Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / b38ce6f78715f8ad4daa383182f1ec324142da64 / . / docs / LangRef.html
blob: c57518befc098a464cf6e96f3c3f9a595f35f52f [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">
8 <meta name="description"
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>
23 <li><a href="#linkage">Linkage Types</a></li>
24 <li><a href="#callingconv">Calling Conventions</a></li>
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]25 <li><a href="#namedtypes">Named Types</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]26 <li><a href="#globalvars">Global Variables</a></li>
27 <li><a href="#functionstructure">Functions</a></li>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]28 <li><a href="#aliasstructure">Aliases</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]29 <li><a href="#paramattrs">Parameter Attributes</a></li>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]30 <li><a href="#fnattrs">Function Attributes</a></li>
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]31 <li><a href="#gc">Garbage Collector Names</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]32 <li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
33 <li><a href="#datalayout">Data Layout</a></li>
34 </ol>
35 </li>
36 <li><a href="#typesystem">Type System</a>
37 <ol>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]38 <li><a href="#t_classifications">Type Classifications</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]39 <li><a href="#t_primitive">Primitive Types</a>
40 <ol>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]41 <li><a href="#t_floating">Floating Point Types</a></li>
42 <li><a href="#t_void">Void Type</a></li>
43 <li><a href="#t_label">Label Type</a></li>
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]44 <li><a href="#t_metadata">Metadata Type</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]45 </ol>
46 </li>
47 <li><a href="#t_derived">Derived Types</a>
48 <ol>
Chris Lattner251ab812007-12-18 06:18:21 +0000[diff] [blame]49 <li><a href="#t_integer">Integer Type</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]50 <li><a href="#t_array">Array Type</a></li>
51 <li><a href="#t_function">Function Type</a></li>
52 <li><a href="#t_pointer">Pointer Type</a></li>
53 <li><a href="#t_struct">Structure Type</a></li>
54 <li><a href="#t_pstruct">Packed Structure Type</a></li>
55 <li><a href="#t_vector">Vector Type</a></li>
56 <li><a href="#t_opaque">Opaque Type</a></li>
57 </ol>
58 </li>
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]59 <li><a href="#t_uprefs">Type Up-references</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]60 </ol>
61 </li>
62 <li><a href="#constants">Constants</a>
63 <ol>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]64 <li><a href="#simpleconstants">Simple Constants</a></li>
Chris Lattner97063852009-02-28 18:32:25 +0000[diff] [blame]65 <li><a href="#complexconstants">Complex Constants</a></li>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]66 <li><a href="#globalconstants">Global Variable and Function Addresses</a></li>
67 <li><a href="#undefvalues">Undefined Values</a></li>
68 <li><a href="#constantexprs">Constant Expressions</a></li>
Nick Lewycky4dcf8102009-04-04 07:22:01 +0000[diff] [blame]69 <li><a href="#metadata">Embedded Metadata</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]70 </ol>
71 </li>
72 <li><a href="#othervalues">Other Values</a>
73 <ol>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]74 <li><a href="#inlineasm">Inline Assembler Expressions</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]75 </ol>
76 </li>
77 <li><a href="#instref">Instruction Reference</a>
78 <ol>
79 <li><a href="#terminators">Terminator Instructions</a>
80 <ol>
81 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
82 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
83 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
84 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
85 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
86 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
87 </ol>
88 </li>
89 <li><a href="#binaryops">Binary Operations</a>
90 <ol>
91 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]92 <li><a href="#i_fadd">'<tt>fadd</tt>' Instruction</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]93 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]94 <li><a href="#i_fsub">'<tt>fsub</tt>' Instruction</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]95 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]96 <li><a href="#i_fmul">'<tt>fmul</tt>' Instruction</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]97 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
98 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
99 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
100 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
101 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
102 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
103 </ol>
104 </li>
105 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
106 <ol>
107 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
108 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
109 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
110 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
111 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
112 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
113 </ol>
114 </li>
115 <li><a href="#vectorops">Vector Operations</a>
116 <ol>
117 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
118 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
119 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
120 </ol>
121 </li>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]122 <li><a href="#aggregateops">Aggregate Operations</a>
123 <ol>
124 <li><a href="#i_extractvalue">'<tt>extractvalue</tt>' Instruction</a></li>
125 <li><a href="#i_insertvalue">'<tt>insertvalue</tt>' Instruction</a></li>
126 </ol>
127 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]128 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
129 <ol>
130 <li><a href="#i_malloc">'<tt>malloc</tt>' Instruction</a></li>
131 <li><a href="#i_free">'<tt>free</tt>' Instruction</a></li>
132 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
133 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
134 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
135 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
136 </ol>
137 </li>
138 <li><a href="#convertops">Conversion Operations</a>
139 <ol>
140 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
141 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
142 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
143 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
144 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
145 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
146 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
147 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
148 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
149 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
150 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
151 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
152 </ol>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]153 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]154 <li><a href="#otherops">Other Operations</a>
155 <ol>
156 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
157 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]158 <li><a href="#i_vicmp">'<tt>vicmp</tt>' Instruction</a></li>
159 <li><a href="#i_vfcmp">'<tt>vfcmp</tt>' Instruction</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]160 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
161 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
162 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
163 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
164 </ol>
165 </li>
166 </ol>
167 </li>
168 <li><a href="#intrinsics">Intrinsic Functions</a>
169 <ol>
170 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
171 <ol>
172 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
173 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
174 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
175 </ol>
176 </li>
177 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
178 <ol>
179 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
180 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
181 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
182 </ol>
183 </li>
184 <li><a href="#int_codegen">Code Generator Intrinsics</a>
185 <ol>
186 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
187 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
188 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
189 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
190 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
191 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
192 <li><a href="#int_readcyclecounter"><tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
193 </ol>
194 </li>
195 <li><a href="#int_libc">Standard C Library Intrinsics</a>
196 <ol>
197 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
198 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
199 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
200 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
201 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]202 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
203 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
204 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]205 </ol>
206 </li>
207 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
208 <ol>
209 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
210 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
211 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
212 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
213 <li><a href="#int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic </a></li>
214 <li><a href="#int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic </a></li>
215 </ol>
216 </li>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]217 <li><a href="#int_overflow">Arithmetic with Overflow Intrinsics</a>
218 <ol>
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]219 <li><a href="#int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt> Intrinsics</a></li>
220 <li><a href="#int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt> Intrinsics</a></li>
221 <li><a href="#int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt> Intrinsics</a></li>
222 <li><a href="#int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt> Intrinsics</a></li>
223 <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]224 <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]225 </ol>
226 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]227 <li><a href="#int_debugger">Debugger intrinsics</a></li>
228 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]229 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]230 <ol>
231 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]232 </ol>
233 </li>
Bill Wendling9127adb2008-11-18 22:10:53 +0000[diff] [blame]234 <li><a href="#int_atomics">Atomic intrinsics</a>
235 <ol>
236 <li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
237 <li><a href="#int_atomic_cmp_swap"><tt>llvm.atomic.cmp.swap</tt></a></li>
238 <li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
239 <li><a href="#int_atomic_load_add"><tt>llvm.atomic.load.add</tt></a></li>
240 <li><a href="#int_atomic_load_sub"><tt>llvm.atomic.load.sub</tt></a></li>
241 <li><a href="#int_atomic_load_and"><tt>llvm.atomic.load.and</tt></a></li>
242 <li><a href="#int_atomic_load_nand"><tt>llvm.atomic.load.nand</tt></a></li>
243 <li><a href="#int_atomic_load_or"><tt>llvm.atomic.load.or</tt></a></li>
244 <li><a href="#int_atomic_load_xor"><tt>llvm.atomic.load.xor</tt></a></li>
245 <li><a href="#int_atomic_load_max"><tt>llvm.atomic.load.max</tt></a></li>
246 <li><a href="#int_atomic_load_min"><tt>llvm.atomic.load.min</tt></a></li>
247 <li><a href="#int_atomic_load_umax"><tt>llvm.atomic.load.umax</tt></a></li>
248 <li><a href="#int_atomic_load_umin"><tt>llvm.atomic.load.umin</tt></a></li>
249 </ol>
250 </li>
Reid Spencerb043f672007-07-20 19:59:11 +0000[diff] [blame]251 <li><a href="#int_general">General intrinsics</a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]252 <ol>
Reid Spencerb043f672007-07-20 19:59:11 +0000[diff] [blame]253 <li><a href="#int_var_annotation">
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]254 '<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]255 <li><a href="#int_annotation">
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]256 '<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]257 <li><a href="#int_trap">
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]258 '<tt>llvm.trap</tt>' Intrinsic</a></li>
259 <li><a href="#int_stackprotector">
260 '<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]261 </ol>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]262 </li>
263 </ol>
264 </li>
265</ol>
266
267<div class="doc_author">
268 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
269 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
270</div>
271
272<!-- *********************************************************************** -->
273<div class="doc_section"> <a name="abstract">Abstract </a></div>
274<!-- *********************************************************************** -->
275
276<div class="doc_text">
277<p>This document is a reference manual for the LLVM assembly language.
Bill Wendlinge7846a52008-08-05 22:29:16 +0000[diff] [blame]278LLVM is a Static Single Assignment (SSA) based representation that provides
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]279type safety, low-level operations, flexibility, and the capability of
280representing 'all' high-level languages cleanly. It is the common code
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]281representation used throughout all phases of the LLVM compilation
282strategy.</p>
283</div>
284
285<!-- *********************************************************************** -->
286<div class="doc_section"> <a name="introduction">Introduction</a> </div>
287<!-- *********************************************************************** -->
288
289<div class="doc_text">
290
291<p>The LLVM code representation is designed to be used in three
292different forms: as an in-memory compiler IR, as an on-disk bitcode
293representation (suitable for fast loading by a Just-In-Time compiler),
294and as a human readable assembly language representation. This allows
295LLVM to provide a powerful intermediate representation for efficient
296compiler transformations and analysis, while providing a natural means
297to debug and visualize the transformations. The three different forms
298of LLVM are all equivalent. This document describes the human readable
299representation and notation.</p>
300
301<p>The LLVM representation aims to be light-weight and low-level
302while being expressive, typed, and extensible at the same time. It
303aims to be a "universal IR" of sorts, by being at a low enough level
304that high-level ideas may be cleanly mapped to it (similar to how
305microprocessors are "universal IR's", allowing many source languages to
306be mapped to them). By providing type information, LLVM can be used as
307the target of optimizations: for example, through pointer analysis, it
308can be proven that a C automatic variable is never accessed outside of
309the current function... allowing it to be promoted to a simple SSA
310value instead of a memory location.</p>
311
312</div>
313
314<!-- _______________________________________________________________________ -->
315<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
316
317<div class="doc_text">
318
319<p>It is important to note that this document describes 'well formed'
320LLVM assembly language. There is a difference between what the parser
321accepts and what is considered 'well formed'. For example, the
322following instruction is syntactically okay, but not well formed:</p>
323
324<div class="doc_code">
325<pre>
326%x = <a href="#i_add">add</a> i32 1, %x
327</pre>
328</div>
329
330<p>...because the definition of <tt>%x</tt> does not dominate all of
331its uses. The LLVM infrastructure provides a verification pass that may
332be used to verify that an LLVM module is well formed. This pass is
333automatically run by the parser after parsing input assembly and by
334the optimizer before it outputs bitcode. The violations pointed out
335by the verifier pass indicate bugs in transformation passes or input to
336the parser.</p>
337</div>
338
Chris Lattnera83fdc02007-10-03 17:34:29 +0000[diff] [blame]339<!-- Describe the typesetting conventions here. -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]340
341<!-- *********************************************************************** -->
342<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
343<!-- *********************************************************************** -->
344
345<div class="doc_text">
346
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]347 <p>LLVM identifiers come in two basic types: global and local. Global
348 identifiers (functions, global variables) begin with the @ character. Local
349 identifiers (register names, types) begin with the % character. Additionally,
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]350 there are three different formats for identifiers, for different purposes:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]351
352<ol>
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]353 <li>Named values are represented as a string of characters with their prefix.
354 For example, %foo, @DivisionByZero, %a.really.long.identifier. The actual
355 regular expression used is '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]356 Identifiers which require other characters in their names can be surrounded
Daniel Dunbar73f9a772008-10-14 23:51:43 +0000[diff] [blame]357 with quotes. Special characters may be escaped using "\xx" where xx is the
358 ASCII code for the character in hexadecimal. In this way, any character can
359 be used in a name value, even quotes themselves.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]360
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]361 <li>Unnamed values are represented as an unsigned numeric value with their
362 prefix. For example, %12, @2, %44.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]363
364 <li>Constants, which are described in a <a href="#constants">section about
365 constants</a>, below.</li>
366</ol>
367
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]368<p>LLVM requires that values start with a prefix for two reasons: Compilers
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]369don't need to worry about name clashes with reserved words, and the set of
370reserved words may be expanded in the future without penalty. Additionally,
371unnamed identifiers allow a compiler to quickly come up with a temporary
372variable without having to avoid symbol table conflicts.</p>
373
374<p>Reserved words in LLVM are very similar to reserved words in other
375languages. There are keywords for different opcodes
376('<tt><a href="#i_add">add</a></tt>',
377 '<tt><a href="#i_bitcast">bitcast</a></tt>',
378 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
379href="#t_void">void</a></tt>', '<tt><a href="#t_primitive">i32</a></tt>', etc...),
380and others. These reserved words cannot conflict with variable names, because
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]381none of them start with a prefix character ('%' or '@').</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]382
383<p>Here is an example of LLVM code to multiply the integer variable
384'<tt>%X</tt>' by 8:</p>
385
386<p>The easy way:</p>
387
388<div class="doc_code">
389<pre>
390%result = <a href="#i_mul">mul</a> i32 %X, 8
391</pre>
392</div>
393
394<p>After strength reduction:</p>
395
396<div class="doc_code">
397<pre>
398%result = <a href="#i_shl">shl</a> i32 %X, i8 3
399</pre>
400</div>
401
402<p>And the hard way:</p>
403
404<div class="doc_code">
405<pre>
406<a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
407<a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
408%result = <a href="#i_add">add</a> i32 %1, %1
409</pre>
410</div>
411
412<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
413important lexical features of LLVM:</p>
414
415<ol>
416
417 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
418 line.</li>
419
420 <li>Unnamed temporaries are created when the result of a computation is not
421 assigned to a named value.</li>
422
423 <li>Unnamed temporaries are numbered sequentially</li>
424
425</ol>
426
427<p>...and it also shows a convention that we follow in this document. When
428demonstrating instructions, we will follow an instruction with a comment that
429defines the type and name of value produced. Comments are shown in italic
430text.</p>
431
432</div>
433
434<!-- *********************************************************************** -->
435<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
436<!-- *********************************************************************** -->
437
438<!-- ======================================================================= -->
439<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
440</div>
441
442<div class="doc_text">
443
444<p>LLVM programs are composed of "Module"s, each of which is a
445translation unit of the input programs. Each module consists of
446functions, global variables, and symbol table entries. Modules may be
447combined together with the LLVM linker, which merges function (and
448global variable) definitions, resolves forward declarations, and merges
449symbol table entries. Here is an example of the "hello world" module:</p>
450
451<div class="doc_code">
452<pre><i>; Declare the string constant as a global constant...</i>
453<a href="#identifiers">@.LC0</a> = <a href="#linkage_internal">internal</a> <a
454 href="#globalvars">constant</a> <a href="#t_array">[13 x i8]</a> c"hello world\0A\00" <i>; [13 x i8]*</i>
455
456<i>; External declaration of the puts function</i>
457<a href="#functionstructure">declare</a> i32 @puts(i8 *) <i>; i32(i8 *)* </i>
458
459<i>; Definition of main function</i>
460define i32 @main() { <i>; i32()* </i>
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]461 <i>; Convert [13 x i8]* to i8 *...</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]462 %cast210 = <a
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]463 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]464
465 <i>; Call puts function to write out the string to stdout...</i>
466 <a
467 href="#i_call">call</a> i32 @puts(i8 * %cast210) <i>; i32</i>
468 <a
469 href="#i_ret">ret</a> i32 0<br>}<br>
470</pre>
471</div>
472
473<p>This example is made up of a <a href="#globalvars">global variable</a>
474named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
475function, and a <a href="#functionstructure">function definition</a>
476for "<tt>main</tt>".</p>
477
478<p>In general, a module is made up of a list of global values,
479where both functions and global variables are global values. Global values are
480represented by a pointer to a memory location (in this case, a pointer to an
481array of char, and a pointer to a function), and have one of the following <a
482href="#linkage">linkage types</a>.</p>
483
484</div>
485
486<!-- ======================================================================= -->
487<div class="doc_subsection">
488 <a name="linkage">Linkage Types</a>
489</div>
490
491<div class="doc_text">
492
493<p>
494All Global Variables and Functions have one of the following types of linkage:
495</p>
496
497<dl>
498
Rafael Espindolaa168fc92009-01-15 20:18:42 +0000[diff] [blame]499 <dt><tt><b><a name="linkage_private">private</a></b></tt>: </dt>
500
501 <dd>Global values with private linkage are only directly accessible by
502 objects in the current module. In particular, linking code into a module with
503 an private global value may cause the private to be renamed as necessary to
504 avoid collisions. Because the symbol is private to the module, all
505 references can be updated. This doesn't show up in any symbol table in the
506 object file.
507 </dd>
508
Dale Johannesen96e7e092008-05-23 23:13:41 +0000[diff] [blame]509 <dt><tt><b><a name="linkage_internal">internal</a></b></tt>: </dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]510
Duncan Sandsa75223a2009-01-16 09:29:46 +0000[diff] [blame]511 <dd> Similar to private, but the value shows as a local symbol (STB_LOCAL in
Rafael Espindolaa168fc92009-01-15 20:18:42 +0000[diff] [blame]512 the case of ELF) in the object file. This corresponds to the notion of the
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]513 '<tt>static</tt>' keyword in C.
514 </dd>
515
Chris Lattner68433442009-04-13 05:44:34 +0000[diff] [blame]516 <dt><tt><b><a name="available_externally">available_externally</a></b></tt>:
517 </dt>
518
519 <dd>Globals with "<tt>available_externally</tt>" linkage are never emitted
520 into the object file corresponding to the LLVM module. They exist to
521 allow inlining and other optimizations to take place given knowledge of the
522 definition of the global, which is known to be somewhere outside the module.
523 Globals with <tt>available_externally</tt> linkage are allowed to be discarded
524 at will, and are otherwise the same as <tt>linkonce_odr</tt>. This linkage
525 type is only allowed on definitions, not declarations.</dd>
526
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]527 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
528
529 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
530 the same name when linkage occurs. This is typically used to implement
531 inline functions, templates, or other code which must be generated in each
532 translation unit that uses it. Unreferenced <tt>linkonce</tt> globals are
533 allowed to be discarded.
534 </dd>
535
Dale Johannesen96e7e092008-05-23 23:13:41 +0000[diff] [blame]536 <dt><tt><b><a name="linkage_common">common</a></b></tt>: </dt>
537
538 <dd>"<tt>common</tt>" linkage is exactly the same as <tt>linkonce</tt>
539 linkage, except that unreferenced <tt>common</tt> globals may not be
540 discarded. This is used for globals that may be emitted in multiple
541 translation units, but that are not guaranteed to be emitted into every
542 translation unit that uses them. One example of this is tentative
543 definitions in C, such as "<tt>int X;</tt>" at global scope.
544 </dd>
545
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]546 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
547
Dale Johannesen96e7e092008-05-23 23:13:41 +0000[diff] [blame]548 <dd>"<tt>weak</tt>" linkage is the same as <tt>common</tt> linkage, except
549 that some targets may choose to emit different assembly sequences for them
550 for target-dependent reasons. This is used for globals that are declared
551 "weak" in C source code.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]552 </dd>
553
554 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
555
556 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
557 pointer to array type. When two global variables with appending linkage are
558 linked together, the two global arrays are appended together. This is the
559 LLVM, typesafe, equivalent of having the system linker append together
560 "sections" with identical names when .o files are linked.
561 </dd>
562
563 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt>: </dt>
Duncan Sands19d161f2009-03-07 15:45:40 +0000[diff] [blame]564
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]565 <dd>The semantics of this linkage follow the ELF object file model: the
566 symbol is weak until linked, if not linked, the symbol becomes null instead
567 of being an undefined reference.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]568 </dd>
569
Duncan Sands19d161f2009-03-07 15:45:40 +0000[diff] [blame]570 <dt><tt><b><a name="linkage_linkonce">linkonce_odr</a></b></tt>: </dt>
Duncan Sands19d161f2009-03-07 15:45:40 +0000[diff] [blame]571 <dt><tt><b><a name="linkage_weak">weak_odr</a></b></tt>: </dt>
Chris Lattner68433442009-04-13 05:44:34 +0000[diff] [blame]572 <dd>Some languages allow differing globals to be merged, such as two
Duncan Sands19d161f2009-03-07 15:45:40 +0000[diff] [blame]573 functions with different semantics. Other languages, such as <tt>C++</tt>,
574 ensure that only equivalent globals are ever merged (the "one definition
Chris Lattner68433442009-04-13 05:44:34 +0000[diff] [blame]575 rule" - "ODR"). Such languages can use the <tt>linkonce_odr</tt>
Duncan Sandsb95df792009-03-11 20:14:15 +0000[diff] [blame]576 and <tt>weak_odr</tt> linkage types to indicate that the global will only
577 be merged with equivalent globals. These linkage types are otherwise the
578 same as their non-<tt>odr</tt> versions.
Duncan Sands19d161f2009-03-07 15:45:40 +0000[diff] [blame]579 </dd>
580
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]581 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
582
583 <dd>If none of the above identifiers are used, the global is externally
584 visible, meaning that it participates in linkage and can be used to resolve
585 external symbol references.
586 </dd>
587</dl>
588
589 <p>
590 The next two types of linkage are targeted for Microsoft Windows platform
591 only. They are designed to support importing (exporting) symbols from (to)
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]592 DLLs (Dynamic Link Libraries).
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]593 </p>
594
595 <dl>
596 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt>: </dt>
597
598 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
599 or variable via a global pointer to a pointer that is set up by the DLL
600 exporting the symbol. On Microsoft Windows targets, the pointer name is
Dan Gohman5ec99832009-01-12 21:35:55 +0000[diff] [blame]601 formed by combining <code>__imp_</code> and the function or variable name.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]602 </dd>
603
604 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt>: </dt>
605
606 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
607 pointer to a pointer in a DLL, so that it can be referenced with the
608 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
Dan Gohman5ec99832009-01-12 21:35:55 +0000[diff] [blame]609 name is formed by combining <code>__imp_</code> and the function or variable
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]610 name.
611 </dd>
612
613</dl>
614
Dan Gohman4dfac702008-11-24 17:18:39 +0000[diff] [blame]615<p>For example, since the "<tt>.LC0</tt>"
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]616variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
617variable and was linked with this one, one of the two would be renamed,
618preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
619external (i.e., lacking any linkage declarations), they are accessible
620outside of the current module.</p>
621<p>It is illegal for a function <i>declaration</i>
Duncan Sands565f65d2009-03-11 08:08:06 +0000[diff] [blame]622to have any linkage type other than "externally visible", <tt>dllimport</tt>
623or <tt>extern_weak</tt>.</p>
Duncan Sands19d161f2009-03-07 15:45:40 +0000[diff] [blame]624<p>Aliases can have only <tt>external</tt>, <tt>internal</tt>, <tt>weak</tt>
625or <tt>weak_odr</tt> linkages.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]626</div>
627
628<!-- ======================================================================= -->
629<div class="doc_subsection">
630 <a name="callingconv">Calling Conventions</a>
631</div>
632
633<div class="doc_text">
634
635<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
636and <a href="#i_invoke">invokes</a> can all have an optional calling convention
637specified for the call. The calling convention of any pair of dynamic
638caller/callee must match, or the behavior of the program is undefined. The
639following calling conventions are supported by LLVM, and more may be added in
640the future:</p>
641
642<dl>
643 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
644
645 <dd>This calling convention (the default if no other calling convention is
646 specified) matches the target C calling conventions. This calling convention
647 supports varargs function calls and tolerates some mismatch in the declared
648 prototype and implemented declaration of the function (as does normal C).
649 </dd>
650
651 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
652
653 <dd>This calling convention attempts to make calls as fast as possible
654 (e.g. by passing things in registers). This calling convention allows the
655 target to use whatever tricks it wants to produce fast code for the target,
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]656 without having to conform to an externally specified ABI (Application Binary
657 Interface). Implementations of this convention should allow arbitrary
Arnold Schwaighofer07444922008-05-14 09:17:12 +0000[diff] [blame]658 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> to be
659 supported. This calling convention does not support varargs and requires the
660 prototype of all callees to exactly match the prototype of the function
661 definition.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]662 </dd>
663
664 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
665
666 <dd>This calling convention attempts to make code in the caller as efficient
667 as possible under the assumption that the call is not commonly executed. As
668 such, these calls often preserve all registers so that the call does not break
669 any live ranges in the caller side. This calling convention does not support
670 varargs and requires the prototype of all callees to exactly match the
671 prototype of the function definition.
672 </dd>
673
674 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
675
676 <dd>Any calling convention may be specified by number, allowing
677 target-specific calling conventions to be used. Target specific calling
678 conventions start at 64.
679 </dd>
680</dl>
681
682<p>More calling conventions can be added/defined on an as-needed basis, to
683support pascal conventions or any other well-known target-independent
684convention.</p>
685
686</div>
687
688<!-- ======================================================================= -->
689<div class="doc_subsection">
690 <a name="visibility">Visibility Styles</a>
691</div>
692
693<div class="doc_text">
694
695<p>
696All Global Variables and Functions have one of the following visibility styles:
697</p>
698
699<dl>
700 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
701
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]702 <dd>On targets that use the ELF object file format, default visibility means
703 that the declaration is visible to other
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]704 modules and, in shared libraries, means that the declared entity may be
705 overridden. On Darwin, default visibility means that the declaration is
706 visible to other modules. Default visibility corresponds to "external
707 linkage" in the language.
708 </dd>
709
710 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
711
712 <dd>Two declarations of an object with hidden visibility refer to the same
713 object if they are in the same shared object. Usually, hidden visibility
714 indicates that the symbol will not be placed into the dynamic symbol table,
715 so no other module (executable or shared library) can reference it
716 directly.
717 </dd>
718
719 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
720
721 <dd>On ELF, protected visibility indicates that the symbol will be placed in
722 the dynamic symbol table, but that references within the defining module will
723 bind to the local symbol. That is, the symbol cannot be overridden by another
724 module.
725 </dd>
726</dl>
727
728</div>
729
730<!-- ======================================================================= -->
731<div class="doc_subsection">
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]732 <a name="namedtypes">Named Types</a>
733</div>
734
735<div class="doc_text">
736
737<p>LLVM IR allows you to specify name aliases for certain types. This can make
738it easier to read the IR and make the IR more condensed (particularly when
739recursive types are involved). An example of a name specification is:
740</p>
741
742<div class="doc_code">
743<pre>
744%mytype = type { %mytype*, i32 }
745</pre>
746</div>
747
748<p>You may give a name to any <a href="#typesystem">type</a> except "<a
749href="t_void">void</a>". Type name aliases may be used anywhere a type is
750expected with the syntax "%mytype".</p>
751
752<p>Note that type names are aliases for the structural type that they indicate,
753and that you can therefore specify multiple names for the same type. This often
754leads to confusing behavior when dumping out a .ll file. Since LLVM IR uses
755structural typing, the name is not part of the type. When printing out LLVM IR,
756the printer will pick <em>one name</em> to render all types of a particular
757shape. This means that if you have code where two different source types end up
758having the same LLVM type, that the dumper will sometimes print the "wrong" or
759unexpected type. This is an important design point and isn't going to
760change.</p>
761
762</div>
763
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]764<!-- ======================================================================= -->
765<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]766 <a name="globalvars">Global Variables</a>
767</div>
768
769<div class="doc_text">
770
771<p>Global variables define regions of memory allocated at compilation time
772instead of run-time. Global variables may optionally be initialized, may have
773an explicit section to be placed in, and may have an optional explicit alignment
774specified. A variable may be defined as "thread_local", which means that it
775will not be shared by threads (each thread will have a separated copy of the
776variable). A variable may be defined as a global "constant," which indicates
777that the contents of the variable will <b>never</b> be modified (enabling better
778optimization, allowing the global data to be placed in the read-only section of
779an executable, etc). Note that variables that need runtime initialization
780cannot be marked "constant" as there is a store to the variable.</p>
781
782<p>
783LLVM explicitly allows <em>declarations</em> of global variables to be marked
784constant, even if the final definition of the global is not. This capability
785can be used to enable slightly better optimization of the program, but requires
786the language definition to guarantee that optimizations based on the
787'constantness' are valid for the translation units that do not include the
788definition.
789</p>
790
791<p>As SSA values, global variables define pointer values that are in
792scope (i.e. they dominate) all basic blocks in the program. Global
793variables always define a pointer to their "content" type because they
794describe a region of memory, and all memory objects in LLVM are
795accessed through pointers.</p>
796
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]797<p>A global variable may be declared to reside in a target-specifc numbered
798address space. For targets that support them, address spaces may affect how
799optimizations are performed and/or what target instructions are used to access
Christopher Lamb20a39e92007-12-12 08:44:39 +0000[diff] [blame]800the variable. The default address space is zero. The address space qualifier
801must precede any other attributes.</p>
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]802
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]803<p>LLVM allows an explicit section to be specified for globals. If the target
804supports it, it will emit globals to the section specified.</p>
805
806<p>An explicit alignment may be specified for a global. If not present, or if
807the alignment is set to zero, the alignment of the global is set by the target
808to whatever it feels convenient. If an explicit alignment is specified, the
809global is forced to have at least that much alignment. All alignments must be
810a power of 2.</p>
811
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]812<p>For example, the following defines a global in a numbered address space with
813an initializer, section, and alignment:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]814
815<div class="doc_code">
816<pre>
Dan Gohman21ef02c2009-01-11 00:40:00 +0000[diff] [blame]817@G = addrspace(5) constant float 1.0, section "foo", align 4
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]818</pre>
819</div>
820
821</div>
822
823
824<!-- ======================================================================= -->
825<div class="doc_subsection">
826 <a name="functionstructure">Functions</a>
827</div>
828
829<div class="doc_text">
830
831<p>LLVM function definitions consist of the "<tt>define</tt>" keyord,
832an optional <a href="#linkage">linkage type</a>, an optional
833<a href="#visibility">visibility style</a>, an optional
834<a href="#callingconv">calling convention</a>, a return type, an optional
835<a href="#paramattrs">parameter attribute</a> for the return type, a function
836name, a (possibly empty) argument list (each with optional
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]837<a href="#paramattrs">parameter attributes</a>), optional
838<a href="#fnattrs">function attributes</a>, an optional section,
839an optional alignment, an optional <a href="#gc">garbage collector name</a>,
Chris Lattneree202542008-10-04 18:10:21 +0000[diff] [blame]840an opening curly brace, a list of basic blocks, and a closing curly brace.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]841
842LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
843optional <a href="#linkage">linkage type</a>, an optional
844<a href="#visibility">visibility style</a>, an optional
845<a href="#callingconv">calling convention</a>, a return type, an optional
846<a href="#paramattrs">parameter attribute</a> for the return type, a function
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]847name, a possibly empty list of arguments, an optional alignment, and an optional
Gordon Henriksend606f5b2007-12-10 03:30:21 +0000[diff] [blame]848<a href="#gc">garbage collector name</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]849
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]850<p>A function definition contains a list of basic blocks, forming the CFG
851(Control Flow Graph) for
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]852the function. Each basic block may optionally start with a label (giving the
853basic block a symbol table entry), contains a list of instructions, and ends
854with a <a href="#terminators">terminator</a> instruction (such as a branch or
855function return).</p>
856
857<p>The first basic block in a function is special in two ways: it is immediately
858executed on entrance to the function, and it is not allowed to have predecessor
859basic blocks (i.e. there can not be any branches to the entry block of a
860function). Because the block can have no predecessors, it also cannot have any
861<a href="#i_phi">PHI nodes</a>.</p>
862
863<p>LLVM allows an explicit section to be specified for functions. If the target
864supports it, it will emit functions to the section specified.</p>
865
866<p>An explicit alignment may be specified for a function. If not present, or if
867the alignment is set to zero, the alignment of the function is set by the target
868to whatever it feels convenient. If an explicit alignment is specified, the
869function is forced to have at least that much alignment. All alignments must be
870a power of 2.</p>
871
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]872 <h5>Syntax:</h5>
873
874<div class="doc_code">
Chris Lattner1e5c5cd02008-10-13 16:55:18 +0000[diff] [blame]875<tt>
876define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
877 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
878 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
879 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
880 [<a href="#gc">gc</a>] { ... }
881</tt>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]882</div>
883
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]884</div>
885
886
887<!-- ======================================================================= -->
888<div class="doc_subsection">
889 <a name="aliasstructure">Aliases</a>
890</div>
891<div class="doc_text">
892 <p>Aliases act as "second name" for the aliasee value (which can be either
Anton Korobeynikov96822812008-03-22 08:36:14 +0000[diff] [blame]893 function, global variable, another alias or bitcast of global value). Aliases
894 may have an optional <a href="#linkage">linkage type</a>, and an
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]895 optional <a href="#visibility">visibility style</a>.</p>
896
897 <h5>Syntax:</h5>
898
899<div class="doc_code">
900<pre>
Duncan Sandsd7bfabf2008-09-12 20:48:21 +0000[diff] [blame]901@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]902</pre>
903</div>
904
905</div>
906
907
908
909<!-- ======================================================================= -->
910<div class="doc_subsection"><a name="paramattrs">Parameter Attributes</a></div>
911<div class="doc_text">
912 <p>The return type and each parameter of a function type may have a set of
913 <i>parameter attributes</i> associated with them. Parameter attributes are
914 used to communicate additional information about the result or parameters of
Duncan Sandsf5588dc2007-11-27 13:23:08 +0000[diff] [blame]915 a function. Parameter attributes are considered to be part of the function,
916 not of the function type, so functions with different parameter attributes
917 can have the same function type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]918
919 <p>Parameter attributes are simple keywords that follow the type specified. If
920 multiple parameter attributes are needed, they are space separated. For
921 example:</p>
922
923<div class="doc_code">
924<pre>
Nick Lewycky3022a742009-02-15 23:06:14 +0000[diff] [blame]925declare i32 @printf(i8* noalias nocapture, ...)
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]926declare i32 @atoi(i8 zeroext)
927declare signext i8 @returns_signed_char()
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]928</pre>
929</div>
930
Duncan Sandsf5588dc2007-11-27 13:23:08 +0000[diff] [blame]931 <p>Note that any attributes for the function result (<tt>nounwind</tt>,
932 <tt>readonly</tt>) come immediately after the argument list.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]933
934 <p>Currently, only the following parameter attributes are defined:</p>
935 <dl>
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]936 <dt><tt>zeroext</tt></dt>
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]937 <dd>This indicates to the code generator that the parameter or return value
938 should be zero-extended to a 32-bit value by the caller (for a parameter)
939 or the callee (for a return value).</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]940
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]941 <dt><tt>signext</tt></dt>
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]942 <dd>This indicates to the code generator that the parameter or return value
943 should be sign-extended to a 32-bit value by the caller (for a parameter)
944 or the callee (for a return value).</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]945
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]946 <dt><tt>inreg</tt></dt>
Dale Johannesenc08a0e22008-09-25 20:47:45 +0000[diff] [blame]947 <dd>This indicates that this parameter or return value should be treated
948 in a special target-dependent fashion during while emitting code for a
949 function call or return (usually, by putting it in a register as opposed
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]950 to memory, though some targets use it to distinguish between two different
951 kinds of registers). Use of this attribute is target-specific.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]952
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]953 <dt><tt><a name="byval">byval</a></tt></dt>
Chris Lattner04c86182008-01-15 04:34:22 +0000[diff] [blame]954 <dd>This indicates that the pointer parameter should really be passed by
955 value to the function. The attribute implies that a hidden copy of the
956 pointee is made between the caller and the callee, so the callee is unable
Chris Lattner6a9f3c42008-08-05 18:21:08 +0000[diff] [blame]957 to modify the value in the callee. This attribute is only valid on LLVM
Chris Lattner04c86182008-01-15 04:34:22 +0000[diff] [blame]958 pointer arguments. It is generally used to pass structs and arrays by
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]959 value, but is also valid on pointers to scalars. The copy is considered to
960 belong to the caller not the callee (for example,
961 <tt><a href="#readonly">readonly</a></tt> functions should not write to
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]962 <tt>byval</tt> parameters). This is not a valid attribute for return
Chris Lattnere39c5bf2009-02-05 05:42:28 +0000[diff] [blame]963 values. The byval attribute also supports specifying an alignment with the
964 align attribute. This has a target-specific effect on the code generator
965 that usually indicates a desired alignment for the synthesized stack
966 slot.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]967
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]968 <dt><tt>sret</tt></dt>
Duncan Sands616cc032008-02-18 04:19:38 +0000[diff] [blame]969 <dd>This indicates that the pointer parameter specifies the address of a
970 structure that is the return value of the function in the source program.
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]971 This pointer must be guaranteed by the caller to be valid: loads and stores
972 to the structure may be assumed by the callee to not to trap. This may only
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]973 be applied to the first parameter. This is not a valid attribute for
974 return values. </dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]975
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]976 <dt><tt>noalias</tt></dt>
Nick Lewyckyff384472008-11-24 03:41:24 +0000[diff] [blame]977 <dd>This indicates that the pointer does not alias any global or any other
978 parameter. The caller is responsible for ensuring that this is the
Nick Lewyckya604a942008-11-24 05:00:44 +0000[diff] [blame]979 case. On a function return value, <tt>noalias</tt> additionally indicates
980 that the pointer does not alias any other pointers visible to the
Nick Lewycky40fb6f22008-12-19 06:39:12 +0000[diff] [blame]981 caller. For further details, please see the discussion of the NoAlias
982 response in
983 <a href="http://llvm.org/docs/AliasAnalysis.html#MustMayNo">alias
984 analysis</a>.</dd>
985
986 <dt><tt>nocapture</tt></dt>
987 <dd>This indicates that the callee does not make any copies of the pointer
988 that outlive the callee itself. This is not a valid attribute for return
989 values.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]990
Duncan Sands4ee46812007-07-27 19:57:41 +0000[diff] [blame]991 <dt><tt>nest</tt></dt>
Duncan Sandsf1a7d4c2008-07-08 09:27:25 +0000[diff] [blame]992 <dd>This indicates that the pointer parameter can be excised using the
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]993 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
994 attribute for return values.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]995 </dl>
996
997</div>
998
999<!-- ======================================================================= -->
1000<div class="doc_subsection">
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]1001 <a name="gc">Garbage Collector Names</a>
1002</div>
1003
1004<div class="doc_text">
1005<p>Each function may specify a garbage collector name, which is simply a
1006string.</p>
1007
1008<div class="doc_code"><pre
1009>define void @f() gc "name" { ...</pre></div>
1010
1011<p>The compiler declares the supported values of <i>name</i>. Specifying a
1012collector which will cause the compiler to alter its output in order to support
1013the named garbage collection algorithm.</p>
1014</div>
1015
1016<!-- ======================================================================= -->
1017<div class="doc_subsection">
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1018 <a name="fnattrs">Function Attributes</a>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1019</div>
1020
1021<div class="doc_text">
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1022
1023<p>Function attributes are set to communicate additional information about
1024 a function. Function attributes are considered to be part of the function,
1025 not of the function type, so functions with different parameter attributes
1026 can have the same function type.</p>
1027
1028 <p>Function attributes are simple keywords that follow the type specified. If
1029 multiple attributes are needed, they are space separated. For
1030 example:</p>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1031
1032<div class="doc_code">
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1033<pre>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1034define void @f() noinline { ... }
1035define void @f() alwaysinline { ... }
1036define void @f() alwaysinline optsize { ... }
1037define void @f() optsize
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1038</pre>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1039</div>
1040
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1041<dl>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1042<dt><tt>alwaysinline</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]1043<dd>This attribute indicates that the inliner should attempt to inline this
1044function into callers whenever possible, ignoring any active inlining size
1045threshold for this caller.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1046
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1047<dt><tt>noinline</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]1048<dd>This attribute indicates that the inliner should never inline this function
Chris Lattner377f3ae2008-10-05 17:14:59 +0000[diff] [blame]1049in any situation. This attribute may not be used together with the
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]1050<tt>alwaysinline</tt> attribute.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1051
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1052<dt><tt>optsize</tt></dt>
Devang Patelc29099b2008-09-29 18:34:44 +0000[diff] [blame]1053<dd>This attribute suggests that optimization passes and code generator passes
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]1054make choices that keep the code size of this function low, and otherwise do
1055optimizations specifically to reduce code size.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1056
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1057<dt><tt>noreturn</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]1058<dd>This function attribute indicates that the function never returns normally.
1059This produces undefined behavior at runtime if the function ever does
1060dynamically return.</dd>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1061
1062<dt><tt>nounwind</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]1063<dd>This function attribute indicates that the function never returns with an
1064unwind or exceptional control flow. If the function does unwind, its runtime
1065behavior is undefined.</dd>
1066
1067<dt><tt>readnone</tt></dt>
Duncan Sands2f500832009-05-06 06:49:50 +0000[diff] [blame]1068<dd>This attribute indicates that the function computes its result (or decides to
1069unwind an exception) based strictly on its arguments, without dereferencing any
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]1070pointer arguments or otherwise accessing any mutable state (e.g. memory, control
1071registers, etc) visible to caller functions. It does not write through any
1072pointer arguments (including <tt><a href="#byval">byval</a></tt> arguments) and
Duncan Sands2f500832009-05-06 06:49:50 +0000[diff] [blame]1073never changes any state visible to callers. This means that it cannot unwind
1074exceptions by calling the <tt>C++</tt> exception throwing methods, but could
1075use the <tt>unwind</tt> instruction.</dd>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1076
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]1077<dt><tt><a name="readonly">readonly</a></tt></dt>
1078<dd>This attribute indicates that the function does not write through any
1079pointer arguments (including <tt><a href="#byval">byval</a></tt> arguments)
1080or otherwise modify any state (e.g. memory, control registers, etc) visible to
1081caller functions. It may dereference pointer arguments and read state that may
Duncan Sands2f500832009-05-06 06:49:50 +0000[diff] [blame]1082be set in the caller. A readonly function always returns the same value (or
1083unwinds an exception identically) when called with the same set of arguments
1084and global state. It cannot unwind an exception by calling the <tt>C++</tt>
1085exception throwing methods, but may use the <tt>unwind</tt> instruction.</dd>
Bill Wendlingdfaabba2008-11-13 01:02:51 +0000[diff] [blame]1086
1087<dt><tt><a name="ssp">ssp</a></tt></dt>
Bill Wendling34f70d82008-11-26 19:19:05 +0000[diff] [blame]1088<dd>This attribute indicates that the function should emit a stack smashing
Bill Wendlingdfaabba2008-11-13 01:02:51 +0000[diff] [blame]1089protector. It is in the form of a "canary"&mdash;a random value placed on the
1090stack before the local variables that's checked upon return from the function to
1091see if it has been overwritten. A heuristic is used to determine if a function
Bill Wendling34f70d82008-11-26 19:19:05 +0000[diff] [blame]1092needs stack protectors or not.
Bill Wendlingdfaabba2008-11-13 01:02:51 +0000[diff] [blame]1093
Devang Patela2f9f412009-06-12 19:45:19 +0000[diff] [blame]1094<br><br>If a function that has an <tt>ssp</tt> attribute is inlined into a function
Bill Wendlingbe9ec3f2008-11-26 19:07:40 +0000[diff] [blame]1095that doesn't have an <tt>ssp</tt> attribute, then the resulting function will
Devang Patela2f9f412009-06-12 19:45:19 +0000[diff] [blame]1096have an <tt>ssp</tt> attribute.</dd>
Bill Wendlingbe9ec3f2008-11-26 19:07:40 +0000[diff] [blame]1097
1098<dt><tt>sspreq</tt></dt>
Bill Wendling34f70d82008-11-26 19:19:05 +0000[diff] [blame]1099<dd>This attribute indicates that the function should <em>always</em> emit a
Bill Wendlingdfaabba2008-11-13 01:02:51 +0000[diff] [blame]1100stack smashing protector. This overrides the <tt><a href="#ssp">ssp</a></tt>
Bill Wendling34f70d82008-11-26 19:19:05 +0000[diff] [blame]1101function attribute.
Bill Wendlingbe9ec3f2008-11-26 19:07:40 +0000[diff] [blame]1102
Devang Patela2f9f412009-06-12 19:45:19 +0000[diff] [blame]1103If a function that has an <tt>sspreq</tt> attribute is inlined into a
Bill Wendlingbe9ec3f2008-11-26 19:07:40 +0000[diff] [blame]1104function that doesn't have an <tt>sspreq</tt> attribute or which has
1105an <tt>ssp</tt> attribute, then the resulting function will have
Devang Patela2f9f412009-06-12 19:45:19 +0000[diff] [blame]1106an <tt>sspreq</tt> attribute.</dd>
1107
1108<dt><tt>noredzone</tt></dt>
Dan Gohman06c9b732009-06-15 17:37:09 +0000[diff] [blame]1109<dd>This attribute indicates that the code generator should not use a
Dan Gohmanf958d5c2009-06-15 21:18:01 +0000[diff] [blame]1110red zone, even if the target-specific ABI normally permits it.
Dan Gohman06c9b732009-06-15 17:37:09 +0000[diff] [blame]1111</dd>
Devang Patela2f9f412009-06-12 19:45:19 +0000[diff] [blame]1112
1113<dt><tt>noimplicitfloat</tt></dt>
1114<dd>This attributes disables implicit floating point instructions.</dd>
1115
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1116</dl>
1117
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1118</div>
1119
1120<!-- ======================================================================= -->
1121<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1122 <a name="moduleasm">Module-Level Inline Assembly</a>
1123</div>
1124
1125<div class="doc_text">
1126<p>
1127Modules may contain "module-level inline asm" blocks, which corresponds to the
1128GCC "file scope inline asm" blocks. These blocks are internally concatenated by
1129LLVM and treated as a single unit, but may be separated in the .ll file if
1130desired. The syntax is very simple:
1131</p>
1132
1133<div class="doc_code">
1134<pre>
1135module asm "inline asm code goes here"
1136module asm "more can go here"
1137</pre>
1138</div>
1139
1140<p>The strings can contain any character by escaping non-printable characters.
1141 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
1142 for the number.
1143</p>
1144
1145<p>
1146 The inline asm code is simply printed to the machine code .s file when
1147 assembly code is generated.
1148</p>
1149</div>
1150
1151<!-- ======================================================================= -->
1152<div class="doc_subsection">
1153 <a name="datalayout">Data Layout</a>
1154</div>
1155
1156<div class="doc_text">
1157<p>A module may specify a target specific data layout string that specifies how
1158data is to be laid out in memory. The syntax for the data layout is simply:</p>
1159<pre> target datalayout = "<i>layout specification</i>"</pre>
1160<p>The <i>layout specification</i> consists of a list of specifications
1161separated by the minus sign character ('-'). Each specification starts with a
1162letter and may include other information after the letter to define some
1163aspect of the data layout. The specifications accepted are as follows: </p>
1164<dl>
1165 <dt><tt>E</tt></dt>
1166 <dd>Specifies that the target lays out data in big-endian form. That is, the
1167 bits with the most significance have the lowest address location.</dd>
1168 <dt><tt>e</tt></dt>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1169 <dd>Specifies that the target lays out data in little-endian form. That is,
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1170 the bits with the least significance have the lowest address location.</dd>
1171 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1172 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
1173 <i>preferred</i> alignments. All sizes are in bits. Specifying the <i>pref</i>
1174 alignment is optional. If omitted, the preceding <tt>:</tt> should be omitted
1175 too.</dd>
1176 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1177 <dd>This specifies the alignment for an integer type of a given bit
1178 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1179 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1180 <dd>This specifies the alignment for a vector type of a given bit
1181 <i>size</i>.</dd>
1182 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1183 <dd>This specifies the alignment for a floating point type of a given bit
1184 <i>size</i>. The value of <i>size</i> must be either 32 (float) or 64
1185 (double).</dd>
1186 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1187 <dd>This specifies the alignment for an aggregate type of a given bit
1188 <i>size</i>.</dd>
Daniel Dunbard88a97b2009-06-08 22:17:53 +0000[diff] [blame]1189 <dt><tt>s<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1190 <dd>This specifies the alignment for a stack object of a given bit
1191 <i>size</i>.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1192</dl>
1193<p>When constructing the data layout for a given target, LLVM starts with a
1194default set of specifications which are then (possibly) overriden by the
1195specifications in the <tt>datalayout</tt> keyword. The default specifications
1196are given in this list:</p>
1197<ul>
1198 <li><tt>E</tt> - big endian</li>
1199 <li><tt>p:32:64:64</tt> - 32-bit pointers with 64-bit alignment</li>
1200 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1201 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1202 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1203 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1204 <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]1205 alignment of 64-bits</li>
1206 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1207 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1208 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1209 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1210 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
Daniel Dunbard88a97b2009-06-08 22:17:53 +0000[diff] [blame]1211 <li><tt>s0:64:64</tt> - stack objects are 64-bit aligned</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1212</ul>
Chris Lattner6a9f3c42008-08-05 18:21:08 +0000[diff] [blame]1213<p>When LLVM is determining the alignment for a given type, it uses the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]1214following rules:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1215<ol>
1216 <li>If the type sought is an exact match for one of the specifications, that
1217 specification is used.</li>
1218 <li>If no match is found, and the type sought is an integer type, then the
1219 smallest integer type that is larger than the bitwidth of the sought type is
1220 used. If none of the specifications are larger than the bitwidth then the the
1221 largest integer type is used. For example, given the default specifications
1222 above, the i7 type will use the alignment of i8 (next largest) while both
1223 i65 and i256 will use the alignment of i64 (largest specified).</li>
1224 <li>If no match is found, and the type sought is a vector type, then the
1225 largest vector type that is smaller than the sought vector type will be used
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]1226 as a fall back. This happens because &lt;128 x double&gt; can be implemented
1227 in terms of 64 &lt;2 x double&gt;, for example.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1228</ol>
1229</div>
1230
1231<!-- *********************************************************************** -->
1232<div class="doc_section"> <a name="typesystem">Type System</a> </div>
1233<!-- *********************************************************************** -->
1234
1235<div class="doc_text">
1236
1237<p>The LLVM type system is one of the most important features of the
1238intermediate representation. Being typed enables a number of
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1239optimizations to be performed on the intermediate representation directly,
1240without having to do
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1241extra analyses on the side before the transformation. A strong type
1242system makes it easier to read the generated code and enables novel
1243analyses and transformations that are not feasible to perform on normal
1244three address code representations.</p>
1245
1246</div>
1247
1248<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1249<div class="doc_subsection"> <a name="t_classifications">Type
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1250Classifications</a> </div>
1251<div class="doc_text">
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1252<p>The types fall into a few useful
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1253classifications:</p>
1254
1255<table border="1" cellspacing="0" cellpadding="4">
1256 <tbody>
1257 <tr><th>Classification</th><th>Types</th></tr>
1258 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1259 <td><a href="#t_integer">integer</a></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1260 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
1261 </tr>
1262 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1263 <td><a href="#t_floating">floating point</a></td>
1264 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1265 </tr>
1266 <tr>
1267 <td><a name="t_firstclass">first class</a></td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1268 <td><a href="#t_integer">integer</a>,
1269 <a href="#t_floating">floating point</a>,
1270 <a href="#t_pointer">pointer</a>,
Dan Gohmanf6237db2008-06-18 18:42:13 +0000[diff] [blame]1271 <a href="#t_vector">vector</a>,
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]1272 <a href="#t_struct">structure</a>,
1273 <a href="#t_array">array</a>,
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]1274 <a href="#t_label">label</a>,
1275 <a href="#t_metadata">metadata</a>.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1276 </td>
1277 </tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1278 <tr>
1279 <td><a href="#t_primitive">primitive</a></td>
1280 <td><a href="#t_label">label</a>,
1281 <a href="#t_void">void</a>,
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]1282 <a href="#t_floating">floating point</a>,
1283 <a href="#t_metadata">metadata</a>.</td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1284 </tr>
1285 <tr>
1286 <td><a href="#t_derived">derived</a></td>
1287 <td><a href="#t_integer">integer</a>,
1288 <a href="#t_array">array</a>,
1289 <a href="#t_function">function</a>,
1290 <a href="#t_pointer">pointer</a>,
1291 <a href="#t_struct">structure</a>,
1292 <a href="#t_pstruct">packed structure</a>,
1293 <a href="#t_vector">vector</a>,
1294 <a href="#t_opaque">opaque</a>.
Dan Gohman032ba852008-10-14 16:32:04 +0000[diff] [blame]1295 </td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1296 </tr>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1297 </tbody>
1298</table>
1299
1300<p>The <a href="#t_firstclass">first class</a> types are perhaps the
1301most important. Values of these types are the only ones which can be
1302produced by instructions, passed as arguments, or used as operands to
Dan Gohman4f29e422008-05-23 21:53:15 +0000[diff] [blame]1303instructions.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1304</div>
1305
1306<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1307<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1308
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1309<div class="doc_text">
1310<p>The primitive types are the fundamental building blocks of the LLVM
1311system.</p>
1312
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1313</div>
1314
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1315<!-- _______________________________________________________________________ -->
1316<div class="doc_subsubsection"> <a name="t_floating">Floating Point Types</a> </div>
1317
1318<div class="doc_text">
1319 <table>
1320 <tbody>
1321 <tr><th>Type</th><th>Description</th></tr>
1322 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1323 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1324 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1325 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1326 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1327 </tbody>
1328 </table>
1329</div>
1330
1331<!-- _______________________________________________________________________ -->
1332<div class="doc_subsubsection"> <a name="t_void">Void Type</a> </div>
1333
1334<div class="doc_text">
1335<h5>Overview:</h5>
1336<p>The void type does not represent any value and has no size.</p>
1337
1338<h5>Syntax:</h5>
1339
1340<pre>
1341 void
1342</pre>
1343</div>
1344
1345<!-- _______________________________________________________________________ -->
1346<div class="doc_subsubsection"> <a name="t_label">Label Type</a> </div>
1347
1348<div class="doc_text">
1349<h5>Overview:</h5>
1350<p>The label type represents code labels.</p>
1351
1352<h5>Syntax:</h5>
1353
1354<pre>
1355 label
1356</pre>
1357</div>
1358
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]1359<!-- _______________________________________________________________________ -->
1360<div class="doc_subsubsection"> <a name="t_metadata">Metadata Type</a> </div>
1361
1362<div class="doc_text">
1363<h5>Overview:</h5>
1364<p>The metadata type represents embedded metadata. The only derived type that
1365may contain metadata is <tt>metadata*</tt> or a function type that returns or
1366takes metadata typed parameters, but not pointer to metadata types.</p>
1367
1368<h5>Syntax:</h5>
1369
1370<pre>
1371 metadata
1372</pre>
1373</div>
1374
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1375
1376<!-- ======================================================================= -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1377<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
1378
1379<div class="doc_text">
1380
1381<p>The real power in LLVM comes from the derived types in the system.
1382This is what allows a programmer to represent arrays, functions,
1383pointers, and other useful types. Note that these derived types may be
1384recursive: For example, it is possible to have a two dimensional array.</p>
1385
1386</div>
1387
1388<!-- _______________________________________________________________________ -->
1389<div class="doc_subsubsection"> <a name="t_integer">Integer Type</a> </div>
1390
1391<div class="doc_text">
1392
1393<h5>Overview:</h5>
1394<p>The integer type is a very simple derived type that simply specifies an
1395arbitrary bit width for the integer type desired. Any bit width from 1 bit to
13962^23-1 (about 8 million) can be specified.</p>
1397
1398<h5>Syntax:</h5>
1399
1400<pre>
1401 iN
1402</pre>
1403
1404<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1405value.</p>
1406
1407<h5>Examples:</h5>
1408<table class="layout">
Nick Lewycky39382d62009-05-24 02:46:06 +0000[diff] [blame]1409 <tr class="layout">
1410 <td class="left"><tt>i1</tt></td>
1411 <td class="left">a single-bit integer.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1412 </tr>
Nick Lewycky39382d62009-05-24 02:46:06 +0000[diff] [blame]1413 <tr class="layout">
1414 <td class="left"><tt>i32</tt></td>
1415 <td class="left">a 32-bit integer.</td>
1416 </tr>
1417 <tr class="layout">
1418 <td class="left"><tt>i1942652</tt></td>
1419 <td class="left">a really big integer of over 1 million bits.</td>
1420 </tr>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1421</table>
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1422
1423<p>Note that the code generator does not yet support large integer types
1424to be used as function return types. The specific limit on how large a
1425return type the code generator can currently handle is target-dependent;
1426currently it's often 64 bits for 32-bit targets and 128 bits for 64-bit
1427targets.</p>
1428
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1429</div>
1430
1431<!-- _______________________________________________________________________ -->
1432<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
1433
1434<div class="doc_text">
1435
1436<h5>Overview:</h5>
1437
1438<p>The array type is a very simple derived type that arranges elements
1439sequentially in memory. The array type requires a size (number of
1440elements) and an underlying data type.</p>
1441
1442<h5>Syntax:</h5>
1443
1444<pre>
1445 [&lt;# elements&gt; x &lt;elementtype&gt;]
1446</pre>
1447
1448<p>The number of elements is a constant integer value; elementtype may
1449be any type with a size.</p>
1450
1451<h5>Examples:</h5>
1452<table class="layout">
1453 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1454 <td class="left"><tt>[40 x i32]</tt></td>
1455 <td class="left">Array of 40 32-bit integer values.</td>
1456 </tr>
1457 <tr class="layout">
1458 <td class="left"><tt>[41 x i32]</tt></td>
1459 <td class="left">Array of 41 32-bit integer values.</td>
1460 </tr>
1461 <tr class="layout">
1462 <td class="left"><tt>[4 x i8]</tt></td>
1463 <td class="left">Array of 4 8-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1464 </tr>
1465</table>
1466<p>Here are some examples of multidimensional arrays:</p>
1467<table class="layout">
1468 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1469 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1470 <td class="left">3x4 array of 32-bit integer values.</td>
1471 </tr>
1472 <tr class="layout">
1473 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1474 <td class="left">12x10 array of single precision floating point values.</td>
1475 </tr>
1476 <tr class="layout">
1477 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1478 <td class="left">2x3x4 array of 16-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1479 </tr>
1480</table>
1481
1482<p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
1483length array. Normally, accesses past the end of an array are undefined in
1484LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
1485As a special case, however, zero length arrays are recognized to be variable
1486length. This allows implementation of 'pascal style arrays' with the LLVM
1487type "{ i32, [0 x float]}", for example.</p>
1488
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1489<p>Note that the code generator does not yet support large aggregate types
1490to be used as function return types. The specific limit on how large an
1491aggregate return type the code generator can currently handle is
1492target-dependent, and also dependent on the aggregate element types.</p>
1493
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1494</div>
1495
1496<!-- _______________________________________________________________________ -->
1497<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
1498<div class="doc_text">
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1499
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1500<h5>Overview:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1501
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1502<p>The function type can be thought of as a function signature. It
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1503consists of a return type and a list of formal parameter types. The
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1504return type of a function type is a scalar type, a void type, or a struct type.
Devang Pateld5404c02008-03-24 20:52:42 +0000[diff] [blame]1505If the return type is a struct type then all struct elements must be of first
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1506class types, and the struct must have at least one element.</p>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1507
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1508<h5>Syntax:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1509
1510<pre>
1511 &lt;returntype list&gt; (&lt;parameter list&gt;)
1512</pre>
1513
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1514<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
1515specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1516which indicates that the function takes a variable number of arguments.
1517Variable argument functions can access their arguments with the <a
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1518 href="#int_varargs">variable argument handling intrinsic</a> functions.
1519'<tt>&lt;returntype list&gt;</tt>' is a comma-separated list of
1520<a href="#t_firstclass">first class</a> type specifiers.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1521
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1522<h5>Examples:</h5>
1523<table class="layout">
1524 <tr class="layout">
1525 <td class="left"><tt>i32 (i32)</tt></td>
1526 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
1527 </td>
1528 </tr><tr class="layout">
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]1529 <td class="left"><tt>float&nbsp;(i16&nbsp;signext,&nbsp;i32&nbsp;*)&nbsp;*
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1530 </tt></td>
1531 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
1532 an <tt>i16</tt> that should be sign extended and a
1533 <a href="#t_pointer">pointer</a> to <tt>i32</tt>, returning
1534 <tt>float</tt>.
1535 </td>
1536 </tr><tr class="layout">
1537 <td class="left"><tt>i32 (i8*, ...)</tt></td>
1538 <td class="left">A vararg function that takes at least one
1539 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1540 which returns an integer. This is the signature for <tt>printf</tt> in
1541 LLVM.
1542 </td>
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1543 </tr><tr class="layout">
1544 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Misha Brukmanafc88b02008-11-27 06:41:20 +0000[diff] [blame]1545 <td class="left">A function taking an <tt>i32</tt>, returning two
1546 <tt>i32</tt> values as an aggregate of type <tt>{ i32, i32 }</tt>
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1547 </td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1548 </tr>
1549</table>
1550
1551</div>
1552<!-- _______________________________________________________________________ -->
1553<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
1554<div class="doc_text">
1555<h5>Overview:</h5>
1556<p>The structure type is used to represent a collection of data members
1557together in memory. The packing of the field types is defined to match
1558the ABI of the underlying processor. The elements of a structure may
1559be any type that has a size.</p>
1560<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1561and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1562field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1563instruction.</p>
1564<h5>Syntax:</h5>
1565<pre> { &lt;type list&gt; }<br></pre>
1566<h5>Examples:</h5>
1567<table class="layout">
1568 <tr class="layout">
1569 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1570 <td class="left">A triple of three <tt>i32</tt> values</td>
1571 </tr><tr class="layout">
1572 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1573 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1574 second element is a <a href="#t_pointer">pointer</a> to a
1575 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1576 an <tt>i32</tt>.</td>
1577 </tr>
1578</table>
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1579
1580<p>Note that the code generator does not yet support large aggregate types
1581to be used as function return types. The specific limit on how large an
1582aggregate return type the code generator can currently handle is
1583target-dependent, and also dependent on the aggregate element types.</p>
1584
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1585</div>
1586
1587<!-- _______________________________________________________________________ -->
1588<div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
1589</div>
1590<div class="doc_text">
1591<h5>Overview:</h5>
1592<p>The packed structure type is used to represent a collection of data members
1593together in memory. There is no padding between fields. Further, the alignment
1594of a packed structure is 1 byte. The elements of a packed structure may
1595be any type that has a size.</p>
1596<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1597and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1598field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1599instruction.</p>
1600<h5>Syntax:</h5>
1601<pre> &lt; { &lt;type list&gt; } &gt; <br></pre>
1602<h5>Examples:</h5>
1603<table class="layout">
1604 <tr class="layout">
1605 <td class="left"><tt>&lt; { i32, i32, i32 } &gt;</tt></td>
1606 <td class="left">A triple of three <tt>i32</tt> values</td>
1607 </tr><tr class="layout">
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1608 <td class="left">
1609<tt>&lt;&nbsp;{&nbsp;float,&nbsp;i32&nbsp;(i32)*&nbsp;}&nbsp;&gt;</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1610 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1611 second element is a <a href="#t_pointer">pointer</a> to a
1612 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1613 an <tt>i32</tt>.</td>
1614 </tr>
1615</table>
1616</div>
1617
1618<!-- _______________________________________________________________________ -->
1619<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
1620<div class="doc_text">
1621<h5>Overview:</h5>
1622<p>As in many languages, the pointer type represents a pointer or
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]1623reference to another object, which must live in memory. Pointer types may have
1624an optional address space attribute defining the target-specific numbered
1625address space where the pointed-to object resides. The default address space is
1626zero.</p>
Chris Lattner96edbd32009-02-08 19:53:29 +0000[diff] [blame]1627
1628<p>Note that LLVM does not permit pointers to void (<tt>void*</tt>) nor does
Chris Lattner0d879e02009-02-08 22:21:28 +0000[diff] [blame]1629it permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</tt> instead.</p>
Chris Lattner96edbd32009-02-08 19:53:29 +0000[diff] [blame]1630
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1631<h5>Syntax:</h5>
1632<pre> &lt;type&gt; *<br></pre>
1633<h5>Examples:</h5>
1634<table class="layout">
1635 <tr class="layout">
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]1636 <td class="left"><tt>[4 x i32]*</tt></td>
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1637 <td class="left">A <a href="#t_pointer">pointer</a> to <a
1638 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
1639 </tr>
1640 <tr class="layout">
1641 <td class="left"><tt>i32 (i32 *) *</tt></td>
1642 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1643 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1644 <tt>i32</tt>.</td>
1645 </tr>
1646 <tr class="layout">
1647 <td class="left"><tt>i32 addrspace(5)*</tt></td>
1648 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
1649 that resides in address space #5.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1650 </tr>
1651</table>
1652</div>
1653
1654<!-- _______________________________________________________________________ -->
1655<div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
1656<div class="doc_text">
1657
1658<h5>Overview:</h5>
1659
1660<p>A vector type is a simple derived type that represents a vector
1661of elements. Vector types are used when multiple primitive data
1662are operated in parallel using a single instruction (SIMD).
1663A vector type requires a size (number of
1664elements) and an underlying primitive data type. Vectors must have a power
1665of two length (1, 2, 4, 8, 16 ...). Vector types are
1666considered <a href="#t_firstclass">first class</a>.</p>
1667
1668<h5>Syntax:</h5>
1669
1670<pre>
1671 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1672</pre>
1673
1674<p>The number of elements is a constant integer value; elementtype may
1675be any integer or floating point type.</p>
1676
1677<h5>Examples:</h5>
1678
1679<table class="layout">
1680 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1681 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
1682 <td class="left">Vector of 4 32-bit integer values.</td>
1683 </tr>
1684 <tr class="layout">
1685 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
1686 <td class="left">Vector of 8 32-bit floating-point values.</td>
1687 </tr>
1688 <tr class="layout">
1689 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
1690 <td class="left">Vector of 2 64-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1691 </tr>
1692</table>
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1693
1694<p>Note that the code generator does not yet support large vector types
1695to be used as function return types. The specific limit on how large a
1696vector return type codegen can currently handle is target-dependent;
1697currently it's often a few times longer than a hardware vector register.</p>
1698
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1699</div>
1700
1701<!-- _______________________________________________________________________ -->
1702<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1703<div class="doc_text">
1704
1705<h5>Overview:</h5>
1706
1707<p>Opaque types are used to represent unknown types in the system. This
Gordon Henriksenda0706e2007-10-14 00:34:53 +0000[diff] [blame]1708corresponds (for example) to the C notion of a forward declared structure type.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1709In LLVM, opaque types can eventually be resolved to any type (not just a
1710structure type).</p>
1711
1712<h5>Syntax:</h5>
1713
1714<pre>
1715 opaque
1716</pre>
1717
1718<h5>Examples:</h5>
1719
1720<table class="layout">
1721 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1722 <td class="left"><tt>opaque</tt></td>
1723 <td class="left">An opaque type.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1724 </tr>
1725</table>
1726</div>
1727
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]1728<!-- ======================================================================= -->
1729<div class="doc_subsection">
1730 <a name="t_uprefs">Type Up-references</a>
1731</div>
1732
1733<div class="doc_text">
1734<h5>Overview:</h5>
1735<p>
1736An "up reference" allows you to refer to a lexically enclosing type without
1737requiring it to have a name. For instance, a structure declaration may contain a
1738pointer to any of the types it is lexically a member of. Example of up
1739references (with their equivalent as named type declarations) include:</p>
1740
1741<pre>
Chris Lattner5ad632d2009-02-09 10:00:56 +0000[diff] [blame]1742 { \2 * } %x = type { %x* }
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]1743 { \2 }* %y = type { %y }*
1744 \1* %z = type %z*
1745</pre>
1746
1747<p>
1748An up reference is needed by the asmprinter for printing out cyclic types when
1749there is no declared name for a type in the cycle. Because the asmprinter does
1750not want to print out an infinite type string, it needs a syntax to handle
1751recursive types that have no names (all names are optional in llvm IR).
1752</p>
1753
1754<h5>Syntax:</h5>
1755<pre>
1756 \&lt;level&gt;
1757</pre>
1758
1759<p>
1760The level is the count of the lexical type that is being referred to.
1761</p>
1762
1763<h5>Examples:</h5>
1764
1765<table class="layout">
1766 <tr class="layout">
1767 <td class="left"><tt>\1*</tt></td>
1768 <td class="left">Self-referential pointer.</td>
1769 </tr>
1770 <tr class="layout">
1771 <td class="left"><tt>{ { \3*, i8 }, i32 }</tt></td>
1772 <td class="left">Recursive structure where the upref refers to the out-most
1773 structure.</td>
1774 </tr>
1775</table>
1776</div>
1777
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1778
1779<!-- *********************************************************************** -->
1780<div class="doc_section"> <a name="constants">Constants</a> </div>
1781<!-- *********************************************************************** -->
1782
1783<div class="doc_text">
1784
1785<p>LLVM has several different basic types of constants. This section describes
1786them all and their syntax.</p>
1787
1788</div>
1789
1790<!-- ======================================================================= -->
1791<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
1792
1793<div class="doc_text">
1794
1795<dl>
1796 <dt><b>Boolean constants</b></dt>
1797
1798 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
1799 constants of the <tt><a href="#t_primitive">i1</a></tt> type.
1800 </dd>
1801
1802 <dt><b>Integer constants</b></dt>
1803
1804 <dd>Standard integers (such as '4') are constants of the <a
1805 href="#t_integer">integer</a> type. Negative numbers may be used with
1806 integer types.
1807 </dd>
1808
1809 <dt><b>Floating point constants</b></dt>
1810
1811 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
1812 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]1813 notation (see below). The assembler requires the exact decimal value of
1814 a floating-point constant. For example, the assembler accepts 1.25 but
1815 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
1816 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1817
1818 <dt><b>Null pointer constants</b></dt>
1819
1820 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
1821 and must be of <a href="#t_pointer">pointer type</a>.</dd>
1822
1823</dl>
1824
Dale Johannesenf82a52f2009-02-11 22:14:51 +0000[diff] [blame]1825<p>The one non-intuitive notation for constants is the hexadecimal form
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1826of floating point constants. For example, the form '<tt>double
18270x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
18284.5e+15</tt>'. The only time hexadecimal floating point constants are required
1829(and the only time that they are generated by the disassembler) is when a
1830floating point constant must be emitted but it cannot be represented as a
Dale Johannesenf82a52f2009-02-11 22:14:51 +0000[diff] [blame]1831decimal floating point number in a reasonable number of digits. For example,
1832NaN's, infinities, and other
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1833special values are represented in their IEEE hexadecimal format so that
1834assembly and disassembly do not cause any bits to change in the constants.</p>
Dale Johannesenf82a52f2009-02-11 22:14:51 +0000[diff] [blame]1835<p>When using the hexadecimal form, constants of types float and double are
1836represented using the 16-digit form shown above (which matches the IEEE754
1837representation for double); float values must, however, be exactly representable
1838as IEE754 single precision.
1839Hexadecimal format is always used for long
1840double, and there are three forms of long double. The 80-bit
1841format used by x86 is represented as <tt>0xK</tt>
1842followed by 20 hexadecimal digits.
1843The 128-bit format used by PowerPC (two adjacent doubles) is represented
1844by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit
1845format is represented
1846by <tt>0xL</tt> followed by 32 hexadecimal digits; no currently supported
1847target uses this format. Long doubles will only work if they match
1848the long double format on your target. All hexadecimal formats are big-endian
1849(sign bit at the left).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1850</div>
1851
1852<!-- ======================================================================= -->
Chris Lattner97063852009-02-28 18:32:25 +0000[diff] [blame]1853<div class="doc_subsection">
1854<a name="aggregateconstants"> <!-- old anchor -->
1855<a name="complexconstants">Complex Constants</a></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1856</div>
1857
1858<div class="doc_text">
Chris Lattner97063852009-02-28 18:32:25 +0000[diff] [blame]1859<p>Complex constants are a (potentially recursive) combination of simple
1860constants and smaller complex constants.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1861
1862<dl>
1863 <dt><b>Structure constants</b></dt>
1864
1865 <dd>Structure constants are represented with notation similar to structure
1866 type definitions (a comma separated list of elements, surrounded by braces
Chris Lattner6c8de962007-12-25 20:34:52 +0000[diff] [blame]1867 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
1868 where "<tt>@G</tt>" is declared as "<tt>@G = external global i32</tt>". Structure constants
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1869 must have <a href="#t_struct">structure type</a>, and the number and
1870 types of elements must match those specified by the type.
1871 </dd>
1872
1873 <dt><b>Array constants</b></dt>
1874
1875 <dd>Array constants are represented with notation similar to array type
1876 definitions (a comma separated list of elements, surrounded by square brackets
1877 (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74 ]</tt>". Array
1878 constants must have <a href="#t_array">array type</a>, and the number and
1879 types of elements must match those specified by the type.
1880 </dd>
1881
1882 <dt><b>Vector constants</b></dt>
1883
1884 <dd>Vector constants are represented with notation similar to vector type
1885 definitions (a comma separated list of elements, surrounded by
1886 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32 42,
1887 i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must have <a
1888 href="#t_vector">vector type</a>, and the number and types of elements must
1889 match those specified by the type.
1890 </dd>
1891
1892 <dt><b>Zero initialization</b></dt>
1893
1894 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1895 value to zero of <em>any</em> type, including scalar and aggregate types.
1896 This is often used to avoid having to print large zero initializers (e.g. for
1897 large arrays) and is always exactly equivalent to using explicit zero
1898 initializers.
1899 </dd>
Nick Lewycky4dcf8102009-04-04 07:22:01 +0000[diff] [blame]1900
1901 <dt><b>Metadata node</b></dt>
1902
Nick Lewyckyf122c7e2009-05-30 16:08:30 +0000[diff] [blame]1903 <dd>A metadata node is a structure-like constant with
1904 <a href="#t_metadata">metadata type</a>. For example:
1905 "<tt>metadata !{ i32 0, metadata !"test" }</tt>". Unlike other constants
1906 that are meant to be interpreted as part of the instruction stream, metadata
1907 is a place to attach additional information such as debug info.
Nick Lewycky4dcf8102009-04-04 07:22:01 +0000[diff] [blame]1908 </dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1909</dl>
1910
1911</div>
1912
1913<!-- ======================================================================= -->
1914<div class="doc_subsection">
1915 <a name="globalconstants">Global Variable and Function Addresses</a>
1916</div>
1917
1918<div class="doc_text">
1919
1920<p>The addresses of <a href="#globalvars">global variables</a> and <a
1921href="#functionstructure">functions</a> are always implicitly valid (link-time)
1922constants. These constants are explicitly referenced when the <a
1923href="#identifiers">identifier for the global</a> is used and always have <a
1924href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1925file:</p>
1926
1927<div class="doc_code">
1928<pre>
1929@X = global i32 17
1930@Y = global i32 42
1931@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
1932</pre>
1933</div>
1934
1935</div>
1936
1937<!-- ======================================================================= -->
1938<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
1939<div class="doc_text">
1940 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
1941 no specific value. Undefined values may be of any type and be used anywhere
1942 a constant is permitted.</p>
1943
1944 <p>Undefined values indicate to the compiler that the program is well defined
1945 no matter what value is used, giving the compiler more freedom to optimize.
1946 </p>
1947</div>
1948
1949<!-- ======================================================================= -->
1950<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1951</div>
1952
1953<div class="doc_text">
1954
1955<p>Constant expressions are used to allow expressions involving other constants
1956to be used as constants. Constant expressions may be of any <a
1957href="#t_firstclass">first class</a> type and may involve any LLVM operation
1958that does not have side effects (e.g. load and call are not supported). The
1959following is the syntax for constant expressions:</p>
1960
1961<dl>
1962 <dt><b><tt>trunc ( CST to TYPE )</tt></b></dt>
1963 <dd>Truncate a constant to another type. The bit size of CST must be larger
1964 than the bit size of TYPE. Both types must be integers.</dd>
1965
1966 <dt><b><tt>zext ( CST to TYPE )</tt></b></dt>
1967 <dd>Zero extend a constant to another type. The bit size of CST must be
1968 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1969
1970 <dt><b><tt>sext ( CST to TYPE )</tt></b></dt>
1971 <dd>Sign extend a constant to another type. The bit size of CST must be
1972 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1973
1974 <dt><b><tt>fptrunc ( CST to TYPE )</tt></b></dt>
1975 <dd>Truncate a floating point constant to another floating point type. The
1976 size of CST must be larger than the size of TYPE. Both types must be
1977 floating point.</dd>
1978
1979 <dt><b><tt>fpext ( CST to TYPE )</tt></b></dt>
1980 <dd>Floating point extend a constant to another type. The size of CST must be
1981 smaller or equal to the size of TYPE. Both types must be floating point.</dd>
1982
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]1983 <dt><b><tt>fptoui ( CST to TYPE )</tt></b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1984 <dd>Convert a floating point constant to the corresponding unsigned integer
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1985 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1986 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1987 of the same number of elements. If the value won't fit in the integer type,
1988 the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1989
1990 <dt><b><tt>fptosi ( CST to TYPE )</tt></b></dt>
1991 <dd>Convert a floating point constant to the corresponding signed integer
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1992 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1993 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1994 of the same number of elements. If the value won't fit in the integer type,
1995 the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1996
1997 <dt><b><tt>uitofp ( CST to TYPE )</tt></b></dt>
1998 <dd>Convert an unsigned integer constant to the corresponding floating point
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1999 constant. TYPE must be a scalar or vector floating point type. CST must be of
2000 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
2001 of the same number of elements. If the value won't fit in the floating point
2002 type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2003
2004 <dt><b><tt>sitofp ( CST to TYPE )</tt></b></dt>
2005 <dd>Convert a signed integer constant to the corresponding floating point
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]2006 constant. TYPE must be a scalar or vector floating point type. CST must be of
2007 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
2008 of the same number of elements. If the value won't fit in the floating point
2009 type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2010
2011 <dt><b><tt>ptrtoint ( CST to TYPE )</tt></b></dt>
2012 <dd>Convert a pointer typed constant to the corresponding integer constant
2013 TYPE must be an integer type. CST must be of pointer type. The CST value is
2014 zero extended, truncated, or unchanged to make it fit in TYPE.</dd>
2015
2016 <dt><b><tt>inttoptr ( CST to TYPE )</tt></b></dt>
2017 <dd>Convert a integer constant to a pointer constant. TYPE must be a
2018 pointer type. CST must be of integer type. The CST value is zero extended,
2019 truncated, or unchanged to make it fit in a pointer size. This one is
2020 <i>really</i> dangerous!</dd>
2021
2022 <dt><b><tt>bitcast ( CST to TYPE )</tt></b></dt>
Chris Lattner557bc5d2009-02-28 18:27:03 +0000[diff] [blame]2023 <dd>Convert a constant, CST, to another TYPE. The constraints of the operands
2024 are the same as those for the <a href="#i_bitcast">bitcast
2025 instruction</a>.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2026
2027 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
2028
2029 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
2030 constants. As with the <a href="#i_getelementptr">getelementptr</a>
2031 instruction, the index list may have zero or more indexes, which are required
2032 to make sense for the type of "CSTPTR".</dd>
2033
2034 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
2035
2036 <dd>Perform the <a href="#i_select">select operation</a> on
2037 constants.</dd>
2038
2039 <dt><b><tt>icmp COND ( VAL1, VAL2 )</tt></b></dt>
2040 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
2041
2042 <dt><b><tt>fcmp COND ( VAL1, VAL2 )</tt></b></dt>
2043 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
2044
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]2045 <dt><b><tt>vicmp COND ( VAL1, VAL2 )</tt></b></dt>
2046 <dd>Performs the <a href="#i_vicmp">vicmp operation</a> on constants.</dd>
2047
2048 <dt><b><tt>vfcmp COND ( VAL1, VAL2 )</tt></b></dt>
2049 <dd>Performs the <a href="#i_vfcmp">vfcmp operation</a> on constants.</dd>
2050
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2051 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
2052
2053 <dd>Perform the <a href="#i_extractelement">extractelement
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]2054 operation</a> on constants.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2055
2056 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
2057
2058 <dd>Perform the <a href="#i_insertelement">insertelement
2059 operation</a> on constants.</dd>
2060
2061
2062 <dt><b><tt>shufflevector ( VEC1, VEC2, IDXMASK )</tt></b></dt>
2063
2064 <dd>Perform the <a href="#i_shufflevector">shufflevector
2065 operation</a> on constants.</dd>
2066
2067 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
2068
2069 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
2070 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
2071 binary</a> operations. The constraints on operands are the same as those for
2072 the corresponding instruction (e.g. no bitwise operations on floating point
2073 values are allowed).</dd>
2074</dl>
2075</div>
2076
Nick Lewycky4dcf8102009-04-04 07:22:01 +0000[diff] [blame]2077<!-- ======================================================================= -->
2078<div class="doc_subsection"><a name="metadata">Embedded Metadata</a>
2079</div>
2080
2081<div class="doc_text">
2082
2083<p>Embedded metadata provides a way to attach arbitrary data to the
2084instruction stream without affecting the behaviour of the program. There are
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]2085two metadata primitives, strings and nodes. All metadata has the
2086<tt>metadata</tt> type and is identified in syntax by a preceding exclamation
2087point ('<tt>!</tt>').
Nick Lewycky4dcf8102009-04-04 07:22:01 +0000[diff] [blame]2088</p>
2089
2090<p>A metadata string is a string surrounded by double quotes. It can contain
2091any character by escaping non-printable characters with "\xx" where "xx" is
2092the two digit hex code. For example: "<tt>!"test\00"</tt>".
2093</p>
2094
2095<p>Metadata nodes are represented with notation similar to structure constants
2096(a comma separated list of elements, surrounded by braces and preceeded by an
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]2097exclamation point). For example: "<tt>!{ metadata !"test\00", i32 10}</tt>".
Nick Lewycky4dcf8102009-04-04 07:22:01 +0000[diff] [blame]2098</p>
2099
Nick Lewycky117f4382009-05-10 20:57:05 +0000[diff] [blame]2100<p>A metadata node will attempt to track changes to the values it holds. In
2101the event that a value is deleted, it will be replaced with a typeless
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]2102"<tt>null</tt>", such as "<tt>metadata !{null, i32 10}</tt>".</p>
Nick Lewycky117f4382009-05-10 20:57:05 +0000[diff] [blame]2103
Nick Lewycky4dcf8102009-04-04 07:22:01 +0000[diff] [blame]2104<p>Optimizations may rely on metadata to provide additional information about
2105the program that isn't available in the instructions, or that isn't easily
2106computable. Similarly, the code generator may expect a certain metadata format
2107to be used to express debugging information.</p>
2108</div>
2109
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2110<!-- *********************************************************************** -->
2111<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
2112<!-- *********************************************************************** -->
2113
2114<!-- ======================================================================= -->
2115<div class="doc_subsection">
2116<a name="inlineasm">Inline Assembler Expressions</a>
2117</div>
2118
2119<div class="doc_text">
2120
2121<p>
2122LLVM supports inline assembler expressions (as opposed to <a href="#moduleasm">
2123Module-Level Inline Assembly</a>) through the use of a special value. This
2124value represents the inline assembler as a string (containing the instructions
2125to emit), a list of operand constraints (stored as a string), and a flag that
2126indicates whether or not the inline asm expression has side effects. An example
2127inline assembler expression is:
2128</p>
2129
2130<div class="doc_code">
2131<pre>
2132i32 (i32) asm "bswap $0", "=r,r"
2133</pre>
2134</div>
2135
2136<p>
2137Inline assembler expressions may <b>only</b> be used as the callee operand of
2138a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we have:
2139</p>
2140
2141<div class="doc_code">
2142<pre>
2143%X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
2144</pre>
2145</div>
2146
2147<p>
2148Inline asms with side effects not visible in the constraint list must be marked
2149as having side effects. This is done through the use of the
2150'<tt>sideeffect</tt>' keyword, like so:
2151</p>
2152
2153<div class="doc_code">
2154<pre>
2155call void asm sideeffect "eieio", ""()
2156</pre>
2157</div>
2158
2159<p>TODO: The format of the asm and constraints string still need to be
2160documented here. Constraints on what can be done (e.g. duplication, moving, etc
Chris Lattner1cdd64e2008-10-04 18:36:02 +0000[diff] [blame]2161need to be documented). This is probably best done by reference to another
2162document that covers inline asm from a holistic perspective.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2163</p>
2164
2165</div>
2166
2167<!-- *********************************************************************** -->
2168<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
2169<!-- *********************************************************************** -->
2170
2171<div class="doc_text">
2172
2173<p>The LLVM instruction set consists of several different
2174classifications of instructions: <a href="#terminators">terminator
2175instructions</a>, <a href="#binaryops">binary instructions</a>,
2176<a href="#bitwiseops">bitwise binary instructions</a>, <a
2177 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
2178instructions</a>.</p>
2179
2180</div>
2181
2182<!-- ======================================================================= -->
2183<div class="doc_subsection"> <a name="terminators">Terminator
2184Instructions</a> </div>
2185
2186<div class="doc_text">
2187
2188<p>As mentioned <a href="#functionstructure">previously</a>, every
2189basic block in a program ends with a "Terminator" instruction, which
2190indicates which block should be executed after the current block is
2191finished. These terminator instructions typically yield a '<tt>void</tt>'
2192value: they produce control flow, not values (the one exception being
2193the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
2194<p>There are six different terminator instructions: the '<a
2195 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
2196instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
2197the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
2198 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
2199 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
2200
2201</div>
2202
2203<!-- _______________________________________________________________________ -->
2204<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
2205Instruction</a> </div>
2206<div class="doc_text">
2207<h5>Syntax:</h5>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2208<pre>
2209 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]2210 ret void <i>; Return from void function</i>
2211</pre>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2212
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2213<h5>Overview:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2214
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2215<p>The '<tt>ret</tt>' instruction is used to return control flow (and
2216optionally a value) from a function back to the caller.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2217<p>There are two forms of the '<tt>ret</tt>' instruction: one that
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2218returns a value and then causes control flow, and one that just causes
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2219control flow to occur.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2220
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2221<h5>Arguments:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2222
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2223<p>The '<tt>ret</tt>' instruction optionally accepts a single argument,
2224the return value. The type of the return value must be a
2225'<a href="#t_firstclass">first class</a>' type.</p>
2226
2227<p>A function is not <a href="#wellformed">well formed</a> if
2228it it has a non-void return type and contains a '<tt>ret</tt>'
2229instruction with no return value or a return value with a type that
2230does not match its type, or if it has a void return type and contains
2231a '<tt>ret</tt>' instruction with a return value.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2232
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2233<h5>Semantics:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2234
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2235<p>When the '<tt>ret</tt>' instruction is executed, control flow
2236returns back to the calling function's context. If the caller is a "<a
2237 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
2238the instruction after the call. If the caller was an "<a
2239 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
2240at the beginning of the "normal" destination block. If the instruction
2241returns a value, that value shall set the call or invoke instruction's
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]2242return value.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2243
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2244<h5>Example:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2245
2246<pre>
2247 ret i32 5 <i>; Return an integer value of 5</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2248 ret void <i>; Return from a void function</i>
Bill Wendlingd163e2d2009-02-28 22:12:54 +0000[diff] [blame]2249 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]2250</pre>
Dan Gohman60967192009-01-12 23:12:39 +0000[diff] [blame]2251
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]2252<p>Note that the code generator does not yet fully support large
2253 return values. The specific sizes that are currently supported are
2254 dependent on the target. For integers, on 32-bit targets the limit
2255 is often 64 bits, and on 64-bit targets the limit is often 128 bits.
2256 For aggregate types, the current limits are dependent on the element
2257 types; for example targets are often limited to 2 total integer
2258 elements and 2 total floating-point elements.</p>
Dan Gohman60967192009-01-12 23:12:39 +0000[diff] [blame]2259
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2260</div>
2261<!-- _______________________________________________________________________ -->
2262<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
2263<div class="doc_text">
2264<h5>Syntax:</h5>
2265<pre> br i1 &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;<br> br label &lt;dest&gt; <i>; Unconditional branch</i>
2266</pre>
2267<h5>Overview:</h5>
2268<p>The '<tt>br</tt>' instruction is used to cause control flow to
2269transfer to a different basic block in the current function. There are
2270two forms of this instruction, corresponding to a conditional branch
2271and an unconditional branch.</p>
2272<h5>Arguments:</h5>
2273<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
2274single '<tt>i1</tt>' value and two '<tt>label</tt>' values. The
2275unconditional form of the '<tt>br</tt>' instruction takes a single
2276'<tt>label</tt>' value as a target.</p>
2277<h5>Semantics:</h5>
2278<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
2279argument is evaluated. If the value is <tt>true</tt>, control flows
2280to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
2281control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
2282<h5>Example:</h5>
Chris Lattner95127832009-05-09 18:11:50 +0000[diff] [blame]2283<pre>Test:<br> %cond = <a href="#i_icmp">icmp</a> eq i32 %a, %b<br> br i1 %cond, label %IfEqual, label %IfUnequal<br>IfEqual:<br> <a
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2284 href="#i_ret">ret</a> i32 1<br>IfUnequal:<br> <a href="#i_ret">ret</a> i32 0<br></pre>
2285</div>
2286<!-- _______________________________________________________________________ -->
2287<div class="doc_subsubsection">
2288 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
2289</div>
2290
2291<div class="doc_text">
2292<h5>Syntax:</h5>
2293
2294<pre>
2295 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
2296</pre>
2297
2298<h5>Overview:</h5>
2299
2300<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
2301several different places. It is a generalization of the '<tt>br</tt>'
2302instruction, allowing a branch to occur to one of many possible
2303destinations.</p>
2304
2305
2306<h5>Arguments:</h5>
2307
2308<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
2309comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
2310an array of pairs of comparison value constants and '<tt>label</tt>'s. The
2311table is not allowed to contain duplicate constant entries.</p>
2312
2313<h5>Semantics:</h5>
2314
2315<p>The <tt>switch</tt> instruction specifies a table of values and
2316destinations. When the '<tt>switch</tt>' instruction is executed, this
2317table is searched for the given value. If the value is found, control flow is
2318transfered to the corresponding destination; otherwise, control flow is
2319transfered to the default destination.</p>
2320
2321<h5>Implementation:</h5>
2322
2323<p>Depending on properties of the target machine and the particular
2324<tt>switch</tt> instruction, this instruction may be code generated in different
2325ways. For example, it could be generated as a series of chained conditional
2326branches or with a lookup table.</p>
2327
2328<h5>Example:</h5>
2329
2330<pre>
2331 <i>; Emulate a conditional br instruction</i>
2332 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]2333 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2334
2335 <i>; Emulate an unconditional br instruction</i>
2336 switch i32 0, label %dest [ ]
2337
2338 <i>; Implement a jump table:</i>
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]2339 switch i32 %val, label %otherwise [ i32 0, label %onzero
2340 i32 1, label %onone
2341 i32 2, label %ontwo ]
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2342</pre>
2343</div>
2344
2345<!-- _______________________________________________________________________ -->
2346<div class="doc_subsubsection">
2347 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
2348</div>
2349
2350<div class="doc_text">
2351
2352<h5>Syntax:</h5>
2353
2354<pre>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]2355 &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]2356 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
2357</pre>
2358
2359<h5>Overview:</h5>
2360
2361<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
2362function, with the possibility of control flow transfer to either the
2363'<tt>normal</tt>' label or the
2364'<tt>exception</tt>' label. If the callee function returns with the
2365"<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
2366"normal" label. If the callee (or any indirect callees) returns with the "<a
2367href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]2368continued at the dynamically nearest "exception" label.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2369
2370<h5>Arguments:</h5>
2371
2372<p>This instruction requires several arguments:</p>
2373
2374<ol>
2375 <li>
2376 The optional "cconv" marker indicates which <a href="#callingconv">calling
2377 convention</a> the call should use. If none is specified, the call defaults
2378 to using C calling conventions.
2379 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]2380
2381 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
2382 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>',
2383 and '<tt>inreg</tt>' attributes are valid here.</li>
2384
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2385 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
2386 function value being invoked. In most cases, this is a direct function
2387 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
2388 an arbitrary pointer to function value.
2389 </li>
2390
2391 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
2392 function to be invoked. </li>
2393
2394 <li>'<tt>function args</tt>': argument list whose types match the function
2395 signature argument types. If the function signature indicates the function
2396 accepts a variable number of arguments, the extra arguments can be
2397 specified. </li>
2398
2399 <li>'<tt>normal label</tt>': the label reached when the called function
2400 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
2401
2402 <li>'<tt>exception label</tt>': the label reached when a callee returns with
2403 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
2404
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]2405 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]2406 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
2407 '<tt>readnone</tt>' attributes are valid here.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2408</ol>
2409
2410<h5>Semantics:</h5>
2411
2412<p>This instruction is designed to operate as a standard '<tt><a
2413href="#i_call">call</a></tt>' instruction in most regards. The primary
2414difference is that it establishes an association with a label, which is used by
2415the runtime library to unwind the stack.</p>
2416
2417<p>This instruction is used in languages with destructors to ensure that proper
2418cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
2419exception. Additionally, this is important for implementation of
2420'<tt>catch</tt>' clauses in high-level languages that support them.</p>
2421
Jay Foad8e2fd2c2009-06-03 10:20:10 +0000[diff] [blame]2422<p>For the purposes of the SSA form, the definition of the value
2423returned by the '<tt>invoke</tt>' instruction is deemed to occur on
2424the edge from the current block to the "normal" label. If the callee
2425unwinds then no return value is available.</p>
Dan Gohman140ba5d2009-05-22 21:47:08 +0000[diff] [blame]2426
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2427<h5>Example:</h5>
2428<pre>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2429 %retval = invoke i32 @Test(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2430 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2431 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2432 unwind label %TestCleanup <i>; {i32}:retval set</i>
2433</pre>
2434</div>
2435
2436
2437<!-- _______________________________________________________________________ -->
2438
2439<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
2440Instruction</a> </div>
2441
2442<div class="doc_text">
2443
2444<h5>Syntax:</h5>
2445<pre>
2446 unwind
2447</pre>
2448
2449<h5>Overview:</h5>
2450
2451<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
2452at the first callee in the dynamic call stack which used an <a
2453href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
2454primarily used to implement exception handling.</p>
2455
2456<h5>Semantics:</h5>
2457
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]2458<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2459immediately halt. The dynamic call stack is then searched for the first <a
2460href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
2461execution continues at the "exceptional" destination block specified by the
2462<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
2463dynamic call chain, undefined behavior results.</p>
2464</div>
2465
2466<!-- _______________________________________________________________________ -->
2467
2468<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
2469Instruction</a> </div>
2470
2471<div class="doc_text">
2472
2473<h5>Syntax:</h5>
2474<pre>
2475 unreachable
2476</pre>
2477
2478<h5>Overview:</h5>
2479
2480<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
2481instruction is used to inform the optimizer that a particular portion of the
2482code is not reachable. This can be used to indicate that the code after a
2483no-return function cannot be reached, and other facts.</p>
2484
2485<h5>Semantics:</h5>
2486
2487<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
2488</div>
2489
2490
2491
2492<!-- ======================================================================= -->
2493<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
2494<div class="doc_text">
2495<p>Binary operators are used to do most of the computation in a
Chris Lattnerab596d92008-04-01 18:47:32 +0000[diff] [blame]2496program. They require two operands of the same type, execute an operation on them, and
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2497produce a single value. The operands might represent
2498multiple data, as is the case with the <a href="#t_vector">vector</a> data type.
Chris Lattnerab596d92008-04-01 18:47:32 +0000[diff] [blame]2499The result value has the same type as its operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2500<p>There are several different binary operators:</p>
2501</div>
2502<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2503<div class="doc_subsubsection">
2504 <a name="i_add">'<tt>add</tt>' Instruction</a>
2505</div>
2506
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2507<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2508
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2509<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2510
2511<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2512 &lt;result&gt; = add &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2513</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2514
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2515<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2516
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2517<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2518
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2519<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2520
2521<p>The two arguments to the '<tt>add</tt>' instruction must be <a
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2522 href="#t_integer">integer</a> or
2523 <a href="#t_vector">vector</a> of integer values. Both arguments must
2524 have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2525
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2526<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2527
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2528<p>The value produced is the integer sum of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2529
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2530<p>If the sum has unsigned overflow, the result returned is the
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2531mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2532the result.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2533
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2534<p>Because LLVM integers use a two's complement representation, this
2535instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2536
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2537<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2538
2539<pre>
2540 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2541</pre>
2542</div>
2543<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2544<div class="doc_subsubsection">
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2545 <a name="i_fadd">'<tt>fadd</tt>' Instruction</a>
2546</div>
2547
2548<div class="doc_text">
2549
2550<h5>Syntax:</h5>
2551
2552<pre>
2553 &lt;result&gt; = fadd &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
2554</pre>
2555
2556<h5>Overview:</h5>
2557
2558<p>The '<tt>fadd</tt>' instruction returns the sum of its two operands.</p>
2559
2560<h5>Arguments:</h5>
2561
2562<p>The two arguments to the '<tt>fadd</tt>' instruction must be
2563<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
2564floating point values. Both arguments must have identical types.</p>
2565
2566<h5>Semantics:</h5>
2567
2568<p>The value produced is the floating point sum of the two operands.</p>
2569
2570<h5>Example:</h5>
2571
2572<pre>
2573 &lt;result&gt; = fadd float 4.0, %var <i>; yields {float}:result = 4.0 + %var</i>
2574</pre>
2575</div>
2576<!-- _______________________________________________________________________ -->
2577<div class="doc_subsubsection">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2578 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
2579</div>
2580
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2581<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2582
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2583<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2584
2585<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2586 &lt;result&gt; = sub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2587</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2588
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2589<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2590
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2591<p>The '<tt>sub</tt>' instruction returns the difference of its two
2592operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2593
2594<p>Note that the '<tt>sub</tt>' instruction is used to represent the
2595'<tt>neg</tt>' instruction present in most other intermediate
2596representations.</p>
2597
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2598<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2599
2600<p>The two arguments to the '<tt>sub</tt>' instruction must be <a
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2601 href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
2602 integer values. Both arguments must have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2603
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2604<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2605
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2606<p>The value produced is the integer difference of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2607
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2608<p>If the difference has unsigned overflow, the result returned is the
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2609mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2610the result.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2611
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2612<p>Because LLVM integers use a two's complement representation, this
2613instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2614
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2615<h5>Example:</h5>
2616<pre>
2617 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
2618 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
2619</pre>
2620</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2621
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2622<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2623<div class="doc_subsubsection">
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2624 <a name="i_fsub">'<tt>fsub</tt>' Instruction</a>
2625</div>
2626
2627<div class="doc_text">
2628
2629<h5>Syntax:</h5>
2630
2631<pre>
2632 &lt;result&gt; = fsub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
2633</pre>
2634
2635<h5>Overview:</h5>
2636
2637<p>The '<tt>fsub</tt>' instruction returns the difference of its two
2638operands.</p>
2639
2640<p>Note that the '<tt>fsub</tt>' instruction is used to represent the
2641'<tt>fneg</tt>' instruction present in most other intermediate
2642representations.</p>
2643
2644<h5>Arguments:</h5>
2645
2646<p>The two arguments to the '<tt>fsub</tt>' instruction must be <a
2647 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2648 of floating point values. Both arguments must have identical types.</p>
2649
2650<h5>Semantics:</h5>
2651
2652<p>The value produced is the floating point difference of the two operands.</p>
2653
2654<h5>Example:</h5>
2655<pre>
2656 &lt;result&gt; = fsub float 4.0, %var <i>; yields {float}:result = 4.0 - %var</i>
2657 &lt;result&gt; = fsub float -0.0, %val <i>; yields {float}:result = -%var</i>
2658</pre>
2659</div>
2660
2661<!-- _______________________________________________________________________ -->
2662<div class="doc_subsubsection">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2663 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
2664</div>
2665
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2666<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2667
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2668<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2669<pre> &lt;result&gt; = mul &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2670</pre>
2671<h5>Overview:</h5>
2672<p>The '<tt>mul</tt>' instruction returns the product of its two
2673operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2674
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2675<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2676
2677<p>The two arguments to the '<tt>mul</tt>' instruction must be <a
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2678href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2679values. Both arguments must have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2680
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2681<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2682
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2683<p>The value produced is the integer product of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2684
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2685<p>If the result of the multiplication has unsigned overflow,
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2686the result returned is the mathematical result modulo
26872<sup>n</sup>, where n is the bit width of the result.</p>
2688<p>Because LLVM integers use a two's complement representation, and the
2689result is the same width as the operands, this instruction returns the
2690correct result for both signed and unsigned integers. If a full product
2691(e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands
2692should be sign-extended or zero-extended as appropriate to the
2693width of the full product.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2694<h5>Example:</h5>
2695<pre> &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
2696</pre>
2697</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2698
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2699<!-- _______________________________________________________________________ -->
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2700<div class="doc_subsubsection">
2701 <a name="i_fmul">'<tt>fmul</tt>' Instruction</a>
2702</div>
2703
2704<div class="doc_text">
2705
2706<h5>Syntax:</h5>
2707<pre> &lt;result&gt; = fmul &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
2708</pre>
2709<h5>Overview:</h5>
2710<p>The '<tt>fmul</tt>' instruction returns the product of its two
2711operands.</p>
2712
2713<h5>Arguments:</h5>
2714
2715<p>The two arguments to the '<tt>fmul</tt>' instruction must be
2716<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2717of floating point values. Both arguments must have identical types.</p>
2718
2719<h5>Semantics:</h5>
2720
2721<p>The value produced is the floating point product of the two operands.</p>
2722
2723<h5>Example:</h5>
2724<pre> &lt;result&gt; = fmul float 4.0, %var <i>; yields {float}:result = 4.0 * %var</i>
2725</pre>
2726</div>
2727
2728<!-- _______________________________________________________________________ -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2729<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
2730</a></div>
2731<div class="doc_text">
2732<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2733<pre> &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]2734</pre>
2735<h5>Overview:</h5>
2736<p>The '<tt>udiv</tt>' instruction returns the quotient of its two
2737operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2738
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2739<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2740
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2741<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2742<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2743values. Both arguments must have identical types.</p>
2744
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2745<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2746
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2747<p>The value produced is the unsigned integer quotient of the two operands.</p>
2748<p>Note that unsigned integer division and signed integer division are distinct
2749operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
2750<p>Division by zero leads to undefined behavior.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2751<h5>Example:</h5>
2752<pre> &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2753</pre>
2754</div>
2755<!-- _______________________________________________________________________ -->
2756<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
2757</a> </div>
2758<div class="doc_text">
2759<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2760<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2761 &lt;result&gt; = sdiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2762</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2763
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2764<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2765
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2766<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two
2767operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2768
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2769<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2770
2771<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
2772<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2773values. Both arguments must have identical types.</p>
2774
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2775<h5>Semantics:</h5>
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2776<p>The value produced is the signed integer quotient of the two operands rounded towards zero.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2777<p>Note that signed integer division and unsigned integer division are distinct
2778operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
2779<p>Division by zero leads to undefined behavior. Overflow also leads to
2780undefined behavior; this is a rare case, but can occur, for example,
2781by doing a 32-bit division of -2147483648 by -1.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2782<h5>Example:</h5>
2783<pre> &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2784</pre>
2785</div>
2786<!-- _______________________________________________________________________ -->
2787<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
2788Instruction</a> </div>
2789<div class="doc_text">
2790<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2791<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2792 &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]2793</pre>
2794<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2795
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2796<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two
2797operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2798
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2799<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2800
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2801<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2802<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2803of floating point values. Both arguments must have identical types.</p>
2804
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2805<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2806
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2807<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2808
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2809<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2810
2811<pre>
2812 &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]2813</pre>
2814</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2815
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2816<!-- _______________________________________________________________________ -->
2817<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
2818</div>
2819<div class="doc_text">
2820<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2821<pre> &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]2822</pre>
2823<h5>Overview:</h5>
2824<p>The '<tt>urem</tt>' instruction returns the remainder from the
2825unsigned division of its two arguments.</p>
2826<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2827<p>The two arguments to the '<tt>urem</tt>' instruction must be
2828<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2829values. Both arguments must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2830<h5>Semantics:</h5>
2831<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2832This instruction always performs an unsigned division to get the remainder.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2833<p>Note that unsigned integer remainder and signed integer remainder are
2834distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
2835<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2836<h5>Example:</h5>
2837<pre> &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2838</pre>
2839
2840</div>
2841<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2842<div class="doc_subsubsection">
2843 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
2844</div>
2845
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2846<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2847
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2848<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2849
2850<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2851 &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]2852</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2853
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2854<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2855
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2856<p>The '<tt>srem</tt>' instruction returns the remainder from the
Dan Gohman3e3fd8c2007-11-05 23:35:22 +0000[diff] [blame]2857signed division of its two operands. This instruction can also take
2858<a href="#t_vector">vector</a> versions of the values in which case
2859the elements must be integers.</p>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]2860
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2861<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2862
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2863<p>The two arguments to the '<tt>srem</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2864<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2865values. Both arguments must have identical types.</p>
2866
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2867<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2868
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2869<p>This instruction returns the <i>remainder</i> of a division (where the result
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2870has the same sign as the dividend, <tt>op1</tt>), not the <i>modulo</i>
2871operator (where the result has the same sign as the divisor, <tt>op2</tt>) of
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2872a value. For more information about the difference, see <a
2873 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
2874Math Forum</a>. For a table of how this is implemented in various languages,
2875please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
2876Wikipedia: modulo operation</a>.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2877<p>Note that signed integer remainder and unsigned integer remainder are
2878distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
2879<p>Taking the remainder of a division by zero leads to undefined behavior.
2880Overflow also leads to undefined behavior; this is a rare case, but can occur,
2881for example, by taking the remainder of a 32-bit division of -2147483648 by -1.
2882(The remainder doesn't actually overflow, but this rule lets srem be
2883implemented using instructions that return both the result of the division
2884and the remainder.)</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2885<h5>Example:</h5>
2886<pre> &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2887</pre>
2888
2889</div>
2890<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2891<div class="doc_subsubsection">
2892 <a name="i_frem">'<tt>frem</tt>' Instruction</a> </div>
2893
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2894<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2895
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2896<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2897<pre> &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]2898</pre>
2899<h5>Overview:</h5>
2900<p>The '<tt>frem</tt>' instruction returns the remainder from the
2901division of its two operands.</p>
2902<h5>Arguments:</h5>
2903<p>The two arguments to the '<tt>frem</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2904<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2905of floating point values. Both arguments must have identical types.</p>
2906
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2907<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2908
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2909<p>This instruction returns the <i>remainder</i> of a division.
2910The remainder has the same sign as the dividend.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2911
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2912<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2913
2914<pre>
2915 &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]2916</pre>
2917</div>
2918
2919<!-- ======================================================================= -->
2920<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
2921Operations</a> </div>
2922<div class="doc_text">
2923<p>Bitwise binary operators are used to do various forms of
2924bit-twiddling in a program. They are generally very efficient
2925instructions and can commonly be strength reduced from other
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2926instructions. They require two operands of the same type, execute an operation on them,
2927and produce a single value. The resulting value is the same type as its operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2928</div>
2929
2930<!-- _______________________________________________________________________ -->
2931<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
2932Instruction</a> </div>
2933<div class="doc_text">
2934<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2935<pre> &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]2936</pre>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2937
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2938<h5>Overview:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2939
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2940<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
2941the left a specified number of bits.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2942
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2943<h5>Arguments:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2944
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2945<p>Both arguments to the '<tt>shl</tt>' instruction must be the same <a
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]2946 href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2947type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2948
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2949<h5>Semantics:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2950
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2951<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod 2<sup>n</sup>,
2952where n is the width of the result. If <tt>op2</tt> is (statically or dynamically) negative or
Mon P Wangb0f51822008-12-10 08:55:09 +0000[diff] [blame]2953equal to or larger than the number of bits in <tt>op1</tt>, the result is undefined.
2954If the arguments are vectors, each vector element of <tt>op1</tt> is shifted by the
2955corresponding shift amount in <tt>op2</tt>.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2956
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2957<h5>Example:</h5><pre>
2958 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
2959 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
2960 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2961 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Mon P Wang9901e732008-12-09 05:46:39 +0000[diff] [blame]2962 &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]2963</pre>
2964</div>
2965<!-- _______________________________________________________________________ -->
2966<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
2967Instruction</a> </div>
2968<div class="doc_text">
2969<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2970<pre> &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]2971</pre>
2972
2973<h5>Overview:</h5>
2974<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
2975operand shifted to the right a specified number of bits with zero fill.</p>
2976
2977<h5>Arguments:</h5>
2978<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]2979<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2980type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2981
2982<h5>Semantics:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2983
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2984<p>This instruction always performs a logical shift right operation. The most
2985significant bits of the result will be filled with zero bits after the
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2986shift. If <tt>op2</tt> is (statically or dynamically) equal to or larger than
Mon P Wangb0f51822008-12-10 08:55:09 +0000[diff] [blame]2987the number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
2988vectors, each vector element of <tt>op1</tt> is shifted by the corresponding shift
2989amount in <tt>op2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2990
2991<h5>Example:</h5>
2992<pre>
2993 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
2994 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
2995 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
2996 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2997 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Mon P Wang9901e732008-12-09 05:46:39 +0000[diff] [blame]2998 &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]2999</pre>
3000</div>
3001
3002<!-- _______________________________________________________________________ -->
3003<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
3004Instruction</a> </div>
3005<div class="doc_text">
3006
3007<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3008<pre> &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]3009</pre>
3010
3011<h5>Overview:</h5>
3012<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
3013operand shifted to the right a specified number of bits with sign extension.</p>
3014
3015<h5>Arguments:</h5>
3016<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]3017<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3018type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3019
3020<h5>Semantics:</h5>
3021<p>This instruction always performs an arithmetic shift right operation,
3022The most significant bits of the result will be filled with the sign bit
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3023of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
Mon P Wangb0f51822008-12-10 08:55:09 +0000[diff] [blame]3024larger than the number of bits in <tt>op1</tt>, the result is undefined. If the
3025arguments are vectors, each vector element of <tt>op1</tt> is shifted by the
3026corresponding shift amount in <tt>op2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3027
3028<h5>Example:</h5>
3029<pre>
3030 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
3031 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
3032 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
3033 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]3034 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Mon P Wang9901e732008-12-09 05:46:39 +0000[diff] [blame]3035 &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]3036</pre>
3037</div>
3038
3039<!-- _______________________________________________________________________ -->
3040<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
3041Instruction</a> </div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3042
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3043<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3044
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3045<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3046
3047<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3048 &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]3049</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3050
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3051<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3052
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3053<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
3054its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3055
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3056<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3057
3058<p>The two arguments to the '<tt>and</tt>' instruction must be
3059<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3060values. Both arguments must have identical types.</p>
3061
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3062<h5>Semantics:</h5>
3063<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
3064<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]3065<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3066<table border="1" cellspacing="0" cellpadding="4">
3067 <tbody>
3068 <tr>
3069 <td>In0</td>
3070 <td>In1</td>
3071 <td>Out</td>
3072 </tr>
3073 <tr>
3074 <td>0</td>
3075 <td>0</td>
3076 <td>0</td>
3077 </tr>
3078 <tr>
3079 <td>0</td>
3080 <td>1</td>
3081 <td>0</td>
3082 </tr>
3083 <tr>
3084 <td>1</td>
3085 <td>0</td>
3086 <td>0</td>
3087 </tr>
3088 <tr>
3089 <td>1</td>
3090 <td>1</td>
3091 <td>1</td>
3092 </tr>
3093 </tbody>
3094</table>
3095</div>
3096<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3097<pre>
3098 &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]3099 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
3100 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
3101</pre>
3102</div>
3103<!-- _______________________________________________________________________ -->
3104<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
3105<div class="doc_text">
3106<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3107<pre> &lt;result&gt; = or &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3108</pre>
3109<h5>Overview:</h5>
3110<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
3111or of its two operands.</p>
3112<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3113
3114<p>The two arguments to the '<tt>or</tt>' instruction must be
3115<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3116values. Both arguments must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3117<h5>Semantics:</h5>
3118<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
3119<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]3120<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3121<table border="1" cellspacing="0" cellpadding="4">
3122 <tbody>
3123 <tr>
3124 <td>In0</td>
3125 <td>In1</td>
3126 <td>Out</td>
3127 </tr>
3128 <tr>
3129 <td>0</td>
3130 <td>0</td>
3131 <td>0</td>
3132 </tr>
3133 <tr>
3134 <td>0</td>
3135 <td>1</td>
3136 <td>1</td>
3137 </tr>
3138 <tr>
3139 <td>1</td>
3140 <td>0</td>
3141 <td>1</td>
3142 </tr>
3143 <tr>
3144 <td>1</td>
3145 <td>1</td>
3146 <td>1</td>
3147 </tr>
3148 </tbody>
3149</table>
3150</div>
3151<h5>Example:</h5>
3152<pre> &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
3153 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
3154 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
3155</pre>
3156</div>
3157<!-- _______________________________________________________________________ -->
3158<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
3159Instruction</a> </div>
3160<div class="doc_text">
3161<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3162<pre> &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]3163</pre>
3164<h5>Overview:</h5>
3165<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
3166or of its two operands. The <tt>xor</tt> is used to implement the
3167"one's complement" operation, which is the "~" operator in C.</p>
3168<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3169<p>The two arguments to the '<tt>xor</tt>' instruction must be
3170<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3171values. Both arguments must have identical types.</p>
3172
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3173<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3174
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3175<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
3176<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]3177<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3178<table border="1" cellspacing="0" cellpadding="4">
3179 <tbody>
3180 <tr>
3181 <td>In0</td>
3182 <td>In1</td>
3183 <td>Out</td>
3184 </tr>
3185 <tr>
3186 <td>0</td>
3187 <td>0</td>
3188 <td>0</td>
3189 </tr>
3190 <tr>
3191 <td>0</td>
3192 <td>1</td>
3193 <td>1</td>
3194 </tr>
3195 <tr>
3196 <td>1</td>
3197 <td>0</td>
3198 <td>1</td>
3199 </tr>
3200 <tr>
3201 <td>1</td>
3202 <td>1</td>
3203 <td>0</td>
3204 </tr>
3205 </tbody>
3206</table>
3207</div>
3208<p> </p>
3209<h5>Example:</h5>
3210<pre> &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
3211 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
3212 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
3213 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
3214</pre>
3215</div>
3216
3217<!-- ======================================================================= -->
3218<div class="doc_subsection">
3219 <a name="vectorops">Vector Operations</a>
3220</div>
3221
3222<div class="doc_text">
3223
3224<p>LLVM supports several instructions to represent vector operations in a
3225target-independent manner. These instructions cover the element-access and
3226vector-specific operations needed to process vectors effectively. While LLVM
3227does directly support these vector operations, many sophisticated algorithms
3228will want to use target-specific intrinsics to take full advantage of a specific
3229target.</p>
3230
3231</div>
3232
3233<!-- _______________________________________________________________________ -->
3234<div class="doc_subsubsection">
3235 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
3236</div>
3237
3238<div class="doc_text">
3239
3240<h5>Syntax:</h5>
3241
3242<pre>
3243 &lt;result&gt; = extractelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, i32 &lt;idx&gt; <i>; yields &lt;ty&gt;</i>
3244</pre>
3245
3246<h5>Overview:</h5>
3247
3248<p>
3249The '<tt>extractelement</tt>' instruction extracts a single scalar
3250element from a vector at a specified index.
3251</p>
3252
3253
3254<h5>Arguments:</h5>
3255
3256<p>
3257The first operand of an '<tt>extractelement</tt>' instruction is a
3258value of <a href="#t_vector">vector</a> type. The second operand is
3259an index indicating the position from which to extract the element.
3260The index may be a variable.</p>
3261
3262<h5>Semantics:</h5>
3263
3264<p>
3265The result is a scalar of the same type as the element type of
3266<tt>val</tt>. Its value is the value at position <tt>idx</tt> of
3267<tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
3268results are undefined.
3269</p>
3270
3271<h5>Example:</h5>
3272
3273<pre>
3274 %result = extractelement &lt;4 x i32&gt; %vec, i32 0 <i>; yields i32</i>
3275</pre>
3276</div>
3277
3278
3279<!-- _______________________________________________________________________ -->
3280<div class="doc_subsubsection">
3281 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
3282</div>
3283
3284<div class="doc_text">
3285
3286<h5>Syntax:</h5>
3287
3288<pre>
Dan Gohmanbcc3c502008-05-12 23:38:42 +0000[diff] [blame]3289 &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]3290</pre>
3291
3292<h5>Overview:</h5>
3293
3294<p>
3295The '<tt>insertelement</tt>' instruction inserts a scalar
3296element into a vector at a specified index.
3297</p>
3298
3299
3300<h5>Arguments:</h5>
3301
3302<p>
3303The first operand of an '<tt>insertelement</tt>' instruction is a
3304value of <a href="#t_vector">vector</a> type. The second operand is a
3305scalar value whose type must equal the element type of the first
3306operand. The third operand is an index indicating the position at
3307which to insert the value. The index may be a variable.</p>
3308
3309<h5>Semantics:</h5>
3310
3311<p>
3312The result is a vector of the same type as <tt>val</tt>. Its
3313element values are those of <tt>val</tt> except at position
3314<tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
3315exceeds the length of <tt>val</tt>, the results are undefined.
3316</p>
3317
3318<h5>Example:</h5>
3319
3320<pre>
3321 %result = insertelement &lt;4 x i32&gt; %vec, i32 1, i32 0 <i>; yields &lt;4 x i32&gt;</i>
3322</pre>
3323</div>
3324
3325<!-- _______________________________________________________________________ -->
3326<div class="doc_subsubsection">
3327 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
3328</div>
3329
3330<div class="doc_text">
3331
3332<h5>Syntax:</h5>
3333
3334<pre>
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3335 &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]3336</pre>
3337
3338<h5>Overview:</h5>
3339
3340<p>
3341The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3342from two input vectors, returning a vector with the same element type as
3343the input and length that is the same as the shuffle mask.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3344</p>
3345
3346<h5>Arguments:</h5>
3347
3348<p>
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3349The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
3350with types that match each other. The third argument is a shuffle mask whose
3351element type is always 'i32'. The result of the instruction is a vector whose
3352length is the same as the shuffle mask and whose element type is the same as
3353the element type of the first two operands.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3354</p>
3355
3356<p>
3357The shuffle mask operand is required to be a constant vector with either
3358constant integer or undef values.
3359</p>
3360
3361<h5>Semantics:</h5>
3362
3363<p>
3364The elements of the two input vectors are numbered from left to right across
3365both of the vectors. The shuffle mask operand specifies, for each element of
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3366the result vector, which element of the two input vectors the result element
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3367gets. The element selector may be undef (meaning "don't care") and the second
3368operand may be undef if performing a shuffle from only one vector.
3369</p>
3370
3371<h5>Example:</h5>
3372
3373<pre>
3374 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
3375 &lt;4 x i32&gt; &lt;i32 0, i32 4, i32 1, i32 5&gt; <i>; yields &lt;4 x i32&gt;</i>
3376 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; undef,
3377 &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.
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3378 %result = shufflevector &lt;8 x i32&gt; %v1, &lt;8 x i32&gt; undef,
3379 &lt;4 x i32&gt; &lt;i32 0, i32 1, i32 2, i32 3&gt; <i>; yields &lt;4 x i32&gt;</i>
3380 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
3381 &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]3382</pre>
3383</div>
3384
3385
3386<!-- ======================================================================= -->
3387<div class="doc_subsection">
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3388 <a name="aggregateops">Aggregate Operations</a>
3389</div>
3390
3391<div class="doc_text">
3392
3393<p>LLVM supports several instructions for working with aggregate values.
3394</p>
3395
3396</div>
3397
3398<!-- _______________________________________________________________________ -->
3399<div class="doc_subsubsection">
3400 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
3401</div>
3402
3403<div class="doc_text">
3404
3405<h5>Syntax:</h5>
3406
3407<pre>
3408 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
3409</pre>
3410
3411<h5>Overview:</h5>
3412
3413<p>
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3414The '<tt>extractvalue</tt>' instruction extracts the value of a struct field
3415or array element from an aggregate value.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3416</p>
3417
3418
3419<h5>Arguments:</h5>
3420
3421<p>
3422The first operand of an '<tt>extractvalue</tt>' instruction is a
3423value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a>
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3424type. The operands are constant indices to specify which value to extract
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3425in a similar manner as indices in a
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3426'<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
3427</p>
3428
3429<h5>Semantics:</h5>
3430
3431<p>
3432The result is the value at the position in the aggregate specified by
3433the index operands.
3434</p>
3435
3436<h5>Example:</h5>
3437
3438<pre>
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3439 %result = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3440</pre>
3441</div>
3442
3443
3444<!-- _______________________________________________________________________ -->
3445<div class="doc_subsubsection">
3446 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
3447</div>
3448
3449<div class="doc_text">
3450
3451<h5>Syntax:</h5>
3452
3453<pre>
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3454 &lt;result&gt; = insertvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;ty&gt; &lt;val&gt;, &lt;idx&gt; <i>; yields &lt;n x &lt;ty&gt;&gt;</i>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3455</pre>
3456
3457<h5>Overview:</h5>
3458
3459<p>
3460The '<tt>insertvalue</tt>' instruction inserts a value
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3461into a struct field or array element in an aggregate.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3462</p>
3463
3464
3465<h5>Arguments:</h5>
3466
3467<p>
3468The first operand of an '<tt>insertvalue</tt>' instruction is a
3469value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a> type.
3470The second operand is a first-class value to insert.
Dan Gohman4f29e422008-05-23 21:53:15 +0000[diff] [blame]3471The following operands are constant indices
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3472indicating the position at which to insert the value in a similar manner as
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3473indices in a
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3474'<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
3475The value to insert must have the same type as the value identified
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3476by the indices.
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3477</p>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3478
3479<h5>Semantics:</h5>
3480
3481<p>
3482The result is an aggregate of the same type as <tt>val</tt>. Its
3483value is that of <tt>val</tt> except that the value at the position
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3484specified by the indices is that of <tt>elt</tt>.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3485</p>
3486
3487<h5>Example:</h5>
3488
3489<pre>
Dan Gohmanb1aab4e2008-06-23 15:26:37 +0000[diff] [blame]3490 %result = insertvalue {i32, float} %agg, i32 1, 0 <i>; yields {i32, float}</i>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3491</pre>
3492</div>
3493
3494
3495<!-- ======================================================================= -->
3496<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3497 <a name="memoryops">Memory Access and Addressing Operations</a>
3498</div>
3499
3500<div class="doc_text">
3501
3502<p>A key design point of an SSA-based representation is how it
3503represents memory. In LLVM, no memory locations are in SSA form, which
3504makes things very simple. This section describes how to read, write,
3505allocate, and free memory in LLVM.</p>
3506
3507</div>
3508
3509<!-- _______________________________________________________________________ -->
3510<div class="doc_subsubsection">
3511 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
3512</div>
3513
3514<div class="doc_text">
3515
3516<h5>Syntax:</h5>
3517
3518<pre>
3519 &lt;result&gt; = malloc &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
3520</pre>
3521
3522<h5>Overview:</h5>
3523
3524<p>The '<tt>malloc</tt>' instruction allocates memory from the system
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]3525heap and returns a pointer to it. The object is always allocated in the generic
3526address space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3527
3528<h5>Arguments:</h5>
3529
3530<p>The '<tt>malloc</tt>' instruction allocates
3531<tt>sizeof(&lt;type&gt;)*NumElements</tt>
3532bytes of memory from the operating system and returns a pointer of the
3533appropriate type to the program. If "NumElements" is specified, it is the
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3534number of elements allocated, otherwise "NumElements" is defaulted to be one.
Duncan Sandsb38ce6f2009-06-20 13:26:06 +0000[diff] [blame^]3535If a constant alignment is specified, the value result of the allocation is
3536guaranteed to be aligned to at least that boundary. If not specified, or if
3537zero, the target can choose to align the allocation on any convenient boundary
3538compatible with the type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3539
3540<p>'<tt>type</tt>' must be a sized type.</p>
3541
3542<h5>Semantics:</h5>
3543
3544<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
Nick Lewyckyff384472008-11-24 03:41:24 +0000[diff] [blame]3545a pointer is returned. The result of a zero byte allocation is undefined. The
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]3546result is null if there is insufficient memory available.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3547
3548<h5>Example:</h5>
3549
3550<pre>
Dan Gohmanf54f50a2009-01-04 23:49:44 +0000[diff] [blame]3551 %array = malloc [4 x i8] <i>; yields {[%4 x i8]*}:array</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3552
3553 %size = <a href="#i_add">add</a> i32 2, 2 <i>; yields {i32}:size = i32 4</i>
3554 %array1 = malloc i8, i32 4 <i>; yields {i8*}:array1</i>
3555 %array2 = malloc [12 x i8], i32 %size <i>; yields {[12 x i8]*}:array2</i>
3556 %array3 = malloc i32, i32 4, align 1024 <i>; yields {i32*}:array3</i>
3557 %array4 = malloc i32, align 1024 <i>; yields {i32*}:array4</i>
3558</pre>
Dan Gohman60967192009-01-12 23:12:39 +0000[diff] [blame]3559
3560<p>Note that the code generator does not yet respect the
3561 alignment value.</p>
3562
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3563</div>
3564
3565<!-- _______________________________________________________________________ -->
3566<div class="doc_subsubsection">
3567 <a name="i_free">'<tt>free</tt>' Instruction</a>
3568</div>
3569
3570<div class="doc_text">
3571
3572<h5>Syntax:</h5>
3573
3574<pre>
Dan Gohmanf54f50a2009-01-04 23:49:44 +0000[diff] [blame]3575 free &lt;type&gt; &lt;value&gt; <i>; yields {void}</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3576</pre>
3577
3578<h5>Overview:</h5>
3579
3580<p>The '<tt>free</tt>' instruction returns memory back to the unused
3581memory heap to be reallocated in the future.</p>
3582
3583<h5>Arguments:</h5>
3584
3585<p>'<tt>value</tt>' shall be a pointer value that points to a value
3586that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
3587instruction.</p>
3588
3589<h5>Semantics:</h5>
3590
3591<p>Access to the memory pointed to by the pointer is no longer defined
Chris Lattner329d6532008-04-19 22:41:32 +0000[diff] [blame]3592after this instruction executes. If the pointer is null, the operation
3593is a noop.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3594
3595<h5>Example:</h5>
3596
3597<pre>
Dan Gohmanf54f50a2009-01-04 23:49:44 +0000[diff] [blame]3598 %array = <a href="#i_malloc">malloc</a> [4 x i8] <i>; yields {[4 x i8]*}:array</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3599 free [4 x i8]* %array
3600</pre>
3601</div>
3602
3603<!-- _______________________________________________________________________ -->
3604<div class="doc_subsubsection">
3605 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
3606</div>
3607
3608<div class="doc_text">
3609
3610<h5>Syntax:</h5>
3611
3612<pre>
3613 &lt;result&gt; = alloca &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
3614</pre>
3615
3616<h5>Overview:</h5>
3617
3618<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
3619currently executing function, to be automatically released when this function
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]3620returns to its caller. The object is always allocated in the generic address
3621space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3622
3623<h5>Arguments:</h5>
3624
3625<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
3626bytes of memory on the runtime stack, returning a pointer of the
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3627appropriate type to the program. If "NumElements" is specified, it is the
3628number of elements allocated, otherwise "NumElements" is defaulted to be one.
Duncan Sandsb38ce6f2009-06-20 13:26:06 +0000[diff] [blame^]3629If a constant alignment is specified, the value result of the allocation is
3630guaranteed to be aligned to at least that boundary. If not specified, or if
3631zero, the target can choose to align the allocation on any convenient boundary
3632compatible with the type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3633
3634<p>'<tt>type</tt>' may be any sized type.</p>
3635
3636<h5>Semantics:</h5>
3637
Bill Wendling2a454572009-05-08 20:49:29 +0000[diff] [blame]3638<p>Memory is allocated; a pointer is returned. The operation is undefined if
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]3639there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3640memory is automatically released when the function returns. The '<tt>alloca</tt>'
3641instruction is commonly used to represent automatic variables that must
3642have an address available. When the function returns (either with the <tt><a
3643 href="#i_ret">ret</a></tt> or <tt><a href="#i_unwind">unwind</a></tt>
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3644instructions), the memory is reclaimed. Allocating zero bytes
3645is legal, but the result is undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3646
3647<h5>Example:</h5>
3648
3649<pre>
Dan Gohmanf54f50a2009-01-04 23:49:44 +0000[diff] [blame]3650 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
3651 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
3652 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
3653 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3654</pre>
3655</div>
3656
3657<!-- _______________________________________________________________________ -->
3658<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
3659Instruction</a> </div>
3660<div class="doc_text">
3661<h5>Syntax:</h5>
3662<pre> &lt;result&gt; = load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;]<br> &lt;result&gt; = volatile load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;]<br></pre>
3663<h5>Overview:</h5>
3664<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
3665<h5>Arguments:</h5>
3666<p>The argument to the '<tt>load</tt>' instruction specifies the memory
3667address from which to load. The pointer must point to a <a
3668 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
3669marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
3670the number or order of execution of this <tt>load</tt> with other
3671volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
3672instructions. </p>
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3673<p>
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3674The optional constant "align" argument specifies the alignment of the operation
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3675(that is, the alignment of the memory address). A value of 0 or an
3676omitted "align" argument means that the operation has the preferential
3677alignment for the target. It is the responsibility of the code emitter
3678to ensure that the alignment information is correct. Overestimating
3679the alignment results in an undefined behavior. Underestimating the
3680alignment may produce less efficient code. An alignment of 1 is always
3681safe.
3682</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3683<h5>Semantics:</h5>
Duncan Sandsfe87fff2009-03-22 11:33:16 +0000[diff] [blame]3684<p>The location of memory pointed to is loaded. If the value being loaded
3685is of scalar type then the number of bytes read does not exceed the minimum
3686number of bytes needed to hold all bits of the type. For example, loading an
3687<tt>i24</tt> reads at most three bytes. When loading a value of a type like
3688<tt>i20</tt> with a size that is not an integral number of bytes, the result
3689is undefined if the value was not originally written using a store of the
3690same type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3691<h5>Examples:</h5>
3692<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
3693 <a
3694 href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
3695 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
3696</pre>
3697</div>
3698<!-- _______________________________________________________________________ -->
3699<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
3700Instruction</a> </div>
3701<div class="doc_text">
3702<h5>Syntax:</h5>
3703<pre> store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
3704 volatile store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
3705</pre>
3706<h5>Overview:</h5>
3707<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
3708<h5>Arguments:</h5>
3709<p>There are two arguments to the '<tt>store</tt>' instruction: a value
3710to store and an address at which to store it. The type of the '<tt>&lt;pointer&gt;</tt>'
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3711operand must be a pointer to the <a href="#t_firstclass">first class</a> type
3712of the '<tt>&lt;value&gt;</tt>'
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3713operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
3714optimizer is not allowed to modify the number or order of execution of
3715this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
3716 href="#i_store">store</a></tt> instructions.</p>
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3717<p>
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3718The optional constant "align" argument specifies the alignment of the operation
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3719(that is, the alignment of the memory address). A value of 0 or an
3720omitted "align" argument means that the operation has the preferential
3721alignment for the target. It is the responsibility of the code emitter
3722to ensure that the alignment information is correct. Overestimating
3723the alignment results in an undefined behavior. Underestimating the
3724alignment may produce less efficient code. An alignment of 1 is always
3725safe.
3726</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3727<h5>Semantics:</h5>
3728<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>'
Duncan Sandsfe87fff2009-03-22 11:33:16 +0000[diff] [blame]3729at the location specified by the '<tt>&lt;pointer&gt;</tt>' operand.
3730If '<tt>&lt;value&gt;</tt>' is of scalar type then the number of bytes
3731written does not exceed the minimum number of bytes needed to hold all
3732bits of the type. For example, storing an <tt>i24</tt> writes at most
3733three bytes. When writing a value of a type like <tt>i20</tt> with a
3734size that is not an integral number of bytes, it is unspecified what
3735happens to the extra bits that do not belong to the type, but they will
3736typically be overwritten.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3737<h5>Example:</h5>
3738<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling63ffa142007-10-22 05:10:05 +0000[diff] [blame]3739 store i32 3, i32* %ptr <i>; yields {void}</i>
3740 %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]3741</pre>
3742</div>
3743
3744<!-- _______________________________________________________________________ -->
3745<div class="doc_subsubsection">
3746 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
3747</div>
3748
3749<div class="doc_text">
3750<h5>Syntax:</h5>
3751<pre>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3752 &lt;result&gt; = getelementptr &lt;pty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3753</pre>
3754
3755<h5>Overview:</h5>
3756
3757<p>
3758The '<tt>getelementptr</tt>' instruction is used to get the address of a
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3759subelement of an aggregate data structure. It performs address calculation only
3760and does not access memory.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3761
3762<h5>Arguments:</h5>
3763
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3764<p>The first argument is always a pointer, and forms the basis of the
3765calculation. The remaining arguments are indices, that indicate which of the
3766elements of the aggregate object are indexed. The interpretation of each index
3767is dependent on the type being indexed into. The first index always indexes the
3768pointer value given as the first argument, the second index indexes a value of
3769the type pointed to (not necessarily the value directly pointed to, since the
3770first index can be non-zero), etc. The first type indexed into must be a pointer
3771value, subsequent types can be arrays, vectors and structs. Note that subsequent
3772types being indexed into can never be pointers, since that would require loading
3773the pointer before continuing calculation.</p>
3774
3775<p>The type of each index argument depends on the type it is indexing into.
3776When indexing into a (packed) structure, only <tt>i32</tt> integer
3777<b>constants</b> are allowed. When indexing into an array, pointer or vector,
Sanjiv Guptaa2e72d92009-04-27 03:21:00 +0000[diff] [blame]3778integers of any width are allowed (also non-constants).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3779
3780<p>For example, let's consider a C code fragment and how it gets
3781compiled to LLVM:</p>
3782
3783<div class="doc_code">
3784<pre>
3785struct RT {
3786 char A;
3787 int B[10][20];
3788 char C;
3789};
3790struct ST {
3791 int X;
3792 double Y;
3793 struct RT Z;
3794};
3795
3796int *foo(struct ST *s) {
3797 return &amp;s[1].Z.B[5][13];
3798}
3799</pre>
3800</div>
3801
3802<p>The LLVM code generated by the GCC frontend is:</p>
3803
3804<div class="doc_code">
3805<pre>
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]3806%RT = <a href="#namedtypes">type</a> { i8 , [10 x [20 x i32]], i8 }
3807%ST = <a href="#namedtypes">type</a> { i32, double, %RT }
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3808
3809define i32* %foo(%ST* %s) {
3810entry:
3811 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
3812 ret i32* %reg
3813}
3814</pre>
3815</div>
3816
3817<h5>Semantics:</h5>
3818
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3819<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
3820type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
3821}</tt>' type, a structure. The second index indexes into the third element of
3822the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
3823i8 }</tt>' type, another structure. The third index indexes into the second
3824element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
3825array. The two dimensions of the array are subscripted into, yielding an
3826'<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
3827to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
3828
3829<p>Note that it is perfectly legal to index partially through a
3830structure, returning a pointer to an inner element. Because of this,
3831the LLVM code for the given testcase is equivalent to:</p>
3832
3833<pre>
3834 define i32* %foo(%ST* %s) {
3835 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
3836 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
3837 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
3838 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
3839 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
3840 ret i32* %t5
3841 }
3842</pre>
3843
Chris Lattner50609942009-03-09 20:55:18 +0000[diff] [blame]3844<p>Note that it is undefined to access an array out of bounds: array
3845and pointer indexes must always be within the defined bounds of the
3846array type when accessed with an instruction that dereferences the
3847pointer (e.g. a load or store instruction). The one exception for
3848this rule is zero length arrays. These arrays are defined to be
3849accessible as variable length arrays, which requires access beyond the
3850zero'th element.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3851
3852<p>The getelementptr instruction is often confusing. For some more insight
3853into how it works, see <a href="GetElementPtr.html">the getelementptr
3854FAQ</a>.</p>
3855
3856<h5>Example:</h5>
3857
3858<pre>
3859 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3860 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
3861 <i>; yields i8*:vptr</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3862 %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]3863 <i>; yields i8*:eptr</i>
3864 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Sanjiv Gupta4f9a0dc2009-04-25 07:27:44 +0000[diff] [blame]3865 <i>; yields i32*:iptr</i>
Sanjiv Gupta1e46c582009-04-24 16:38:13 +0000[diff] [blame]3866 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3867</pre>
3868</div>
3869
3870<!-- ======================================================================= -->
3871<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
3872</div>
3873<div class="doc_text">
3874<p>The instructions in this category are the conversion instructions (casting)
3875which all take a single operand and a type. They perform various bit conversions
3876on the operand.</p>
3877</div>
3878
3879<!-- _______________________________________________________________________ -->
3880<div class="doc_subsubsection">
3881 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
3882</div>
3883<div class="doc_text">
3884
3885<h5>Syntax:</h5>
3886<pre>
3887 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3888</pre>
3889
3890<h5>Overview:</h5>
3891<p>
3892The '<tt>trunc</tt>' instruction truncates its operand to the type <tt>ty2</tt>.
3893</p>
3894
3895<h5>Arguments:</h5>
3896<p>
3897The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
3898be an <a href="#t_integer">integer</a> type, and a type that specifies the size
3899and type of the result, which must be an <a href="#t_integer">integer</a>
3900type. The bit size of <tt>value</tt> must be larger than the bit size of
3901<tt>ty2</tt>. Equal sized types are not allowed.</p>
3902
3903<h5>Semantics:</h5>
3904<p>
3905The '<tt>trunc</tt>' instruction truncates the high order bits in <tt>value</tt>
3906and converts the remaining bits to <tt>ty2</tt>. Since the source size must be
3907larger than the destination size, <tt>trunc</tt> cannot be a <i>no-op cast</i>.
3908It will always truncate bits.</p>
3909
3910<h5>Example:</h5>
3911<pre>
3912 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
3913 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
3914 %Y = trunc i32 122 to i1 <i>; yields i1:false</i>
3915</pre>
3916</div>
3917
3918<!-- _______________________________________________________________________ -->
3919<div class="doc_subsubsection">
3920 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
3921</div>
3922<div class="doc_text">
3923
3924<h5>Syntax:</h5>
3925<pre>
3926 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3927</pre>
3928
3929<h5>Overview:</h5>
3930<p>The '<tt>zext</tt>' instruction zero extends its operand to type
3931<tt>ty2</tt>.</p>
3932
3933
3934<h5>Arguments:</h5>
3935<p>The '<tt>zext</tt>' instruction takes a value to cast, which must be of
3936<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3937also be of <a href="#t_integer">integer</a> type. The bit size of the
3938<tt>value</tt> must be smaller than the bit size of the destination type,
3939<tt>ty2</tt>.</p>
3940
3941<h5>Semantics:</h5>
3942<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
3943bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
3944
3945<p>When zero extending from i1, the result will always be either 0 or 1.</p>
3946
3947<h5>Example:</h5>
3948<pre>
3949 %X = zext i32 257 to i64 <i>; yields i64:257</i>
3950 %Y = zext i1 true to i32 <i>; yields i32:1</i>
3951</pre>
3952</div>
3953
3954<!-- _______________________________________________________________________ -->
3955<div class="doc_subsubsection">
3956 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
3957</div>
3958<div class="doc_text">
3959
3960<h5>Syntax:</h5>
3961<pre>
3962 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3963</pre>
3964
3965<h5>Overview:</h5>
3966<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
3967
3968<h5>Arguments:</h5>
3969<p>
3970The '<tt>sext</tt>' instruction takes a value to cast, which must be of
3971<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3972also be of <a href="#t_integer">integer</a> type. The bit size of the
3973<tt>value</tt> must be smaller than the bit size of the destination type,
3974<tt>ty2</tt>.</p>
3975
3976<h5>Semantics:</h5>
3977<p>
3978The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
3979bit (highest order bit) of the <tt>value</tt> until it reaches the bit size of
3980the type <tt>ty2</tt>.</p>
3981
3982<p>When sign extending from i1, the extension always results in -1 or 0.</p>
3983
3984<h5>Example:</h5>
3985<pre>
3986 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
3987 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
3988</pre>
3989</div>
3990
3991<!-- _______________________________________________________________________ -->
3992<div class="doc_subsubsection">
3993 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
3994</div>
3995
3996<div class="doc_text">
3997
3998<h5>Syntax:</h5>
3999
4000<pre>
4001 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4002</pre>
4003
4004<h5>Overview:</h5>
4005<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
4006<tt>ty2</tt>.</p>
4007
4008
4009<h5>Arguments:</h5>
4010<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
4011 point</a> value to cast and a <a href="#t_floating">floating point</a> type to
4012cast it to. The size of <tt>value</tt> must be larger than the size of
4013<tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
4014<i>no-op cast</i>.</p>
4015
4016<h5>Semantics:</h5>
4017<p> The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
4018<a href="#t_floating">floating point</a> type to a smaller
4019<a href="#t_floating">floating point</a> type. If the value cannot fit within
4020the destination type, <tt>ty2</tt>, then the results are undefined.</p>
4021
4022<h5>Example:</h5>
4023<pre>
4024 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
4025 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
4026</pre>
4027</div>
4028
4029<!-- _______________________________________________________________________ -->
4030<div class="doc_subsubsection">
4031 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
4032</div>
4033<div class="doc_text">
4034
4035<h5>Syntax:</h5>
4036<pre>
4037 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4038</pre>
4039
4040<h5>Overview:</h5>
4041<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
4042floating point value.</p>
4043
4044<h5>Arguments:</h5>
4045<p>The '<tt>fpext</tt>' instruction takes a
4046<a href="#t_floating">floating point</a> <tt>value</tt> to cast,
4047and a <a href="#t_floating">floating point</a> type to cast it to. The source
4048type must be smaller than the destination type.</p>
4049
4050<h5>Semantics:</h5>
4051<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
4052<a href="#t_floating">floating point</a> type to a larger
4053<a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
4054used to make a <i>no-op cast</i> because it always changes bits. Use
4055<tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
4056
4057<h5>Example:</h5>
4058<pre>
4059 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
4060 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
4061</pre>
4062</div>
4063
4064<!-- _______________________________________________________________________ -->
4065<div class="doc_subsubsection">
4066 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
4067</div>
4068<div class="doc_text">
4069
4070<h5>Syntax:</h5>
4071<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]4072 &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]4073</pre>
4074
4075<h5>Overview:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]4076<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4077unsigned integer equivalent of type <tt>ty2</tt>.
4078</p>
4079
4080<h5>Arguments:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]4081<p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]4082scalar or vector <a href="#t_floating">floating point</a> value, and a type
4083to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4084type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4085vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4086
4087<h5>Semantics:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]4088<p> The '<tt>fptoui</tt>' instruction converts its
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4089<a href="#t_floating">floating point</a> operand into the nearest (rounding
4090towards zero) unsigned integer value. If the value cannot fit in <tt>ty2</tt>,
4091the results are undefined.</p>
4092
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4093<h5>Example:</h5>
4094<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]4095 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]4096 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]4097 %X = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4098</pre>
4099</div>
4100
4101<!-- _______________________________________________________________________ -->
4102<div class="doc_subsubsection">
4103 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
4104</div>
4105<div class="doc_text">
4106
4107<h5>Syntax:</h5>
4108<pre>
4109 &lt;result&gt; = fptosi &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4110</pre>
4111
4112<h5>Overview:</h5>
4113<p>The '<tt>fptosi</tt>' instruction converts
4114<a href="#t_floating">floating point</a> <tt>value</tt> to type <tt>ty2</tt>.
4115</p>
4116
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4117<h5>Arguments:</h5>
4118<p> The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]4119scalar or vector <a href="#t_floating">floating point</a> value, and a type
4120to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4121type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4122vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4123
4124<h5>Semantics:</h5>
4125<p>The '<tt>fptosi</tt>' instruction converts its
4126<a href="#t_floating">floating point</a> operand into the nearest (rounding
4127towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
4128the results are undefined.</p>
4129
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4130<h5>Example:</h5>
4131<pre>
4132 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]4133 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4134 %X = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
4135</pre>
4136</div>
4137
4138<!-- _______________________________________________________________________ -->
4139<div class="doc_subsubsection">
4140 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
4141</div>
4142<div class="doc_text">
4143
4144<h5>Syntax:</h5>
4145<pre>
4146 &lt;result&gt; = uitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4147</pre>
4148
4149<h5>Overview:</h5>
4150<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
4151integer and converts that value to the <tt>ty2</tt> type.</p>
4152
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4153<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]4154<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be a
4155scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
4156to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4157type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4158floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4159
4160<h5>Semantics:</h5>
4161<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
4162integer quantity and converts it to the corresponding floating point value. If
4163the value cannot fit in the floating point value, the results are undefined.</p>
4164
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4165<h5>Example:</h5>
4166<pre>
4167 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4168 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4169</pre>
4170</div>
4171
4172<!-- _______________________________________________________________________ -->
4173<div class="doc_subsubsection">
4174 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
4175</div>
4176<div class="doc_text">
4177
4178<h5>Syntax:</h5>
4179<pre>
4180 &lt;result&gt; = sitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4181</pre>
4182
4183<h5>Overview:</h5>
4184<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed
4185integer and converts that value to the <tt>ty2</tt> type.</p>
4186
4187<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]4188<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be a
4189scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
4190to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4191type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4192floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4193
4194<h5>Semantics:</h5>
4195<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed
4196integer quantity and converts it to the corresponding floating point value. If
4197the value cannot fit in the floating point value, the results are undefined.</p>
4198
4199<h5>Example:</h5>
4200<pre>
4201 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4202 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4203</pre>
4204</div>
4205
4206<!-- _______________________________________________________________________ -->
4207<div class="doc_subsubsection">
4208 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
4209</div>
4210<div class="doc_text">
4211
4212<h5>Syntax:</h5>
4213<pre>
4214 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4215</pre>
4216
4217<h5>Overview:</h5>
4218<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
4219the integer type <tt>ty2</tt>.</p>
4220
4221<h5>Arguments:</h5>
4222<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
4223must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4224<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]4225
4226<h5>Semantics:</h5>
4227<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
4228<tt>ty2</tt> by interpreting the pointer value as an integer and either
4229truncating or zero extending that value to the size of the integer type. If
4230<tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
4231<tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
4232are the same size, then nothing is done (<i>no-op cast</i>) other than a type
4233change.</p>
4234
4235<h5>Example:</h5>
4236<pre>
4237 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
4238 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
4239</pre>
4240</div>
4241
4242<!-- _______________________________________________________________________ -->
4243<div class="doc_subsubsection">
4244 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
4245</div>
4246<div class="doc_text">
4247
4248<h5>Syntax:</h5>
4249<pre>
4250 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4251</pre>
4252
4253<h5>Overview:</h5>
4254<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to
4255a pointer type, <tt>ty2</tt>.</p>
4256
4257<h5>Arguments:</h5>
4258<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
4259value to cast, and a type to cast it to, which must be a
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4260<a href="#t_pointer">pointer</a> type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4261
4262<h5>Semantics:</h5>
4263<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
4264<tt>ty2</tt> by applying either a zero extension or a truncation depending on
4265the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
4266size of a pointer then a truncation is done. If <tt>value</tt> is smaller than
4267the size of a pointer then a zero extension is done. If they are the same size,
4268nothing is done (<i>no-op cast</i>).</p>
4269
4270<h5>Example:</h5>
4271<pre>
4272 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
4273 %X = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
4274 %Y = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
4275</pre>
4276</div>
4277
4278<!-- _______________________________________________________________________ -->
4279<div class="doc_subsubsection">
4280 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
4281</div>
4282<div class="doc_text">
4283
4284<h5>Syntax:</h5>
4285<pre>
4286 &lt;result&gt; = bitcast &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4287</pre>
4288
4289<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4290
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4291<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
4292<tt>ty2</tt> without changing any bits.</p>
4293
4294<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4295
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4296<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be
Dan Gohman7305fa02008-09-08 16:45:59 +0000[diff] [blame]4297a non-aggregate first class value, and a type to cast it to, which must also be
4298a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes of
4299<tt>value</tt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4300and the destination type, <tt>ty2</tt>, must be identical. If the source
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4301type is a pointer, the destination type must also be a pointer. This
4302instruction supports bitwise conversion of vectors to integers and to vectors
4303of other types (as long as they have the same size).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4304
4305<h5>Semantics:</h5>
4306<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
4307<tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
4308this conversion. The conversion is done as if the <tt>value</tt> had been
4309stored to memory and read back as type <tt>ty2</tt>. Pointer types may only be
4310converted to other pointer types with this instruction. To convert pointers to
4311other types, use the <a href="#i_inttoptr">inttoptr</a> or
4312<a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
4313
4314<h5>Example:</h5>
4315<pre>
4316 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
4317 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4318 %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]4319</pre>
4320</div>
4321
4322<!-- ======================================================================= -->
4323<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
4324<div class="doc_text">
4325<p>The instructions in this category are the "miscellaneous"
4326instructions, which defy better classification.</p>
4327</div>
4328
4329<!-- _______________________________________________________________________ -->
4330<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
4331</div>
4332<div class="doc_text">
4333<h5>Syntax:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4334<pre> &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]4335</pre>
4336<h5>Overview:</h5>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4337<p>The '<tt>icmp</tt>' instruction returns a boolean value or
4338a vector of boolean values based on comparison
4339of its two integer, integer vector, or pointer operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4340<h5>Arguments:</h5>
4341<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
4342the condition code indicating the kind of comparison to perform. It is not
4343a value, just a keyword. The possible condition code are:
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4344</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4345<ol>
4346 <li><tt>eq</tt>: equal</li>
4347 <li><tt>ne</tt>: not equal </li>
4348 <li><tt>ugt</tt>: unsigned greater than</li>
4349 <li><tt>uge</tt>: unsigned greater or equal</li>
4350 <li><tt>ult</tt>: unsigned less than</li>
4351 <li><tt>ule</tt>: unsigned less or equal</li>
4352 <li><tt>sgt</tt>: signed greater than</li>
4353 <li><tt>sge</tt>: signed greater or equal</li>
4354 <li><tt>slt</tt>: signed less than</li>
4355 <li><tt>sle</tt>: signed less or equal</li>
4356</ol>
4357<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4358<a href="#t_pointer">pointer</a>
4359or integer <a href="#t_vector">vector</a> typed.
4360They must also be identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4361<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4362<p>The '<tt>icmp</tt>' compares <tt>op1</tt> and <tt>op2</tt> according to
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4363the condition code given as <tt>cond</tt>. The comparison performed always
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4364yields either an <a href="#t_primitive"><tt>i1</tt></a> or vector of <tt>i1</tt> result, as follows:
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4365</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4366<ol>
4367 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
4368 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
4369 </li>
4370 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4371 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4372 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4373 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4374 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4375 <tt>true</tt> if <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4376 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4377 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4378 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4379 <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]4380 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4381 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4382 <li><tt>sge</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4383 <tt>true</tt> if <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4384 <li><tt>slt</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4385 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4386 <li><tt>sle</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4387 <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]4388</ol>
4389<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
4390values are compared as if they were integers.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4391<p>If the operands are integer vectors, then they are compared
4392element by element. The result is an <tt>i1</tt> vector with
4393the same number of elements as the values being compared.
4394Otherwise, the result is an <tt>i1</tt>.
4395</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4396
4397<h5>Example:</h5>
4398<pre> &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
4399 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
4400 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
4401 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
4402 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
4403 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
4404</pre>
Dan Gohmana53eb382009-01-22 01:39:38 +0000[diff] [blame]4405
4406<p>Note that the code generator does not yet support vector types with
4407 the <tt>icmp</tt> instruction.</p>
4408
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4409</div>
4410
4411<!-- _______________________________________________________________________ -->
4412<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
4413</div>
4414<div class="doc_text">
4415<h5>Syntax:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4416<pre> &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]4417</pre>
4418<h5>Overview:</h5>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4419<p>The '<tt>fcmp</tt>' instruction returns a boolean value
4420or vector of boolean values based on comparison
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4421of its operands.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4422<p>
4423If the operands are floating point scalars, then the result
4424type is a boolean (<a href="#t_primitive"><tt>i1</tt></a>).
4425</p>
4426<p>If the operands are floating point vectors, then the result type
4427is a vector of boolean with the same number of elements as the
4428operands being compared.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4429<h5>Arguments:</h5>
4430<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
4431the condition code indicating the kind of comparison to perform. It is not
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4432a value, just a keyword. The possible condition code are:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4433<ol>
4434 <li><tt>false</tt>: no comparison, always returns false</li>
4435 <li><tt>oeq</tt>: ordered and equal</li>
4436 <li><tt>ogt</tt>: ordered and greater than </li>
4437 <li><tt>oge</tt>: ordered and greater than or equal</li>
4438 <li><tt>olt</tt>: ordered and less than </li>
4439 <li><tt>ole</tt>: ordered and less than or equal</li>
4440 <li><tt>one</tt>: ordered and not equal</li>
4441 <li><tt>ord</tt>: ordered (no nans)</li>
4442 <li><tt>ueq</tt>: unordered or equal</li>
4443 <li><tt>ugt</tt>: unordered or greater than </li>
4444 <li><tt>uge</tt>: unordered or greater than or equal</li>
4445 <li><tt>ult</tt>: unordered or less than </li>
4446 <li><tt>ule</tt>: unordered or less than or equal</li>
4447 <li><tt>une</tt>: unordered or not equal</li>
4448 <li><tt>uno</tt>: unordered (either nans)</li>
4449 <li><tt>true</tt>: no comparison, always returns true</li>
4450</ol>
4451<p><i>Ordered</i> means that neither operand is a QNAN while
4452<i>unordered</i> means that either operand may be a QNAN.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4453<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be
4454either a <a href="#t_floating">floating point</a> type
4455or a <a href="#t_vector">vector</a> of floating point type.
4456They must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4457<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4458<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4459according to the condition code given as <tt>cond</tt>.
4460If the operands are vectors, then the vectors are compared
4461element by element.
4462Each comparison performed
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4463always yields an <a href="#t_primitive">i1</a> result, as follows:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4464<ol>
4465 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
4466 <li><tt>oeq</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4467 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4468 <li><tt>ogt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4469 <tt>op1</tt> is greather than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4470 <li><tt>oge</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4471 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4472 <li><tt>olt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4473 <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4474 <li><tt>ole</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4475 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4476 <li><tt>one</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4477 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4478 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
4479 <li><tt>ueq</tt>: yields <tt>true</tt> if either operand is a QNAN or
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4480 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4481 <li><tt>ugt</tt>: yields <tt>true</tt> if either operand is a QNAN or
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4482 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4483 <li><tt>uge</tt>: yields <tt>true</tt> if either operand is a QNAN or
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4484 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4485 <li><tt>ult</tt>: yields <tt>true</tt> if either operand is a QNAN or
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4486 <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4487 <li><tt>ule</tt>: yields <tt>true</tt> if either operand is a QNAN or
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4488 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4489 <li><tt>une</tt>: yields <tt>true</tt> if either operand is a QNAN or
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4490 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4491 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
4492 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
4493</ol>
4494
4495<h5>Example:</h5>
4496<pre> &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]4497 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
4498 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
4499 &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]4500</pre>
Dan Gohmana53eb382009-01-22 01:39:38 +0000[diff] [blame]4501
4502<p>Note that the code generator does not yet support vector types with
4503 the <tt>fcmp</tt> instruction.</p>
4504
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4505</div>
4506
4507<!-- _______________________________________________________________________ -->
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4508<div class="doc_subsubsection">
4509 <a name="i_vicmp">'<tt>vicmp</tt>' Instruction</a>
4510</div>
4511<div class="doc_text">
4512<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4513<pre> &lt;result&gt; = vicmp &lt;cond&gt; &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4514</pre>
4515<h5>Overview:</h5>
4516<p>The '<tt>vicmp</tt>' instruction returns an integer vector value based on
4517element-wise comparison of its two integer vector operands.</p>
4518<h5>Arguments:</h5>
4519<p>The '<tt>vicmp</tt>' instruction takes three operands. The first operand is
4520the condition code indicating the kind of comparison to perform. It is not
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4521a value, just a keyword. The possible condition code are:</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4522<ol>
4523 <li><tt>eq</tt>: equal</li>
4524 <li><tt>ne</tt>: not equal </li>
4525 <li><tt>ugt</tt>: unsigned greater than</li>
4526 <li><tt>uge</tt>: unsigned greater or equal</li>
4527 <li><tt>ult</tt>: unsigned less than</li>
4528 <li><tt>ule</tt>: unsigned less or equal</li>
4529 <li><tt>sgt</tt>: signed greater than</li>
4530 <li><tt>sge</tt>: signed greater or equal</li>
4531 <li><tt>slt</tt>: signed less than</li>
4532 <li><tt>sle</tt>: signed less or equal</li>
4533</ol>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4534<p>The remaining two arguments must be <a href="#t_vector">vector</a> or
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4535<a href="#t_integer">integer</a> typed. They must also be identical types.</p>
4536<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4537<p>The '<tt>vicmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4538according to the condition code given as <tt>cond</tt>. The comparison yields a
4539<a href="#t_vector">vector</a> of <a href="#t_integer">integer</a> result, of
4540identical type as the values being compared. The most significant bit in each
4541element is 1 if the element-wise comparison evaluates to true, and is 0
4542otherwise. All other bits of the result are undefined. The condition codes
4543are evaluated identically to the <a href="#i_icmp">'<tt>icmp</tt>'
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4544instruction</a>.</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4545
4546<h5>Example:</h5>
4547<pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4548 &lt;result&gt; = vicmp eq &lt;2 x i32&gt; &lt; i32 4, i32 0&gt;, &lt; i32 5, i32 0&gt; <i>; yields: result=&lt;2 x i32&gt; &lt; i32 0, i32 -1 &gt;</i>
4549 &lt;result&gt; = vicmp ult &lt;2 x i8 &gt; &lt; i8 1, i8 2&gt;, &lt; i8 2, i8 2 &gt; <i>; yields: result=&lt;2 x i8&gt; &lt; i8 -1, i8 0 &gt;</i>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4550</pre>
4551</div>
4552
4553<!-- _______________________________________________________________________ -->
4554<div class="doc_subsubsection">
4555 <a name="i_vfcmp">'<tt>vfcmp</tt>' Instruction</a>
4556</div>
4557<div class="doc_text">
4558<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4559<pre> &lt;result&gt; = vfcmp &lt;cond&gt; &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt;</pre>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4560<h5>Overview:</h5>
4561<p>The '<tt>vfcmp</tt>' instruction returns an integer vector value based on
4562element-wise comparison of its two floating point vector operands. The output
4563elements have the same width as the input elements.</p>
4564<h5>Arguments:</h5>
4565<p>The '<tt>vfcmp</tt>' instruction takes three operands. The first operand is
4566the condition code indicating the kind of comparison to perform. It is not
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4567a value, just a keyword. The possible condition code are:</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4568<ol>
4569 <li><tt>false</tt>: no comparison, always returns false</li>
4570 <li><tt>oeq</tt>: ordered and equal</li>
4571 <li><tt>ogt</tt>: ordered and greater than </li>
4572 <li><tt>oge</tt>: ordered and greater than or equal</li>
4573 <li><tt>olt</tt>: ordered and less than </li>
4574 <li><tt>ole</tt>: ordered and less than or equal</li>
4575 <li><tt>one</tt>: ordered and not equal</li>
4576 <li><tt>ord</tt>: ordered (no nans)</li>
4577 <li><tt>ueq</tt>: unordered or equal</li>
4578 <li><tt>ugt</tt>: unordered or greater than </li>
4579 <li><tt>uge</tt>: unordered or greater than or equal</li>
4580 <li><tt>ult</tt>: unordered or less than </li>
4581 <li><tt>ule</tt>: unordered or less than or equal</li>
4582 <li><tt>une</tt>: unordered or not equal</li>
4583 <li><tt>uno</tt>: unordered (either nans)</li>
4584 <li><tt>true</tt>: no comparison, always returns true</li>
4585</ol>
4586<p>The remaining two arguments must be <a href="#t_vector">vector</a> of
4587<a href="#t_floating">floating point</a> typed. They must also be identical
4588types.</p>
4589<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4590<p>The '<tt>vfcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4591according to the condition code given as <tt>cond</tt>. The comparison yields a
4592<a href="#t_vector">vector</a> of <a href="#t_integer">integer</a> result, with
4593an identical number of elements as the values being compared, and each element
4594having identical with to the width of the floating point elements. The most
4595significant bit in each element is 1 if the element-wise comparison evaluates to
4596true, and is 0 otherwise. All other bits of the result are undefined. The
4597condition codes are evaluated identically to the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4598<a href="#i_fcmp">'<tt>fcmp</tt>' instruction</a>.</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4599
4600<h5>Example:</h5>
4601<pre>
Chris Lattner1e5c5cd02008-10-13 16:55:18 +0000[diff] [blame]4602 <i>; yields: result=&lt;2 x i32&gt; &lt; i32 0, i32 -1 &gt;</i>
4603 &lt;result&gt; = vfcmp oeq &lt;2 x float&gt; &lt; float 4, float 0 &gt;, &lt; float 5, float 0 &gt;
4604
4605 <i>; yields: result=&lt;2 x i64&gt; &lt; i64 -1, i64 0 &gt;</i>
4606 &lt;result&gt; = vfcmp ult &lt;2 x double&gt; &lt; double 1, double 2 &gt;, &lt; double 2, double 2&gt;
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4607</pre>
4608</div>
4609
4610<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4611<div class="doc_subsubsection">
4612 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
4613</div>
4614
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4615<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4616
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4617<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4618
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4619<pre> &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...<br></pre>
4620<h5>Overview:</h5>
4621<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in
4622the SSA graph representing the function.</p>
4623<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4624
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4625<p>The type of the incoming values is specified with the first type
4626field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
4627as arguments, with one pair for each predecessor basic block of the
4628current block. Only values of <a href="#t_firstclass">first class</a>
4629type may be used as the value arguments to the PHI node. Only labels
4630may be used as the label arguments.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4631
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4632<p>There must be no non-phi instructions between the start of a basic
4633block and the PHI instructions: i.e. PHI instructions must be first in
4634a basic block.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4635
Jay Foad8e2fd2c2009-06-03 10:20:10 +0000[diff] [blame]4636<p>For the purposes of the SSA form, the use of each incoming value is
4637deemed to occur on the edge from the corresponding predecessor block
4638to the current block (but after any definition of an '<tt>invoke</tt>'
4639instruction's return value on the same edge).</p>
4640
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4641<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4642
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4643<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
4644specified by the pair corresponding to the predecessor basic block that executed
4645just prior to the current block.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4646
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4647<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4648<pre>
4649Loop: ; Infinite loop that counts from 0 on up...
4650 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
4651 %nextindvar = add i32 %indvar, 1
4652 br label %Loop
4653</pre>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4654</div>
4655
4656<!-- _______________________________________________________________________ -->
4657<div class="doc_subsubsection">
4658 <a name="i_select">'<tt>select</tt>' Instruction</a>
4659</div>
4660
4661<div class="doc_text">
4662
4663<h5>Syntax:</h5>
4664
4665<pre>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4666 &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>
4667
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4668 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4669</pre>
4670
4671<h5>Overview:</h5>
4672
4673<p>
4674The '<tt>select</tt>' instruction is used to choose one value based on a
4675condition, without branching.
4676</p>
4677
4678
4679<h5>Arguments:</h5>
4680
4681<p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4682The '<tt>select</tt>' instruction requires an 'i1' value or
4683a vector of 'i1' values indicating the
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4684condition, and two values of the same <a href="#t_firstclass">first class</a>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4685type. If the val1/val2 are vectors and
4686the condition is a scalar, then entire vectors are selected, not
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4687individual elements.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4688</p>
4689
4690<h5>Semantics:</h5>
4691
4692<p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4693If the condition is an i1 and it evaluates to 1, the instruction returns the first
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4694value argument; otherwise, it returns the second value argument.
4695</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4696<p>
4697If the condition is a vector of i1, then the value arguments must
4698be vectors of the same size, and the selection is done element
4699by element.
4700</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4701
4702<h5>Example:</h5>
4703
4704<pre>
4705 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
4706</pre>
Dan Gohmana53eb382009-01-22 01:39:38 +0000[diff] [blame]4707
4708<p>Note that the code generator does not yet support conditions
4709 with vector type.</p>
4710
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4711</div>
4712
4713
4714<!-- _______________________________________________________________________ -->
4715<div class="doc_subsubsection">
4716 <a name="i_call">'<tt>call</tt>' Instruction</a>
4717</div>
4718
4719<div class="doc_text">
4720
4721<h5>Syntax:</h5>
4722<pre>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]4723 &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]4724</pre>
4725
4726<h5>Overview:</h5>
4727
4728<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
4729
4730<h5>Arguments:</h5>
4731
4732<p>This instruction requires several arguments:</p>
4733
4734<ol>
4735 <li>
4736 <p>The optional "tail" marker indicates whether the callee function accesses
4737 any allocas or varargs in the caller. If the "tail" marker is present, the
4738 function call is eligible for tail call optimization. Note that calls may
4739 be marked "tail" even if they do not occur before a <a
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4740 href="#i_ret"><tt>ret</tt></a> instruction.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4741 </li>
4742 <li>
4743 <p>The optional "cconv" marker indicates which <a href="#callingconv">calling
4744 convention</a> the call should use. If none is specified, the call defaults
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4745 to using C calling conventions.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4746 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4747
4748 <li>
4749 <p>The optional <a href="#paramattrs">Parameter Attributes</a> list for
4750 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>',
4751 and '<tt>inreg</tt>' attributes are valid here.</p>
4752 </li>
4753
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4754 <li>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]4755 <p>'<tt>ty</tt>': the type of the call instruction itself which is also
4756 the type of the return value. Functions that return no value are marked
4757 <tt><a href="#t_void">void</a></tt>.</p>
4758 </li>
4759 <li>
4760 <p>'<tt>fnty</tt>': shall be the signature of the pointer to function
4761 value being invoked. The argument types must match the types implied by
4762 this signature. This type can be omitted if the function is not varargs
4763 and if the function type does not return a pointer to a function.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4764 </li>
4765 <li>
4766 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
4767 be invoked. In most cases, this is a direct function invocation, but
4768 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
4769 to function value.</p>
4770 </li>
4771 <li>
4772 <p>'<tt>function args</tt>': argument list whose types match the
4773 function signature argument types. All arguments must be of
4774 <a href="#t_firstclass">first class</a> type. If the function signature
4775 indicates the function accepts a variable number of arguments, the extra
4776 arguments can be specified.</p>
4777 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4778 <li>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]4779 <p>The optional <a href="#fnattrs">function attributes</a> list. Only
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4780 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
4781 '<tt>readnone</tt>' attributes are valid here.</p>
4782 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4783</ol>
4784
4785<h5>Semantics:</h5>
4786
4787<p>The '<tt>call</tt>' instruction is used to cause control flow to
4788transfer to a specified function, with its incoming arguments bound to
4789the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
4790instruction in the called function, control flow continues with the
4791instruction after the function call, and the return value of the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4792function is bound to the result argument.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4793
4794<h5>Example:</h5>
4795
4796<pre>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]4797 %retval = call i32 @test(i32 %argc)
Chris Lattner5e893ef2008-03-21 17:24:17 +0000[diff] [blame]4798 call i32 (i8 *, ...)* @printf(i8 * %msg, i32 12, i8 42) <i>; yields i32</i>
4799 %X = tail call i32 @foo() <i>; yields i32</i>
4800 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
4801 call void %foo(i8 97 signext)
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]4802
4803 %struct.A = type { i32, i8 }
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4804 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]4805 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
4806 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattnerac454b32008-10-08 06:26:11 +0000[diff] [blame]4807 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijman2c4e05a2008-10-07 10:03:45 +0000[diff] [blame]4808 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4809</pre>
4810
4811</div>
4812
4813<!-- _______________________________________________________________________ -->
4814<div class="doc_subsubsection">
4815 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
4816</div>
4817
4818<div class="doc_text">
4819
4820<h5>Syntax:</h5>
4821
4822<pre>
4823 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
4824</pre>
4825
4826<h5>Overview:</h5>
4827
4828<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
4829the "variable argument" area of a function call. It is used to implement the
4830<tt>va_arg</tt> macro in C.</p>
4831
4832<h5>Arguments:</h5>
4833
4834<p>This instruction takes a <tt>va_list*</tt> value and the type of
4835the argument. It returns a value of the specified argument type and
4836increments the <tt>va_list</tt> to point to the next argument. The
4837actual type of <tt>va_list</tt> is target specific.</p>
4838
4839<h5>Semantics:</h5>
4840
4841<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
4842type from the specified <tt>va_list</tt> and causes the
4843<tt>va_list</tt> to point to the next argument. For more information,
4844see the variable argument handling <a href="#int_varargs">Intrinsic
4845Functions</a>.</p>
4846
4847<p>It is legal for this instruction to be called in a function which does not
4848take a variable number of arguments, for example, the <tt>vfprintf</tt>
4849function.</p>
4850
4851<p><tt>va_arg</tt> is an LLVM instruction instead of an <a
4852href="#intrinsics">intrinsic function</a> because it takes a type as an
4853argument.</p>
4854
4855<h5>Example:</h5>
4856
4857<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
4858
Dan Gohman60967192009-01-12 23:12:39 +0000[diff] [blame]4859<p>Note that the code generator does not yet fully support va_arg
4860 on many targets. Also, it does not currently support va_arg with
4861 aggregate types on any target.</p>
4862
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4863</div>
4864
4865<!-- *********************************************************************** -->
4866<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
4867<!-- *********************************************************************** -->
4868
4869<div class="doc_text">
4870
4871<p>LLVM supports the notion of an "intrinsic function". These functions have
4872well known names and semantics and are required to follow certain restrictions.
4873Overall, these intrinsics represent an extension mechanism for the LLVM
4874language that does not require changing all of the transformations in LLVM when
4875adding to the language (or the bitcode reader/writer, the parser, etc...).</p>
4876
4877<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
4878prefix is reserved in LLVM for intrinsic names; thus, function names may not
4879begin with this prefix. Intrinsic functions must always be external functions:
4880you cannot define the body of intrinsic functions. Intrinsic functions may
4881only be used in call or invoke instructions: it is illegal to take the address
4882of an intrinsic function. Additionally, because intrinsic functions are part
4883of the LLVM language, it is required if any are added that they be documented
4884here.</p>
4885
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4886<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents
4887a family of functions that perform the same operation but on different data
4888types. Because LLVM can represent over 8 million different integer types,
4889overloading is used commonly to allow an intrinsic function to operate on any
4890integer type. One or more of the argument types or the result type can be
4891overloaded to accept any integer type. Argument types may also be defined as
4892exactly matching a previous argument's type or the result type. This allows an
4893intrinsic function which accepts multiple arguments, but needs all of them to
4894be of the same type, to only be overloaded with respect to a single argument or
4895the result.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4896
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4897<p>Overloaded intrinsics will have the names of its overloaded argument types
4898encoded into its function name, each preceded by a period. Only those types
4899which are overloaded result in a name suffix. Arguments whose type is matched
4900against another type do not. For example, the <tt>llvm.ctpop</tt> function can
4901take an integer of any width and returns an integer of exactly the same integer
4902width. This leads to a family of functions such as
4903<tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29 %val)</tt>.
4904Only one type, the return type, is overloaded, and only one type suffix is
4905required. Because the argument's type is matched against the return type, it
4906does not require its own name suffix.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4907
4908<p>To learn how to add an intrinsic function, please see the
4909<a href="ExtendingLLVM.html">Extending LLVM Guide</a>.
4910</p>
4911
4912</div>
4913
4914<!-- ======================================================================= -->
4915<div class="doc_subsection">
4916 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
4917</div>
4918
4919<div class="doc_text">
4920
4921<p>Variable argument support is defined in LLVM with the <a
4922 href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
4923intrinsic functions. These functions are related to the similarly
4924named macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
4925
4926<p>All of these functions operate on arguments that use a
4927target-specific value type "<tt>va_list</tt>". The LLVM assembly
4928language reference manual does not define what this type is, so all
4929transformations should be prepared to handle these functions regardless of
4930the type used.</p>
4931
4932<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
4933instruction and the variable argument handling intrinsic functions are
4934used.</p>
4935
4936<div class="doc_code">
4937<pre>
4938define i32 @test(i32 %X, ...) {
4939 ; Initialize variable argument processing
4940 %ap = alloca i8*
4941 %ap2 = bitcast i8** %ap to i8*
4942 call void @llvm.va_start(i8* %ap2)
4943
4944 ; Read a single integer argument
4945 %tmp = va_arg i8** %ap, i32
4946
4947 ; Demonstrate usage of llvm.va_copy and llvm.va_end
4948 %aq = alloca i8*
4949 %aq2 = bitcast i8** %aq to i8*
4950 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
4951 call void @llvm.va_end(i8* %aq2)
4952
4953 ; Stop processing of arguments.
4954 call void @llvm.va_end(i8* %ap2)
4955 ret i32 %tmp
4956}
4957
4958declare void @llvm.va_start(i8*)
4959declare void @llvm.va_copy(i8*, i8*)
4960declare void @llvm.va_end(i8*)
4961</pre>
4962</div>
4963
4964</div>
4965
4966<!-- _______________________________________________________________________ -->
4967<div class="doc_subsubsection">
4968 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
4969</div>
4970
4971
4972<div class="doc_text">
4973<h5>Syntax:</h5>
4974<pre> declare void %llvm.va_start(i8* &lt;arglist&gt;)<br></pre>
4975<h5>Overview:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4976<p>The '<tt>llvm.va_start</tt>' intrinsic initializes
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4977<tt>*&lt;arglist&gt;</tt> for subsequent use by <tt><a
4978href="#i_va_arg">va_arg</a></tt>.</p>
4979
4980<h5>Arguments:</h5>
4981
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4982<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]4983
4984<h5>Semantics:</h5>
4985
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4986<p>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4987macro available in C. In a target-dependent way, it initializes the
4988<tt>va_list</tt> element to which the argument points, so that the next call to
4989<tt>va_arg</tt> will produce the first variable argument passed to the function.
4990Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
4991last argument of the function as the compiler can figure that out.</p>
4992
4993</div>
4994
4995<!-- _______________________________________________________________________ -->
4996<div class="doc_subsubsection">
4997 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
4998</div>
4999
5000<div class="doc_text">
5001<h5>Syntax:</h5>
5002<pre> declare void @llvm.va_end(i8* &lt;arglist&gt;)<br></pre>
5003<h5>Overview:</h5>
5004
5005<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
5006which has been initialized previously with <tt><a href="#int_va_start">llvm.va_start</a></tt>
5007or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
5008
5009<h5>Arguments:</h5>
5010
5011<p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
5012
5013<h5>Semantics:</h5>
5014
5015<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
5016macro available in C. In a target-dependent way, it destroys the
5017<tt>va_list</tt> element to which the argument points. Calls to <a
5018href="#int_va_start"><tt>llvm.va_start</tt></a> and <a href="#int_va_copy">
5019<tt>llvm.va_copy</tt></a> must be matched exactly with calls to
5020<tt>llvm.va_end</tt>.</p>
5021
5022</div>
5023
5024<!-- _______________________________________________________________________ -->
5025<div class="doc_subsubsection">
5026 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
5027</div>
5028
5029<div class="doc_text">
5030
5031<h5>Syntax:</h5>
5032
5033<pre>
5034 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
5035</pre>
5036
5037<h5>Overview:</h5>
5038
5039<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
5040from the source argument list to the destination argument list.</p>
5041
5042<h5>Arguments:</h5>
5043
5044<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
5045The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
5046
5047
5048<h5>Semantics:</h5>
5049
5050<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
5051macro available in C. In a target-dependent way, it copies the source
5052<tt>va_list</tt> element into the destination <tt>va_list</tt> element. This
5053intrinsic is necessary because the <tt><a href="#int_va_start">
5054llvm.va_start</a></tt> intrinsic may be arbitrarily complex and require, for
5055example, memory allocation.</p>
5056
5057</div>
5058
5059<!-- ======================================================================= -->
5060<div class="doc_subsection">
5061 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
5062</div>
5063
5064<div class="doc_text">
5065
5066<p>
5067LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]5068Collection</a> (GC) requires the implementation and generation of these
5069intrinsics.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5070These intrinsics allow identification of <a href="#int_gcroot">GC roots on the
5071stack</a>, as well as garbage collector implementations that require <a
5072href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a> barriers.
5073Front-ends for type-safe garbage collected languages should generate these
5074intrinsics to make use of the LLVM garbage collectors. For more details, see <a
5075href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
5076</p>
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]5077
5078<p>The garbage collection intrinsics only operate on objects in the generic
5079 address space (address space zero).</p>
5080
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5081</div>
5082
5083<!-- _______________________________________________________________________ -->
5084<div class="doc_subsubsection">
5085 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
5086</div>
5087
5088<div class="doc_text">
5089
5090<h5>Syntax:</h5>
5091
5092<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5093 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5094</pre>
5095
5096<h5>Overview:</h5>
5097
5098<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
5099the code generator, and allows some metadata to be associated with it.</p>
5100
5101<h5>Arguments:</h5>
5102
5103<p>The first argument specifies the address of a stack object that contains the
5104root pointer. The second pointer (which must be either a constant or a global
5105value address) contains the meta-data to be associated with the root.</p>
5106
5107<h5>Semantics:</h5>
5108
Chris Lattnera7d94ba2008-04-24 05:59:56 +0000[diff] [blame]5109<p>At runtime, a call to this intrinsic stores a null pointer into the "ptrloc"
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5110location. At compile-time, the code generator generates information to allow
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]5111the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
5112intrinsic may only be used in a function which <a href="#gc">specifies a GC
5113algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5114
5115</div>
5116
5117
5118<!-- _______________________________________________________________________ -->
5119<div class="doc_subsubsection">
5120 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
5121</div>
5122
5123<div class="doc_text">
5124
5125<h5>Syntax:</h5>
5126
5127<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5128 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5129</pre>
5130
5131<h5>Overview:</h5>
5132
5133<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
5134locations, allowing garbage collector implementations that require read
5135barriers.</p>
5136
5137<h5>Arguments:</h5>
5138
5139<p>The second argument is the address to read from, which should be an address
5140allocated from the garbage collector. The first object is a pointer to the
5141start of the referenced object, if needed by the language runtime (otherwise
5142null).</p>
5143
5144<h5>Semantics:</h5>
5145
5146<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
5147instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]5148garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
5149may only be used in a function which <a href="#gc">specifies a GC
5150algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5151
5152</div>
5153
5154
5155<!-- _______________________________________________________________________ -->
5156<div class="doc_subsubsection">
5157 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
5158</div>
5159
5160<div class="doc_text">
5161
5162<h5>Syntax:</h5>
5163
5164<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5165 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5166</pre>
5167
5168<h5>Overview:</h5>
5169
5170<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
5171locations, allowing garbage collector implementations that require write
5172barriers (such as generational or reference counting collectors).</p>
5173
5174<h5>Arguments:</h5>
5175
5176<p>The first argument is the reference to store, the second is the start of the
5177object to store it to, and the third is the address of the field of Obj to
5178store to. If the runtime does not require a pointer to the object, Obj may be
5179null.</p>
5180
5181<h5>Semantics:</h5>
5182
5183<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
5184instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]5185garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
5186may only be used in a function which <a href="#gc">specifies a GC
5187algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5188
5189</div>
5190
5191
5192
5193<!-- ======================================================================= -->
5194<div class="doc_subsection">
5195 <a name="int_codegen">Code Generator Intrinsics</a>
5196</div>
5197
5198<div class="doc_text">
5199<p>
5200These intrinsics are provided by LLVM to expose special features that may only
5201be implemented with code generator support.
5202</p>
5203
5204</div>
5205
5206<!-- _______________________________________________________________________ -->
5207<div class="doc_subsubsection">
5208 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
5209</div>
5210
5211<div class="doc_text">
5212
5213<h5>Syntax:</h5>
5214<pre>
5215 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
5216</pre>
5217
5218<h5>Overview:</h5>
5219
5220<p>
5221The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
5222target-specific value indicating the return address of the current function
5223or one of its callers.
5224</p>
5225
5226<h5>Arguments:</h5>
5227
5228<p>
5229The argument to this intrinsic indicates which function to return the address
5230for. Zero indicates the calling function, one indicates its caller, etc. The
5231argument is <b>required</b> to be a constant integer value.
5232</p>
5233
5234<h5>Semantics:</h5>
5235
5236<p>
5237The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
5238the return address of the specified call frame, or zero if it cannot be
5239identified. The value returned by this intrinsic is likely to be incorrect or 0
5240for arguments other than zero, so it should only be used for debugging purposes.
5241</p>
5242
5243<p>
5244Note that calling this intrinsic does not prevent function inlining or other
5245aggressive transformations, so the value returned may not be that of the obvious
5246source-language caller.
5247</p>
5248</div>
5249
5250
5251<!-- _______________________________________________________________________ -->
5252<div class="doc_subsubsection">
5253 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
5254</div>
5255
5256<div class="doc_text">
5257
5258<h5>Syntax:</h5>
5259<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5260 declare i8 *@llvm.frameaddress(i32 &lt;level&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5261</pre>
5262
5263<h5>Overview:</h5>
5264
5265<p>
5266The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
5267target-specific frame pointer value for the specified stack frame.
5268</p>
5269
5270<h5>Arguments:</h5>
5271
5272<p>
5273The argument to this intrinsic indicates which function to return the frame
5274pointer for. Zero indicates the calling function, one indicates its caller,
5275etc. The argument is <b>required</b> to be a constant integer value.
5276</p>
5277
5278<h5>Semantics:</h5>
5279
5280<p>
5281The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
5282the frame address of the specified call frame, or zero if it cannot be
5283identified. The value returned by this intrinsic is likely to be incorrect or 0
5284for arguments other than zero, so it should only be used for debugging purposes.
5285</p>
5286
5287<p>
5288Note that calling this intrinsic does not prevent function inlining or other
5289aggressive transformations, so the value returned may not be that of the obvious
5290source-language caller.
5291</p>
5292</div>
5293
5294<!-- _______________________________________________________________________ -->
5295<div class="doc_subsubsection">
5296 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
5297</div>
5298
5299<div class="doc_text">
5300
5301<h5>Syntax:</h5>
5302<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5303 declare i8 *@llvm.stacksave()
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5304</pre>
5305
5306<h5>Overview:</h5>
5307
5308<p>
5309The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
5310the function stack, for use with <a href="#int_stackrestore">
5311<tt>llvm.stackrestore</tt></a>. This is useful for implementing language
5312features like scoped automatic variable sized arrays in C99.
5313</p>
5314
5315<h5>Semantics:</h5>
5316
5317<p>
5318This intrinsic returns a opaque pointer value that can be passed to <a
5319href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
5320<tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
5321<tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
5322state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
5323practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
5324that were allocated after the <tt>llvm.stacksave</tt> was executed.
5325</p>
5326
5327</div>
5328
5329<!-- _______________________________________________________________________ -->
5330<div class="doc_subsubsection">
5331 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
5332</div>
5333
5334<div class="doc_text">
5335
5336<h5>Syntax:</h5>
5337<pre>
5338 declare void @llvm.stackrestore(i8 * %ptr)
5339</pre>
5340
5341<h5>Overview:</h5>
5342
5343<p>
5344The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
5345the function stack to the state it was in when the corresponding <a
5346href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
5347useful for implementing language features like scoped automatic variable sized
5348arrays in C99.
5349</p>
5350
5351<h5>Semantics:</h5>
5352
5353<p>
5354See the description for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.
5355</p>
5356
5357</div>
5358
5359
5360<!-- _______________________________________________________________________ -->
5361<div class="doc_subsubsection">
5362 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
5363</div>
5364
5365<div class="doc_text">
5366
5367<h5>Syntax:</h5>
5368<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5369 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]5370</pre>
5371
5372<h5>Overview:</h5>
5373
5374
5375<p>
5376The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
5377a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
5378no
5379effect on the behavior of the program but can change its performance
5380characteristics.
5381</p>
5382
5383<h5>Arguments:</h5>
5384
5385<p>
5386<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
5387determining if the fetch should be for a read (0) or write (1), and
5388<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
5389locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
5390<tt>locality</tt> arguments must be constant integers.
5391</p>
5392
5393<h5>Semantics:</h5>
5394
5395<p>
5396This intrinsic does not modify the behavior of the program. In particular,
5397prefetches cannot trap and do not produce a value. On targets that support this
5398intrinsic, the prefetch can provide hints to the processor cache for better
5399performance.
5400</p>
5401
5402</div>
5403
5404<!-- _______________________________________________________________________ -->
5405<div class="doc_subsubsection">
5406 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
5407</div>
5408
5409<div class="doc_text">
5410
5411<h5>Syntax:</h5>
5412<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5413 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5414</pre>
5415
5416<h5>Overview:</h5>
5417
5418
5419<p>
5420The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]5421(PC) in a region of
5422code to simulators and other tools. The method is target specific, but it is
5423expected that the marker will use exported symbols to transmit the PC of the
5424marker.
5425The marker makes no guarantees that it will remain with any specific instruction
5426after optimizations. It is possible that the presence of a marker will inhibit
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5427optimizations. The intended use is to be inserted after optimizations to allow
5428correlations of simulation runs.
5429</p>
5430
5431<h5>Arguments:</h5>
5432
5433<p>
5434<tt>id</tt> is a numerical id identifying the marker.
5435</p>
5436
5437<h5>Semantics:</h5>
5438
5439<p>
5440This intrinsic does not modify the behavior of the program. Backends that do not
5441support this intrinisic may ignore it.
5442</p>
5443
5444</div>
5445
5446<!-- _______________________________________________________________________ -->
5447<div class="doc_subsubsection">
5448 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
5449</div>
5450
5451<div class="doc_text">
5452
5453<h5>Syntax:</h5>
5454<pre>
5455 declare i64 @llvm.readcyclecounter( )
5456</pre>
5457
5458<h5>Overview:</h5>
5459
5460
5461<p>
5462The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
5463counter register (or similar low latency, high accuracy clocks) on those targets
5464that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
5465As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
5466should only be used for small timings.
5467</p>
5468
5469<h5>Semantics:</h5>
5470
5471<p>
5472When directly supported, reading the cycle counter should not modify any memory.
5473Implementations are allowed to either return a application specific value or a
5474system wide value. On backends without support, this is lowered to a constant 0.
5475</p>
5476
5477</div>
5478
5479<!-- ======================================================================= -->
5480<div class="doc_subsection">
5481 <a name="int_libc">Standard C Library Intrinsics</a>
5482</div>
5483
5484<div class="doc_text">
5485<p>
5486LLVM provides intrinsics for a few important standard C library functions.
5487These intrinsics allow source-language front-ends to pass information about the
5488alignment of the pointer arguments to the code generator, providing opportunity
5489for more efficient code generation.
5490</p>
5491
5492</div>
5493
5494<!-- _______________________________________________________________________ -->
5495<div class="doc_subsubsection">
5496 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
5497</div>
5498
5499<div class="doc_text">
5500
5501<h5>Syntax:</h5>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5502<p>This is an overloaded intrinsic. You can use llvm.memcpy on any integer bit
5503width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5504<pre>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5505 declare void @llvm.memcpy.i8(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5506 i8 &lt;len&gt;, i32 &lt;align&gt;)
5507 declare void @llvm.memcpy.i16(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5508 i16 &lt;len&gt;, i32 &lt;align&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5509 declare void @llvm.memcpy.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5510 i32 &lt;len&gt;, i32 &lt;align&gt;)
5511 declare void @llvm.memcpy.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5512 i64 &lt;len&gt;, i32 &lt;align&gt;)
5513</pre>
5514
5515<h5>Overview:</h5>
5516
5517<p>
5518The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
5519location to the destination location.
5520</p>
5521
5522<p>
5523Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
5524intrinsics do not return a value, and takes an extra alignment argument.
5525</p>
5526
5527<h5>Arguments:</h5>
5528
5529<p>
5530The first argument is a pointer to the destination, the second is a pointer to
5531the source. The third argument is an integer argument
5532specifying the number of bytes to copy, and the fourth argument is the alignment
5533of the source and destination locations.
5534</p>
5535
5536<p>
5537If the call to this intrinisic has an alignment value that is not 0 or 1, then
5538the caller guarantees that both the source and destination pointers are aligned
5539to that boundary.
5540</p>
5541
5542<h5>Semantics:</h5>
5543
5544<p>
5545The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
5546location to the destination location, which are not allowed to overlap. It
5547copies "len" bytes of memory over. If the argument is known to be aligned to
5548some boundary, this can be specified as the fourth argument, otherwise it should
5549be set to 0 or 1.
5550</p>
5551</div>
5552
5553
5554<!-- _______________________________________________________________________ -->
5555<div class="doc_subsubsection">
5556 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
5557</div>
5558
5559<div class="doc_text">
5560
5561<h5>Syntax:</h5>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5562<p>This is an overloaded intrinsic. You can use llvm.memmove on any integer bit
5563width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5564<pre>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5565 declare void @llvm.memmove.i8(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5566 i8 &lt;len&gt;, i32 &lt;align&gt;)
5567 declare void @llvm.memmove.i16(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5568 i16 &lt;len&gt;, i32 &lt;align&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5569 declare void @llvm.memmove.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5570 i32 &lt;len&gt;, i32 &lt;align&gt;)
5571 declare void @llvm.memmove.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5572 i64 &lt;len&gt;, i32 &lt;align&gt;)
5573</pre>
5574
5575<h5>Overview:</h5>
5576
5577<p>
5578The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the source
5579location to the destination location. It is similar to the
Chris Lattnerdba16ea2008-01-06 19:51:52 +0000[diff] [blame]5580'<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to overlap.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5581</p>
5582
5583<p>
5584Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
5585intrinsics do not return a value, and takes an extra alignment argument.
5586</p>
5587
5588<h5>Arguments:</h5>
5589
5590<p>
5591The first argument is a pointer to the destination, the second is a pointer to
5592the source. The third argument is an integer argument
5593specifying the number of bytes to copy, and the fourth argument is the alignment
5594of the source and destination locations.
5595</p>
5596
5597<p>
5598If the call to this intrinisic has an alignment value that is not 0 or 1, then
5599the caller guarantees that the source and destination pointers are aligned to
5600that boundary.
5601</p>
5602
5603<h5>Semantics:</h5>
5604
5605<p>
5606The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the source
5607location to the destination location, which may overlap. It
5608copies "len" bytes of memory over. If the argument is known to be aligned to
5609some boundary, this can be specified as the fourth argument, otherwise it should
5610be set to 0 or 1.
5611</p>
5612</div>
5613
5614
5615<!-- _______________________________________________________________________ -->
5616<div class="doc_subsubsection">
5617 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
5618</div>
5619
5620<div class="doc_text">
5621
5622<h5>Syntax:</h5>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5623<p>This is an overloaded intrinsic. You can use llvm.memset on any integer bit
5624width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5625<pre>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5626 declare void @llvm.memset.i8(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5627 i8 &lt;len&gt;, i32 &lt;align&gt;)
5628 declare void @llvm.memset.i16(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5629 i16 &lt;len&gt;, i32 &lt;align&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5630 declare void @llvm.memset.i32(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5631 i32 &lt;len&gt;, i32 &lt;align&gt;)
5632 declare void @llvm.memset.i64(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5633 i64 &lt;len&gt;, i32 &lt;align&gt;)
5634</pre>
5635
5636<h5>Overview:</h5>
5637
5638<p>
5639The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a particular
5640byte value.
5641</p>
5642
5643<p>
5644Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
5645does not return a value, and takes an extra alignment argument.
5646</p>
5647
5648<h5>Arguments:</h5>
5649
5650<p>
5651The first argument is a pointer to the destination to fill, the second is the
5652byte value to fill it with, the third argument is an integer
5653argument specifying the number of bytes to fill, and the fourth argument is the
5654known alignment of destination location.
5655</p>
5656
5657<p>
5658If the call to this intrinisic has an alignment value that is not 0 or 1, then
5659the caller guarantees that the destination pointer is aligned to that boundary.
5660</p>
5661
5662<h5>Semantics:</h5>
5663
5664<p>
5665The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting at
5666the
5667destination location. If the argument is known to be aligned to some boundary,
5668this can be specified as the fourth argument, otherwise it should be set to 0 or
56691.
5670</p>
5671</div>
5672
5673
5674<!-- _______________________________________________________________________ -->
5675<div class="doc_subsubsection">
5676 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
5677</div>
5678
5679<div class="doc_text">
5680
5681<h5>Syntax:</h5>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5682<p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5683floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5684types however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5685<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5686 declare float @llvm.sqrt.f32(float %Val)
5687 declare double @llvm.sqrt.f64(double %Val)
5688 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
5689 declare fp128 @llvm.sqrt.f128(fp128 %Val)
5690 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5691</pre>
5692
5693<h5>Overview:</h5>
5694
5695<p>
5696The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5697returning the same value as the libm '<tt>sqrt</tt>' functions would. Unlike
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5698<tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
Chris Lattnerf914a002008-01-29 07:00:44 +0000[diff] [blame]5699negative numbers other than -0.0 (which allows for better optimization, because
5700there is no need to worry about errno being set). <tt>llvm.sqrt(-0.0)</tt> is
5701defined to return -0.0 like IEEE sqrt.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5702</p>
5703
5704<h5>Arguments:</h5>
5705
5706<p>
5707The argument and return value are floating point numbers of the same type.
5708</p>
5709
5710<h5>Semantics:</h5>
5711
5712<p>
5713This function returns the sqrt of the specified operand if it is a nonnegative
5714floating point number.
5715</p>
5716</div>
5717
5718<!-- _______________________________________________________________________ -->
5719<div class="doc_subsubsection">
5720 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
5721</div>
5722
5723<div class="doc_text">
5724
5725<h5>Syntax:</h5>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5726<p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5727floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5728types however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5729<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5730 declare float @llvm.powi.f32(float %Val, i32 %power)
5731 declare double @llvm.powi.f64(double %Val, i32 %power)
5732 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
5733 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
5734 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5735</pre>
5736
5737<h5>Overview:</h5>
5738
5739<p>
5740The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
5741specified (positive or negative) power. The order of evaluation of
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5742multiplications is not defined. When a vector of floating point type is
5743used, the second argument remains a scalar integer value.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5744</p>
5745
5746<h5>Arguments:</h5>
5747
5748<p>
5749The second argument is an integer power, and the first is a value to raise to
5750that power.
5751</p>
5752
5753<h5>Semantics:</h5>
5754
5755<p>
5756This function returns the first value raised to the second power with an
5757unspecified sequence of rounding operations.</p>
5758</div>
5759
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5760<!-- _______________________________________________________________________ -->
5761<div class="doc_subsubsection">
5762 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
5763</div>
5764
5765<div class="doc_text">
5766
5767<h5>Syntax:</h5>
5768<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
5769floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5770types however.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5771<pre>
5772 declare float @llvm.sin.f32(float %Val)
5773 declare double @llvm.sin.f64(double %Val)
5774 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
5775 declare fp128 @llvm.sin.f128(fp128 %Val)
5776 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
5777</pre>
5778
5779<h5>Overview:</h5>
5780
5781<p>
5782The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.
5783</p>
5784
5785<h5>Arguments:</h5>
5786
5787<p>
5788The argument and return value are floating point numbers of the same type.
5789</p>
5790
5791<h5>Semantics:</h5>
5792
5793<p>
5794This function returns the sine of the specified operand, returning the
5795same values as the libm <tt>sin</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5796conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5797</div>
5798
5799<!-- _______________________________________________________________________ -->
5800<div class="doc_subsubsection">
5801 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
5802</div>
5803
5804<div class="doc_text">
5805
5806<h5>Syntax:</h5>
5807<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
5808floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5809types however.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5810<pre>
5811 declare float @llvm.cos.f32(float %Val)
5812 declare double @llvm.cos.f64(double %Val)
5813 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
5814 declare fp128 @llvm.cos.f128(fp128 %Val)
5815 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
5816</pre>
5817
5818<h5>Overview:</h5>
5819
5820<p>
5821The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.
5822</p>
5823
5824<h5>Arguments:</h5>
5825
5826<p>
5827The argument and return value are floating point numbers of the same type.
5828</p>
5829
5830<h5>Semantics:</h5>
5831
5832<p>
5833This function returns the cosine of the specified operand, returning the
5834same values as the libm <tt>cos</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5835conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5836</div>
5837
5838<!-- _______________________________________________________________________ -->
5839<div class="doc_subsubsection">
5840 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
5841</div>
5842
5843<div class="doc_text">
5844
5845<h5>Syntax:</h5>
5846<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
5847floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5848types however.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5849<pre>
5850 declare float @llvm.pow.f32(float %Val, float %Power)
5851 declare double @llvm.pow.f64(double %Val, double %Power)
5852 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
5853 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
5854 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
5855</pre>
5856
5857<h5>Overview:</h5>
5858
5859<p>
5860The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
5861specified (positive or negative) power.
5862</p>
5863
5864<h5>Arguments:</h5>
5865
5866<p>
5867The second argument is a floating point power, and the first is a value to
5868raise to that power.
5869</p>
5870
5871<h5>Semantics:</h5>
5872
5873<p>
5874This function returns the first value raised to the second power,
5875returning the
5876same values as the libm <tt>pow</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5877conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5878</div>
5879
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5880
5881<!-- ======================================================================= -->
5882<div class="doc_subsection">
5883 <a name="int_manip">Bit Manipulation Intrinsics</a>
5884</div>
5885
5886<div class="doc_text">
5887<p>
5888LLVM provides intrinsics for a few important bit manipulation operations.
5889These allow efficient code generation for some algorithms.
5890</p>
5891
5892</div>
5893
5894<!-- _______________________________________________________________________ -->
5895<div class="doc_subsubsection">
5896 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
5897</div>
5898
5899<div class="doc_text">
5900
5901<h5>Syntax:</h5>
5902<p>This is an overloaded intrinsic function. You can use bswap on any integer
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5903type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5904<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5905 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
5906 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
5907 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5908</pre>
5909
5910<h5>Overview:</h5>
5911
5912<p>
5913The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
5914values with an even number of bytes (positive multiple of 16 bits). These are
5915useful for performing operations on data that is not in the target's native
5916byte order.
5917</p>
5918
5919<h5>Semantics:</h5>
5920
5921<p>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5922The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5923and low byte of the input i16 swapped. Similarly, the <tt>llvm.bswap.i32</tt>
5924intrinsic returns an i32 value that has the four bytes of the input i32
5925swapped, so that if the input bytes are numbered 0, 1, 2, 3 then the returned
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5926i32 will have its bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i48</tt>,
5927<tt>llvm.bswap.i64</tt> and other intrinsics extend this concept to
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5928additional even-byte lengths (6 bytes, 8 bytes and more, respectively).
5929</p>
5930
5931</div>
5932
5933<!-- _______________________________________________________________________ -->
5934<div class="doc_subsubsection">
5935 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
5936</div>
5937
5938<div class="doc_text">
5939
5940<h5>Syntax:</h5>
5941<p>This is an overloaded intrinsic. You can use llvm.ctpop on any integer bit
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5942width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5943<pre>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5944 declare i8 @llvm.ctpop.i8(i8 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5945 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5946 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5947 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
5948 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5949</pre>
5950
5951<h5>Overview:</h5>
5952
5953<p>
5954The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
5955value.
5956</p>
5957
5958<h5>Arguments:</h5>
5959
5960<p>
5961The only argument is the value to be counted. The argument may be of any
5962integer type. The return type must match the argument type.
5963</p>
5964
5965<h5>Semantics:</h5>
5966
5967<p>
5968The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
5969</p>
5970</div>
5971
5972<!-- _______________________________________________________________________ -->
5973<div class="doc_subsubsection">
5974 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
5975</div>
5976
5977<div class="doc_text">
5978
5979<h5>Syntax:</h5>
5980<p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5981integer bit width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5982<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5983 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
5984 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5985 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5986 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
5987 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5988</pre>
5989
5990<h5>Overview:</h5>
5991
5992<p>
5993The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
5994leading zeros in a variable.
5995</p>
5996
5997<h5>Arguments:</h5>
5998
5999<p>
6000The only argument is the value to be counted. The argument may be of any
6001integer type. The return type must match the argument type.
6002</p>
6003
6004<h5>Semantics:</h5>
6005
6006<p>
6007The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
6008in a variable. If the src == 0 then the result is the size in bits of the type
6009of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.
6010</p>
6011</div>
6012
6013
6014
6015<!-- _______________________________________________________________________ -->
6016<div class="doc_subsubsection">
6017 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
6018</div>
6019
6020<div class="doc_text">
6021
6022<h5>Syntax:</h5>
6023<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]6024integer bit width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6025<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]6026 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
6027 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6028 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]6029 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
6030 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6031</pre>
6032
6033<h5>Overview:</h5>
6034
6035<p>
6036The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
6037trailing zeros.
6038</p>
6039
6040<h5>Arguments:</h5>
6041
6042<p>
6043The only argument is the value to be counted. The argument may be of any
6044integer type. The return type must match the argument type.
6045</p>
6046
6047<h5>Semantics:</h5>
6048
6049<p>
6050The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
6051in a variable. If the src == 0 then the result is the size in bits of the type
6052of src. For example, <tt>llvm.cttz(2) = 1</tt>.
6053</p>
6054</div>
6055
6056<!-- _______________________________________________________________________ -->
6057<div class="doc_subsubsection">
6058 <a name="int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic</a>
6059</div>
6060
6061<div class="doc_text">
6062
6063<h5>Syntax:</h5>
6064<p>This is an overloaded intrinsic. You can use <tt>llvm.part.select</tt>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]6065on any integer bit width.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6066<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]6067 declare i17 @llvm.part.select.i17 (i17 %val, i32 %loBit, i32 %hiBit)
6068 declare i29 @llvm.part.select.i29 (i29 %val, i32 %loBit, i32 %hiBit)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6069</pre>
6070
6071<h5>Overview:</h5>
6072<p>The '<tt>llvm.part.select</tt>' family of intrinsic functions selects a
6073range of bits from an integer value and returns them in the same bit width as
6074the original value.</p>
6075
6076<h5>Arguments:</h5>
6077<p>The first argument, <tt>%val</tt> and the result may be integer types of
6078any bit width but they must have the same bit width. The second and third
6079arguments must be <tt>i32</tt> type since they specify only a bit index.</p>
6080
6081<h5>Semantics:</h5>
6082<p>The operation of the '<tt>llvm.part.select</tt>' intrinsic has two modes
6083of operation: forwards and reverse. If <tt>%loBit</tt> is greater than
6084<tt>%hiBits</tt> then the intrinsic operates in reverse mode. Otherwise it
6085operates in forward mode.</p>
6086<p>In forward mode, this intrinsic is the equivalent of shifting <tt>%val</tt>
6087right by <tt>%loBit</tt> bits and then ANDing it with a mask with
6088only the <tt>%hiBit - %loBit</tt> bits set, as follows:</p>
6089<ol>
6090 <li>The <tt>%val</tt> is shifted right (LSHR) by the number of bits specified
6091 by <tt>%loBits</tt>. This normalizes the value to the low order bits.</li>
6092 <li>The <tt>%loBits</tt> value is subtracted from the <tt>%hiBits</tt> value
6093 to determine the number of bits to retain.</li>
6094 <li>A mask of the retained bits is created by shifting a -1 value.</li>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]6095 <li>The mask is ANDed with <tt>%val</tt> to produce the result.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6096</ol>
6097<p>In reverse mode, a similar computation is made except that the bits are
6098returned in the reverse order. So, for example, if <tt>X</tt> has the value
6099<tt>i16 0x0ACF (101011001111)</tt> and we apply
6100<tt>part.select(i16 X, 8, 3)</tt> to it, we get back the value
6101<tt>i16 0x0026 (000000100110)</tt>.</p>
6102</div>
6103
6104<div class="doc_subsubsection">
6105 <a name="int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic</a>
6106</div>
6107
6108<div class="doc_text">
6109
6110<h5>Syntax:</h5>
6111<p>This is an overloaded intrinsic. You can use <tt>llvm.part.set</tt>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]6112on any integer bit width.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6113<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]6114 declare i17 @llvm.part.set.i17.i9 (i17 %val, i9 %repl, i32 %lo, i32 %hi)
6115 declare i29 @llvm.part.set.i29.i9 (i29 %val, i9 %repl, i32 %lo, i32 %hi)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6116</pre>
6117
6118<h5>Overview:</h5>
6119<p>The '<tt>llvm.part.set</tt>' family of intrinsic functions replaces a range
6120of bits in an integer value with another integer value. It returns the integer
6121with the replaced bits.</p>
6122
6123<h5>Arguments:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6124<p>The first argument, <tt>%val</tt>, and the result may be integer types of
6125any bit width, but they must have the same bit width. <tt>%val</tt> is the value
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6126whose bits will be replaced. The second argument, <tt>%repl</tt> may be an
6127integer of any bit width. The third and fourth arguments must be <tt>i32</tt>
6128type since they specify only a bit index.</p>
6129
6130<h5>Semantics:</h5>
6131<p>The operation of the '<tt>llvm.part.set</tt>' intrinsic has two modes
6132of operation: forwards and reverse. If <tt>%lo</tt> is greater than
6133<tt>%hi</tt> then the intrinsic operates in reverse mode. Otherwise it
6134operates in forward mode.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6135
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6136<p>For both modes, the <tt>%repl</tt> value is prepared for use by either
6137truncating it down to the size of the replacement area or zero extending it
6138up to that size.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6139
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6140<p>In forward mode, the bits between <tt>%lo</tt> and <tt>%hi</tt> (inclusive)
6141are replaced with corresponding bits from <tt>%repl</tt>. That is the 0th bit
6142in <tt>%repl</tt> replaces the <tt>%lo</tt>th bit in <tt>%val</tt> and etc. up
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]6143to the <tt>%hi</tt>th bit.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6144
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6145<p>In reverse mode, a similar computation is made except that the bits are
6146reversed. That is, the <tt>0</tt>th bit in <tt>%repl</tt> replaces the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]6147<tt>%hi</tt> bit in <tt>%val</tt> and etc. down to the <tt>%lo</tt>th bit.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6148
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6149<h5>Examples:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6150
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6151<pre>
6152 llvm.part.set(0xFFFF, 0, 4, 7) -&gt; 0xFF0F
6153 llvm.part.set(0xFFFF, 0, 7, 4) -&gt; 0xFF0F
6154 llvm.part.set(0xFFFF, 1, 7, 4) -&gt; 0xFF8F
6155 llvm.part.set(0xFFFF, F, 8, 3) -&gt; 0xFFE7
6156 llvm.part.set(0xFFFF, 0, 3, 8) -&gt; 0xFE07
6157</pre>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6158
6159</div>
6160
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6161<!-- ======================================================================= -->
6162<div class="doc_subsection">
6163 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
6164</div>
6165
6166<div class="doc_text">
6167<p>
6168LLVM provides intrinsics for some arithmetic with overflow operations.
6169</p>
6170
6171</div>
6172
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6173<!-- _______________________________________________________________________ -->
6174<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6175 <a name="int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6176</div>
6177
6178<div class="doc_text">
6179
6180<h5>Syntax:</h5>
6181
6182<p>This is an overloaded intrinsic. You can use <tt>llvm.sadd.with.overflow</tt>
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6183on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6184
6185<pre>
6186 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
6187 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6188 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
6189</pre>
6190
6191<h5>Overview:</h5>
6192
6193<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
6194a signed addition of the two arguments, and indicate whether an overflow
6195occurred during the signed summation.</p>
6196
6197<h5>Arguments:</h5>
6198
6199<p>The arguments (%a and %b) and the first element of the result structure may
6200be of integer types of any bit width, but they must have the same bit width. The
6201second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6202and <tt>%b</tt> are the two values that will undergo signed addition.</p>
6203
6204<h5>Semantics:</h5>
6205
6206<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
6207a signed addition of the two variables. They return a structure &mdash; the
6208first element of which is the signed summation, and the second element of which
6209is a bit specifying if the signed summation resulted in an overflow.</p>
6210
6211<h5>Examples:</h5>
6212<pre>
6213 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6214 %sum = extractvalue {i32, i1} %res, 0
6215 %obit = extractvalue {i32, i1} %res, 1
6216 br i1 %obit, label %overflow, label %normal
6217</pre>
6218
6219</div>
6220
6221<!-- _______________________________________________________________________ -->
6222<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6223 <a name="int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6224</div>
6225
6226<div class="doc_text">
6227
6228<h5>Syntax:</h5>
6229
6230<p>This is an overloaded intrinsic. You can use <tt>llvm.uadd.with.overflow</tt>
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6231on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6232
6233<pre>
6234 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
6235 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6236 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
6237</pre>
6238
6239<h5>Overview:</h5>
6240
6241<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
6242an unsigned addition of the two arguments, and indicate whether a carry occurred
6243during the unsigned summation.</p>
6244
6245<h5>Arguments:</h5>
6246
6247<p>The arguments (%a and %b) and the first element of the result structure may
6248be of integer types of any bit width, but they must have the same bit width. The
6249second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6250and <tt>%b</tt> are the two values that will undergo unsigned addition.</p>
6251
6252<h5>Semantics:</h5>
6253
6254<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
6255an unsigned addition of the two arguments. They return a structure &mdash; the
6256first element of which is the sum, and the second element of which is a bit
6257specifying if the unsigned summation resulted in a carry.</p>
6258
6259<h5>Examples:</h5>
6260<pre>
6261 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6262 %sum = extractvalue {i32, i1} %res, 0
6263 %obit = extractvalue {i32, i1} %res, 1
6264 br i1 %obit, label %carry, label %normal
6265</pre>
6266
6267</div>
6268
6269<!-- _______________________________________________________________________ -->
6270<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6271 <a name="int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6272</div>
6273
6274<div class="doc_text">
6275
6276<h5>Syntax:</h5>
6277
6278<p>This is an overloaded intrinsic. You can use <tt>llvm.ssub.with.overflow</tt>
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6279on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6280
6281<pre>
6282 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
6283 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6284 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
6285</pre>
6286
6287<h5>Overview:</h5>
6288
6289<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
6290a signed subtraction of the two arguments, and indicate whether an overflow
6291occurred during the signed subtraction.</p>
6292
6293<h5>Arguments:</h5>
6294
6295<p>The arguments (%a and %b) and the first element of the result structure may
6296be of integer types of any bit width, but they must have the same bit width. The
6297second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6298and <tt>%b</tt> are the two values that will undergo signed subtraction.</p>
6299
6300<h5>Semantics:</h5>
6301
6302<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
6303a signed subtraction of the two arguments. They return a structure &mdash; the
6304first element of which is the subtraction, and the second element of which is a bit
6305specifying if the signed subtraction resulted in an overflow.</p>
6306
6307<h5>Examples:</h5>
6308<pre>
6309 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6310 %sum = extractvalue {i32, i1} %res, 0
6311 %obit = extractvalue {i32, i1} %res, 1
6312 br i1 %obit, label %overflow, label %normal
6313</pre>
6314
6315</div>
6316
6317<!-- _______________________________________________________________________ -->
6318<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6319 <a name="int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6320</div>
6321
6322<div class="doc_text">
6323
6324<h5>Syntax:</h5>
6325
6326<p>This is an overloaded intrinsic. You can use <tt>llvm.usub.with.overflow</tt>
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6327on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6328
6329<pre>
6330 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
6331 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6332 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
6333</pre>
6334
6335<h5>Overview:</h5>
6336
6337<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
6338an unsigned subtraction of the two arguments, and indicate whether an overflow
6339occurred during the unsigned subtraction.</p>
6340
6341<h5>Arguments:</h5>
6342
6343<p>The arguments (%a and %b) and the first element of the result structure may
6344be of integer types of any bit width, but they must have the same bit width. The
6345second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6346and <tt>%b</tt> are the two values that will undergo unsigned subtraction.</p>
6347
6348<h5>Semantics:</h5>
6349
6350<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
6351an unsigned subtraction of the two arguments. They return a structure &mdash; the
6352first element of which is the subtraction, and the second element of which is a bit
6353specifying if the unsigned subtraction resulted in an overflow.</p>
6354
6355<h5>Examples:</h5>
6356<pre>
6357 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6358 %sum = extractvalue {i32, i1} %res, 0
6359 %obit = extractvalue {i32, i1} %res, 1
6360 br i1 %obit, label %overflow, label %normal
6361</pre>
6362
6363</div>
6364
6365<!-- _______________________________________________________________________ -->
6366<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6367 <a name="int_smul_overflow">'<tt>llvm.smul.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6368</div>
6369
6370<div class="doc_text">
6371
6372<h5>Syntax:</h5>
6373
6374<p>This is an overloaded intrinsic. You can use <tt>llvm.smul.with.overflow</tt>
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6375on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6376
6377<pre>
6378 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
6379 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6380 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
6381</pre>
6382
6383<h5>Overview:</h5>
6384
6385<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
6386a signed multiplication of the two arguments, and indicate whether an overflow
6387occurred during the signed multiplication.</p>
6388
6389<h5>Arguments:</h5>
6390
6391<p>The arguments (%a and %b) and the first element of the result structure may
6392be of integer types of any bit width, but they must have the same bit width. The
6393second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6394and <tt>%b</tt> are the two values that will undergo signed multiplication.</p>
6395
6396<h5>Semantics:</h5>
6397
6398<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
6399a signed multiplication of the two arguments. They return a structure &mdash;
6400the first element of which is the multiplication, and the second element of
6401which is a bit specifying if the signed multiplication resulted in an
6402overflow.</p>
6403
6404<h5>Examples:</h5>
6405<pre>
6406 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6407 %sum = extractvalue {i32, i1} %res, 0
6408 %obit = extractvalue {i32, i1} %res, 1
6409 br i1 %obit, label %overflow, label %normal
6410</pre>
6411
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6412</div>
6413
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]6414<!-- _______________________________________________________________________ -->
6415<div class="doc_subsubsection">
6416 <a name="int_umul_overflow">'<tt>llvm.umul.with.overflow.*</tt>' Intrinsics</a>
6417</div>
6418
6419<div class="doc_text">
6420
6421<h5>Syntax:</h5>
6422
6423<p>This is an overloaded intrinsic. You can use <tt>llvm.umul.with.overflow</tt>
6424on any integer bit width.</p>
6425
6426<pre>
6427 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
6428 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6429 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
6430</pre>
6431
6432<h5>Overview:</h5>
6433
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]6434<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
6435a unsigned multiplication of the two arguments, and indicate whether an overflow
6436occurred during the unsigned multiplication.</p>
6437
6438<h5>Arguments:</h5>
6439
6440<p>The arguments (%a and %b) and the first element of the result structure may
6441be of integer types of any bit width, but they must have the same bit width. The
6442second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6443and <tt>%b</tt> are the two values that will undergo unsigned
6444multiplication.</p>
6445
6446<h5>Semantics:</h5>
6447
6448<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
6449an unsigned multiplication of the two arguments. They return a structure &mdash;
6450the first element of which is the multiplication, and the second element of
6451which is a bit specifying if the unsigned multiplication resulted in an
6452overflow.</p>
6453
6454<h5>Examples:</h5>
6455<pre>
6456 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6457 %sum = extractvalue {i32, i1} %res, 0
6458 %obit = extractvalue {i32, i1} %res, 1
6459 br i1 %obit, label %overflow, label %normal
6460</pre>
6461
6462</div>
6463
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6464<!-- ======================================================================= -->
6465<div class="doc_subsection">
6466 <a name="int_debugger">Debugger Intrinsics</a>
6467</div>
6468
6469<div class="doc_text">
6470<p>
6471The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
6472are described in the <a
6473href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
6474Debugging</a> document.
6475</p>
6476</div>
6477
6478
6479<!-- ======================================================================= -->
6480<div class="doc_subsection">
6481 <a name="int_eh">Exception Handling Intrinsics</a>
6482</div>
6483
6484<div class="doc_text">
6485<p> The LLVM exception handling intrinsics (which all start with
6486<tt>llvm.eh.</tt> prefix), are described in the <a
6487href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
6488Handling</a> document. </p>
6489</div>
6490
6491<!-- ======================================================================= -->
6492<div class="doc_subsection">
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6493 <a name="int_trampoline">Trampoline Intrinsic</a>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6494</div>
6495
6496<div class="doc_text">
6497<p>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6498 This intrinsic makes it possible to excise one parameter, marked with
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6499 the <tt>nest</tt> attribute, from a function. The result is a callable
6500 function pointer lacking the nest parameter - the caller does not need
6501 to provide a value for it. Instead, the value to use is stored in
6502 advance in a "trampoline", a block of memory usually allocated
6503 on the stack, which also contains code to splice the nest value into the
6504 argument list. This is used to implement the GCC nested function address
6505 extension.
6506</p>
6507<p>
6508 For example, if the function is
6509 <tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
Bill Wendlinge262b4c2007-09-22 09:23:55 +0000[diff] [blame]6510 pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as follows:</p>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6511<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6512 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
6513 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
6514 %p = call i8* @llvm.init.trampoline( i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8*), i8* %nval )
6515 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6516</pre>
Bill Wendlinge262b4c2007-09-22 09:23:55 +0000[diff] [blame]6517 <p>The call <tt>%val = call i32 %fp( i32 %x, i32 %y )</tt> is then equivalent
6518 to <tt>%val = call i32 %f( i8* %nval, i32 %x, i32 %y )</tt>.</p>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6519</div>
6520
6521<!-- _______________________________________________________________________ -->
6522<div class="doc_subsubsection">
6523 <a name="int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a>
6524</div>
6525<div class="doc_text">
6526<h5>Syntax:</h5>
6527<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6528declare 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]6529</pre>
6530<h5>Overview:</h5>
6531<p>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6532 This fills the memory pointed to by <tt>tramp</tt> with code
6533 and returns a function pointer suitable for executing it.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6534</p>
6535<h5>Arguments:</h5>
6536<p>
6537 The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
6538 pointers. The <tt>tramp</tt> argument must point to a sufficiently large
6539 and sufficiently aligned block of memory; this memory is written to by the
Duncan Sands35012212007-08-22 23:39:54 +0000[diff] [blame]6540 intrinsic. Note that the size and the alignment are target-specific - LLVM
6541 currently provides no portable way of determining them, so a front-end that
6542 generates this intrinsic needs to have some target-specific knowledge.
6543 The <tt>func</tt> argument must hold a function bitcast to an <tt>i8*</tt>.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6544</p>
6545<h5>Semantics:</h5>
6546<p>
6547 The block of memory pointed to by <tt>tramp</tt> is filled with target
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6548 dependent code, turning it into a function. A pointer to this function is
6549 returned, but needs to be bitcast to an
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6550 <a href="#int_trampoline">appropriate function pointer type</a>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6551 before being called. The new function's signature is the same as that of
6552 <tt>func</tt> with any arguments marked with the <tt>nest</tt> attribute
6553 removed. At most one such <tt>nest</tt> argument is allowed, and it must be
6554 of pointer type. Calling the new function is equivalent to calling
6555 <tt>func</tt> with the same argument list, but with <tt>nval</tt> used for the
6556 missing <tt>nest</tt> argument. If, after calling
6557 <tt>llvm.init.trampoline</tt>, the memory pointed to by <tt>tramp</tt> is
6558 modified, then the effect of any later call to the returned function pointer is
6559 undefined.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6560</p>
6561</div>
6562
6563<!-- ======================================================================= -->
6564<div class="doc_subsection">
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6565 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
6566</div>
6567
6568<div class="doc_text">
6569<p>
6570 These intrinsic functions expand the "universal IR" of LLVM to represent
6571 hardware constructs for atomic operations and memory synchronization. This
6572 provides an interface to the hardware, not an interface to the programmer. It
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]6573 is aimed at a low enough level to allow any programming models or APIs
6574 (Application Programming Interfaces) which
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6575 need atomic behaviors to map cleanly onto it. It is also modeled primarily on
6576 hardware behavior. Just as hardware provides a "universal IR" for source
6577 languages, it also provides a starting point for developing a "universal"
6578 atomic operation and synchronization IR.
6579</p>
6580<p>
6581 These do <em>not</em> form an API such as high-level threading libraries,
6582 software transaction memory systems, atomic primitives, and intrinsic
6583 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
6584 application libraries. The hardware interface provided by LLVM should allow
6585 a clean implementation of all of these APIs and parallel programming models.
6586 No one model or paradigm should be selected above others unless the hardware
6587 itself ubiquitously does so.
6588
6589</p>
6590</div>
6591
6592<!-- _______________________________________________________________________ -->
6593<div class="doc_subsubsection">
6594 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
6595</div>
6596<div class="doc_text">
6597<h5>Syntax:</h5>
6598<pre>
6599declare void @llvm.memory.barrier( i1 &lt;ll&gt;, i1 &lt;ls&gt;, i1 &lt;sl&gt;, i1 &lt;ss&gt;,
6600i1 &lt;device&gt; )
6601
6602</pre>
6603<h5>Overview:</h5>
6604<p>
6605 The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
6606 specific pairs of memory access types.
6607</p>
6608<h5>Arguments:</h5>
6609<p>
6610 The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
6611 The first four arguments enables a specific barrier as listed below. The fith
6612 argument specifies that the barrier applies to io or device or uncached memory.
6613
6614</p>
6615 <ul>
6616 <li><tt>ll</tt>: load-load barrier</li>
6617 <li><tt>ls</tt>: load-store barrier</li>
6618 <li><tt>sl</tt>: store-load barrier</li>
6619 <li><tt>ss</tt>: store-store barrier</li>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]6620 <li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6621 </ul>
6622<h5>Semantics:</h5>
6623<p>
6624 This intrinsic causes the system to enforce some ordering constraints upon
6625 the loads and stores of the program. This barrier does not indicate
6626 <em>when</em> any events will occur, it only enforces an <em>order</em> in
6627 which they occur. For any of the specified pairs of load and store operations
6628 (f.ex. load-load, or store-load), all of the first operations preceding the
6629 barrier will complete before any of the second operations succeeding the
6630 barrier begin. Specifically the semantics for each pairing is as follows:
6631</p>
6632 <ul>
6633 <li><tt>ll</tt>: All loads before the barrier must complete before any load
6634 after the barrier begins.</li>
6635
6636 <li><tt>ls</tt>: All loads before the barrier must complete before any
6637 store after the barrier begins.</li>
6638 <li><tt>ss</tt>: All stores before the barrier must complete before any
6639 store after the barrier begins.</li>
6640 <li><tt>sl</tt>: All stores before the barrier must complete before any
6641 load after the barrier begins.</li>
6642 </ul>
6643<p>
6644 These semantics are applied with a logical "and" behavior when more than one
6645 is enabled in a single memory barrier intrinsic.
6646</p>
6647<p>
6648 Backends may implement stronger barriers than those requested when they do not
6649 support as fine grained a barrier as requested. Some architectures do not
6650 need all types of barriers and on such architectures, these become noops.
6651</p>
6652<h5>Example:</h5>
6653<pre>
6654%ptr = malloc i32
6655 store i32 4, %ptr
6656
6657%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
6658 call void @llvm.memory.barrier( i1 false, i1 true, i1 false, i1 false )
6659 <i>; guarantee the above finishes</i>
6660 store i32 8, %ptr <i>; before this begins</i>
6661</pre>
6662</div>
6663
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6664<!-- _______________________________________________________________________ -->
6665<div class="doc_subsubsection">
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6666 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6667</div>
6668<div class="doc_text">
6669<h5>Syntax:</h5>
6670<p>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6671 This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
6672 any integer bit width and for different address spaces. Not all targets
6673 support all bit widths however.</p>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6674
6675<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6676declare i8 @llvm.atomic.cmp.swap.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt; )
6677declare i16 @llvm.atomic.cmp.swap.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt; )
6678declare i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt; )
6679declare 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]6680
6681</pre>
6682<h5>Overview:</h5>
6683<p>
6684 This loads a value in memory and compares it to a given value. If they are
6685 equal, it stores a new value into the memory.
6686</p>
6687<h5>Arguments:</h5>
6688<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6689 The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result as
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6690 well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
6691 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
6692 this integer type. While any bit width integer may be used, targets may only
6693 lower representations they support in hardware.
6694
6695</p>
6696<h5>Semantics:</h5>
6697<p>
6698 This entire intrinsic must be executed atomically. It first loads the value
6699 in memory pointed to by <tt>ptr</tt> and compares it with the value
6700 <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the memory. The
6701 loaded value is yielded in all cases. This provides the equivalent of an
6702 atomic compare-and-swap operation within the SSA framework.
6703</p>
6704<h5>Examples:</h5>
6705
6706<pre>
6707%ptr = malloc i32
6708 store i32 4, %ptr
6709
6710%val1 = add i32 4, 4
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6711%result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 4, %val1 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6712 <i>; yields {i32}:result1 = 4</i>
6713%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
6714%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
6715
6716%val2 = add i32 1, 1
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6717%result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 5, %val2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6718 <i>; yields {i32}:result2 = 8</i>
6719%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
6720
6721%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
6722</pre>
6723</div>
6724
6725<!-- _______________________________________________________________________ -->
6726<div class="doc_subsubsection">
6727 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
6728</div>
6729<div class="doc_text">
6730<h5>Syntax:</h5>
6731
6732<p>
6733 This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
6734 integer bit width. Not all targets support all bit widths however.</p>
6735<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6736declare i8 @llvm.atomic.swap.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;val&gt; )
6737declare i16 @llvm.atomic.swap.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;val&gt; )
6738declare i32 @llvm.atomic.swap.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;val&gt; )
6739declare i64 @llvm.atomic.swap.i64.p0i64( i64* &lt;ptr&gt;, i64 &lt;val&gt; )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6740
6741</pre>
6742<h5>Overview:</h5>
6743<p>
6744 This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
6745 the value from memory. It then stores the value in <tt>val</tt> in the memory
6746 at <tt>ptr</tt>.
6747</p>
6748<h5>Arguments:</h5>
6749
6750<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6751 The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both the
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6752 <tt>val</tt> argument and the result must be integers of the same bit width.
6753 The first argument, <tt>ptr</tt>, must be a pointer to a value of this
6754 integer type. The targets may only lower integer representations they
6755 support.
6756</p>
6757<h5>Semantics:</h5>
6758<p>
6759 This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
6760 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
6761 equivalent of an atomic swap operation within the SSA framework.
6762
6763</p>
6764<h5>Examples:</h5>
6765<pre>
6766%ptr = malloc i32
6767 store i32 4, %ptr
6768
6769%val1 = add i32 4, 4
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6770%result1 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val1 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6771 <i>; yields {i32}:result1 = 4</i>
6772%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
6773%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
6774
6775%val2 = add i32 1, 1
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6776%result2 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6777 <i>; yields {i32}:result2 = 8</i>
6778
6779%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
6780%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
6781</pre>
6782</div>
6783
6784<!-- _______________________________________________________________________ -->
6785<div class="doc_subsubsection">
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6786 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6787
6788</div>
6789<div class="doc_text">
6790<h5>Syntax:</h5>
6791<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6792 This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.add</tt> on any
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6793 integer bit width. Not all targets support all bit widths however.</p>
6794<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6795declare i8 @llvm.atomic.load.add.i8..p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6796declare i16 @llvm.atomic.load.add.i16..p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6797declare i32 @llvm.atomic.load.add.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6798declare 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]6799
6800</pre>
6801<h5>Overview:</h5>
6802<p>
6803 This intrinsic adds <tt>delta</tt> to the value stored in memory at
6804 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
6805</p>
6806<h5>Arguments:</h5>
6807<p>
6808
6809 The intrinsic takes two arguments, the first a pointer to an integer value
6810 and the second an integer value. The result is also an integer value. These
6811 integer types can have any bit width, but they must all have the same bit
6812 width. The targets may only lower integer representations they support.
6813</p>
6814<h5>Semantics:</h5>
6815<p>
6816 This intrinsic does a series of operations atomically. It first loads the
6817 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
6818 to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
6819</p>
6820
6821<h5>Examples:</h5>
6822<pre>
6823%ptr = malloc i32
6824 store i32 4, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6825%result1 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 4 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6826 <i>; yields {i32}:result1 = 4</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6827%result2 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6828 <i>; yields {i32}:result2 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6829%result3 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 5 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6830 <i>; yields {i32}:result3 = 10</i>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6831%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6832</pre>
6833</div>
6834
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6835<!-- _______________________________________________________________________ -->
6836<div class="doc_subsubsection">
6837 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
6838
6839</div>
6840<div class="doc_text">
6841<h5>Syntax:</h5>
6842<p>
6843 This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.sub</tt> on
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6844 any integer bit width and for different address spaces. Not all targets
6845 support all bit widths however.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6846<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6847declare i8 @llvm.atomic.load.sub.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6848declare i16 @llvm.atomic.load.sub.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6849declare i32 @llvm.atomic.load.sub.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6850declare 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]6851
6852</pre>
6853<h5>Overview:</h5>
6854<p>
6855 This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
6856 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
6857</p>
6858<h5>Arguments:</h5>
6859<p>
6860
6861 The intrinsic takes two arguments, the first a pointer to an integer value
6862 and the second an integer value. The result is also an integer value. These
6863 integer types can have any bit width, but they must all have the same bit
6864 width. The targets may only lower integer representations they support.
6865</p>
6866<h5>Semantics:</h5>
6867<p>
6868 This intrinsic does a series of operations atomically. It first loads the
6869 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
6870 result to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
6871</p>
6872
6873<h5>Examples:</h5>
6874<pre>
6875%ptr = malloc i32
6876 store i32 8, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6877%result1 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 4 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6878 <i>; yields {i32}:result1 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6879%result2 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 2 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6880 <i>; yields {i32}:result2 = 4</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6881%result3 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 5 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6882 <i>; yields {i32}:result3 = 2</i>
6883%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
6884</pre>
6885</div>
6886
6887<!-- _______________________________________________________________________ -->
6888<div class="doc_subsubsection">
6889 <a name="int_atomic_load_and">'<tt>llvm.atomic.load.and.*</tt>' Intrinsic</a><br>
6890 <a name="int_atomic_load_nand">'<tt>llvm.atomic.load.nand.*</tt>' Intrinsic</a><br>
6891 <a name="int_atomic_load_or">'<tt>llvm.atomic.load.or.*</tt>' Intrinsic</a><br>
6892 <a name="int_atomic_load_xor">'<tt>llvm.atomic.load.xor.*</tt>' Intrinsic</a><br>
6893
6894</div>
6895<div class="doc_text">
6896<h5>Syntax:</h5>
6897<p>
6898 These are overloaded intrinsics. You can use <tt>llvm.atomic.load_and</tt>,
6899 <tt>llvm.atomic.load_nand</tt>, <tt>llvm.atomic.load_or</tt>, and
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6900 <tt>llvm.atomic.load_xor</tt> on any integer bit width and for different
6901 address spaces. Not all targets support all bit widths however.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6902<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6903declare i8 @llvm.atomic.load.and.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6904declare i16 @llvm.atomic.load.and.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6905declare i32 @llvm.atomic.load.and.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6906declare 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]6907
6908</pre>
6909
6910<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6911declare i8 @llvm.atomic.load.or.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6912declare i16 @llvm.atomic.load.or.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6913declare i32 @llvm.atomic.load.or.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6914declare 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]6915
6916</pre>
6917
6918<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6919declare i8 @llvm.atomic.load.nand.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6920declare i16 @llvm.atomic.load.nand.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6921declare i32 @llvm.atomic.load.nand.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6922declare 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]6923
6924</pre>
6925
6926<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6927declare i8 @llvm.atomic.load.xor.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6928declare i16 @llvm.atomic.load.xor.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6929declare i32 @llvm.atomic.load.xor.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6930declare 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]6931
6932</pre>
6933<h5>Overview:</h5>
6934<p>
6935 These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
6936 the value stored in memory at <tt>ptr</tt>. It yields the original value
6937 at <tt>ptr</tt>.
6938</p>
6939<h5>Arguments:</h5>
6940<p>
6941
6942 These intrinsics take two arguments, the first a pointer to an integer value
6943 and the second an integer value. The result is also an integer value. These
6944 integer types can have any bit width, but they must all have the same bit
6945 width. The targets may only lower integer representations they support.
6946</p>
6947<h5>Semantics:</h5>
6948<p>
6949 These intrinsics does a series of operations atomically. They first load the
6950 value stored at <tt>ptr</tt>. They then do the bitwise operation
6951 <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the original
6952 value stored at <tt>ptr</tt>.
6953</p>
6954
6955<h5>Examples:</h5>
6956<pre>
6957%ptr = malloc i32
6958 store i32 0x0F0F, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6959%result0 = call i32 @llvm.atomic.load.nand.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6960 <i>; yields {i32}:result0 = 0x0F0F</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6961%result1 = call i32 @llvm.atomic.load.and.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6962 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6963%result2 = call i32 @llvm.atomic.load.or.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6964 <i>; yields {i32}:result2 = 0xF0</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6965%result3 = call i32 @llvm.atomic.load.xor.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6966 <i>; yields {i32}:result3 = FF</i>
6967%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
6968</pre>
6969</div>
6970
6971
6972<!-- _______________________________________________________________________ -->
6973<div class="doc_subsubsection">
6974 <a name="int_atomic_load_max">'<tt>llvm.atomic.load.max.*</tt>' Intrinsic</a><br>
6975 <a name="int_atomic_load_min">'<tt>llvm.atomic.load.min.*</tt>' Intrinsic</a><br>
6976 <a name="int_atomic_load_umax">'<tt>llvm.atomic.load.umax.*</tt>' Intrinsic</a><br>
6977 <a name="int_atomic_load_umin">'<tt>llvm.atomic.load.umin.*</tt>' Intrinsic</a><br>
6978
6979</div>
6980<div class="doc_text">
6981<h5>Syntax:</h5>
6982<p>
6983 These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
6984 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6985 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
6986 address spaces. Not all targets
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6987 support all bit widths however.</p>
6988<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6989declare i8 @llvm.atomic.load.max.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6990declare i16 @llvm.atomic.load.max.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6991declare i32 @llvm.atomic.load.max.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6992declare 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]6993
6994</pre>
6995
6996<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6997declare i8 @llvm.atomic.load.min.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6998declare i16 @llvm.atomic.load.min.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6999declare i32 @llvm.atomic.load.min.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
7000declare 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]7001
7002</pre>
7003
7004<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7005declare i8 @llvm.atomic.load.umax.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
7006declare i16 @llvm.atomic.load.umax.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
7007declare i32 @llvm.atomic.load.umax.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
7008declare 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]7009
7010</pre>
7011
7012<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7013declare i8 @llvm.atomic.load.umin.i8..p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
7014declare i16 @llvm.atomic.load.umin.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
7015declare i32 @llvm.atomic.load.umin.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
7016declare 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]7017
7018</pre>
7019<h5>Overview:</h5>
7020<p>
7021 These intrinsics takes the signed or unsigned minimum or maximum of
7022 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
7023 original value at <tt>ptr</tt>.
7024</p>
7025<h5>Arguments:</h5>
7026<p>
7027
7028 These intrinsics take two arguments, the first a pointer to an integer value
7029 and the second an integer value. The result is also an integer value. These
7030 integer types can have any bit width, but they must all have the same bit
7031 width. The targets may only lower integer representations they support.
7032</p>
7033<h5>Semantics:</h5>
7034<p>
7035 These intrinsics does a series of operations atomically. They first load the
7036 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or max
7037 <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They yield
7038 the original value stored at <tt>ptr</tt>.
7039</p>
7040
7041<h5>Examples:</h5>
7042<pre>
7043%ptr = malloc i32
7044 store i32 7, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7045%result0 = call i32 @llvm.atomic.load.min.i32.p0i32( i32* %ptr, i32 -2 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7046 <i>; yields {i32}:result0 = 7</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7047%result1 = call i32 @llvm.atomic.load.max.i32.p0i32( i32* %ptr, i32 8 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7048 <i>; yields {i32}:result1 = -2</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7049%result2 = call i32 @llvm.atomic.load.umin.i32.p0i32( i32* %ptr, i32 10 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7050 <i>; yields {i32}:result2 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7051%result3 = call i32 @llvm.atomic.load.umax.i32.p0i32( i32* %ptr, i32 30 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7052 <i>; yields {i32}:result3 = 8</i>
7053%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
7054</pre>
7055</div>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]7056
7057<!-- ======================================================================= -->
7058<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7059 <a name="int_general">General Intrinsics</a>
7060</div>
7061
7062<div class="doc_text">
7063<p> This class of intrinsics is designed to be generic and has
7064no specific purpose. </p>
7065</div>
7066
7067<!-- _______________________________________________________________________ -->
7068<div class="doc_subsubsection">
7069 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
7070</div>
7071
7072<div class="doc_text">
7073
7074<h5>Syntax:</h5>
7075<pre>
7076 declare void @llvm.var.annotation(i8* &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
7077</pre>
7078
7079<h5>Overview:</h5>
7080
7081<p>
7082The '<tt>llvm.var.annotation</tt>' intrinsic
7083</p>
7084
7085<h5>Arguments:</h5>
7086
7087<p>
7088The first argument is a pointer to a value, the second is a pointer to a
7089global string, the third is a pointer to a global string which is the source
7090file name, and the last argument is the line number.
7091</p>
7092
7093<h5>Semantics:</h5>
7094
7095<p>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]7096This intrinsic allows annotation of local variables with arbitrary strings.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7097This can be useful for special purpose optimizations that want to look for these
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]7098annotations. These have no other defined use, they are ignored by code
7099generation and optimization.
7100</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7101</div>
7102
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7103<!-- _______________________________________________________________________ -->
7104<div class="doc_subsubsection">
Tanya Lattnerc9869b12007-09-21 23:57:59 +0000[diff] [blame]7105 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7106</div>
7107
7108<div class="doc_text">
7109
7110<h5>Syntax:</h5>
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]7111<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
7112any integer bit width.
7113</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7114<pre>
Tanya Lattner09161fe2007-09-22 00:03:01 +0000[diff] [blame]7115 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
7116 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
7117 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
7118 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
7119 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]7120</pre>
7121
7122<h5>Overview:</h5>
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]7123
7124<p>
7125The '<tt>llvm.annotation</tt>' intrinsic.
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7126</p>
7127
7128<h5>Arguments:</h5>
7129
7130<p>
7131The first argument is an integer value (result of some expression),
7132the second is a pointer to a global string, the third is a pointer to a global
7133string which is the source file name, and the last argument is the line number.
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]7134It returns the value of the first argument.
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7135</p>
7136
7137<h5>Semantics:</h5>
7138
7139<p>
7140This intrinsic allows annotations to be put on arbitrary expressions
7141with arbitrary strings. This can be useful for special purpose optimizations
7142that want to look for these annotations. These have no other defined use, they
7143are ignored by code generation and optimization.
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]7144</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7145</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7146
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]7147<!-- _______________________________________________________________________ -->
7148<div class="doc_subsubsection">
7149 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
7150</div>
7151
7152<div class="doc_text">
7153
7154<h5>Syntax:</h5>
7155<pre>
7156 declare void @llvm.trap()
7157</pre>
7158
7159<h5>Overview:</h5>
7160
7161<p>
7162The '<tt>llvm.trap</tt>' intrinsic
7163</p>
7164
7165<h5>Arguments:</h5>
7166
7167<p>
7168None
7169</p>
7170
7171<h5>Semantics:</h5>
7172
7173<p>
7174This intrinsics is lowered to the target dependent trap instruction. If the
7175target does not have a trap instruction, this intrinsic will be lowered to the
7176call of the abort() function.
7177</p>
7178</div>
7179
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]7180<!-- _______________________________________________________________________ -->
7181<div class="doc_subsubsection">
Misha Brukman5dd7f4d2008-11-22 23:55:29 +0000[diff] [blame]7182 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]7183</div>
7184<div class="doc_text">
7185<h5>Syntax:</h5>
7186<pre>
7187declare void @llvm.stackprotector( i8* &lt;guard&gt;, i8** &lt;slot&gt; )
7188
7189</pre>
7190<h5>Overview:</h5>
7191<p>
7192 The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and stores
7193 it onto the stack at <tt>slot</tt>. The stack slot is adjusted to ensure that
7194 it is placed on the stack before local variables.
7195</p>
7196<h5>Arguments:</h5>
7197<p>
7198 The <tt>llvm.stackprotector</tt> intrinsic requires two pointer arguments. The
7199 first argument is the value loaded from the stack guard
7200 <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt> that
7201 has enough space to hold the value of the guard.
7202</p>
7203<h5>Semantics:</h5>
7204<p>
7205 This intrinsic causes the prologue/epilogue inserter to force the position of
7206 the <tt>AllocaInst</tt> stack slot to be before local variables on the
7207 stack. This is to ensure that if a local variable on the stack is overwritten,
7208 it will destroy the value of the guard. When the function exits, the guard on
7209 the stack is checked against the original guard. If they're different, then
7210 the program aborts by calling the <tt>__stack_chk_fail()</tt> function.
7211</p>
7212</div>
7213
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7214<!-- *********************************************************************** -->
7215<hr>
7216<address>
7217 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman947321d2008-12-11 17:34:48 +0000[diff] [blame]7218 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7219 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman947321d2008-12-11 17:34:48 +0000[diff] [blame]7220 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]7221
7222 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
7223 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
7224 Last modified: $Date$
7225</address>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]7226
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7227</body>
7228</html>