary or
4823 performed.</li>
4824
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4825 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4826 <tt>false</tt> otherwise. No sign interpretation is necessary or
4827 performed.</li>
4828
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4829 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4830 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
4831
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4832 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4833 <tt>true</tt> if <tt>op1</tt> is greater than or equal
4834 to <tt>op2</tt>.</li>
4835
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4836 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4837 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
4838
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4839 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4840 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
4841
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4842 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4843 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
4844
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4845 <li><tt>sge</tt>: interprets the operands as signed values and yields
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4846 <tt>true</tt> if <tt>op1</tt> is greater than or equal
4847 to <tt>op2</tt>.</li>
4848
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4849 <li><tt>slt</tt>: interprets the operands as signed values and yields
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4850 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
4851
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4852 <li><tt>sle</tt>: interprets the operands as signed values and yields
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4853 <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]4854</ol>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4855
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4856<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4857 values are compared as if they were integers.</p>
4858
4859<p>If the operands are integer vectors, then they are compared element by
4860 element. The result is an <tt>i1</tt> vector with the same number of elements
4861 as the values being compared. Otherwise, the result is an <tt>i1</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4862
4863<h5>Example:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4864<pre>
4865 &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4866 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
4867 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
4868 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
4869 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
4870 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
4871</pre>
Dan Gohmana53eb382009-01-22 01:39:38 +0000[diff] [blame]4872
4873<p>Note that the code generator does not yet support vector types with
4874 the <tt>icmp</tt> instruction.</p>
4875
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4876</div>
4877
4878<!-- _______________________________________________________________________ -->
4879<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
4880</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4881
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4882<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4883
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4884<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4885<pre>
4886 &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]4887</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4888
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4889<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4890<p>The '<tt>fcmp</tt>' instruction returns a boolean value or vector of boolean
4891 values based on comparison of its operands.</p>
4892
4893<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]4894(<a href="#t_integer"><tt>i1</tt></a>).</p>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4895
4896<p>If the operands are floating point vectors, then the result type is a vector
4897 of boolean with the same number of elements as the operands being
4898 compared.</p>
4899
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4900<h5>Arguments:</h5>
4901<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4902 the condition code indicating the kind of comparison to perform. It is not a
4903 value, just a keyword. The possible condition code are:</p>
4904
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4905<ol>
4906 <li><tt>false</tt>: no comparison, always returns false</li>
4907 <li><tt>oeq</tt>: ordered and equal</li>
4908 <li><tt>ogt</tt>: ordered and greater than </li>
4909 <li><tt>oge</tt>: ordered and greater than or equal</li>
4910 <li><tt>olt</tt>: ordered and less than </li>
4911 <li><tt>ole</tt>: ordered and less than or equal</li>
4912 <li><tt>one</tt>: ordered and not equal</li>
4913 <li><tt>ord</tt>: ordered (no nans)</li>
4914 <li><tt>ueq</tt>: unordered or equal</li>
4915 <li><tt>ugt</tt>: unordered or greater than </li>
4916 <li><tt>uge</tt>: unordered or greater than or equal</li>
4917 <li><tt>ult</tt>: unordered or less than </li>
4918 <li><tt>ule</tt>: unordered or less than or equal</li>
4919 <li><tt>une</tt>: unordered or not equal</li>
4920 <li><tt>uno</tt>: unordered (either nans)</li>
4921 <li><tt>true</tt>: no comparison, always returns true</li>
4922</ol>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4923
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4924<p><i>Ordered</i> means that neither operand is a QNAN while
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4925 <i>unordered</i> means that either operand may be a QNAN.</p>
4926
4927<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be either
4928 a <a href="#t_floating">floating point</a> type or
4929 a <a href="#t_vector">vector</a> of floating point type. They must have
4930 identical types.</p>
4931
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4932<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4933<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4934 according to the condition code given as <tt>cond</tt>. If the operands are
4935 vectors, then the vectors are compared element by element. Each comparison
Nick Lewycky244cf482009-09-27 00:45:11 +0000[diff] [blame]4936 performed always yields an <a href="#t_integer">i1</a> result, as
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4937 follows:</p>
4938
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4939<ol>
4940 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4941
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4942 <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]4943 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
4944
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4945 <li><tt>ogt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4946 <tt>op1</tt> is greather than <tt>op2</tt>.</li>
4947
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4948 <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]4949 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
4950
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4951 <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]4952 <tt>op1</tt> is less than <tt>op2</tt>.</li>
4953
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4954 <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]4955 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
4956
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4957 <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]4958 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
4959
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4960 <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]4961
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4962 <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]4963 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
4964
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4965 <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]4966 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
4967
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4968 <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]4969 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
4970
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4971 <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]4972 <tt>op1</tt> is less than <tt>op2</tt>.</li>
4973
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4974 <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]4975 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
4976
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]4977 <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]4978 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
4979
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4980 <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]4981
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4982 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
4983</ol>
4984
4985<h5>Example:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]4986<pre>
4987 &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]4988 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
4989 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
4990 &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]4991</pre>
Dan Gohmana53eb382009-01-22 01:39:38 +0000[diff] [blame]4992
4993<p>Note that the code generator does not yet support vector types with
4994 the <tt>fcmp</tt> instruction.</p>
4995
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4996</div>
4997
4998<!-- _______________________________________________________________________ -->
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4999<div class="doc_subsubsection">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]5000 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
5001</div>
5002
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5003<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]5004
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5005<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5006<pre>
5007 &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...
5008</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]5009
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5010<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5011<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in the
5012 SSA graph representing the function.</p>
5013
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5014<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5015<p>The type of the incoming values is specified with the first type field. After
5016 this, the '<tt>phi</tt>' instruction takes a list of pairs as arguments, with
5017 one pair for each predecessor basic block of the current block. Only values
5018 of <a href="#t_firstclass">first class</a> type may be used as the value
5019 arguments to the PHI node. Only labels may be used as the label
5020 arguments.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]5021
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5022<p>There must be no non-phi instructions between the start of a basic block and
5023 the PHI instructions: i.e. PHI instructions must be first in a basic
5024 block.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]5025
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5026<p>For the purposes of the SSA form, the use of each incoming value is deemed to
5027 occur on the edge from the corresponding predecessor block to the current
5028 block (but after any definition of an '<tt>invoke</tt>' instruction's return
5029 value on the same edge).</p>
Jay Foad8e2fd2c2009-06-03 10:20:10 +0000[diff] [blame]5030
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5031<h5>Semantics:</h5>
5032<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5033 specified by the pair corresponding to the predecessor basic block that
5034 executed just prior to the current block.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]5035
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5036<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]5037<pre>
5038Loop: ; Infinite loop that counts from 0 on up...
5039 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
5040 %nextindvar = add i32 %indvar, 1
5041 br label %Loop
5042</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5043
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5044</div>
5045
5046<!-- _______________________________________________________________________ -->
5047<div class="doc_subsubsection">
5048 <a name="i_select">'<tt>select</tt>' Instruction</a>
5049</div>
5050
5051<div class="doc_text">
5052
5053<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5054<pre>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]5055 &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>
5056
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5057 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5058</pre>
5059
5060<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5061<p>The '<tt>select</tt>' instruction is used to choose one value based on a
5062 condition, without branching.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5063
5064
5065<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5066<p>The '<tt>select</tt>' instruction requires an 'i1' value or a vector of 'i1'
5067 values indicating the condition, and two values of the
5068 same <a href="#t_firstclass">first class</a> type. If the val1/val2 are
5069 vectors and the condition is a scalar, then entire vectors are selected, not
5070 individual elements.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5071
5072<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5073<p>If the condition is an i1 and it evaluates to 1, the instruction returns the
5074 first value argument; otherwise, it returns the second value argument.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5075
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5076<p>If the condition is a vector of i1, then the value arguments must be vectors
5077 of the same size, and the selection is done element by element.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5078
5079<h5>Example:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5080<pre>
5081 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
5082</pre>
Dan Gohmana53eb382009-01-22 01:39:38 +0000[diff] [blame]5083
5084<p>Note that the code generator does not yet support conditions
5085 with vector type.</p>
5086
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5087</div>
5088
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5089<!-- _______________________________________________________________________ -->
5090<div class="doc_subsubsection">
5091 <a name="i_call">'<tt>call</tt>' Instruction</a>
5092</div>
5093
5094<div class="doc_text">
5095
5096<h5>Syntax:</h5>
5097<pre>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]5098 &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]5099</pre>
5100
5101<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5102<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
5103
5104<h5>Arguments:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5105<p>This instruction requires several arguments:</p>
5106
5107<ol>
Jeffrey Yasskin3e1a3a02010-01-09 19:44:16 +0000[diff] [blame]5108 <li>The optional "tail" marker indicates that the callee function does not
5109 access any allocas or varargs in the caller. Note that calls may be
5110 marked "tail" even if they do not occur before
5111 a <a href="#i_ret"><tt>ret</tt></a> instruction. If the "tail" marker is
5112 present, the function call is eligible for tail call optimization,
5113 but <a href="CodeGenerator.html#tailcallopt">might not in fact be
5114 optimized into a jump</a>. As of this writing, the extra requirements for
5115 a call to actually be optimized are:
5116 <ul>
5117 <li>Caller and callee both have the calling
5118 convention <tt>fastcc</tt>.</li>
5119 <li>The call is in tail position (ret immediately follows call and ret
5120 uses value of call or is void).</li>
5121 <li>Option <tt>-tailcallopt</tt> is enabled,
5122 or <code>llvm::PerformTailCallOpt</code> is <code>true</code>.</li>
5123 <li><a href="CodeGenerator.html#tailcallopt">Platform specific
5124 constraints are met.</a></li>
5125 </ul>
5126 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]5127
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5128 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
5129 convention</a> the call should use. If none is specified, the call
Jeffrey Yasskin3e1a3a02010-01-09 19:44:16 +0000[diff] [blame]5130 defaults to using C calling conventions. The calling convention of the
5131 call must match the calling convention of the target function, or else the
5132 behavior is undefined.</li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]5133
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5134 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
5135 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
5136 '<tt>inreg</tt>' attributes are valid here.</li>
5137
5138 <li>'<tt>ty</tt>': the type of the call instruction itself which is also the
5139 type of the return value. Functions that return no value are marked
5140 <tt><a href="#t_void">void</a></tt>.</li>
5141
5142 <li>'<tt>fnty</tt>': shall be the signature of the pointer to function value
5143 being invoked. The argument types must match the types implied by this
5144 signature. This type can be omitted if the function is not varargs and if
5145 the function type does not return a pointer to a function.</li>
5146
5147 <li>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
5148 be invoked. In most cases, this is a direct function invocation, but
5149 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
5150 to function value.</li>
5151
5152 <li>'<tt>function args</tt>': argument list whose types match the function
5153 signature argument types. All arguments must be of
5154 <a href="#t_firstclass">first class</a> type. If the function signature
5155 indicates the function accepts a variable number of arguments, the extra
5156 arguments can be specified.</li>
5157
5158 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
5159 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
5160 '<tt>readnone</tt>' attributes are valid here.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5161</ol>
5162
5163<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5164<p>The '<tt>call</tt>' instruction is used to cause control flow to transfer to
5165 a specified function, with its incoming arguments bound to the specified
5166 values. Upon a '<tt><a href="#i_ret">ret</a></tt>' instruction in the called
5167 function, control flow continues with the instruction after the function
5168 call, and the return value of the function is bound to the result
5169 argument.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5170
5171<h5>Example:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5172<pre>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]5173 %retval = call i32 @test(i32 %argc)
Chris Lattner5e893ef2008-03-21 17:24:17 +0000[diff] [blame]5174 call i32 (i8 *, ...)* @printf(i8 * %msg, i32 12, i8 42) <i>; yields i32</i>
5175 %X = tail call i32 @foo() <i>; yields i32</i>
5176 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
5177 call void %foo(i8 97 signext)
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]5178
5179 %struct.A = type { i32, i8 }
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]5180 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]5181 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
5182 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattnerac454b32008-10-08 06:26:11 +0000[diff] [blame]5183 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijman2c4e05a2008-10-07 10:03:45 +0000[diff] [blame]5184 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5185</pre>
5186
Dale Johannesen9fcf4a92009-09-24 18:38:21 +0000[diff] [blame]5187<p>llvm treats calls to some functions with names and arguments that match the
Dale Johannesen4730cd32009-09-25 17:04:42 +0000[diff] [blame]5188standard C99 library as being the C99 library functions, and may perform
5189optimizations or generate code for them under that assumption. This is
5190something we'd like to change in the future to provide better support for
5191freestanding environments and non-C-based langauges.</p>
Dale Johannesen9fcf4a92009-09-24 18:38:21 +0000[diff] [blame]5192
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5193</div>
5194
5195<!-- _______________________________________________________________________ -->
5196<div class="doc_subsubsection">
5197 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
5198</div>
5199
5200<div class="doc_text">
5201
5202<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5203<pre>
5204 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
5205</pre>
5206
5207<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5208<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5209 the "variable argument" area of a function call. It is used to implement the
5210 <tt>va_arg</tt> macro in C.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5211
5212<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5213<p>This instruction takes a <tt>va_list*</tt> value and the type of the
5214 argument. It returns a value of the specified argument type and increments
5215 the <tt>va_list</tt> to point to the next argument. The actual type
5216 of <tt>va_list</tt> is target specific.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5217
5218<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5219<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified type
5220 from the specified <tt>va_list</tt> and causes the <tt>va_list</tt> to point
5221 to the next argument. For more information, see the variable argument
5222 handling <a href="#int_varargs">Intrinsic Functions</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5223
5224<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]5225 take a variable number of arguments, for example, the <tt>vfprintf</tt>
5226 function.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5227
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5228<p><tt>va_arg</tt> is an LLVM instruction instead of
5229 an <a href="#intrinsics">intrinsic function</a> because it takes a type as an
5230 argument.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5231
5232<h5>Example:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5233<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
5234
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5235<p>Note that the code generator does not yet fully support va_arg on many
5236 targets. Also, it does not currently support va_arg with aggregate types on
5237 any target.</p>
Dan Gohman60967192009-01-12 23:12:39 +0000[diff] [blame]5238
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5239</div>
5240
5241<!-- *********************************************************************** -->
5242<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
5243<!-- *********************************************************************** -->
5244
5245<div class="doc_text">
5246
5247<p>LLVM supports the notion of an "intrinsic function". These functions have
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5248 well known names and semantics and are required to follow certain
5249 restrictions. Overall, these intrinsics represent an extension mechanism for
5250 the LLVM language that does not require changing all of the transformations
5251 in LLVM when adding to the language (or the bitcode reader/writer, the
5252 parser, etc...).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5253
5254<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5255 prefix is reserved in LLVM for intrinsic names; thus, function names may not
5256 begin with this prefix. Intrinsic functions must always be external
5257 functions: you cannot define the body of intrinsic functions. Intrinsic
5258 functions may only be used in call or invoke instructions: it is illegal to
5259 take the address of an intrinsic function. Additionally, because intrinsic
5260 functions are part of the LLVM language, it is required if any are added that
5261 they be documented here.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5262
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5263<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents a
5264 family of functions that perform the same operation but on different data
5265 types. Because LLVM can represent over 8 million different integer types,
5266 overloading is used commonly to allow an intrinsic function to operate on any
5267 integer type. One or more of the argument types or the result type can be
5268 overloaded to accept any integer type. Argument types may also be defined as
5269 exactly matching a previous argument's type or the result type. This allows
5270 an intrinsic function which accepts multiple arguments, but needs all of them
5271 to be of the same type, to only be overloaded with respect to a single
5272 argument or the result.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5273
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5274<p>Overloaded intrinsics will have the names of its overloaded argument types
5275 encoded into its function name, each preceded by a period. Only those types
5276 which are overloaded result in a name suffix. Arguments whose type is matched
5277 against another type do not. For example, the <tt>llvm.ctpop</tt> function
5278 can take an integer of any width and returns an integer of exactly the same
5279 integer width. This leads to a family of functions such as
5280 <tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29
5281 %val)</tt>. Only one type, the return type, is overloaded, and only one type
5282 suffix is required. Because the argument's type is matched against the return
5283 type, it does not require its own name suffix.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5284
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]5285<p>To learn how to add an intrinsic function, please see the
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5286 <a href="ExtendingLLVM.html">Extending LLVM Guide</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5287
5288</div>
5289
5290<!-- ======================================================================= -->
5291<div class="doc_subsection">
5292 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
5293</div>
5294
5295<div class="doc_text">
5296
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5297<p>Variable argument support is defined in LLVM with
5298 the <a href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
5299 intrinsic functions. These functions are related to the similarly named
5300 macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5301
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5302<p>All of these functions operate on arguments that use a target-specific value
5303 type "<tt>va_list</tt>". The LLVM assembly language reference manual does
5304 not define what this type is, so all transformations should be prepared to
5305 handle these functions regardless of the type used.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5306
5307<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]5308 instruction and the variable argument handling intrinsic functions are
5309 used.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5310
5311<div class="doc_code">
5312<pre>
5313define i32 @test(i32 %X, ...) {
5314 ; Initialize variable argument processing
5315 %ap = alloca i8*
5316 %ap2 = bitcast i8** %ap to i8*
5317 call void @llvm.va_start(i8* %ap2)
5318
5319 ; Read a single integer argument
5320 %tmp = va_arg i8** %ap, i32
5321
5322 ; Demonstrate usage of llvm.va_copy and llvm.va_end
5323 %aq = alloca i8*
5324 %aq2 = bitcast i8** %aq to i8*
5325 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
5326 call void @llvm.va_end(i8* %aq2)
5327
5328 ; Stop processing of arguments.
5329 call void @llvm.va_end(i8* %ap2)
5330 ret i32 %tmp
5331}
5332
5333declare void @llvm.va_start(i8*)
5334declare void @llvm.va_copy(i8*, i8*)
5335declare void @llvm.va_end(i8*)
5336</pre>
5337</div>
5338
5339</div>
5340
5341<!-- _______________________________________________________________________ -->
5342<div class="doc_subsubsection">
5343 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
5344</div>
5345
5346
5347<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5348
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5349<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5350<pre>
5351 declare void %llvm.va_start(i8* &lt;arglist&gt;)
5352</pre>
5353
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5354<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5355<p>The '<tt>llvm.va_start</tt>' intrinsic initializes <tt>*&lt;arglist&gt;</tt>
5356 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]5357
5358<h5>Arguments:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5359<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]5360
5361<h5>Semantics:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5362<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]5363 macro available in C. In a target-dependent way, it initializes
5364 the <tt>va_list</tt> element to which the argument points, so that the next
5365 call to <tt>va_arg</tt> will produce the first variable argument passed to
5366 the function. Unlike the C <tt>va_start</tt> macro, this intrinsic does not
5367 need to know the last argument of the function as the compiler can figure
5368 that out.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5369
5370</div>
5371
5372<!-- _______________________________________________________________________ -->
5373<div class="doc_subsubsection">
5374 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
5375</div>
5376
5377<div class="doc_text">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5378
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5379<h5>Syntax:</h5>
5380<pre>
5381 declare void @llvm.va_end(i8* &lt;arglist&gt;)
5382</pre>
5383
5384<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5385<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5386 which has been initialized previously
5387 with <tt><a href="#int_va_start">llvm.va_start</a></tt>
5388 or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5389
5390<h5>Arguments:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5391<p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
5392
5393<h5>Semantics:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5394<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]5395 macro available in C. In a target-dependent way, it destroys
5396 the <tt>va_list</tt> element to which the argument points. Calls
5397 to <a href="#int_va_start"><tt>llvm.va_start</tt></a>
5398 and <a href="#int_va_copy"> <tt>llvm.va_copy</tt></a> must be matched exactly
5399 with calls to <tt>llvm.va_end</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5400
5401</div>
5402
5403<!-- _______________________________________________________________________ -->
5404<div class="doc_subsubsection">
5405 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
5406</div>
5407
5408<div class="doc_text">
5409
5410<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5411<pre>
5412 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
5413</pre>
5414
5415<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5416<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5417 from the source argument list to the destination argument list.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5418
5419<h5>Arguments:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5420<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]5421 The second argument is a pointer to a <tt>va_list</tt> element to copy
5422 from.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5423
5424<h5>Semantics:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5425<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]5426 macro available in C. In a target-dependent way, it copies the
5427 source <tt>va_list</tt> element into the destination <tt>va_list</tt>
5428 element. This intrinsic is necessary because
5429 the <tt><a href="#int_va_start"> llvm.va_start</a></tt> intrinsic may be
5430 arbitrarily complex and require, for example, memory allocation.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5431
5432</div>
5433
5434<!-- ======================================================================= -->
5435<div class="doc_subsection">
5436 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
5437</div>
5438
5439<div class="doc_text">
5440
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5441<p>LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]5442Collection</a> (GC) requires the implementation and generation of these
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5443intrinsics. These intrinsics allow identification of <a href="#int_gcroot">GC
5444roots on the stack</a>, as well as garbage collector implementations that
5445require <a href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a>
5446barriers. Front-ends for type-safe garbage collected languages should generate
5447these intrinsics to make use of the LLVM garbage collectors. For more details,
5448see <a href="GarbageCollection.html">Accurate Garbage Collection with
5449LLVM</a>.</p>
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]5450
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5451<p>The garbage collection intrinsics only operate on objects in the generic
5452 address space (address space zero).</p>
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]5453
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5454</div>
5455
5456<!-- _______________________________________________________________________ -->
5457<div class="doc_subsubsection">
5458 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
5459</div>
5460
5461<div class="doc_text">
5462
5463<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5464<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5465 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5466</pre>
5467
5468<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5469<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]5470 the code generator, and allows some metadata to be associated with it.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5471
5472<h5>Arguments:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5473<p>The first argument specifies the address of a stack object that contains the
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5474 root pointer. The second pointer (which must be either a constant or a
5475 global value address) contains the meta-data to be associated with the
5476 root.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5477
5478<h5>Semantics:</h5>
Chris Lattnera7d94ba2008-04-24 05:59:56 +0000[diff] [blame]5479<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]5480 location. At compile-time, the code generator generates information to allow
5481 the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
5482 intrinsic may only be used in a function which <a href="#gc">specifies a GC
5483 algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5484
5485</div>
5486
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5487<!-- _______________________________________________________________________ -->
5488<div class="doc_subsubsection">
5489 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
5490</div>
5491
5492<div class="doc_text">
5493
5494<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5495<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5496 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5497</pre>
5498
5499<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5500<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5501 locations, allowing garbage collector implementations that require read
5502 barriers.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5503
5504<h5>Arguments:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5505<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]5506 allocated from the garbage collector. The first object is a pointer to the
5507 start of the referenced object, if needed by the language runtime (otherwise
5508 null).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5509
5510<h5>Semantics:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5511<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5512 instruction, but may be replaced with substantially more complex code by the
5513 garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
5514 may only be used in a function which <a href="#gc">specifies a GC
5515 algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5516
5517</div>
5518
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5519<!-- _______________________________________________________________________ -->
5520<div class="doc_subsubsection">
5521 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
5522</div>
5523
5524<div class="doc_text">
5525
5526<h5>Syntax:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5527<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5528 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5529</pre>
5530
5531<h5>Overview:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5532<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5533 locations, allowing garbage collector implementations that require write
5534 barriers (such as generational or reference counting collectors).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5535
5536<h5>Arguments:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5537<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]5538 object to store it to, and the third is the address of the field of Obj to
5539 store to. If the runtime does not require a pointer to the object, Obj may
5540 be null.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5541
5542<h5>Semantics:</h5>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5543<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5544 instruction, but may be replaced with substantially more complex code by the
5545 garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
5546 may only be used in a function which <a href="#gc">specifies a GC
5547 algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5548
5549</div>
5550
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5551<!-- ======================================================================= -->
5552<div class="doc_subsection">
5553 <a name="int_codegen">Code Generator Intrinsics</a>
5554</div>
5555
5556<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5557
5558<p>These intrinsics are provided by LLVM to expose special features that may
5559 only be implemented with code generator support.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5560
5561</div>
5562
5563<!-- _______________________________________________________________________ -->
5564<div class="doc_subsubsection">
5565 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
5566</div>
5567
5568<div class="doc_text">
5569
5570<h5>Syntax:</h5>
5571<pre>
5572 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
5573</pre>
5574
5575<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5576<p>The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
5577 target-specific value indicating the return address of the current function
5578 or one of its callers.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5579
5580<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5581<p>The argument to this intrinsic indicates which function to return the address
5582 for. Zero indicates the calling function, one indicates its caller, etc.
5583 The argument is <b>required</b> to be a constant integer value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5584
5585<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5586<p>The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer
5587 indicating the return address of the specified call frame, or zero if it
5588 cannot be identified. The value returned by this intrinsic is likely to be
5589 incorrect or 0 for arguments other than zero, so it should only be used for
5590 debugging purposes.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5591
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5592<p>Note that calling this intrinsic does not prevent function inlining or other
5593 aggressive transformations, so the value returned may not be that of the
5594 obvious source-language caller.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5595
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5596</div>
5597
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5598<!-- _______________________________________________________________________ -->
5599<div class="doc_subsubsection">
5600 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
5601</div>
5602
5603<div class="doc_text">
5604
5605<h5>Syntax:</h5>
5606<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5607 declare i8 *@llvm.frameaddress(i32 &lt;level&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5608</pre>
5609
5610<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5611<p>The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
5612 target-specific frame pointer value for the specified stack frame.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5613
5614<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5615<p>The argument to this intrinsic indicates which function to return the frame
5616 pointer for. Zero indicates the calling function, one indicates its caller,
5617 etc. The argument is <b>required</b> to be a constant integer value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5618
5619<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5620<p>The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer
5621 indicating the frame address of the specified call frame, or zero if it
5622 cannot be identified. The value returned by this intrinsic is likely to be
5623 incorrect or 0 for arguments other than zero, so it should only be used for
5624 debugging purposes.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5625
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5626<p>Note that calling this intrinsic does not prevent function inlining or other
5627 aggressive transformations, so the value returned may not be that of the
5628 obvious source-language caller.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5629
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5630</div>
5631
5632<!-- _______________________________________________________________________ -->
5633<div class="doc_subsubsection">
5634 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
5635</div>
5636
5637<div class="doc_text">
5638
5639<h5>Syntax:</h5>
5640<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5641 declare i8 *@llvm.stacksave()
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5642</pre>
5643
5644<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5645<p>The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state
5646 of the function stack, for use
5647 with <a href="#int_stackrestore"> <tt>llvm.stackrestore</tt></a>. This is
5648 useful for implementing language features like scoped automatic variable
5649 sized arrays in C99.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5650
5651<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5652<p>This intrinsic returns a opaque pointer value that can be passed
5653 to <a href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When
5654 an <tt>llvm.stackrestore</tt> intrinsic is executed with a value saved
5655 from <tt>llvm.stacksave</tt>, it effectively restores the state of the stack
5656 to the state it was in when the <tt>llvm.stacksave</tt> intrinsic executed.
5657 In practice, this pops any <a href="#i_alloca">alloca</a> blocks from the
5658 stack that were allocated after the <tt>llvm.stacksave</tt> was executed.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5659
5660</div>
5661
5662<!-- _______________________________________________________________________ -->
5663<div class="doc_subsubsection">
5664 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
5665</div>
5666
5667<div class="doc_text">
5668
5669<h5>Syntax:</h5>
5670<pre>
5671 declare void @llvm.stackrestore(i8 * %ptr)
5672</pre>
5673
5674<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5675<p>The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
5676 the function stack to the state it was in when the
5677 corresponding <a href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic
5678 executed. This is useful for implementing language features like scoped
5679 automatic variable sized arrays in C99.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5680
5681<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5682<p>See the description
5683 for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5684
5685</div>
5686
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5687<!-- _______________________________________________________________________ -->
5688<div class="doc_subsubsection">
5689 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
5690</div>
5691
5692<div class="doc_text">
5693
5694<h5>Syntax:</h5>
5695<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5696 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]5697</pre>
5698
5699<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5700<p>The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to
5701 insert a prefetch instruction if supported; otherwise, it is a noop.
5702 Prefetches have no effect on the behavior of the program but can change its
5703 performance characteristics.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5704
5705<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5706<p><tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the
5707 specifier determining if the fetch should be for a read (0) or write (1),
5708 and <tt>locality</tt> is a temporal locality specifier ranging from (0) - no
5709 locality, to (3) - extremely local keep in cache. The <tt>rw</tt>
5710 and <tt>locality</tt> arguments must be constant integers.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5711
5712<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5713<p>This intrinsic does not modify the behavior of the program. In particular,
5714 prefetches cannot trap and do not produce a value. On targets that support
5715 this intrinsic, the prefetch can provide hints to the processor cache for
5716 better performance.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5717
5718</div>
5719
5720<!-- _______________________________________________________________________ -->
5721<div class="doc_subsubsection">
5722 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
5723</div>
5724
5725<div class="doc_text">
5726
5727<h5>Syntax:</h5>
5728<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5729 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5730</pre>
5731
5732<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5733<p>The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program
5734 Counter (PC) in a region of code to simulators and other tools. The method
5735 is target specific, but it is expected that the marker will use exported
5736 symbols to transmit the PC of the marker. The marker makes no guarantees
5737 that it will remain with any specific instruction after optimizations. It is
5738 possible that the presence of a marker will inhibit optimizations. The
5739 intended use is to be inserted after optimizations to allow correlations of
5740 simulation runs.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5741
5742<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5743<p><tt>id</tt> is a numerical id identifying the marker.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5744
5745<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5746<p>This intrinsic does not modify the behavior of the program. Backends that do
5747 not support this intrinisic may ignore it.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5748
5749</div>
5750
5751<!-- _______________________________________________________________________ -->
5752<div class="doc_subsubsection">
5753 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
5754</div>
5755
5756<div class="doc_text">
5757
5758<h5>Syntax:</h5>
5759<pre>
5760 declare i64 @llvm.readcyclecounter( )
5761</pre>
5762
5763<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5764<p>The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
5765 counter register (or similar low latency, high accuracy clocks) on those
5766 targets that support it. On X86, it should map to RDTSC. On Alpha, it
5767 should map to RPCC. As the backing counters overflow quickly (on the order
5768 of 9 seconds on alpha), this should only be used for small timings.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5769
5770<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5771<p>When directly supported, reading the cycle counter should not modify any
5772 memory. Implementations are allowed to either return a application specific
5773 value or a system wide value. On backends without support, this is lowered
5774 to a constant 0.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5775
5776</div>
5777
5778<!-- ======================================================================= -->
5779<div class="doc_subsection">
5780 <a name="int_libc">Standard C Library Intrinsics</a>
5781</div>
5782
5783<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5784
5785<p>LLVM provides intrinsics for a few important standard C library functions.
5786 These intrinsics allow source-language front-ends to pass information about
5787 the alignment of the pointer arguments to the code generator, providing
5788 opportunity for more efficient code generation.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5789
5790</div>
5791
5792<!-- _______________________________________________________________________ -->
5793<div class="doc_subsubsection">
5794 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
5795</div>
5796
5797<div class="doc_text">
5798
5799<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5800<p>This is an overloaded intrinsic. You can use <tt>llvm.memcpy</tt> on any
5801 integer bit width. Not all targets support all bit widths however.</p>
5802
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5803<pre>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5804 declare void @llvm.memcpy.i8(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5805 i8 &lt;len&gt;, i32 &lt;align&gt;)
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5806 declare void @llvm.memcpy.i16(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5807 i16 &lt;len&gt;, i32 &lt;align&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5808 declare void @llvm.memcpy.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5809 i32 &lt;len&gt;, i32 &lt;align&gt;)
5810 declare void @llvm.memcpy.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5811 i64 &lt;len&gt;, i32 &lt;align&gt;)
5812</pre>
5813
5814<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5815<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
5816 source location to the destination location.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5817
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5818<p>Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
5819 intrinsics do not return a value, and takes an extra alignment argument.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5820
5821<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5822<p>The first argument is a pointer to the destination, the second is a pointer
5823 to the source. The third argument is an integer argument specifying the
5824 number of bytes to copy, and the fourth argument is the alignment of the
5825 source and destination locations.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5826
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5827<p>If the call to this intrinisic has an alignment value that is not 0 or 1,
5828 then the caller guarantees that both the source and destination pointers are
5829 aligned to that boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5830
5831<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5832<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
5833 source location to the destination location, which are not allowed to
5834 overlap. It copies "len" bytes of memory over. If the argument is known to
5835 be aligned to some boundary, this can be specified as the fourth argument,
5836 otherwise it should be set to 0 or 1.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5837
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5838</div>
5839
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5840<!-- _______________________________________________________________________ -->
5841<div class="doc_subsubsection">
5842 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
5843</div>
5844
5845<div class="doc_text">
5846
5847<h5>Syntax:</h5>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5848<p>This is an overloaded intrinsic. You can use llvm.memmove on any integer bit
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5849 width. Not all targets support all bit widths however.</p>
5850
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5851<pre>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5852 declare void @llvm.memmove.i8(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5853 i8 &lt;len&gt;, i32 &lt;align&gt;)
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5854 declare void @llvm.memmove.i16(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5855 i16 &lt;len&gt;, i32 &lt;align&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5856 declare void @llvm.memmove.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5857 i32 &lt;len&gt;, i32 &lt;align&gt;)
5858 declare void @llvm.memmove.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5859 i64 &lt;len&gt;, i32 &lt;align&gt;)
5860</pre>
5861
5862<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5863<p>The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the
5864 source location to the destination location. It is similar to the
5865 '<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to
5866 overlap.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5867
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5868<p>Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
5869 intrinsics do not return a value, and takes an extra alignment argument.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5870
5871<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5872<p>The first argument is a pointer to the destination, the second is a pointer
5873 to the source. The third argument is an integer argument specifying the
5874 number of bytes to copy, and the fourth argument is the alignment of the
5875 source and destination locations.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5876
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5877<p>If the call to this intrinisic has an alignment value that is not 0 or 1,
5878 then the caller guarantees that the source and destination pointers are
5879 aligned to that boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5880
5881<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5882<p>The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the
5883 source location to the destination location, which may overlap. It copies
5884 "len" bytes of memory over. If the argument is known to be aligned to some
5885 boundary, this can be specified as the fourth argument, otherwise it should
5886 be set to 0 or 1.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5887
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5888</div>
5889
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5890<!-- _______________________________________________________________________ -->
5891<div class="doc_subsubsection">
5892 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
5893</div>
5894
5895<div class="doc_text">
5896
5897<h5>Syntax:</h5>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5898<p>This is an overloaded intrinsic. You can use llvm.memset on any integer bit
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5899 width. Not all targets support all bit widths however.</p>
5900
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5901<pre>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5902 declare void @llvm.memset.i8(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5903 i8 &lt;len&gt;, i32 &lt;align&gt;)
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5904 declare void @llvm.memset.i16(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5905 i16 &lt;len&gt;, i32 &lt;align&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5906 declare void @llvm.memset.i32(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5907 i32 &lt;len&gt;, i32 &lt;align&gt;)
5908 declare void @llvm.memset.i64(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5909 i64 &lt;len&gt;, i32 &lt;align&gt;)
5910</pre>
5911
5912<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5913<p>The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a
5914 particular byte value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5915
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5916<p>Note that, unlike the standard libc function, the <tt>llvm.memset</tt>
5917 intrinsic does not return a value, and takes an extra alignment argument.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5918
5919<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5920<p>The first argument is a pointer to the destination to fill, the second is the
5921 byte value to fill it with, the third argument is an integer argument
5922 specifying the number of bytes to fill, and the fourth argument is the known
5923 alignment of destination location.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5924
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5925<p>If the call to this intrinisic has an alignment value that is not 0 or 1,
5926 then the caller guarantees that the destination pointer is aligned to that
5927 boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5928
5929<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5930<p>The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting
5931 at the destination location. If the argument is known to be aligned to some
5932 boundary, this can be specified as the fourth argument, otherwise it should
5933 be set to 0 or 1.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5934
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5935</div>
5936
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5937<!-- _______________________________________________________________________ -->
5938<div class="doc_subsubsection">
5939 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
5940</div>
5941
5942<div class="doc_text">
5943
5944<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5945<p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
5946 floating point or vector of floating point type. Not all targets support all
5947 types however.</p>
5948
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5949<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5950 declare float @llvm.sqrt.f32(float %Val)
5951 declare double @llvm.sqrt.f64(double %Val)
5952 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
5953 declare fp128 @llvm.sqrt.f128(fp128 %Val)
5954 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5955</pre>
5956
5957<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5958<p>The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
5959 returning the same value as the libm '<tt>sqrt</tt>' functions would.
5960 Unlike <tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined
5961 behavior for negative numbers other than -0.0 (which allows for better
5962 optimization, because there is no need to worry about errno being
5963 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]5964
5965<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5966<p>The argument and return value are floating point numbers of the same
5967 type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5968
5969<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5970<p>This function returns the sqrt of the specified operand if it is a
5971 nonnegative floating point number.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5972
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5973</div>
5974
5975<!-- _______________________________________________________________________ -->
5976<div class="doc_subsubsection">
5977 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
5978</div>
5979
5980<div class="doc_text">
5981
5982<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5983<p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
5984 floating point or vector of floating point type. Not all targets support all
5985 types however.</p>
5986
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5987<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5988 declare float @llvm.powi.f32(float %Val, i32 %power)
5989 declare double @llvm.powi.f64(double %Val, i32 %power)
5990 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
5991 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
5992 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5993</pre>
5994
5995<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]5996<p>The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
5997 specified (positive or negative) power. The order of evaluation of
5998 multiplications is not defined. When a vector of floating point type is
5999 used, the second argument remains a scalar integer value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6000
6001<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6002<p>The second argument is an integer power, and the first is a value to raise to
6003 that power.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6004
6005<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6006<p>This function returns the first value raised to the second power with an
6007 unspecified sequence of rounding operations.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6008
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6009</div>
6010
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6011<!-- _______________________________________________________________________ -->
6012<div class="doc_subsubsection">
6013 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
6014</div>
6015
6016<div class="doc_text">
6017
6018<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6019<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
6020 floating point or vector of floating point type. Not all targets support all
6021 types however.</p>
6022
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6023<pre>
6024 declare float @llvm.sin.f32(float %Val)
6025 declare double @llvm.sin.f64(double %Val)
6026 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
6027 declare fp128 @llvm.sin.f128(fp128 %Val)
6028 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
6029</pre>
6030
6031<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6032<p>The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6033
6034<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6035<p>The argument and return value are floating point numbers of the same
6036 type.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6037
6038<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6039<p>This function returns the sine of the specified operand, returning the same
6040 values as the libm <tt>sin</tt> functions would, and handles error conditions
6041 in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6042
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6043</div>
6044
6045<!-- _______________________________________________________________________ -->
6046<div class="doc_subsubsection">
6047 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
6048</div>
6049
6050<div class="doc_text">
6051
6052<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6053<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
6054 floating point or vector of floating point type. Not all targets support all
6055 types however.</p>
6056
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6057<pre>
6058 declare float @llvm.cos.f32(float %Val)
6059 declare double @llvm.cos.f64(double %Val)
6060 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
6061 declare fp128 @llvm.cos.f128(fp128 %Val)
6062 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
6063</pre>
6064
6065<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6066<p>The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6067
6068<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6069<p>The argument and return value are floating point numbers of the same
6070 type.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6071
6072<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6073<p>This function returns the cosine of the specified operand, returning the same
6074 values as the libm <tt>cos</tt> functions would, and handles error conditions
6075 in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6076
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6077</div>
6078
6079<!-- _______________________________________________________________________ -->
6080<div class="doc_subsubsection">
6081 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
6082</div>
6083
6084<div class="doc_text">
6085
6086<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6087<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
6088 floating point or vector of floating point type. Not all targets support all
6089 types however.</p>
6090
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6091<pre>
6092 declare float @llvm.pow.f32(float %Val, float %Power)
6093 declare double @llvm.pow.f64(double %Val, double %Power)
6094 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
6095 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
6096 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
6097</pre>
6098
6099<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6100<p>The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
6101 specified (positive or negative) power.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6102
6103<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6104<p>The second argument is a floating point power, and the first is a value to
6105 raise to that power.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6106
6107<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6108<p>This function returns the first value raised to the second power, returning
6109 the same values as the libm <tt>pow</tt> functions would, and handles error
6110 conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6111
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]6112</div>
6113
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6114<!-- ======================================================================= -->
6115<div class="doc_subsection">
6116 <a name="int_manip">Bit Manipulation Intrinsics</a>
6117</div>
6118
6119<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6120
6121<p>LLVM provides intrinsics for a few important bit manipulation operations.
6122 These allow efficient code generation for some algorithms.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6123
6124</div>
6125
6126<!-- _______________________________________________________________________ -->
6127<div class="doc_subsubsection">
6128 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
6129</div>
6130
6131<div class="doc_text">
6132
6133<h5>Syntax:</h5>
6134<p>This is an overloaded intrinsic function. You can use bswap on any integer
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6135 type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
6136
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6137<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]6138 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
6139 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
6140 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6141</pre>
6142
6143<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6144<p>The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
6145 values with an even number of bytes (positive multiple of 16 bits). These
6146 are useful for performing operations on data that is not in the target's
6147 native byte order.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6148
6149<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6150<p>The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
6151 and low byte of the input i16 swapped. Similarly,
6152 the <tt>llvm.bswap.i32</tt> intrinsic returns an i32 value that has the four
6153 bytes of the input i32 swapped, so that if the input bytes are numbered 0, 1,
6154 2, 3 then the returned i32 will have its bytes in 3, 2, 1, 0 order.
6155 The <tt>llvm.bswap.i48</tt>, <tt>llvm.bswap.i64</tt> and other intrinsics
6156 extend this concept to additional even-byte lengths (6 bytes, 8 bytes and
6157 more, respectively).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6158
6159</div>
6160
6161<!-- _______________________________________________________________________ -->
6162<div class="doc_subsubsection">
6163 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
6164</div>
6165
6166<div class="doc_text">
6167
6168<h5>Syntax:</h5>
6169<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]6170 width. Not all targets support all bit widths however.</p>
6171
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6172<pre>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6173 declare i8 @llvm.ctpop.i8(i8 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]6174 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6175 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]6176 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
6177 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6178</pre>
6179
6180<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6181<p>The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set
6182 in a value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6183
6184<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6185<p>The only argument is the value to be counted. The argument may be of any
6186 integer type. The return type must match the argument type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6187
6188<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6189<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]6190
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6191</div>
6192
6193<!-- _______________________________________________________________________ -->
6194<div class="doc_subsubsection">
6195 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
6196</div>
6197
6198<div class="doc_text">
6199
6200<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6201<p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
6202 integer bit width. Not all targets support all bit widths however.</p>
6203
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6204<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]6205 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
6206 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6207 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]6208 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
6209 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6210</pre>
6211
6212<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6213<p>The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
6214 leading zeros in a variable.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6215
6216<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6217<p>The only argument is the value to be counted. The argument may be of any
6218 integer type. The return type must match the argument type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6219
6220<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6221<p>The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant)
6222 zeros in a variable. If the src == 0 then the result is the size in bits of
6223 the type of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6224
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6225</div>
6226
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6227<!-- _______________________________________________________________________ -->
6228<div class="doc_subsubsection">
6229 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
6230</div>
6231
6232<div class="doc_text">
6233
6234<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6235<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
6236 integer bit width. Not all targets support all bit widths however.</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 i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
6240 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6241 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]6242 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
6243 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6244</pre>
6245
6246<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6247<p>The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
6248 trailing zeros.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6249
6250<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6251<p>The only argument is the value to be counted. The argument may be of any
6252 integer type. The return type must match the argument type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6253
6254<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6255<p>The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant)
6256 zeros in a variable. If the src == 0 then the result is the size in bits of
6257 the type of src. For example, <tt>llvm.cttz(2) = 1</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6258
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6259</div>
6260
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6261<!-- ======================================================================= -->
6262<div class="doc_subsection">
6263 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
6264</div>
6265
6266<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6267
6268<p>LLVM provides intrinsics for some arithmetic with overflow operations.</p>
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6269
6270</div>
6271
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6272<!-- _______________________________________________________________________ -->
6273<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6274 <a name="int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6275</div>
6276
6277<div class="doc_text">
6278
6279<h5>Syntax:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6280<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]6281 on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6282
6283<pre>
6284 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
6285 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6286 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
6287</pre>
6288
6289<h5>Overview:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6290<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6291 a signed addition of the two arguments, and indicate whether an overflow
6292 occurred during the signed summation.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6293
6294<h5>Arguments:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6295<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]6296 be of integer types of any bit width, but they must have the same bit
6297 width. The second element of the result structure must be of
6298 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6299 undergo signed addition.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6300
6301<h5>Semantics:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6302<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6303 a signed addition of the two variables. They return a structure &mdash; the
6304 first element of which is the signed summation, and the second element of
6305 which is a bit specifying if the signed summation resulted in an
6306 overflow.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6307
6308<h5>Examples:</h5>
6309<pre>
6310 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6311 %sum = extractvalue {i32, i1} %res, 0
6312 %obit = extractvalue {i32, i1} %res, 1
6313 br i1 %obit, label %overflow, label %normal
6314</pre>
6315
6316</div>
6317
6318<!-- _______________________________________________________________________ -->
6319<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6320 <a name="int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6321</div>
6322
6323<div class="doc_text">
6324
6325<h5>Syntax:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6326<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]6327 on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6328
6329<pre>
6330 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
6331 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6332 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
6333</pre>
6334
6335<h5>Overview:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6336<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6337 an unsigned addition of the two arguments, and indicate whether a carry
6338 occurred during the unsigned summation.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6339
6340<h5>Arguments:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6341<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]6342 be of integer types of any bit width, but they must have the same bit
6343 width. The second element of the result structure must be of
6344 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6345 undergo unsigned addition.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6346
6347<h5>Semantics:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6348<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6349 an unsigned addition of the two arguments. They return a structure &mdash;
6350 the first element of which is the sum, and the second element of which is a
6351 bit specifying if the unsigned summation resulted in a carry.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6352
6353<h5>Examples:</h5>
6354<pre>
6355 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6356 %sum = extractvalue {i32, i1} %res, 0
6357 %obit = extractvalue {i32, i1} %res, 1
6358 br i1 %obit, label %carry, label %normal
6359</pre>
6360
6361</div>
6362
6363<!-- _______________________________________________________________________ -->
6364<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6365 <a name="int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6366</div>
6367
6368<div class="doc_text">
6369
6370<h5>Syntax:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6371<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]6372 on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6373
6374<pre>
6375 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
6376 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6377 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
6378</pre>
6379
6380<h5>Overview:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6381<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6382 a signed subtraction of the two arguments, and indicate whether an overflow
6383 occurred during the signed subtraction.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6384
6385<h5>Arguments:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6386<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]6387 be of integer types of any bit width, but they must have the same bit
6388 width. The second element of the result structure must be of
6389 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6390 undergo signed subtraction.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6391
6392<h5>Semantics:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6393<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6394 a signed subtraction of the two arguments. They return a structure &mdash;
6395 the first element of which is the subtraction, and the second element of
6396 which is a bit specifying if the signed subtraction resulted in an
6397 overflow.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6398
6399<h5>Examples:</h5>
6400<pre>
6401 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6402 %sum = extractvalue {i32, i1} %res, 0
6403 %obit = extractvalue {i32, i1} %res, 1
6404 br i1 %obit, label %overflow, label %normal
6405</pre>
6406
6407</div>
6408
6409<!-- _______________________________________________________________________ -->
6410<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6411 <a name="int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6412</div>
6413
6414<div class="doc_text">
6415
6416<h5>Syntax:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6417<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]6418 on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6419
6420<pre>
6421 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
6422 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6423 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
6424</pre>
6425
6426<h5>Overview:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6427<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6428 an unsigned subtraction of the two arguments, and indicate whether an
6429 overflow occurred during the unsigned subtraction.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6430
6431<h5>Arguments:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6432<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]6433 be of integer types of any bit width, but they must have the same bit
6434 width. The second element of the result structure must be of
6435 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6436 undergo unsigned subtraction.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6437
6438<h5>Semantics:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6439<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6440 an unsigned subtraction of the two arguments. They return a structure &mdash;
6441 the first element of which is the subtraction, and the second element of
6442 which is a bit specifying if the unsigned subtraction resulted in an
6443 overflow.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6444
6445<h5>Examples:</h5>
6446<pre>
6447 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6448 %sum = extractvalue {i32, i1} %res, 0
6449 %obit = extractvalue {i32, i1} %res, 1
6450 br i1 %obit, label %overflow, label %normal
6451</pre>
6452
6453</div>
6454
6455<!-- _______________________________________________________________________ -->
6456<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6457 <a name="int_smul_overflow">'<tt>llvm.smul.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6458</div>
6459
6460<div class="doc_text">
6461
6462<h5>Syntax:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6463<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]6464 on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6465
6466<pre>
6467 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
6468 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6469 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
6470</pre>
6471
6472<h5>Overview:</h5>
6473
6474<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6475 a signed multiplication of the two arguments, and indicate whether an
6476 overflow occurred during the signed multiplication.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6477
6478<h5>Arguments:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6479<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]6480 be of integer types of any bit width, but they must have the same bit
6481 width. The second element of the result structure must be of
6482 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6483 undergo signed multiplication.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6484
6485<h5>Semantics:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6486<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6487 a signed multiplication of the two arguments. They return a structure &mdash;
6488 the first element of which is the multiplication, and the second element of
6489 which is a bit specifying if the signed multiplication resulted in an
6490 overflow.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6491
6492<h5>Examples:</h5>
6493<pre>
6494 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6495 %sum = extractvalue {i32, i1} %res, 0
6496 %obit = extractvalue {i32, i1} %res, 1
6497 br i1 %obit, label %overflow, label %normal
6498</pre>
6499
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6500</div>
6501
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]6502<!-- _______________________________________________________________________ -->
6503<div class="doc_subsubsection">
6504 <a name="int_umul_overflow">'<tt>llvm.umul.with.overflow.*</tt>' Intrinsics</a>
6505</div>
6506
6507<div class="doc_text">
6508
6509<h5>Syntax:</h5>
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]6510<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]6511 on any integer bit width.</p>
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]6512
6513<pre>
6514 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
6515 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6516 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
6517</pre>
6518
6519<h5>Overview:</h5>
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]6520<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6521 a unsigned multiplication of the two arguments, and indicate whether an
6522 overflow occurred during the unsigned multiplication.</p>
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]6523
6524<h5>Arguments:</h5>
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]6525<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]6526 be of integer types of any bit width, but they must have the same bit
6527 width. The second element of the result structure must be of
6528 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6529 undergo unsigned multiplication.</p>
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]6530
6531<h5>Semantics:</h5>
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]6532<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6533 an unsigned multiplication of the two arguments. They return a structure
6534 &mdash; the first element of which is the multiplication, and the second
6535 element of which is a bit specifying if the unsigned multiplication resulted
6536 in an overflow.</p>
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]6537
6538<h5>Examples:</h5>
6539<pre>
6540 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6541 %sum = extractvalue {i32, i1} %res, 0
6542 %obit = extractvalue {i32, i1} %res, 1
6543 br i1 %obit, label %overflow, label %normal
6544</pre>
6545
6546</div>
6547
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6548<!-- ======================================================================= -->
6549<div class="doc_subsection">
6550 <a name="int_debugger">Debugger Intrinsics</a>
6551</div>
6552
6553<div class="doc_text">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6554
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6555<p>The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt>
6556 prefix), are described in
6557 the <a href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source
6558 Level Debugging</a> document.</p>
6559
6560</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6561
6562<!-- ======================================================================= -->
6563<div class="doc_subsection">
6564 <a name="int_eh">Exception Handling Intrinsics</a>
6565</div>
6566
6567<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6568
6569<p>The LLVM exception handling intrinsics (which all start with
6570 <tt>llvm.eh.</tt> prefix), are described in
6571 the <a href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
6572 Handling</a> document.</p>
6573
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6574</div>
6575
6576<!-- ======================================================================= -->
6577<div class="doc_subsection">
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6578 <a name="int_trampoline">Trampoline Intrinsic</a>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6579</div>
6580
6581<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6582
6583<p>This intrinsic makes it possible to excise one parameter, marked with
6584 the <tt>nest</tt> attribute, from a function. The result is a callable
6585 function pointer lacking the nest parameter - the caller does not need to
6586 provide a value for it. Instead, the value to use is stored in advance in a
6587 "trampoline", a block of memory usually allocated on the stack, which also
6588 contains code to splice the nest value into the argument list. This is used
6589 to implement the GCC nested function address extension.</p>
6590
6591<p>For example, if the function is
6592 <tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
6593 pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as
6594 follows:</p>
6595
6596<div class="doc_code">
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6597<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6598 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
6599 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
6600 %p = call i8* @llvm.init.trampoline( i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8*), i8* %nval )
6601 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6602</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6603</div>
6604
6605<p>The call <tt>%val = call i32 %fp( i32 %x, i32 %y )</tt> is then equivalent
6606 to <tt>%val = call i32 %f( i8* %nval, i32 %x, i32 %y )</tt>.</p>
6607
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6608</div>
6609
6610<!-- _______________________________________________________________________ -->
6611<div class="doc_subsubsection">
6612 <a name="int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a>
6613</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6614
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6615<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6616
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6617<h5>Syntax:</h5>
6618<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6619 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]6620</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6621
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6622<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6623<p>This fills the memory pointed to by <tt>tramp</tt> with code and returns a
6624 function pointer suitable for executing it.</p>
6625
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6626<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6627<p>The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
6628 pointers. The <tt>tramp</tt> argument must point to a sufficiently large and
6629 sufficiently aligned block of memory; this memory is written to by the
6630 intrinsic. Note that the size and the alignment are target-specific - LLVM
6631 currently provides no portable way of determining them, so a front-end that
6632 generates this intrinsic needs to have some target-specific knowledge.
6633 The <tt>func</tt> argument must hold a function bitcast to
6634 an <tt>i8*</tt>.</p>
6635
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6636<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6637<p>The block of memory pointed to by <tt>tramp</tt> is filled with target
6638 dependent code, turning it into a function. A pointer to this function is
6639 returned, but needs to be bitcast to an <a href="#int_trampoline">appropriate
6640 function pointer type</a> before being called. The new function's signature
6641 is the same as that of <tt>func</tt> with any arguments marked with
6642 the <tt>nest</tt> attribute removed. At most one such <tt>nest</tt> argument
6643 is allowed, and it must be of pointer type. Calling the new function is
6644 equivalent to calling <tt>func</tt> with the same argument list, but
6645 with <tt>nval</tt> used for the missing <tt>nest</tt> argument. If, after
6646 calling <tt>llvm.init.trampoline</tt>, the memory pointed to
6647 by <tt>tramp</tt> is modified, then the effect of any later call to the
6648 returned function pointer is undefined.</p>
6649
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6650</div>
6651
6652<!-- ======================================================================= -->
6653<div class="doc_subsection">
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6654 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
6655</div>
6656
6657<div class="doc_text">
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6658
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6659<p>These intrinsic functions expand the "universal IR" of LLVM to represent
6660 hardware constructs for atomic operations and memory synchronization. This
6661 provides an interface to the hardware, not an interface to the programmer. It
6662 is aimed at a low enough level to allow any programming models or APIs
6663 (Application Programming Interfaces) which need atomic behaviors to map
6664 cleanly onto it. It is also modeled primarily on hardware behavior. Just as
6665 hardware provides a "universal IR" for source languages, it also provides a
6666 starting point for developing a "universal" atomic operation and
6667 synchronization IR.</p>
6668
6669<p>These do <em>not</em> form an API such as high-level threading libraries,
6670 software transaction memory systems, atomic primitives, and intrinsic
6671 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
6672 application libraries. The hardware interface provided by LLVM should allow
6673 a clean implementation of all of these APIs and parallel programming models.
6674 No one model or paradigm should be selected above others unless the hardware
6675 itself ubiquitously does so.</p>
6676
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6677</div>
6678
6679<!-- _______________________________________________________________________ -->
6680<div class="doc_subsubsection">
6681 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
6682</div>
6683<div class="doc_text">
6684<h5>Syntax:</h5>
6685<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6686 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]6687</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6688
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6689<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6690<p>The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
6691 specific pairs of memory access types.</p>
6692
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6693<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6694<p>The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
6695 The first four arguments enables a specific barrier as listed below. The
6696 fith argument specifies that the barrier applies to io or device or uncached
6697 memory.</p>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6698
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6699<ul>
6700 <li><tt>ll</tt>: load-load barrier</li>
6701 <li><tt>ls</tt>: load-store barrier</li>
6702 <li><tt>sl</tt>: store-load barrier</li>
6703 <li><tt>ss</tt>: store-store barrier</li>
6704 <li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
6705</ul>
6706
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6707<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6708<p>This intrinsic causes the system to enforce some ordering constraints upon
6709 the loads and stores of the program. This barrier does not
6710 indicate <em>when</em> any events will occur, it only enforces
6711 an <em>order</em> in which they occur. For any of the specified pairs of load
6712 and store operations (f.ex. load-load, or store-load), all of the first
6713 operations preceding the barrier will complete before any of the second
6714 operations succeeding the barrier begin. Specifically the semantics for each
6715 pairing is as follows:</p>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6716
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6717<ul>
6718 <li><tt>ll</tt>: All loads before the barrier must complete before any load
6719 after the barrier begins.</li>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]6720 <li><tt>ls</tt>: All loads before the barrier must complete before any
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6721 store after the barrier begins.</li>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]6722 <li><tt>ss</tt>: All stores before the barrier must complete before any
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6723 store after the barrier begins.</li>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]6724 <li><tt>sl</tt>: All stores before the barrier must complete before any
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6725 load after the barrier begins.</li>
6726</ul>
6727
6728<p>These semantics are applied with a logical "and" behavior when more than one
6729 is enabled in a single memory barrier intrinsic.</p>
6730
6731<p>Backends may implement stronger barriers than those requested when they do
6732 not support as fine grained a barrier as requested. Some architectures do
6733 not need all types of barriers and on such architectures, these become
6734 noops.</p>
6735
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6736<h5>Example:</h5>
6737<pre>
Victor Hernandez6b054092009-10-26 23:44:29 +0000[diff] [blame]6738%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
6739%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6740 store i32 4, %ptr
6741
6742%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
6743 call void @llvm.memory.barrier( i1 false, i1 true, i1 false, i1 false )
6744 <i>; guarantee the above finishes</i>
6745 store i32 8, %ptr <i>; before this begins</i>
6746</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6747
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6748</div>
6749
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6750<!-- _______________________________________________________________________ -->
6751<div class="doc_subsubsection">
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6752 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6753</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6754
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6755<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6756
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6757<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6758<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
6759 any integer bit width and for different address spaces. Not all targets
6760 support all bit widths however.</p>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6761
6762<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6763 declare i8 @llvm.atomic.cmp.swap.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt; )
6764 declare i16 @llvm.atomic.cmp.swap.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt; )
6765 declare i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt; )
6766 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]6767</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6768
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6769<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6770<p>This loads a value in memory and compares it to a given value. If they are
6771 equal, it stores a new value into the memory.</p>
6772
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6773<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6774<p>The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result
6775 as well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
6776 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
6777 this integer type. While any bit width integer may be used, targets may only
6778 lower representations they support in hardware.</p>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6779
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6780<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6781<p>This entire intrinsic must be executed atomically. It first loads the value
6782 in memory pointed to by <tt>ptr</tt> and compares it with the
6783 value <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the
6784 memory. The loaded value is yielded in all cases. This provides the
6785 equivalent of an atomic compare-and-swap operation within the SSA
6786 framework.</p>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6787
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6788<h5>Examples:</h5>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6789<pre>
Victor Hernandez6b054092009-10-26 23:44:29 +0000[diff] [blame]6790%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
6791%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6792 store i32 4, %ptr
6793
6794%val1 = add i32 4, 4
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6795%result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 4, %val1 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6796 <i>; yields {i32}:result1 = 4</i>
6797%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
6798%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
6799
6800%val2 = add i32 1, 1
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6801%result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 5, %val2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6802 <i>; yields {i32}:result2 = 8</i>
6803%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
6804
6805%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
6806</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6807
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6808</div>
6809
6810<!-- _______________________________________________________________________ -->
6811<div class="doc_subsubsection">
6812 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
6813</div>
6814<div class="doc_text">
6815<h5>Syntax:</h5>
6816
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6817<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
6818 integer bit width. Not all targets support all bit widths however.</p>
6819
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6820<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6821 declare i8 @llvm.atomic.swap.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;val&gt; )
6822 declare i16 @llvm.atomic.swap.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;val&gt; )
6823 declare i32 @llvm.atomic.swap.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;val&gt; )
6824 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]6825</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6826
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6827<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6828<p>This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
6829 the value from memory. It then stores the value in <tt>val</tt> in the memory
6830 at <tt>ptr</tt>.</p>
6831
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6832<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6833<p>The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both
6834 the <tt>val</tt> argument and the result must be integers of the same bit
6835 width. The first argument, <tt>ptr</tt>, must be a pointer to a value of this
6836 integer type. The targets may only lower integer representations they
6837 support.</p>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6838
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6839<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6840<p>This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
6841 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
6842 equivalent of an atomic swap operation within the SSA framework.</p>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6843
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6844<h5>Examples:</h5>
6845<pre>
Victor Hernandez6b054092009-10-26 23:44:29 +0000[diff] [blame]6846%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
6847%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6848 store i32 4, %ptr
6849
6850%val1 = add i32 4, 4
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6851%result1 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val1 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6852 <i>; yields {i32}:result1 = 4</i>
6853%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
6854%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
6855
6856%val2 = add i32 1, 1
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6857%result2 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6858 <i>; yields {i32}:result2 = 8</i>
6859
6860%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
6861%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
6862</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6863
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6864</div>
6865
6866<!-- _______________________________________________________________________ -->
6867<div class="doc_subsubsection">
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6868 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6869
6870</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6871
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6872<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6873
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6874<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6875<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.add</tt> on
6876 any integer bit width. Not all targets support all bit widths however.</p>
6877
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6878<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6879 declare i8 @llvm.atomic.load.add.i8..p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6880 declare i16 @llvm.atomic.load.add.i16..p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6881 declare i32 @llvm.atomic.load.add.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6882 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]6883</pre>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6884
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6885<h5>Overview:</h5>
6886<p>This intrinsic adds <tt>delta</tt> to the value stored in memory
6887 at <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
6888
6889<h5>Arguments:</h5>
6890<p>The intrinsic takes two arguments, the first a pointer to an integer value
6891 and the second an integer value. The result is also an integer value. These
6892 integer types can have any bit width, but they must all have the same bit
6893 width. The targets may only lower integer representations they support.</p>
6894
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6895<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6896<p>This intrinsic does a series of operations atomically. It first loads the
6897 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
6898 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]6899
6900<h5>Examples:</h5>
6901<pre>
Victor Hernandez6b054092009-10-26 23:44:29 +0000[diff] [blame]6902%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
6903%ptr = bitcast i8* %mallocP to i32*
6904 store i32 4, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6905%result1 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 4 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6906 <i>; yields {i32}:result1 = 4</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6907%result2 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6908 <i>; yields {i32}:result2 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6909%result3 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 5 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6910 <i>; yields {i32}:result3 = 10</i>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6911%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6912</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6913
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6914</div>
6915
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6916<!-- _______________________________________________________________________ -->
6917<div class="doc_subsubsection">
6918 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
6919
6920</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6921
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6922<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6923
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6924<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6925<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.sub</tt> on
6926 any integer bit width and for different address spaces. Not all targets
6927 support all bit widths however.</p>
6928
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6929<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6930 declare i8 @llvm.atomic.load.sub.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6931 declare i16 @llvm.atomic.load.sub.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6932 declare i32 @llvm.atomic.load.sub.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6933 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]6934</pre>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6935
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6936<h5>Overview:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]6937<p>This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6938 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
6939
6940<h5>Arguments:</h5>
6941<p>The intrinsic takes two arguments, the first a pointer to an integer value
6942 and the second an integer value. The result is also an integer value. These
6943 integer types can have any bit width, but they must all have the same bit
6944 width. The targets may only lower integer representations they support.</p>
6945
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6946<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6947<p>This intrinsic does a series of operations atomically. It first loads the
6948 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
6949 result to <tt>ptr</tt>. It yields the original value stored
6950 at <tt>ptr</tt>.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6951
6952<h5>Examples:</h5>
6953<pre>
Victor Hernandez6b054092009-10-26 23:44:29 +0000[diff] [blame]6954%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
6955%ptr = bitcast i8* %mallocP to i32*
6956 store i32 8, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6957%result1 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 4 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6958 <i>; yields {i32}:result1 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6959%result2 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 2 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6960 <i>; yields {i32}:result2 = 4</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6961%result3 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 5 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6962 <i>; yields {i32}:result3 = 2</i>
6963%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
6964</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6965
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6966</div>
6967
6968<!-- _______________________________________________________________________ -->
6969<div class="doc_subsubsection">
6970 <a name="int_atomic_load_and">'<tt>llvm.atomic.load.and.*</tt>' Intrinsic</a><br>
6971 <a name="int_atomic_load_nand">'<tt>llvm.atomic.load.nand.*</tt>' Intrinsic</a><br>
6972 <a name="int_atomic_load_or">'<tt>llvm.atomic.load.or.*</tt>' Intrinsic</a><br>
6973 <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]6974</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6975
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6976<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6977
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6978<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6979<p>These are overloaded intrinsics. You can
6980 use <tt>llvm.atomic.load_and</tt>, <tt>llvm.atomic.load_nand</tt>,
6981 <tt>llvm.atomic.load_or</tt>, and <tt>llvm.atomic.load_xor</tt> on any integer
6982 bit width and for different address spaces. Not all targets support all bit
6983 widths however.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6984
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6985<pre>
6986 declare i8 @llvm.atomic.load.and.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6987 declare i16 @llvm.atomic.load.and.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6988 declare i32 @llvm.atomic.load.and.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6989 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]6990</pre>
6991
6992<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]6993 declare i8 @llvm.atomic.load.or.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6994 declare i16 @llvm.atomic.load.or.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6995 declare i32 @llvm.atomic.load.or.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6996 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]6997</pre>
6998
6999<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7000 declare i8 @llvm.atomic.load.nand.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
7001 declare i16 @llvm.atomic.load.nand.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
7002 declare i32 @llvm.atomic.load.nand.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
7003 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]7004</pre>
7005
7006<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7007 declare i8 @llvm.atomic.load.xor.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
7008 declare i16 @llvm.atomic.load.xor.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
7009 declare i32 @llvm.atomic.load.xor.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
7010 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]7011</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7012
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7013<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7014<p>These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
7015 the value stored in memory at <tt>ptr</tt>. It yields the original value
7016 at <tt>ptr</tt>.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7017
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7018<h5>Arguments:</h5>
7019<p>These intrinsics take two arguments, the first a pointer to an integer value
7020 and the second an integer value. The result is also an integer value. These
7021 integer types can have any bit width, but they must all have the same bit
7022 width. The targets may only lower integer representations they support.</p>
7023
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7024<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7025<p>These intrinsics does a series of operations atomically. They first load the
7026 value stored at <tt>ptr</tt>. They then do the bitwise
7027 operation <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the
7028 original value stored at <tt>ptr</tt>.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7029
7030<h5>Examples:</h5>
7031<pre>
Victor Hernandez6b054092009-10-26 23:44:29 +0000[diff] [blame]7032%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7033%ptr = bitcast i8* %mallocP to i32*
7034 store i32 0x0F0F, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7035%result0 = call i32 @llvm.atomic.load.nand.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7036 <i>; yields {i32}:result0 = 0x0F0F</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7037%result1 = call i32 @llvm.atomic.load.and.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7038 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7039%result2 = call i32 @llvm.atomic.load.or.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7040 <i>; yields {i32}:result2 = 0xF0</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7041%result3 = call i32 @llvm.atomic.load.xor.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7042 <i>; yields {i32}:result3 = FF</i>
7043%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
7044</pre>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7045
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7046</div>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7047
7048<!-- _______________________________________________________________________ -->
7049<div class="doc_subsubsection">
7050 <a name="int_atomic_load_max">'<tt>llvm.atomic.load.max.*</tt>' Intrinsic</a><br>
7051 <a name="int_atomic_load_min">'<tt>llvm.atomic.load.min.*</tt>' Intrinsic</a><br>
7052 <a name="int_atomic_load_umax">'<tt>llvm.atomic.load.umax.*</tt>' Intrinsic</a><br>
7053 <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]7054</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7055
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7056<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7057
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7058<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7059<p>These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
7060 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
7061 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
7062 address spaces. Not all targets support all bit widths however.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7063
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7064<pre>
7065 declare i8 @llvm.atomic.load.max.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
7066 declare i16 @llvm.atomic.load.max.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
7067 declare i32 @llvm.atomic.load.max.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
7068 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]7069</pre>
7070
7071<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7072 declare i8 @llvm.atomic.load.min.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
7073 declare i16 @llvm.atomic.load.min.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
7074 declare i32 @llvm.atomic.load.min.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
7075 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]7076</pre>
7077
7078<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7079 declare i8 @llvm.atomic.load.umax.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
7080 declare i16 @llvm.atomic.load.umax.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
7081 declare i32 @llvm.atomic.load.umax.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
7082 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]7083</pre>
7084
7085<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7086 declare i8 @llvm.atomic.load.umin.i8..p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
7087 declare i16 @llvm.atomic.load.umin.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
7088 declare i32 @llvm.atomic.load.umin.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
7089 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]7090</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7091
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7092<h5>Overview:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]7093<p>These intrinsics takes the signed or unsigned minimum or maximum of
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7094 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
7095 original value at <tt>ptr</tt>.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7096
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7097<h5>Arguments:</h5>
7098<p>These intrinsics take two arguments, the first a pointer to an integer value
7099 and the second an integer value. The result is also an integer value. These
7100 integer types can have any bit width, but they must all have the same bit
7101 width. The targets may only lower integer representations they support.</p>
7102
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7103<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7104<p>These intrinsics does a series of operations atomically. They first load the
7105 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or
7106 max <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They
7107 yield the original value stored at <tt>ptr</tt>.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7108
7109<h5>Examples:</h5>
7110<pre>
Victor Hernandez6b054092009-10-26 23:44:29 +0000[diff] [blame]7111%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7112%ptr = bitcast i8* %mallocP to i32*
7113 store i32 7, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7114%result0 = call i32 @llvm.atomic.load.min.i32.p0i32( i32* %ptr, i32 -2 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7115 <i>; yields {i32}:result0 = 7</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7116%result1 = call i32 @llvm.atomic.load.max.i32.p0i32( i32* %ptr, i32 8 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7117 <i>; yields {i32}:result1 = -2</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7118%result2 = call i32 @llvm.atomic.load.umin.i32.p0i32( i32* %ptr, i32 10 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7119 <i>; yields {i32}:result2 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7120%result3 = call i32 @llvm.atomic.load.umax.i32.p0i32( i32* %ptr, i32 30 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7121 <i>; yields {i32}:result3 = 8</i>
7122%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
7123</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7124
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7125</div>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]7126
Nick Lewyckyc888d352009-10-13 07:03:23 +0000[diff] [blame]7127
7128<!-- ======================================================================= -->
7129<div class="doc_subsection">
7130 <a name="int_memorymarkers">Memory Use Markers</a>
7131</div>
7132
7133<div class="doc_text">
7134
7135<p>This class of intrinsics exists to information about the lifetime of memory
7136 objects and ranges where variables are immutable.</p>
7137
7138</div>
7139
7140<!-- _______________________________________________________________________ -->
7141<div class="doc_subsubsection">
7142 <a name="int_lifetime_start">'<tt>llvm.lifetime.start</tt>' Intrinsic</a>
7143</div>
7144
7145<div class="doc_text">
7146
7147<h5>Syntax:</h5>
7148<pre>
7149 declare void @llvm.lifetime.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7150</pre>
7151
7152<h5>Overview:</h5>
7153<p>The '<tt>llvm.lifetime.start</tt>' intrinsic specifies the start of a memory
7154 object's lifetime.</p>
7155
7156<h5>Arguments:</h5>
Nick Lewycky23f49982009-10-13 07:57:33 +0000[diff] [blame]7157<p>The first argument is a constant integer representing the size of the
7158 object, or -1 if it is variable sized. The second argument is a pointer to
7159 the object.</p>
Nick Lewyckyc888d352009-10-13 07:03:23 +0000[diff] [blame]7160
7161<h5>Semantics:</h5>
7162<p>This intrinsic indicates that before this point in the code, the value of the
7163 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]7164 never be used and has an undefined value. A load from the pointer that
7165 precedes this intrinsic can be replaced with
Nick Lewyckyc888d352009-10-13 07:03:23 +0000[diff] [blame]7166 <tt>'<a href="#undefvalues">undef</a>'</tt>.</p>
7167
7168</div>
7169
7170<!-- _______________________________________________________________________ -->
7171<div class="doc_subsubsection">
7172 <a name="int_lifetime_end">'<tt>llvm.lifetime.end</tt>' Intrinsic</a>
7173</div>
7174
7175<div class="doc_text">
7176
7177<h5>Syntax:</h5>
7178<pre>
7179 declare void @llvm.lifetime.end(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7180</pre>
7181
7182<h5>Overview:</h5>
7183<p>The '<tt>llvm.lifetime.end</tt>' intrinsic specifies the end of a memory
7184 object's lifetime.</p>
7185
7186<h5>Arguments:</h5>
Nick Lewycky23f49982009-10-13 07:57:33 +0000[diff] [blame]7187<p>The first argument is a constant integer representing the size of the
7188 object, or -1 if it is variable sized. The second argument is a pointer to
7189 the object.</p>
Nick Lewyckyc888d352009-10-13 07:03:23 +0000[diff] [blame]7190
7191<h5>Semantics:</h5>
7192<p>This intrinsic indicates that after this point in the code, the value of the
7193 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
7194 never be used and has an undefined value. Any stores into the memory object
7195 following this intrinsic may be removed as dead.
7196
7197</div>
7198
7199<!-- _______________________________________________________________________ -->
7200<div class="doc_subsubsection">
7201 <a name="int_invariant_start">'<tt>llvm.invariant.start</tt>' Intrinsic</a>
7202</div>
7203
7204<div class="doc_text">
7205
7206<h5>Syntax:</h5>
7207<pre>
7208 declare {}* @llvm.invariant.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;) readonly
7209</pre>
7210
7211<h5>Overview:</h5>
7212<p>The '<tt>llvm.invariant.start</tt>' intrinsic specifies that the contents of
7213 a memory object will not change.</p>
7214
7215<h5>Arguments:</h5>
Nick Lewycky23f49982009-10-13 07:57:33 +0000[diff] [blame]7216<p>The first argument is a constant integer representing the size of the
7217 object, or -1 if it is variable sized. The second argument is a pointer to
7218 the object.</p>
Nick Lewyckyc888d352009-10-13 07:03:23 +0000[diff] [blame]7219
7220<h5>Semantics:</h5>
7221<p>This intrinsic indicates that until an <tt>llvm.invariant.end</tt> that uses
7222 the return value, the referenced memory location is constant and
7223 unchanging.</p>
7224
7225</div>
7226
7227<!-- _______________________________________________________________________ -->
7228<div class="doc_subsubsection">
7229 <a name="int_invariant_end">'<tt>llvm.invariant.end</tt>' Intrinsic</a>
7230</div>
7231
7232<div class="doc_text">
7233
7234<h5>Syntax:</h5>
7235<pre>
7236 declare void @llvm.invariant.end({}* &lt;start&gt;, i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7237</pre>
7238
7239<h5>Overview:</h5>
7240<p>The '<tt>llvm.invariant.end</tt>' intrinsic specifies that the contents of
7241 a memory object are mutable.</p>
7242
7243<h5>Arguments:</h5>
7244<p>The first argument is the matching <tt>llvm.invariant.start</tt> intrinsic.
Nick Lewycky23f49982009-10-13 07:57:33 +0000[diff] [blame]7245 The second argument is a constant integer representing the size of the
7246 object, or -1 if it is variable sized and the third argument is a pointer
7247 to the object.</p>
Nick Lewyckyc888d352009-10-13 07:03:23 +0000[diff] [blame]7248
7249<h5>Semantics:</h5>
7250<p>This intrinsic indicates that the memory is mutable again.</p>
7251
7252</div>
7253
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]7254<!-- ======================================================================= -->
7255<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7256 <a name="int_general">General Intrinsics</a>
7257</div>
7258
7259<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7260
7261<p>This class of intrinsics is designed to be generic and has no specific
7262 purpose.</p>
7263
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7264</div>
7265
7266<!-- _______________________________________________________________________ -->
7267<div class="doc_subsubsection">
7268 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
7269</div>
7270
7271<div class="doc_text">
7272
7273<h5>Syntax:</h5>
7274<pre>
7275 declare void @llvm.var.annotation(i8* &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
7276</pre>
7277
7278<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7279<p>The '<tt>llvm.var.annotation</tt>' intrinsic.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7280
7281<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7282<p>The first argument is a pointer to a value, the second is a pointer to a
7283 global string, the third is a pointer to a global string which is the source
7284 file name, and the last argument is the line number.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7285
7286<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7287<p>This intrinsic allows annotation of local variables with arbitrary strings.
7288 This can be useful for special purpose optimizations that want to look for
7289 these annotations. These have no other defined use, they are ignored by code
7290 generation and optimization.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7291
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7292</div>
7293
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7294<!-- _______________________________________________________________________ -->
7295<div class="doc_subsubsection">
Tanya Lattnerc9869b12007-09-21 23:57:59 +0000[diff] [blame]7296 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7297</div>
7298
7299<div class="doc_text">
7300
7301<h5>Syntax:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7302<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
7303 any integer bit width.</p>
7304
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7305<pre>
Tanya Lattner09161fe2007-09-22 00:03:01 +0000[diff] [blame]7306 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
7307 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
7308 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
7309 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
7310 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]7311</pre>
7312
7313<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7314<p>The '<tt>llvm.annotation</tt>' intrinsic.</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7315
7316<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7317<p>The first argument is an integer value (result of some expression), the
7318 second is a pointer to a global string, the third is a pointer to a global
7319 string which is the source file name, and the last argument is the line
7320 number. It returns the value of the first argument.</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7321
7322<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7323<p>This intrinsic allows annotations to be put on arbitrary expressions with
7324 arbitrary strings. This can be useful for special purpose optimizations that
7325 want to look for these annotations. These have no other defined use, they
7326 are ignored by code generation and optimization.</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7327
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7328</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7329
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]7330<!-- _______________________________________________________________________ -->
7331<div class="doc_subsubsection">
7332 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
7333</div>
7334
7335<div class="doc_text">
7336
7337<h5>Syntax:</h5>
7338<pre>
7339 declare void @llvm.trap()
7340</pre>
7341
7342<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7343<p>The '<tt>llvm.trap</tt>' intrinsic.</p>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]7344
7345<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7346<p>None.</p>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]7347
7348<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7349<p>This intrinsics is lowered to the target dependent trap instruction. If the
7350 target does not have a trap instruction, this intrinsic will be lowered to
7351 the call of the <tt>abort()</tt> function.</p>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]7352
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]7353</div>
7354
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]7355<!-- _______________________________________________________________________ -->
7356<div class="doc_subsubsection">
Misha Brukman5dd7f4d2008-11-22 23:55:29 +0000[diff] [blame]7357 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]7358</div>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7359
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]7360<div class="doc_text">
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7361
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]7362<h5>Syntax:</h5>
7363<pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7364 declare void @llvm.stackprotector( i8* &lt;guard&gt;, i8** &lt;slot&gt; )
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]7365</pre>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7366
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]7367<h5>Overview:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7368<p>The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and
7369 stores it onto the stack at <tt>slot</tt>. The stack slot is adjusted to
7370 ensure that it is placed on the stack before local variables.</p>
7371
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]7372<h5>Arguments:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7373<p>The <tt>llvm.stackprotector</tt> intrinsic requires two pointer
7374 arguments. The first argument is the value loaded from the stack
7375 guard <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt>
7376 that has enough space to hold the value of the guard.</p>
7377
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]7378<h5>Semantics:</h5>
Bill Wendlingf85859d2009-07-20 02:29:24 +0000[diff] [blame]7379<p>This intrinsic causes the prologue/epilogue inserter to force the position of
7380 the <tt>AllocaInst</tt> stack slot to be before local variables on the
7381 stack. This is to ensure that if a local variable on the stack is
7382 overwritten, it will destroy the value of the guard. When the function exits,
7383 the guard on the stack is checked against the original guard. If they're
7384 different, then the program aborts by calling the <tt>__stack_chk_fail()</tt>
7385 function.</p>
7386
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]7387</div>
7388
Eric Christopher767a3722009-11-30 08:03:53 +0000[diff] [blame]7389<!-- _______________________________________________________________________ -->
7390<div class="doc_subsubsection">
7391 <a name="int_objectsize">'<tt>llvm.objectsize</tt>' Intrinsic</a>
7392</div>
7393
7394<div class="doc_text">
7395
7396<h5>Syntax:</h5>
7397<pre>
Eric Christopher0101f9d2009-12-23 00:29:49 +0000[diff] [blame]7398 declare i32 @llvm.objectsize.i32( i8* &lt;object&gt;, i1 &lt;type&gt; )
7399 declare i64 @llvm.objectsize.i64( i8* &lt;object&gt;, i1 &lt;type&gt; )
Eric Christopher767a3722009-11-30 08:03:53 +0000[diff] [blame]7400</pre>
7401
7402<h5>Overview:</h5>
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]7403<p>The <tt>llvm.objectsize</tt> intrinsic is designed to provide information
Eric Christopher184f44d2010-01-08 21:42:39 +0000[diff] [blame]7404 to the optimizers to discover at compile time either a) when an
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]7405 operation like memcpy will either overflow a buffer that corresponds to
7406 an object, or b) to determine that a runtime check for overflow isn't
7407 necessary. An object in this context means an allocation of a
Eric Christopher0101f9d2009-12-23 00:29:49 +0000[diff] [blame]7408 specific class, structure, array, or other object.</p>
Eric Christopher767a3722009-11-30 08:03:53 +0000[diff] [blame]7409
7410<h5>Arguments:</h5>
7411<p>The <tt>llvm.objectsize</tt> intrinsic takes two arguments. The first
Eric Christopher0101f9d2009-12-23 00:29:49 +0000[diff] [blame]7412 argument is a pointer to or into the <tt>object</tt>. The second argument
7413 is a boolean 0 or 1. This argument determines whether you want the
7414 maximum (0) or minimum (1) bytes remaining. This needs to be a literal 0 or
7415 1, variables are not allowed.</p>
7416
Eric Christopher767a3722009-11-30 08:03:53 +0000[diff] [blame]7417<h5>Semantics:</h5>
7418<p>The <tt>llvm.objectsize</tt> intrinsic is lowered to either a constant
Eric Christophera1151bf2009-12-05 02:46:03 +0000[diff] [blame]7419 representing the size of the object concerned or <tt>i32/i64 -1 or 0</tt>
7420 (depending on the <tt>type</tt> argument if the size cannot be determined
7421 at compile time.</p>
Eric Christopher767a3722009-11-30 08:03:53 +0000[diff] [blame]7422
7423</div>
7424
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7425<!-- *********************************************************************** -->
7426<hr>
7427<address>
7428 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman947321d2008-12-11 17:34:48 +0000[diff] [blame]7429 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7430 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman947321d2008-12-11 17:34:48 +0000[diff] [blame]7431 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]7432
7433 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
7434 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
7435 Last modified: $Date$
7436</address>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]7437
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7438</body>
7439</html>