Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / d88a97b4ace26da67dad6e87b555f9a0933e2339 / . / docs / LangRef.html
blob: c4255ba296e685e3780af115b8afe5cfd808a9d7 [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
Bill Wendlingbe9ec3f2008-11-26 19:07:40 +0000[diff] [blame]1094<p>If a function that has an <tt>ssp</tt> attribute is inlined into a function
1095that doesn't have an <tt>ssp</tt> attribute, then the resulting function will
1096have an <tt>ssp</tt> attribute.</p></dd>
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
1103<p>If a function that has an <tt>sspreq</tt> attribute is inlined into a
1104function that doesn't have an <tt>sspreq</tt> attribute or which has
1105an <tt>ssp</tt> attribute, then the resulting function will have
1106an <tt>sspreq</tt> attribute.</p></dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1107</dl>
1108
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1109</div>
1110
1111<!-- ======================================================================= -->
1112<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1113 <a name="moduleasm">Module-Level Inline Assembly</a>
1114</div>
1115
1116<div class="doc_text">
1117<p>
1118Modules may contain "module-level inline asm" blocks, which corresponds to the
1119GCC "file scope inline asm" blocks. These blocks are internally concatenated by
1120LLVM and treated as a single unit, but may be separated in the .ll file if
1121desired. The syntax is very simple:
1122</p>
1123
1124<div class="doc_code">
1125<pre>
1126module asm "inline asm code goes here"
1127module asm "more can go here"
1128</pre>
1129</div>
1130
1131<p>The strings can contain any character by escaping non-printable characters.
1132 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
1133 for the number.
1134</p>
1135
1136<p>
1137 The inline asm code is simply printed to the machine code .s file when
1138 assembly code is generated.
1139</p>
1140</div>
1141
1142<!-- ======================================================================= -->
1143<div class="doc_subsection">
1144 <a name="datalayout">Data Layout</a>
1145</div>
1146
1147<div class="doc_text">
1148<p>A module may specify a target specific data layout string that specifies how
1149data is to be laid out in memory. The syntax for the data layout is simply:</p>
1150<pre> target datalayout = "<i>layout specification</i>"</pre>
1151<p>The <i>layout specification</i> consists of a list of specifications
1152separated by the minus sign character ('-'). Each specification starts with a
1153letter and may include other information after the letter to define some
1154aspect of the data layout. The specifications accepted are as follows: </p>
1155<dl>
1156 <dt><tt>E</tt></dt>
1157 <dd>Specifies that the target lays out data in big-endian form. That is, the
1158 bits with the most significance have the lowest address location.</dd>
1159 <dt><tt>e</tt></dt>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1160 <dd>Specifies that the target lays out data in little-endian form. That is,
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1161 the bits with the least significance have the lowest address location.</dd>
1162 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1163 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
1164 <i>preferred</i> alignments. All sizes are in bits. Specifying the <i>pref</i>
1165 alignment is optional. If omitted, the preceding <tt>:</tt> should be omitted
1166 too.</dd>
1167 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1168 <dd>This specifies the alignment for an integer type of a given bit
1169 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1170 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1171 <dd>This specifies the alignment for a vector type of a given bit
1172 <i>size</i>.</dd>
1173 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1174 <dd>This specifies the alignment for a floating point type of a given bit
1175 <i>size</i>. The value of <i>size</i> must be either 32 (float) or 64
1176 (double).</dd>
1177 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1178 <dd>This specifies the alignment for an aggregate type of a given bit
1179 <i>size</i>.</dd>
Daniel Dunbard88a97b2009-06-08 22:17:53 +0000[diff] [blame^]1180 <dt><tt>s<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1181 <dd>This specifies the alignment for a stack object of a given bit
1182 <i>size</i>.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1183</dl>
1184<p>When constructing the data layout for a given target, LLVM starts with a
1185default set of specifications which are then (possibly) overriden by the
1186specifications in the <tt>datalayout</tt> keyword. The default specifications
1187are given in this list:</p>
1188<ul>
1189 <li><tt>E</tt> - big endian</li>
1190 <li><tt>p:32:64:64</tt> - 32-bit pointers with 64-bit alignment</li>
1191 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1192 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1193 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1194 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1195 <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]1196 alignment of 64-bits</li>
1197 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1198 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1199 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1200 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1201 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
Daniel Dunbard88a97b2009-06-08 22:17:53 +0000[diff] [blame^]1202 <li><tt>s0:64:64</tt> - stack objects are 64-bit aligned</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1203</ul>
Chris Lattner6a9f3c42008-08-05 18:21:08 +0000[diff] [blame]1204<p>When LLVM is determining the alignment for a given type, it uses the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]1205following rules:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1206<ol>
1207 <li>If the type sought is an exact match for one of the specifications, that
1208 specification is used.</li>
1209 <li>If no match is found, and the type sought is an integer type, then the
1210 smallest integer type that is larger than the bitwidth of the sought type is
1211 used. If none of the specifications are larger than the bitwidth then the the
1212 largest integer type is used. For example, given the default specifications
1213 above, the i7 type will use the alignment of i8 (next largest) while both
1214 i65 and i256 will use the alignment of i64 (largest specified).</li>
1215 <li>If no match is found, and the type sought is a vector type, then the
1216 largest vector type that is smaller than the sought vector type will be used
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]1217 as a fall back. This happens because &lt;128 x double&gt; can be implemented
1218 in terms of 64 &lt;2 x double&gt;, for example.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1219</ol>
1220</div>
1221
1222<!-- *********************************************************************** -->
1223<div class="doc_section"> <a name="typesystem">Type System</a> </div>
1224<!-- *********************************************************************** -->
1225
1226<div class="doc_text">
1227
1228<p>The LLVM type system is one of the most important features of the
1229intermediate representation. Being typed enables a number of
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1230optimizations to be performed on the intermediate representation directly,
1231without having to do
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1232extra analyses on the side before the transformation. A strong type
1233system makes it easier to read the generated code and enables novel
1234analyses and transformations that are not feasible to perform on normal
1235three address code representations.</p>
1236
1237</div>
1238
1239<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1240<div class="doc_subsection"> <a name="t_classifications">Type
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1241Classifications</a> </div>
1242<div class="doc_text">
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1243<p>The types fall into a few useful
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1244classifications:</p>
1245
1246<table border="1" cellspacing="0" cellpadding="4">
1247 <tbody>
1248 <tr><th>Classification</th><th>Types</th></tr>
1249 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1250 <td><a href="#t_integer">integer</a></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1251 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
1252 </tr>
1253 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1254 <td><a href="#t_floating">floating point</a></td>
1255 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1256 </tr>
1257 <tr>
1258 <td><a name="t_firstclass">first class</a></td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1259 <td><a href="#t_integer">integer</a>,
1260 <a href="#t_floating">floating point</a>,
1261 <a href="#t_pointer">pointer</a>,
Dan Gohmanf6237db2008-06-18 18:42:13 +0000[diff] [blame]1262 <a href="#t_vector">vector</a>,
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]1263 <a href="#t_struct">structure</a>,
1264 <a href="#t_array">array</a>,
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]1265 <a href="#t_label">label</a>,
1266 <a href="#t_metadata">metadata</a>.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1267 </td>
1268 </tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1269 <tr>
1270 <td><a href="#t_primitive">primitive</a></td>
1271 <td><a href="#t_label">label</a>,
1272 <a href="#t_void">void</a>,
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]1273 <a href="#t_floating">floating point</a>,
1274 <a href="#t_metadata">metadata</a>.</td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1275 </tr>
1276 <tr>
1277 <td><a href="#t_derived">derived</a></td>
1278 <td><a href="#t_integer">integer</a>,
1279 <a href="#t_array">array</a>,
1280 <a href="#t_function">function</a>,
1281 <a href="#t_pointer">pointer</a>,
1282 <a href="#t_struct">structure</a>,
1283 <a href="#t_pstruct">packed structure</a>,
1284 <a href="#t_vector">vector</a>,
1285 <a href="#t_opaque">opaque</a>.
Dan Gohman032ba852008-10-14 16:32:04 +0000[diff] [blame]1286 </td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1287 </tr>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1288 </tbody>
1289</table>
1290
1291<p>The <a href="#t_firstclass">first class</a> types are perhaps the
1292most important. Values of these types are the only ones which can be
1293produced by instructions, passed as arguments, or used as operands to
Dan Gohman4f29e422008-05-23 21:53:15 +0000[diff] [blame]1294instructions.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1295</div>
1296
1297<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1298<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1299
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1300<div class="doc_text">
1301<p>The primitive types are the fundamental building blocks of the LLVM
1302system.</p>
1303
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1304</div>
1305
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1306<!-- _______________________________________________________________________ -->
1307<div class="doc_subsubsection"> <a name="t_floating">Floating Point Types</a> </div>
1308
1309<div class="doc_text">
1310 <table>
1311 <tbody>
1312 <tr><th>Type</th><th>Description</th></tr>
1313 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1314 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1315 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1316 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1317 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1318 </tbody>
1319 </table>
1320</div>
1321
1322<!-- _______________________________________________________________________ -->
1323<div class="doc_subsubsection"> <a name="t_void">Void Type</a> </div>
1324
1325<div class="doc_text">
1326<h5>Overview:</h5>
1327<p>The void type does not represent any value and has no size.</p>
1328
1329<h5>Syntax:</h5>
1330
1331<pre>
1332 void
1333</pre>
1334</div>
1335
1336<!-- _______________________________________________________________________ -->
1337<div class="doc_subsubsection"> <a name="t_label">Label Type</a> </div>
1338
1339<div class="doc_text">
1340<h5>Overview:</h5>
1341<p>The label type represents code labels.</p>
1342
1343<h5>Syntax:</h5>
1344
1345<pre>
1346 label
1347</pre>
1348</div>
1349
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]1350<!-- _______________________________________________________________________ -->
1351<div class="doc_subsubsection"> <a name="t_metadata">Metadata Type</a> </div>
1352
1353<div class="doc_text">
1354<h5>Overview:</h5>
1355<p>The metadata type represents embedded metadata. The only derived type that
1356may contain metadata is <tt>metadata*</tt> or a function type that returns or
1357takes metadata typed parameters, but not pointer to metadata types.</p>
1358
1359<h5>Syntax:</h5>
1360
1361<pre>
1362 metadata
1363</pre>
1364</div>
1365
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1366
1367<!-- ======================================================================= -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1368<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
1369
1370<div class="doc_text">
1371
1372<p>The real power in LLVM comes from the derived types in the system.
1373This is what allows a programmer to represent arrays, functions,
1374pointers, and other useful types. Note that these derived types may be
1375recursive: For example, it is possible to have a two dimensional array.</p>
1376
1377</div>
1378
1379<!-- _______________________________________________________________________ -->
1380<div class="doc_subsubsection"> <a name="t_integer">Integer Type</a> </div>
1381
1382<div class="doc_text">
1383
1384<h5>Overview:</h5>
1385<p>The integer type is a very simple derived type that simply specifies an
1386arbitrary bit width for the integer type desired. Any bit width from 1 bit to
13872^23-1 (about 8 million) can be specified.</p>
1388
1389<h5>Syntax:</h5>
1390
1391<pre>
1392 iN
1393</pre>
1394
1395<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1396value.</p>
1397
1398<h5>Examples:</h5>
1399<table class="layout">
Nick Lewycky39382d62009-05-24 02:46:06 +0000[diff] [blame]1400 <tr class="layout">
1401 <td class="left"><tt>i1</tt></td>
1402 <td class="left">a single-bit integer.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1403 </tr>
Nick Lewycky39382d62009-05-24 02:46:06 +0000[diff] [blame]1404 <tr class="layout">
1405 <td class="left"><tt>i32</tt></td>
1406 <td class="left">a 32-bit integer.</td>
1407 </tr>
1408 <tr class="layout">
1409 <td class="left"><tt>i1942652</tt></td>
1410 <td class="left">a really big integer of over 1 million bits.</td>
1411 </tr>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1412</table>
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1413
1414<p>Note that the code generator does not yet support large integer types
1415to be used as function return types. The specific limit on how large a
1416return type the code generator can currently handle is target-dependent;
1417currently it's often 64 bits for 32-bit targets and 128 bits for 64-bit
1418targets.</p>
1419
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1420</div>
1421
1422<!-- _______________________________________________________________________ -->
1423<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
1424
1425<div class="doc_text">
1426
1427<h5>Overview:</h5>
1428
1429<p>The array type is a very simple derived type that arranges elements
1430sequentially in memory. The array type requires a size (number of
1431elements) and an underlying data type.</p>
1432
1433<h5>Syntax:</h5>
1434
1435<pre>
1436 [&lt;# elements&gt; x &lt;elementtype&gt;]
1437</pre>
1438
1439<p>The number of elements is a constant integer value; elementtype may
1440be any type with a size.</p>
1441
1442<h5>Examples:</h5>
1443<table class="layout">
1444 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1445 <td class="left"><tt>[40 x i32]</tt></td>
1446 <td class="left">Array of 40 32-bit integer values.</td>
1447 </tr>
1448 <tr class="layout">
1449 <td class="left"><tt>[41 x i32]</tt></td>
1450 <td class="left">Array of 41 32-bit integer values.</td>
1451 </tr>
1452 <tr class="layout">
1453 <td class="left"><tt>[4 x i8]</tt></td>
1454 <td class="left">Array of 4 8-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1455 </tr>
1456</table>
1457<p>Here are some examples of multidimensional arrays:</p>
1458<table class="layout">
1459 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1460 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1461 <td class="left">3x4 array of 32-bit integer values.</td>
1462 </tr>
1463 <tr class="layout">
1464 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1465 <td class="left">12x10 array of single precision floating point values.</td>
1466 </tr>
1467 <tr class="layout">
1468 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1469 <td class="left">2x3x4 array of 16-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1470 </tr>
1471</table>
1472
1473<p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
1474length array. Normally, accesses past the end of an array are undefined in
1475LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
1476As a special case, however, zero length arrays are recognized to be variable
1477length. This allows implementation of 'pascal style arrays' with the LLVM
1478type "{ i32, [0 x float]}", for example.</p>
1479
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1480<p>Note that the code generator does not yet support large aggregate types
1481to be used as function return types. The specific limit on how large an
1482aggregate return type the code generator can currently handle is
1483target-dependent, and also dependent on the aggregate element types.</p>
1484
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1485</div>
1486
1487<!-- _______________________________________________________________________ -->
1488<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
1489<div class="doc_text">
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1490
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1491<h5>Overview:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1492
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1493<p>The function type can be thought of as a function signature. It
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1494consists of a return type and a list of formal parameter types. The
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1495return 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]1496If 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]1497class types, and the struct must have at least one element.</p>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1498
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1499<h5>Syntax:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1500
1501<pre>
1502 &lt;returntype list&gt; (&lt;parameter list&gt;)
1503</pre>
1504
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1505<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
1506specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1507which indicates that the function takes a variable number of arguments.
1508Variable argument functions can access their arguments with the <a
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1509 href="#int_varargs">variable argument handling intrinsic</a> functions.
1510'<tt>&lt;returntype list&gt;</tt>' is a comma-separated list of
1511<a href="#t_firstclass">first class</a> type specifiers.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1512
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1513<h5>Examples:</h5>
1514<table class="layout">
1515 <tr class="layout">
1516 <td class="left"><tt>i32 (i32)</tt></td>
1517 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
1518 </td>
1519 </tr><tr class="layout">
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]1520 <td class="left"><tt>float&nbsp;(i16&nbsp;signext,&nbsp;i32&nbsp;*)&nbsp;*
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1521 </tt></td>
1522 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
1523 an <tt>i16</tt> that should be sign extended and a
1524 <a href="#t_pointer">pointer</a> to <tt>i32</tt>, returning
1525 <tt>float</tt>.
1526 </td>
1527 </tr><tr class="layout">
1528 <td class="left"><tt>i32 (i8*, ...)</tt></td>
1529 <td class="left">A vararg function that takes at least one
1530 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1531 which returns an integer. This is the signature for <tt>printf</tt> in
1532 LLVM.
1533 </td>
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1534 </tr><tr class="layout">
1535 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Misha Brukmanafc88b02008-11-27 06:41:20 +0000[diff] [blame]1536 <td class="left">A function taking an <tt>i32</tt>, returning two
1537 <tt>i32</tt> values as an aggregate of type <tt>{ i32, i32 }</tt>
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1538 </td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1539 </tr>
1540</table>
1541
1542</div>
1543<!-- _______________________________________________________________________ -->
1544<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
1545<div class="doc_text">
1546<h5>Overview:</h5>
1547<p>The structure type is used to represent a collection of data members
1548together in memory. The packing of the field types is defined to match
1549the ABI of the underlying processor. The elements of a structure may
1550be any type that has a size.</p>
1551<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1552and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1553field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1554instruction.</p>
1555<h5>Syntax:</h5>
1556<pre> { &lt;type list&gt; }<br></pre>
1557<h5>Examples:</h5>
1558<table class="layout">
1559 <tr class="layout">
1560 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1561 <td class="left">A triple of three <tt>i32</tt> values</td>
1562 </tr><tr class="layout">
1563 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1564 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1565 second element is a <a href="#t_pointer">pointer</a> to a
1566 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1567 an <tt>i32</tt>.</td>
1568 </tr>
1569</table>
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1570
1571<p>Note that the code generator does not yet support large aggregate types
1572to be used as function return types. The specific limit on how large an
1573aggregate return type the code generator can currently handle is
1574target-dependent, and also dependent on the aggregate element types.</p>
1575
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1576</div>
1577
1578<!-- _______________________________________________________________________ -->
1579<div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
1580</div>
1581<div class="doc_text">
1582<h5>Overview:</h5>
1583<p>The packed structure type is used to represent a collection of data members
1584together in memory. There is no padding between fields. Further, the alignment
1585of a packed structure is 1 byte. The elements of a packed structure may
1586be any type that has a size.</p>
1587<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1588and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1589field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1590instruction.</p>
1591<h5>Syntax:</h5>
1592<pre> &lt; { &lt;type list&gt; } &gt; <br></pre>
1593<h5>Examples:</h5>
1594<table class="layout">
1595 <tr class="layout">
1596 <td class="left"><tt>&lt; { i32, i32, i32 } &gt;</tt></td>
1597 <td class="left">A triple of three <tt>i32</tt> values</td>
1598 </tr><tr class="layout">
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1599 <td class="left">
1600<tt>&lt;&nbsp;{&nbsp;float,&nbsp;i32&nbsp;(i32)*&nbsp;}&nbsp;&gt;</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1601 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1602 second element is a <a href="#t_pointer">pointer</a> to a
1603 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1604 an <tt>i32</tt>.</td>
1605 </tr>
1606</table>
1607</div>
1608
1609<!-- _______________________________________________________________________ -->
1610<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
1611<div class="doc_text">
1612<h5>Overview:</h5>
1613<p>As in many languages, the pointer type represents a pointer or
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]1614reference to another object, which must live in memory. Pointer types may have
1615an optional address space attribute defining the target-specific numbered
1616address space where the pointed-to object resides. The default address space is
1617zero.</p>
Chris Lattner96edbd32009-02-08 19:53:29 +0000[diff] [blame]1618
1619<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]1620it permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</tt> instead.</p>
Chris Lattner96edbd32009-02-08 19:53:29 +0000[diff] [blame]1621
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1622<h5>Syntax:</h5>
1623<pre> &lt;type&gt; *<br></pre>
1624<h5>Examples:</h5>
1625<table class="layout">
1626 <tr class="layout">
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]1627 <td class="left"><tt>[4 x i32]*</tt></td>
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1628 <td class="left">A <a href="#t_pointer">pointer</a> to <a
1629 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
1630 </tr>
1631 <tr class="layout">
1632 <td class="left"><tt>i32 (i32 *) *</tt></td>
1633 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1634 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1635 <tt>i32</tt>.</td>
1636 </tr>
1637 <tr class="layout">
1638 <td class="left"><tt>i32 addrspace(5)*</tt></td>
1639 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
1640 that resides in address space #5.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1641 </tr>
1642</table>
1643</div>
1644
1645<!-- _______________________________________________________________________ -->
1646<div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
1647<div class="doc_text">
1648
1649<h5>Overview:</h5>
1650
1651<p>A vector type is a simple derived type that represents a vector
1652of elements. Vector types are used when multiple primitive data
1653are operated in parallel using a single instruction (SIMD).
1654A vector type requires a size (number of
1655elements) and an underlying primitive data type. Vectors must have a power
1656of two length (1, 2, 4, 8, 16 ...). Vector types are
1657considered <a href="#t_firstclass">first class</a>.</p>
1658
1659<h5>Syntax:</h5>
1660
1661<pre>
1662 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1663</pre>
1664
1665<p>The number of elements is a constant integer value; elementtype may
1666be any integer or floating point type.</p>
1667
1668<h5>Examples:</h5>
1669
1670<table class="layout">
1671 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1672 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
1673 <td class="left">Vector of 4 32-bit integer values.</td>
1674 </tr>
1675 <tr class="layout">
1676 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
1677 <td class="left">Vector of 8 32-bit floating-point values.</td>
1678 </tr>
1679 <tr class="layout">
1680 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
1681 <td class="left">Vector of 2 64-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1682 </tr>
1683</table>
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1684
1685<p>Note that the code generator does not yet support large vector types
1686to be used as function return types. The specific limit on how large a
1687vector return type codegen can currently handle is target-dependent;
1688currently it's often a few times longer than a hardware vector register.</p>
1689
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1690</div>
1691
1692<!-- _______________________________________________________________________ -->
1693<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1694<div class="doc_text">
1695
1696<h5>Overview:</h5>
1697
1698<p>Opaque types are used to represent unknown types in the system. This
Gordon Henriksenda0706e2007-10-14 00:34:53 +0000[diff] [blame]1699corresponds (for example) to the C notion of a forward declared structure type.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1700In LLVM, opaque types can eventually be resolved to any type (not just a
1701structure type).</p>
1702
1703<h5>Syntax:</h5>
1704
1705<pre>
1706 opaque
1707</pre>
1708
1709<h5>Examples:</h5>
1710
1711<table class="layout">
1712 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1713 <td class="left"><tt>opaque</tt></td>
1714 <td class="left">An opaque type.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1715 </tr>
1716</table>
1717</div>
1718
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]1719<!-- ======================================================================= -->
1720<div class="doc_subsection">
1721 <a name="t_uprefs">Type Up-references</a>
1722</div>
1723
1724<div class="doc_text">
1725<h5>Overview:</h5>
1726<p>
1727An "up reference" allows you to refer to a lexically enclosing type without
1728requiring it to have a name. For instance, a structure declaration may contain a
1729pointer to any of the types it is lexically a member of. Example of up
1730references (with their equivalent as named type declarations) include:</p>
1731
1732<pre>
Chris Lattner5ad632d2009-02-09 10:00:56 +0000[diff] [blame]1733 { \2 * } %x = type { %x* }
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]1734 { \2 }* %y = type { %y }*
1735 \1* %z = type %z*
1736</pre>
1737
1738<p>
1739An up reference is needed by the asmprinter for printing out cyclic types when
1740there is no declared name for a type in the cycle. Because the asmprinter does
1741not want to print out an infinite type string, it needs a syntax to handle
1742recursive types that have no names (all names are optional in llvm IR).
1743</p>
1744
1745<h5>Syntax:</h5>
1746<pre>
1747 \&lt;level&gt;
1748</pre>
1749
1750<p>
1751The level is the count of the lexical type that is being referred to.
1752</p>
1753
1754<h5>Examples:</h5>
1755
1756<table class="layout">
1757 <tr class="layout">
1758 <td class="left"><tt>\1*</tt></td>
1759 <td class="left">Self-referential pointer.</td>
1760 </tr>
1761 <tr class="layout">
1762 <td class="left"><tt>{ { \3*, i8 }, i32 }</tt></td>
1763 <td class="left">Recursive structure where the upref refers to the out-most
1764 structure.</td>
1765 </tr>
1766</table>
1767</div>
1768
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1769
1770<!-- *********************************************************************** -->
1771<div class="doc_section"> <a name="constants">Constants</a> </div>
1772<!-- *********************************************************************** -->
1773
1774<div class="doc_text">
1775
1776<p>LLVM has several different basic types of constants. This section describes
1777them all and their syntax.</p>
1778
1779</div>
1780
1781<!-- ======================================================================= -->
1782<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
1783
1784<div class="doc_text">
1785
1786<dl>
1787 <dt><b>Boolean constants</b></dt>
1788
1789 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
1790 constants of the <tt><a href="#t_primitive">i1</a></tt> type.
1791 </dd>
1792
1793 <dt><b>Integer constants</b></dt>
1794
1795 <dd>Standard integers (such as '4') are constants of the <a
1796 href="#t_integer">integer</a> type. Negative numbers may be used with
1797 integer types.
1798 </dd>
1799
1800 <dt><b>Floating point constants</b></dt>
1801
1802 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
1803 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]1804 notation (see below). The assembler requires the exact decimal value of
1805 a floating-point constant. For example, the assembler accepts 1.25 but
1806 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
1807 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1808
1809 <dt><b>Null pointer constants</b></dt>
1810
1811 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
1812 and must be of <a href="#t_pointer">pointer type</a>.</dd>
1813
1814</dl>
1815
Dale Johannesenf82a52f2009-02-11 22:14:51 +0000[diff] [blame]1816<p>The one non-intuitive notation for constants is the hexadecimal form
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1817of floating point constants. For example, the form '<tt>double
18180x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
18194.5e+15</tt>'. The only time hexadecimal floating point constants are required
1820(and the only time that they are generated by the disassembler) is when a
1821floating point constant must be emitted but it cannot be represented as a
Dale Johannesenf82a52f2009-02-11 22:14:51 +0000[diff] [blame]1822decimal floating point number in a reasonable number of digits. For example,
1823NaN's, infinities, and other
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1824special values are represented in their IEEE hexadecimal format so that
1825assembly and disassembly do not cause any bits to change in the constants.</p>
Dale Johannesenf82a52f2009-02-11 22:14:51 +0000[diff] [blame]1826<p>When using the hexadecimal form, constants of types float and double are
1827represented using the 16-digit form shown above (which matches the IEEE754
1828representation for double); float values must, however, be exactly representable
1829as IEE754 single precision.
1830Hexadecimal format is always used for long
1831double, and there are three forms of long double. The 80-bit
1832format used by x86 is represented as <tt>0xK</tt>
1833followed by 20 hexadecimal digits.
1834The 128-bit format used by PowerPC (two adjacent doubles) is represented
1835by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit
1836format is represented
1837by <tt>0xL</tt> followed by 32 hexadecimal digits; no currently supported
1838target uses this format. Long doubles will only work if they match
1839the long double format on your target. All hexadecimal formats are big-endian
1840(sign bit at the left).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1841</div>
1842
1843<!-- ======================================================================= -->
Chris Lattner97063852009-02-28 18:32:25 +0000[diff] [blame]1844<div class="doc_subsection">
1845<a name="aggregateconstants"> <!-- old anchor -->
1846<a name="complexconstants">Complex Constants</a></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1847</div>
1848
1849<div class="doc_text">
Chris Lattner97063852009-02-28 18:32:25 +0000[diff] [blame]1850<p>Complex constants are a (potentially recursive) combination of simple
1851constants and smaller complex constants.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1852
1853<dl>
1854 <dt><b>Structure constants</b></dt>
1855
1856 <dd>Structure constants are represented with notation similar to structure
1857 type definitions (a comma separated list of elements, surrounded by braces
Chris Lattner6c8de962007-12-25 20:34:52 +0000[diff] [blame]1858 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
1859 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]1860 must have <a href="#t_struct">structure type</a>, and the number and
1861 types of elements must match those specified by the type.
1862 </dd>
1863
1864 <dt><b>Array constants</b></dt>
1865
1866 <dd>Array constants are represented with notation similar to array type
1867 definitions (a comma separated list of elements, surrounded by square brackets
1868 (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74 ]</tt>". Array
1869 constants must have <a href="#t_array">array type</a>, and the number and
1870 types of elements must match those specified by the type.
1871 </dd>
1872
1873 <dt><b>Vector constants</b></dt>
1874
1875 <dd>Vector constants are represented with notation similar to vector type
1876 definitions (a comma separated list of elements, surrounded by
1877 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32 42,
1878 i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must have <a
1879 href="#t_vector">vector type</a>, and the number and types of elements must
1880 match those specified by the type.
1881 </dd>
1882
1883 <dt><b>Zero initialization</b></dt>
1884
1885 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1886 value to zero of <em>any</em> type, including scalar and aggregate types.
1887 This is often used to avoid having to print large zero initializers (e.g. for
1888 large arrays) and is always exactly equivalent to using explicit zero
1889 initializers.
1890 </dd>
Nick Lewycky4dcf8102009-04-04 07:22:01 +0000[diff] [blame]1891
1892 <dt><b>Metadata node</b></dt>
1893
Nick Lewyckyf122c7e2009-05-30 16:08:30 +0000[diff] [blame]1894 <dd>A metadata node is a structure-like constant with
1895 <a href="#t_metadata">metadata type</a>. For example:
1896 "<tt>metadata !{ i32 0, metadata !"test" }</tt>". Unlike other constants
1897 that are meant to be interpreted as part of the instruction stream, metadata
1898 is a place to attach additional information such as debug info.
Nick Lewycky4dcf8102009-04-04 07:22:01 +0000[diff] [blame]1899 </dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1900</dl>
1901
1902</div>
1903
1904<!-- ======================================================================= -->
1905<div class="doc_subsection">
1906 <a name="globalconstants">Global Variable and Function Addresses</a>
1907</div>
1908
1909<div class="doc_text">
1910
1911<p>The addresses of <a href="#globalvars">global variables</a> and <a
1912href="#functionstructure">functions</a> are always implicitly valid (link-time)
1913constants. These constants are explicitly referenced when the <a
1914href="#identifiers">identifier for the global</a> is used and always have <a
1915href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1916file:</p>
1917
1918<div class="doc_code">
1919<pre>
1920@X = global i32 17
1921@Y = global i32 42
1922@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
1923</pre>
1924</div>
1925
1926</div>
1927
1928<!-- ======================================================================= -->
1929<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
1930<div class="doc_text">
1931 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
1932 no specific value. Undefined values may be of any type and be used anywhere
1933 a constant is permitted.</p>
1934
1935 <p>Undefined values indicate to the compiler that the program is well defined
1936 no matter what value is used, giving the compiler more freedom to optimize.
1937 </p>
1938</div>
1939
1940<!-- ======================================================================= -->
1941<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1942</div>
1943
1944<div class="doc_text">
1945
1946<p>Constant expressions are used to allow expressions involving other constants
1947to be used as constants. Constant expressions may be of any <a
1948href="#t_firstclass">first class</a> type and may involve any LLVM operation
1949that does not have side effects (e.g. load and call are not supported). The
1950following is the syntax for constant expressions:</p>
1951
1952<dl>
1953 <dt><b><tt>trunc ( CST to TYPE )</tt></b></dt>
1954 <dd>Truncate a constant to another type. The bit size of CST must be larger
1955 than the bit size of TYPE. Both types must be integers.</dd>
1956
1957 <dt><b><tt>zext ( CST to TYPE )</tt></b></dt>
1958 <dd>Zero extend a constant to another type. The bit size of CST must be
1959 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1960
1961 <dt><b><tt>sext ( CST to TYPE )</tt></b></dt>
1962 <dd>Sign extend a constant to another type. The bit size of CST must be
1963 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1964
1965 <dt><b><tt>fptrunc ( CST to TYPE )</tt></b></dt>
1966 <dd>Truncate a floating point constant to another floating point type. The
1967 size of CST must be larger than the size of TYPE. Both types must be
1968 floating point.</dd>
1969
1970 <dt><b><tt>fpext ( CST to TYPE )</tt></b></dt>
1971 <dd>Floating point extend a constant to another type. The size of CST must be
1972 smaller or equal to the size of TYPE. Both types must be floating point.</dd>
1973
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]1974 <dt><b><tt>fptoui ( CST to TYPE )</tt></b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1975 <dd>Convert a floating point constant to the corresponding unsigned integer
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1976 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1977 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1978 of the same number of elements. If the value won't fit in the integer type,
1979 the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1980
1981 <dt><b><tt>fptosi ( CST to TYPE )</tt></b></dt>
1982 <dd>Convert a floating point constant to the corresponding signed integer
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1983 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1984 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1985 of the same number of elements. If the value won't fit in the integer type,
1986 the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1987
1988 <dt><b><tt>uitofp ( CST to TYPE )</tt></b></dt>
1989 <dd>Convert an unsigned integer constant to the corresponding floating point
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1990 constant. TYPE must be a scalar or vector floating point type. CST must be of
1991 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1992 of the same number of elements. If the value won't fit in the floating point
1993 type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1994
1995 <dt><b><tt>sitofp ( CST to TYPE )</tt></b></dt>
1996 <dd>Convert a signed integer constant to the corresponding floating point
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1997 constant. TYPE must be a scalar or vector floating point type. CST must be of
1998 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1999 of the same number of elements. If the value won't fit in the floating point
2000 type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2001
2002 <dt><b><tt>ptrtoint ( CST to TYPE )</tt></b></dt>
2003 <dd>Convert a pointer typed constant to the corresponding integer constant
2004 TYPE must be an integer type. CST must be of pointer type. The CST value is
2005 zero extended, truncated, or unchanged to make it fit in TYPE.</dd>
2006
2007 <dt><b><tt>inttoptr ( CST to TYPE )</tt></b></dt>
2008 <dd>Convert a integer constant to a pointer constant. TYPE must be a
2009 pointer type. CST must be of integer type. The CST value is zero extended,
2010 truncated, or unchanged to make it fit in a pointer size. This one is
2011 <i>really</i> dangerous!</dd>
2012
2013 <dt><b><tt>bitcast ( CST to TYPE )</tt></b></dt>
Chris Lattner557bc5d2009-02-28 18:27:03 +0000[diff] [blame]2014 <dd>Convert a constant, CST, to another TYPE. The constraints of the operands
2015 are the same as those for the <a href="#i_bitcast">bitcast
2016 instruction</a>.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2017
2018 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
2019
2020 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
2021 constants. As with the <a href="#i_getelementptr">getelementptr</a>
2022 instruction, the index list may have zero or more indexes, which are required
2023 to make sense for the type of "CSTPTR".</dd>
2024
2025 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
2026
2027 <dd>Perform the <a href="#i_select">select operation</a> on
2028 constants.</dd>
2029
2030 <dt><b><tt>icmp COND ( VAL1, VAL2 )</tt></b></dt>
2031 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
2032
2033 <dt><b><tt>fcmp COND ( VAL1, VAL2 )</tt></b></dt>
2034 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
2035
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]2036 <dt><b><tt>vicmp COND ( VAL1, VAL2 )</tt></b></dt>
2037 <dd>Performs the <a href="#i_vicmp">vicmp operation</a> on constants.</dd>
2038
2039 <dt><b><tt>vfcmp COND ( VAL1, VAL2 )</tt></b></dt>
2040 <dd>Performs the <a href="#i_vfcmp">vfcmp operation</a> on constants.</dd>
2041
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2042 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
2043
2044 <dd>Perform the <a href="#i_extractelement">extractelement
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]2045 operation</a> on constants.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2046
2047 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
2048
2049 <dd>Perform the <a href="#i_insertelement">insertelement
2050 operation</a> on constants.</dd>
2051
2052
2053 <dt><b><tt>shufflevector ( VEC1, VEC2, IDXMASK )</tt></b></dt>
2054
2055 <dd>Perform the <a href="#i_shufflevector">shufflevector
2056 operation</a> on constants.</dd>
2057
2058 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
2059
2060 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
2061 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
2062 binary</a> operations. The constraints on operands are the same as those for
2063 the corresponding instruction (e.g. no bitwise operations on floating point
2064 values are allowed).</dd>
2065</dl>
2066</div>
2067
Nick Lewycky4dcf8102009-04-04 07:22:01 +0000[diff] [blame]2068<!-- ======================================================================= -->
2069<div class="doc_subsection"><a name="metadata">Embedded Metadata</a>
2070</div>
2071
2072<div class="doc_text">
2073
2074<p>Embedded metadata provides a way to attach arbitrary data to the
2075instruction stream without affecting the behaviour of the program. There are
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]2076two metadata primitives, strings and nodes. All metadata has the
2077<tt>metadata</tt> type and is identified in syntax by a preceding exclamation
2078point ('<tt>!</tt>').
Nick Lewycky4dcf8102009-04-04 07:22:01 +0000[diff] [blame]2079</p>
2080
2081<p>A metadata string is a string surrounded by double quotes. It can contain
2082any character by escaping non-printable characters with "\xx" where "xx" is
2083the two digit hex code. For example: "<tt>!"test\00"</tt>".
2084</p>
2085
2086<p>Metadata nodes are represented with notation similar to structure constants
2087(a comma separated list of elements, surrounded by braces and preceeded by an
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]2088exclamation point). For example: "<tt>!{ metadata !"test\00", i32 10}</tt>".
Nick Lewycky4dcf8102009-04-04 07:22:01 +0000[diff] [blame]2089</p>
2090
Nick Lewycky117f4382009-05-10 20:57:05 +0000[diff] [blame]2091<p>A metadata node will attempt to track changes to the values it holds. In
2092the event that a value is deleted, it will be replaced with a typeless
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]2093"<tt>null</tt>", such as "<tt>metadata !{null, i32 10}</tt>".</p>
Nick Lewycky117f4382009-05-10 20:57:05 +0000[diff] [blame]2094
Nick Lewycky4dcf8102009-04-04 07:22:01 +0000[diff] [blame]2095<p>Optimizations may rely on metadata to provide additional information about
2096the program that isn't available in the instructions, or that isn't easily
2097computable. Similarly, the code generator may expect a certain metadata format
2098to be used to express debugging information.</p>
2099</div>
2100
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2101<!-- *********************************************************************** -->
2102<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
2103<!-- *********************************************************************** -->
2104
2105<!-- ======================================================================= -->
2106<div class="doc_subsection">
2107<a name="inlineasm">Inline Assembler Expressions</a>
2108</div>
2109
2110<div class="doc_text">
2111
2112<p>
2113LLVM supports inline assembler expressions (as opposed to <a href="#moduleasm">
2114Module-Level Inline Assembly</a>) through the use of a special value. This
2115value represents the inline assembler as a string (containing the instructions
2116to emit), a list of operand constraints (stored as a string), and a flag that
2117indicates whether or not the inline asm expression has side effects. An example
2118inline assembler expression is:
2119</p>
2120
2121<div class="doc_code">
2122<pre>
2123i32 (i32) asm "bswap $0", "=r,r"
2124</pre>
2125</div>
2126
2127<p>
2128Inline assembler expressions may <b>only</b> be used as the callee operand of
2129a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we have:
2130</p>
2131
2132<div class="doc_code">
2133<pre>
2134%X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
2135</pre>
2136</div>
2137
2138<p>
2139Inline asms with side effects not visible in the constraint list must be marked
2140as having side effects. This is done through the use of the
2141'<tt>sideeffect</tt>' keyword, like so:
2142</p>
2143
2144<div class="doc_code">
2145<pre>
2146call void asm sideeffect "eieio", ""()
2147</pre>
2148</div>
2149
2150<p>TODO: The format of the asm and constraints string still need to be
2151documented here. Constraints on what can be done (e.g. duplication, moving, etc
Chris Lattner1cdd64e2008-10-04 18:36:02 +0000[diff] [blame]2152need to be documented). This is probably best done by reference to another
2153document that covers inline asm from a holistic perspective.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2154</p>
2155
2156</div>
2157
2158<!-- *********************************************************************** -->
2159<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
2160<!-- *********************************************************************** -->
2161
2162<div class="doc_text">
2163
2164<p>The LLVM instruction set consists of several different
2165classifications of instructions: <a href="#terminators">terminator
2166instructions</a>, <a href="#binaryops">binary instructions</a>,
2167<a href="#bitwiseops">bitwise binary instructions</a>, <a
2168 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
2169instructions</a>.</p>
2170
2171</div>
2172
2173<!-- ======================================================================= -->
2174<div class="doc_subsection"> <a name="terminators">Terminator
2175Instructions</a> </div>
2176
2177<div class="doc_text">
2178
2179<p>As mentioned <a href="#functionstructure">previously</a>, every
2180basic block in a program ends with a "Terminator" instruction, which
2181indicates which block should be executed after the current block is
2182finished. These terminator instructions typically yield a '<tt>void</tt>'
2183value: they produce control flow, not values (the one exception being
2184the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
2185<p>There are six different terminator instructions: the '<a
2186 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
2187instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
2188the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
2189 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
2190 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
2191
2192</div>
2193
2194<!-- _______________________________________________________________________ -->
2195<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
2196Instruction</a> </div>
2197<div class="doc_text">
2198<h5>Syntax:</h5>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2199<pre>
2200 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]2201 ret void <i>; Return from void function</i>
2202</pre>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2203
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2204<h5>Overview:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2205
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2206<p>The '<tt>ret</tt>' instruction is used to return control flow (and
2207optionally a value) from a function back to the caller.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2208<p>There are two forms of the '<tt>ret</tt>' instruction: one that
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2209returns a value and then causes control flow, and one that just causes
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2210control flow to occur.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2211
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2212<h5>Arguments:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2213
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2214<p>The '<tt>ret</tt>' instruction optionally accepts a single argument,
2215the return value. The type of the return value must be a
2216'<a href="#t_firstclass">first class</a>' type.</p>
2217
2218<p>A function is not <a href="#wellformed">well formed</a> if
2219it it has a non-void return type and contains a '<tt>ret</tt>'
2220instruction with no return value or a return value with a type that
2221does not match its type, or if it has a void return type and contains
2222a '<tt>ret</tt>' instruction with a return value.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2223
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2224<h5>Semantics:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2225
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2226<p>When the '<tt>ret</tt>' instruction is executed, control flow
2227returns back to the calling function's context. If the caller is a "<a
2228 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
2229the instruction after the call. If the caller was an "<a
2230 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
2231at the beginning of the "normal" destination block. If the instruction
2232returns a value, that value shall set the call or invoke instruction's
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]2233return value.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2234
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2235<h5>Example:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2236
2237<pre>
2238 ret i32 5 <i>; Return an integer value of 5</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2239 ret void <i>; Return from a void function</i>
Bill Wendlingd163e2d2009-02-28 22:12:54 +0000[diff] [blame]2240 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]2241</pre>
Dan Gohman60967192009-01-12 23:12:39 +0000[diff] [blame]2242
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]2243<p>Note that the code generator does not yet fully support large
2244 return values. The specific sizes that are currently supported are
2245 dependent on the target. For integers, on 32-bit targets the limit
2246 is often 64 bits, and on 64-bit targets the limit is often 128 bits.
2247 For aggregate types, the current limits are dependent on the element
2248 types; for example targets are often limited to 2 total integer
2249 elements and 2 total floating-point elements.</p>
Dan Gohman60967192009-01-12 23:12:39 +0000[diff] [blame]2250
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2251</div>
2252<!-- _______________________________________________________________________ -->
2253<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
2254<div class="doc_text">
2255<h5>Syntax:</h5>
2256<pre> br i1 &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;<br> br label &lt;dest&gt; <i>; Unconditional branch</i>
2257</pre>
2258<h5>Overview:</h5>
2259<p>The '<tt>br</tt>' instruction is used to cause control flow to
2260transfer to a different basic block in the current function. There are
2261two forms of this instruction, corresponding to a conditional branch
2262and an unconditional branch.</p>
2263<h5>Arguments:</h5>
2264<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
2265single '<tt>i1</tt>' value and two '<tt>label</tt>' values. The
2266unconditional form of the '<tt>br</tt>' instruction takes a single
2267'<tt>label</tt>' value as a target.</p>
2268<h5>Semantics:</h5>
2269<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
2270argument is evaluated. If the value is <tt>true</tt>, control flows
2271to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
2272control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
2273<h5>Example:</h5>
Chris Lattner95127832009-05-09 18:11:50 +0000[diff] [blame]2274<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]2275 href="#i_ret">ret</a> i32 1<br>IfUnequal:<br> <a href="#i_ret">ret</a> i32 0<br></pre>
2276</div>
2277<!-- _______________________________________________________________________ -->
2278<div class="doc_subsubsection">
2279 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
2280</div>
2281
2282<div class="doc_text">
2283<h5>Syntax:</h5>
2284
2285<pre>
2286 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
2287</pre>
2288
2289<h5>Overview:</h5>
2290
2291<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
2292several different places. It is a generalization of the '<tt>br</tt>'
2293instruction, allowing a branch to occur to one of many possible
2294destinations.</p>
2295
2296
2297<h5>Arguments:</h5>
2298
2299<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
2300comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
2301an array of pairs of comparison value constants and '<tt>label</tt>'s. The
2302table is not allowed to contain duplicate constant entries.</p>
2303
2304<h5>Semantics:</h5>
2305
2306<p>The <tt>switch</tt> instruction specifies a table of values and
2307destinations. When the '<tt>switch</tt>' instruction is executed, this
2308table is searched for the given value. If the value is found, control flow is
2309transfered to the corresponding destination; otherwise, control flow is
2310transfered to the default destination.</p>
2311
2312<h5>Implementation:</h5>
2313
2314<p>Depending on properties of the target machine and the particular
2315<tt>switch</tt> instruction, this instruction may be code generated in different
2316ways. For example, it could be generated as a series of chained conditional
2317branches or with a lookup table.</p>
2318
2319<h5>Example:</h5>
2320
2321<pre>
2322 <i>; Emulate a conditional br instruction</i>
2323 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]2324 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2325
2326 <i>; Emulate an unconditional br instruction</i>
2327 switch i32 0, label %dest [ ]
2328
2329 <i>; Implement a jump table:</i>
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]2330 switch i32 %val, label %otherwise [ i32 0, label %onzero
2331 i32 1, label %onone
2332 i32 2, label %ontwo ]
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2333</pre>
2334</div>
2335
2336<!-- _______________________________________________________________________ -->
2337<div class="doc_subsubsection">
2338 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
2339</div>
2340
2341<div class="doc_text">
2342
2343<h5>Syntax:</h5>
2344
2345<pre>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]2346 &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]2347 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
2348</pre>
2349
2350<h5>Overview:</h5>
2351
2352<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
2353function, with the possibility of control flow transfer to either the
2354'<tt>normal</tt>' label or the
2355'<tt>exception</tt>' label. If the callee function returns with the
2356"<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
2357"normal" label. If the callee (or any indirect callees) returns with the "<a
2358href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]2359continued at the dynamically nearest "exception" label.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2360
2361<h5>Arguments:</h5>
2362
2363<p>This instruction requires several arguments:</p>
2364
2365<ol>
2366 <li>
2367 The optional "cconv" marker indicates which <a href="#callingconv">calling
2368 convention</a> the call should use. If none is specified, the call defaults
2369 to using C calling conventions.
2370 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]2371
2372 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
2373 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>',
2374 and '<tt>inreg</tt>' attributes are valid here.</li>
2375
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2376 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
2377 function value being invoked. In most cases, this is a direct function
2378 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
2379 an arbitrary pointer to function value.
2380 </li>
2381
2382 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
2383 function to be invoked. </li>
2384
2385 <li>'<tt>function args</tt>': argument list whose types match the function
2386 signature argument types. If the function signature indicates the function
2387 accepts a variable number of arguments, the extra arguments can be
2388 specified. </li>
2389
2390 <li>'<tt>normal label</tt>': the label reached when the called function
2391 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
2392
2393 <li>'<tt>exception label</tt>': the label reached when a callee returns with
2394 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
2395
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]2396 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]2397 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
2398 '<tt>readnone</tt>' attributes are valid here.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2399</ol>
2400
2401<h5>Semantics:</h5>
2402
2403<p>This instruction is designed to operate as a standard '<tt><a
2404href="#i_call">call</a></tt>' instruction in most regards. The primary
2405difference is that it establishes an association with a label, which is used by
2406the runtime library to unwind the stack.</p>
2407
2408<p>This instruction is used in languages with destructors to ensure that proper
2409cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
2410exception. Additionally, this is important for implementation of
2411'<tt>catch</tt>' clauses in high-level languages that support them.</p>
2412
Jay Foad8e2fd2c2009-06-03 10:20:10 +0000[diff] [blame]2413<p>For the purposes of the SSA form, the definition of the value
2414returned by the '<tt>invoke</tt>' instruction is deemed to occur on
2415the edge from the current block to the "normal" label. If the callee
2416unwinds then no return value is available.</p>
Dan Gohman140ba5d2009-05-22 21:47:08 +0000[diff] [blame]2417
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2418<h5>Example:</h5>
2419<pre>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2420 %retval = invoke i32 @Test(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2421 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2422 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2423 unwind label %TestCleanup <i>; {i32}:retval set</i>
2424</pre>
2425</div>
2426
2427
2428<!-- _______________________________________________________________________ -->
2429
2430<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
2431Instruction</a> </div>
2432
2433<div class="doc_text">
2434
2435<h5>Syntax:</h5>
2436<pre>
2437 unwind
2438</pre>
2439
2440<h5>Overview:</h5>
2441
2442<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
2443at the first callee in the dynamic call stack which used an <a
2444href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
2445primarily used to implement exception handling.</p>
2446
2447<h5>Semantics:</h5>
2448
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]2449<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2450immediately halt. The dynamic call stack is then searched for the first <a
2451href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
2452execution continues at the "exceptional" destination block specified by the
2453<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
2454dynamic call chain, undefined behavior results.</p>
2455</div>
2456
2457<!-- _______________________________________________________________________ -->
2458
2459<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
2460Instruction</a> </div>
2461
2462<div class="doc_text">
2463
2464<h5>Syntax:</h5>
2465<pre>
2466 unreachable
2467</pre>
2468
2469<h5>Overview:</h5>
2470
2471<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
2472instruction is used to inform the optimizer that a particular portion of the
2473code is not reachable. This can be used to indicate that the code after a
2474no-return function cannot be reached, and other facts.</p>
2475
2476<h5>Semantics:</h5>
2477
2478<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
2479</div>
2480
2481
2482
2483<!-- ======================================================================= -->
2484<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
2485<div class="doc_text">
2486<p>Binary operators are used to do most of the computation in a
Chris Lattnerab596d92008-04-01 18:47:32 +0000[diff] [blame]2487program. They require two operands of the same type, execute an operation on them, and
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2488produce a single value. The operands might represent
2489multiple 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]2490The result value has the same type as its operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2491<p>There are several different binary operators:</p>
2492</div>
2493<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2494<div class="doc_subsubsection">
2495 <a name="i_add">'<tt>add</tt>' Instruction</a>
2496</div>
2497
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2498<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2499
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2500<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2501
2502<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2503 &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]2504</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2505
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2506<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2507
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2508<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2509
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2510<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2511
2512<p>The two arguments to the '<tt>add</tt>' instruction must be <a
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2513 href="#t_integer">integer</a> or
2514 <a href="#t_vector">vector</a> of integer values. Both arguments must
2515 have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2516
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2517<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2518
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2519<p>The value produced is the integer sum of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2520
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2521<p>If the sum has unsigned overflow, the result returned is the
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2522mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2523the result.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2524
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2525<p>Because LLVM integers use a two's complement representation, this
2526instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2527
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2528<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2529
2530<pre>
2531 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2532</pre>
2533</div>
2534<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2535<div class="doc_subsubsection">
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2536 <a name="i_fadd">'<tt>fadd</tt>' Instruction</a>
2537</div>
2538
2539<div class="doc_text">
2540
2541<h5>Syntax:</h5>
2542
2543<pre>
2544 &lt;result&gt; = fadd &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
2545</pre>
2546
2547<h5>Overview:</h5>
2548
2549<p>The '<tt>fadd</tt>' instruction returns the sum of its two operands.</p>
2550
2551<h5>Arguments:</h5>
2552
2553<p>The two arguments to the '<tt>fadd</tt>' instruction must be
2554<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
2555floating point values. Both arguments must have identical types.</p>
2556
2557<h5>Semantics:</h5>
2558
2559<p>The value produced is the floating point sum of the two operands.</p>
2560
2561<h5>Example:</h5>
2562
2563<pre>
2564 &lt;result&gt; = fadd float 4.0, %var <i>; yields {float}:result = 4.0 + %var</i>
2565</pre>
2566</div>
2567<!-- _______________________________________________________________________ -->
2568<div class="doc_subsubsection">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2569 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
2570</div>
2571
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2572<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2573
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2574<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2575
2576<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2577 &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]2578</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2579
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2580<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2581
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2582<p>The '<tt>sub</tt>' instruction returns the difference of its two
2583operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2584
2585<p>Note that the '<tt>sub</tt>' instruction is used to represent the
2586'<tt>neg</tt>' instruction present in most other intermediate
2587representations.</p>
2588
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2589<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2590
2591<p>The two arguments to the '<tt>sub</tt>' instruction must be <a
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2592 href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
2593 integer values. Both arguments must have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2594
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2595<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2596
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2597<p>The value produced is the integer difference of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2598
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2599<p>If the difference has unsigned overflow, the result returned is the
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2600mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2601the result.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2602
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2603<p>Because LLVM integers use a two's complement representation, this
2604instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2605
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2606<h5>Example:</h5>
2607<pre>
2608 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
2609 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
2610</pre>
2611</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2612
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2613<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2614<div class="doc_subsubsection">
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2615 <a name="i_fsub">'<tt>fsub</tt>' Instruction</a>
2616</div>
2617
2618<div class="doc_text">
2619
2620<h5>Syntax:</h5>
2621
2622<pre>
2623 &lt;result&gt; = fsub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
2624</pre>
2625
2626<h5>Overview:</h5>
2627
2628<p>The '<tt>fsub</tt>' instruction returns the difference of its two
2629operands.</p>
2630
2631<p>Note that the '<tt>fsub</tt>' instruction is used to represent the
2632'<tt>fneg</tt>' instruction present in most other intermediate
2633representations.</p>
2634
2635<h5>Arguments:</h5>
2636
2637<p>The two arguments to the '<tt>fsub</tt>' instruction must be <a
2638 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2639 of floating point values. Both arguments must have identical types.</p>
2640
2641<h5>Semantics:</h5>
2642
2643<p>The value produced is the floating point difference of the two operands.</p>
2644
2645<h5>Example:</h5>
2646<pre>
2647 &lt;result&gt; = fsub float 4.0, %var <i>; yields {float}:result = 4.0 - %var</i>
2648 &lt;result&gt; = fsub float -0.0, %val <i>; yields {float}:result = -%var</i>
2649</pre>
2650</div>
2651
2652<!-- _______________________________________________________________________ -->
2653<div class="doc_subsubsection">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2654 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
2655</div>
2656
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2657<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2658
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2659<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2660<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]2661</pre>
2662<h5>Overview:</h5>
2663<p>The '<tt>mul</tt>' instruction returns the product of its two
2664operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2665
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2666<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2667
2668<p>The two arguments to the '<tt>mul</tt>' instruction must be <a
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2669href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2670values. Both arguments must have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2671
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2672<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2673
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2674<p>The value produced is the integer product of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2675
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2676<p>If the result of the multiplication has unsigned overflow,
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2677the result returned is the mathematical result modulo
26782<sup>n</sup>, where n is the bit width of the result.</p>
2679<p>Because LLVM integers use a two's complement representation, and the
2680result is the same width as the operands, this instruction returns the
2681correct result for both signed and unsigned integers. If a full product
2682(e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands
2683should be sign-extended or zero-extended as appropriate to the
2684width of the full product.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2685<h5>Example:</h5>
2686<pre> &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
2687</pre>
2688</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2689
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2690<!-- _______________________________________________________________________ -->
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2691<div class="doc_subsubsection">
2692 <a name="i_fmul">'<tt>fmul</tt>' Instruction</a>
2693</div>
2694
2695<div class="doc_text">
2696
2697<h5>Syntax:</h5>
2698<pre> &lt;result&gt; = fmul &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
2699</pre>
2700<h5>Overview:</h5>
2701<p>The '<tt>fmul</tt>' instruction returns the product of its two
2702operands.</p>
2703
2704<h5>Arguments:</h5>
2705
2706<p>The two arguments to the '<tt>fmul</tt>' instruction must be
2707<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2708of floating point values. Both arguments must have identical types.</p>
2709
2710<h5>Semantics:</h5>
2711
2712<p>The value produced is the floating point product of the two operands.</p>
2713
2714<h5>Example:</h5>
2715<pre> &lt;result&gt; = fmul float 4.0, %var <i>; yields {float}:result = 4.0 * %var</i>
2716</pre>
2717</div>
2718
2719<!-- _______________________________________________________________________ -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2720<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
2721</a></div>
2722<div class="doc_text">
2723<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2724<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]2725</pre>
2726<h5>Overview:</h5>
2727<p>The '<tt>udiv</tt>' instruction returns the quotient of its two
2728operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2729
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2730<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2731
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2732<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2733<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2734values. Both arguments must have identical types.</p>
2735
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2736<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2737
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2738<p>The value produced is the unsigned integer quotient of the two operands.</p>
2739<p>Note that unsigned integer division and signed integer division are distinct
2740operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
2741<p>Division by zero leads to undefined behavior.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2742<h5>Example:</h5>
2743<pre> &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2744</pre>
2745</div>
2746<!-- _______________________________________________________________________ -->
2747<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
2748</a> </div>
2749<div class="doc_text">
2750<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2751<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2752 &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]2753</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2754
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2755<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2756
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2757<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two
2758operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2759
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2760<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2761
2762<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
2763<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2764values. Both arguments must have identical types.</p>
2765
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2766<h5>Semantics:</h5>
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2767<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]2768<p>Note that signed integer division and unsigned integer division are distinct
2769operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
2770<p>Division by zero leads to undefined behavior. Overflow also leads to
2771undefined behavior; this is a rare case, but can occur, for example,
2772by doing a 32-bit division of -2147483648 by -1.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2773<h5>Example:</h5>
2774<pre> &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2775</pre>
2776</div>
2777<!-- _______________________________________________________________________ -->
2778<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
2779Instruction</a> </div>
2780<div class="doc_text">
2781<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2782<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2783 &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]2784</pre>
2785<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2786
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2787<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two
2788operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2789
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2790<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2791
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2792<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2793<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2794of floating point values. Both arguments must have identical types.</p>
2795
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2796<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2797
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2798<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2799
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2800<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2801
2802<pre>
2803 &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]2804</pre>
2805</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2806
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2807<!-- _______________________________________________________________________ -->
2808<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
2809</div>
2810<div class="doc_text">
2811<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2812<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]2813</pre>
2814<h5>Overview:</h5>
2815<p>The '<tt>urem</tt>' instruction returns the remainder from the
2816unsigned division of its two arguments.</p>
2817<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2818<p>The two arguments to the '<tt>urem</tt>' instruction must be
2819<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2820values. Both arguments must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2821<h5>Semantics:</h5>
2822<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2823This instruction always performs an unsigned division to get the remainder.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2824<p>Note that unsigned integer remainder and signed integer remainder are
2825distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
2826<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2827<h5>Example:</h5>
2828<pre> &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2829</pre>
2830
2831</div>
2832<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2833<div class="doc_subsubsection">
2834 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
2835</div>
2836
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2837<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2838
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2839<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2840
2841<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2842 &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]2843</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2844
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2845<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2846
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2847<p>The '<tt>srem</tt>' instruction returns the remainder from the
Dan Gohman3e3fd8c2007-11-05 23:35:22 +0000[diff] [blame]2848signed division of its two operands. This instruction can also take
2849<a href="#t_vector">vector</a> versions of the values in which case
2850the elements must be integers.</p>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]2851
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2852<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2853
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2854<p>The two arguments to the '<tt>srem</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2855<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2856values. Both arguments must have identical types.</p>
2857
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2858<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2859
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2860<p>This instruction returns the <i>remainder</i> of a division (where the result
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2861has the same sign as the dividend, <tt>op1</tt>), not the <i>modulo</i>
2862operator (where the result has the same sign as the divisor, <tt>op2</tt>) of
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2863a value. For more information about the difference, see <a
2864 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
2865Math Forum</a>. For a table of how this is implemented in various languages,
2866please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
2867Wikipedia: modulo operation</a>.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2868<p>Note that signed integer remainder and unsigned integer remainder are
2869distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
2870<p>Taking the remainder of a division by zero leads to undefined behavior.
2871Overflow also leads to undefined behavior; this is a rare case, but can occur,
2872for example, by taking the remainder of a 32-bit division of -2147483648 by -1.
2873(The remainder doesn't actually overflow, but this rule lets srem be
2874implemented using instructions that return both the result of the division
2875and the remainder.)</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2876<h5>Example:</h5>
2877<pre> &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2878</pre>
2879
2880</div>
2881<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2882<div class="doc_subsubsection">
2883 <a name="i_frem">'<tt>frem</tt>' Instruction</a> </div>
2884
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2885<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2886
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2887<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2888<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]2889</pre>
2890<h5>Overview:</h5>
2891<p>The '<tt>frem</tt>' instruction returns the remainder from the
2892division of its two operands.</p>
2893<h5>Arguments:</h5>
2894<p>The two arguments to the '<tt>frem</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2895<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2896of floating point values. Both arguments must have identical types.</p>
2897
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2898<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2899
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2900<p>This instruction returns the <i>remainder</i> of a division.
2901The remainder has the same sign as the dividend.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2902
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2903<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2904
2905<pre>
2906 &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]2907</pre>
2908</div>
2909
2910<!-- ======================================================================= -->
2911<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
2912Operations</a> </div>
2913<div class="doc_text">
2914<p>Bitwise binary operators are used to do various forms of
2915bit-twiddling in a program. They are generally very efficient
2916instructions and can commonly be strength reduced from other
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2917instructions. They require two operands of the same type, execute an operation on them,
2918and 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]2919</div>
2920
2921<!-- _______________________________________________________________________ -->
2922<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
2923Instruction</a> </div>
2924<div class="doc_text">
2925<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2926<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]2927</pre>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2928
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2929<h5>Overview:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2930
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2931<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
2932the left a specified number of bits.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2933
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2934<h5>Arguments:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2935
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2936<p>Both arguments to the '<tt>shl</tt>' instruction must be the same <a
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]2937 href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2938type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2939
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2940<h5>Semantics:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2941
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2942<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod 2<sup>n</sup>,
2943where 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]2944equal to or larger than the number of bits in <tt>op1</tt>, the result is undefined.
2945If the arguments are vectors, each vector element of <tt>op1</tt> is shifted by the
2946corresponding shift amount in <tt>op2</tt>.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2947
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2948<h5>Example:</h5><pre>
2949 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
2950 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
2951 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2952 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Mon P Wang9901e732008-12-09 05:46:39 +0000[diff] [blame]2953 &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]2954</pre>
2955</div>
2956<!-- _______________________________________________________________________ -->
2957<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
2958Instruction</a> </div>
2959<div class="doc_text">
2960<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2961<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]2962</pre>
2963
2964<h5>Overview:</h5>
2965<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
2966operand shifted to the right a specified number of bits with zero fill.</p>
2967
2968<h5>Arguments:</h5>
2969<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]2970<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]2971type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2972
2973<h5>Semantics:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2974
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2975<p>This instruction always performs a logical shift right operation. The most
2976significant bits of the result will be filled with zero bits after the
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2977shift. If <tt>op2</tt> is (statically or dynamically) equal to or larger than
Mon P Wangb0f51822008-12-10 08:55:09 +0000[diff] [blame]2978the number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
2979vectors, each vector element of <tt>op1</tt> is shifted by the corresponding shift
2980amount in <tt>op2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2981
2982<h5>Example:</h5>
2983<pre>
2984 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
2985 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
2986 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
2987 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2988 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Mon P Wang9901e732008-12-09 05:46:39 +0000[diff] [blame]2989 &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]2990</pre>
2991</div>
2992
2993<!-- _______________________________________________________________________ -->
2994<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
2995Instruction</a> </div>
2996<div class="doc_text">
2997
2998<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2999<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]3000</pre>
3001
3002<h5>Overview:</h5>
3003<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
3004operand shifted to the right a specified number of bits with sign extension.</p>
3005
3006<h5>Arguments:</h5>
3007<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]3008<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]3009type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3010
3011<h5>Semantics:</h5>
3012<p>This instruction always performs an arithmetic shift right operation,
3013The most significant bits of the result will be filled with the sign bit
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3014of <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]3015larger than the number of bits in <tt>op1</tt>, the result is undefined. If the
3016arguments are vectors, each vector element of <tt>op1</tt> is shifted by the
3017corresponding shift amount in <tt>op2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3018
3019<h5>Example:</h5>
3020<pre>
3021 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
3022 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
3023 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
3024 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]3025 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Mon P Wang9901e732008-12-09 05:46:39 +0000[diff] [blame]3026 &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]3027</pre>
3028</div>
3029
3030<!-- _______________________________________________________________________ -->
3031<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
3032Instruction</a> </div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3033
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3034<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3035
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3036<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3037
3038<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3039 &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]3040</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3041
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3042<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3043
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3044<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
3045its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3046
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3047<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3048
3049<p>The two arguments to the '<tt>and</tt>' instruction must be
3050<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3051values. Both arguments must have identical types.</p>
3052
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3053<h5>Semantics:</h5>
3054<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
3055<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]3056<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3057<table border="1" cellspacing="0" cellpadding="4">
3058 <tbody>
3059 <tr>
3060 <td>In0</td>
3061 <td>In1</td>
3062 <td>Out</td>
3063 </tr>
3064 <tr>
3065 <td>0</td>
3066 <td>0</td>
3067 <td>0</td>
3068 </tr>
3069 <tr>
3070 <td>0</td>
3071 <td>1</td>
3072 <td>0</td>
3073 </tr>
3074 <tr>
3075 <td>1</td>
3076 <td>0</td>
3077 <td>0</td>
3078 </tr>
3079 <tr>
3080 <td>1</td>
3081 <td>1</td>
3082 <td>1</td>
3083 </tr>
3084 </tbody>
3085</table>
3086</div>
3087<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3088<pre>
3089 &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]3090 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
3091 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
3092</pre>
3093</div>
3094<!-- _______________________________________________________________________ -->
3095<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
3096<div class="doc_text">
3097<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3098<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]3099</pre>
3100<h5>Overview:</h5>
3101<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
3102or of its two operands.</p>
3103<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3104
3105<p>The two arguments to the '<tt>or</tt>' instruction must be
3106<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3107values. Both arguments must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3108<h5>Semantics:</h5>
3109<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
3110<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]3111<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3112<table border="1" cellspacing="0" cellpadding="4">
3113 <tbody>
3114 <tr>
3115 <td>In0</td>
3116 <td>In1</td>
3117 <td>Out</td>
3118 </tr>
3119 <tr>
3120 <td>0</td>
3121 <td>0</td>
3122 <td>0</td>
3123 </tr>
3124 <tr>
3125 <td>0</td>
3126 <td>1</td>
3127 <td>1</td>
3128 </tr>
3129 <tr>
3130 <td>1</td>
3131 <td>0</td>
3132 <td>1</td>
3133 </tr>
3134 <tr>
3135 <td>1</td>
3136 <td>1</td>
3137 <td>1</td>
3138 </tr>
3139 </tbody>
3140</table>
3141</div>
3142<h5>Example:</h5>
3143<pre> &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
3144 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
3145 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
3146</pre>
3147</div>
3148<!-- _______________________________________________________________________ -->
3149<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
3150Instruction</a> </div>
3151<div class="doc_text">
3152<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3153<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]3154</pre>
3155<h5>Overview:</h5>
3156<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
3157or of its two operands. The <tt>xor</tt> is used to implement the
3158"one's complement" operation, which is the "~" operator in C.</p>
3159<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3160<p>The two arguments to the '<tt>xor</tt>' instruction must be
3161<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3162values. Both arguments must have identical types.</p>
3163
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3164<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3165
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3166<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
3167<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]3168<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3169<table border="1" cellspacing="0" cellpadding="4">
3170 <tbody>
3171 <tr>
3172 <td>In0</td>
3173 <td>In1</td>
3174 <td>Out</td>
3175 </tr>
3176 <tr>
3177 <td>0</td>
3178 <td>0</td>
3179 <td>0</td>
3180 </tr>
3181 <tr>
3182 <td>0</td>
3183 <td>1</td>
3184 <td>1</td>
3185 </tr>
3186 <tr>
3187 <td>1</td>
3188 <td>0</td>
3189 <td>1</td>
3190 </tr>
3191 <tr>
3192 <td>1</td>
3193 <td>1</td>
3194 <td>0</td>
3195 </tr>
3196 </tbody>
3197</table>
3198</div>
3199<p> </p>
3200<h5>Example:</h5>
3201<pre> &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
3202 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
3203 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
3204 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
3205</pre>
3206</div>
3207
3208<!-- ======================================================================= -->
3209<div class="doc_subsection">
3210 <a name="vectorops">Vector Operations</a>
3211</div>
3212
3213<div class="doc_text">
3214
3215<p>LLVM supports several instructions to represent vector operations in a
3216target-independent manner. These instructions cover the element-access and
3217vector-specific operations needed to process vectors effectively. While LLVM
3218does directly support these vector operations, many sophisticated algorithms
3219will want to use target-specific intrinsics to take full advantage of a specific
3220target.</p>
3221
3222</div>
3223
3224<!-- _______________________________________________________________________ -->
3225<div class="doc_subsubsection">
3226 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
3227</div>
3228
3229<div class="doc_text">
3230
3231<h5>Syntax:</h5>
3232
3233<pre>
3234 &lt;result&gt; = extractelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, i32 &lt;idx&gt; <i>; yields &lt;ty&gt;</i>
3235</pre>
3236
3237<h5>Overview:</h5>
3238
3239<p>
3240The '<tt>extractelement</tt>' instruction extracts a single scalar
3241element from a vector at a specified index.
3242</p>
3243
3244
3245<h5>Arguments:</h5>
3246
3247<p>
3248The first operand of an '<tt>extractelement</tt>' instruction is a
3249value of <a href="#t_vector">vector</a> type. The second operand is
3250an index indicating the position from which to extract the element.
3251The index may be a variable.</p>
3252
3253<h5>Semantics:</h5>
3254
3255<p>
3256The result is a scalar of the same type as the element type of
3257<tt>val</tt>. Its value is the value at position <tt>idx</tt> of
3258<tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
3259results are undefined.
3260</p>
3261
3262<h5>Example:</h5>
3263
3264<pre>
3265 %result = extractelement &lt;4 x i32&gt; %vec, i32 0 <i>; yields i32</i>
3266</pre>
3267</div>
3268
3269
3270<!-- _______________________________________________________________________ -->
3271<div class="doc_subsubsection">
3272 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
3273</div>
3274
3275<div class="doc_text">
3276
3277<h5>Syntax:</h5>
3278
3279<pre>
Dan Gohmanbcc3c502008-05-12 23:38:42 +0000[diff] [blame]3280 &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]3281</pre>
3282
3283<h5>Overview:</h5>
3284
3285<p>
3286The '<tt>insertelement</tt>' instruction inserts a scalar
3287element into a vector at a specified index.
3288</p>
3289
3290
3291<h5>Arguments:</h5>
3292
3293<p>
3294The first operand of an '<tt>insertelement</tt>' instruction is a
3295value of <a href="#t_vector">vector</a> type. The second operand is a
3296scalar value whose type must equal the element type of the first
3297operand. The third operand is an index indicating the position at
3298which to insert the value. The index may be a variable.</p>
3299
3300<h5>Semantics:</h5>
3301
3302<p>
3303The result is a vector of the same type as <tt>val</tt>. Its
3304element values are those of <tt>val</tt> except at position
3305<tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
3306exceeds the length of <tt>val</tt>, the results are undefined.
3307</p>
3308
3309<h5>Example:</h5>
3310
3311<pre>
3312 %result = insertelement &lt;4 x i32&gt; %vec, i32 1, i32 0 <i>; yields &lt;4 x i32&gt;</i>
3313</pre>
3314</div>
3315
3316<!-- _______________________________________________________________________ -->
3317<div class="doc_subsubsection">
3318 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
3319</div>
3320
3321<div class="doc_text">
3322
3323<h5>Syntax:</h5>
3324
3325<pre>
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3326 &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]3327</pre>
3328
3329<h5>Overview:</h5>
3330
3331<p>
3332The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3333from two input vectors, returning a vector with the same element type as
3334the input and length that is the same as the shuffle mask.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3335</p>
3336
3337<h5>Arguments:</h5>
3338
3339<p>
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3340The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
3341with types that match each other. The third argument is a shuffle mask whose
3342element type is always 'i32'. The result of the instruction is a vector whose
3343length is the same as the shuffle mask and whose element type is the same as
3344the element type of the first two operands.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3345</p>
3346
3347<p>
3348The shuffle mask operand is required to be a constant vector with either
3349constant integer or undef values.
3350</p>
3351
3352<h5>Semantics:</h5>
3353
3354<p>
3355The elements of the two input vectors are numbered from left to right across
3356both of the vectors. The shuffle mask operand specifies, for each element of
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3357the result vector, which element of the two input vectors the result element
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3358gets. The element selector may be undef (meaning "don't care") and the second
3359operand may be undef if performing a shuffle from only one vector.
3360</p>
3361
3362<h5>Example:</h5>
3363
3364<pre>
3365 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
3366 &lt;4 x i32&gt; &lt;i32 0, i32 4, i32 1, i32 5&gt; <i>; yields &lt;4 x i32&gt;</i>
3367 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; undef,
3368 &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]3369 %result = shufflevector &lt;8 x i32&gt; %v1, &lt;8 x i32&gt; undef,
3370 &lt;4 x i32&gt; &lt;i32 0, i32 1, i32 2, i32 3&gt; <i>; yields &lt;4 x i32&gt;</i>
3371 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
3372 &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]3373</pre>
3374</div>
3375
3376
3377<!-- ======================================================================= -->
3378<div class="doc_subsection">
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3379 <a name="aggregateops">Aggregate Operations</a>
3380</div>
3381
3382<div class="doc_text">
3383
3384<p>LLVM supports several instructions for working with aggregate values.
3385</p>
3386
3387</div>
3388
3389<!-- _______________________________________________________________________ -->
3390<div class="doc_subsubsection">
3391 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
3392</div>
3393
3394<div class="doc_text">
3395
3396<h5>Syntax:</h5>
3397
3398<pre>
3399 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
3400</pre>
3401
3402<h5>Overview:</h5>
3403
3404<p>
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3405The '<tt>extractvalue</tt>' instruction extracts the value of a struct field
3406or array element from an aggregate value.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3407</p>
3408
3409
3410<h5>Arguments:</h5>
3411
3412<p>
3413The first operand of an '<tt>extractvalue</tt>' instruction is a
3414value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a>
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3415type. The operands are constant indices to specify which value to extract
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3416in a similar manner as indices in a
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3417'<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
3418</p>
3419
3420<h5>Semantics:</h5>
3421
3422<p>
3423The result is the value at the position in the aggregate specified by
3424the index operands.
3425</p>
3426
3427<h5>Example:</h5>
3428
3429<pre>
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3430 %result = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3431</pre>
3432</div>
3433
3434
3435<!-- _______________________________________________________________________ -->
3436<div class="doc_subsubsection">
3437 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
3438</div>
3439
3440<div class="doc_text">
3441
3442<h5>Syntax:</h5>
3443
3444<pre>
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3445 &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]3446</pre>
3447
3448<h5>Overview:</h5>
3449
3450<p>
3451The '<tt>insertvalue</tt>' instruction inserts a value
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3452into a struct field or array element in an aggregate.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3453</p>
3454
3455
3456<h5>Arguments:</h5>
3457
3458<p>
3459The first operand of an '<tt>insertvalue</tt>' instruction is a
3460value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a> type.
3461The second operand is a first-class value to insert.
Dan Gohman4f29e422008-05-23 21:53:15 +0000[diff] [blame]3462The following operands are constant indices
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3463indicating the position at which to insert the value in a similar manner as
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3464indices in a
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3465'<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
3466The value to insert must have the same type as the value identified
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3467by the indices.
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3468</p>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3469
3470<h5>Semantics:</h5>
3471
3472<p>
3473The result is an aggregate of the same type as <tt>val</tt>. Its
3474value is that of <tt>val</tt> except that the value at the position
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3475specified by the indices is that of <tt>elt</tt>.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3476</p>
3477
3478<h5>Example:</h5>
3479
3480<pre>
Dan Gohmanb1aab4e2008-06-23 15:26:37 +0000[diff] [blame]3481 %result = insertvalue {i32, float} %agg, i32 1, 0 <i>; yields {i32, float}</i>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3482</pre>
3483</div>
3484
3485
3486<!-- ======================================================================= -->
3487<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3488 <a name="memoryops">Memory Access and Addressing Operations</a>
3489</div>
3490
3491<div class="doc_text">
3492
3493<p>A key design point of an SSA-based representation is how it
3494represents memory. In LLVM, no memory locations are in SSA form, which
3495makes things very simple. This section describes how to read, write,
3496allocate, and free memory in LLVM.</p>
3497
3498</div>
3499
3500<!-- _______________________________________________________________________ -->
3501<div class="doc_subsubsection">
3502 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
3503</div>
3504
3505<div class="doc_text">
3506
3507<h5>Syntax:</h5>
3508
3509<pre>
3510 &lt;result&gt; = malloc &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
3511</pre>
3512
3513<h5>Overview:</h5>
3514
3515<p>The '<tt>malloc</tt>' instruction allocates memory from the system
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]3516heap and returns a pointer to it. The object is always allocated in the generic
3517address space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3518
3519<h5>Arguments:</h5>
3520
3521<p>The '<tt>malloc</tt>' instruction allocates
3522<tt>sizeof(&lt;type&gt;)*NumElements</tt>
3523bytes of memory from the operating system and returns a pointer of the
3524appropriate type to the program. If "NumElements" is specified, it is the
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3525number of elements allocated, otherwise "NumElements" is defaulted to be one.
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3526If a constant alignment is specified, the value result of the allocation is guaranteed to
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3527be aligned to at least that boundary. If not specified, or if zero, the target can
3528choose to align the allocation on any convenient boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3529
3530<p>'<tt>type</tt>' must be a sized type.</p>
3531
3532<h5>Semantics:</h5>
3533
3534<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
Nick Lewyckyff384472008-11-24 03:41:24 +0000[diff] [blame]3535a pointer is returned. The result of a zero byte allocation is undefined. The
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]3536result is null if there is insufficient memory available.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3537
3538<h5>Example:</h5>
3539
3540<pre>
Dan Gohmanf54f50a2009-01-04 23:49:44 +0000[diff] [blame]3541 %array = malloc [4 x i8] <i>; yields {[%4 x i8]*}:array</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3542
3543 %size = <a href="#i_add">add</a> i32 2, 2 <i>; yields {i32}:size = i32 4</i>
3544 %array1 = malloc i8, i32 4 <i>; yields {i8*}:array1</i>
3545 %array2 = malloc [12 x i8], i32 %size <i>; yields {[12 x i8]*}:array2</i>
3546 %array3 = malloc i32, i32 4, align 1024 <i>; yields {i32*}:array3</i>
3547 %array4 = malloc i32, align 1024 <i>; yields {i32*}:array4</i>
3548</pre>
Dan Gohman60967192009-01-12 23:12:39 +0000[diff] [blame]3549
3550<p>Note that the code generator does not yet respect the
3551 alignment value.</p>
3552
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3553</div>
3554
3555<!-- _______________________________________________________________________ -->
3556<div class="doc_subsubsection">
3557 <a name="i_free">'<tt>free</tt>' Instruction</a>
3558</div>
3559
3560<div class="doc_text">
3561
3562<h5>Syntax:</h5>
3563
3564<pre>
Dan Gohmanf54f50a2009-01-04 23:49:44 +0000[diff] [blame]3565 free &lt;type&gt; &lt;value&gt; <i>; yields {void}</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3566</pre>
3567
3568<h5>Overview:</h5>
3569
3570<p>The '<tt>free</tt>' instruction returns memory back to the unused
3571memory heap to be reallocated in the future.</p>
3572
3573<h5>Arguments:</h5>
3574
3575<p>'<tt>value</tt>' shall be a pointer value that points to a value
3576that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
3577instruction.</p>
3578
3579<h5>Semantics:</h5>
3580
3581<p>Access to the memory pointed to by the pointer is no longer defined
Chris Lattner329d6532008-04-19 22:41:32 +0000[diff] [blame]3582after this instruction executes. If the pointer is null, the operation
3583is a noop.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3584
3585<h5>Example:</h5>
3586
3587<pre>
Dan Gohmanf54f50a2009-01-04 23:49:44 +0000[diff] [blame]3588 %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]3589 free [4 x i8]* %array
3590</pre>
3591</div>
3592
3593<!-- _______________________________________________________________________ -->
3594<div class="doc_subsubsection">
3595 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
3596</div>
3597
3598<div class="doc_text">
3599
3600<h5>Syntax:</h5>
3601
3602<pre>
3603 &lt;result&gt; = alloca &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
3604</pre>
3605
3606<h5>Overview:</h5>
3607
3608<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
3609currently executing function, to be automatically released when this function
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]3610returns to its caller. The object is always allocated in the generic address
3611space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3612
3613<h5>Arguments:</h5>
3614
3615<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
3616bytes of memory on the runtime stack, returning a pointer of the
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3617appropriate type to the program. If "NumElements" is specified, it is the
3618number of elements allocated, otherwise "NumElements" is defaulted to be one.
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3619If a constant alignment is specified, the value result of the allocation is guaranteed
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3620to be aligned to at least that boundary. If not specified, or if zero, the target
3621can choose to align the allocation on any convenient boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3622
3623<p>'<tt>type</tt>' may be any sized type.</p>
3624
3625<h5>Semantics:</h5>
3626
Bill Wendling2a454572009-05-08 20:49:29 +0000[diff] [blame]3627<p>Memory is allocated; a pointer is returned. The operation is undefined if
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]3628there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3629memory is automatically released when the function returns. The '<tt>alloca</tt>'
3630instruction is commonly used to represent automatic variables that must
3631have an address available. When the function returns (either with the <tt><a
3632 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]3633instructions), the memory is reclaimed. Allocating zero bytes
3634is legal, but the result is undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3635
3636<h5>Example:</h5>
3637
3638<pre>
Dan Gohmanf54f50a2009-01-04 23:49:44 +0000[diff] [blame]3639 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
3640 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
3641 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
3642 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3643</pre>
3644</div>
3645
3646<!-- _______________________________________________________________________ -->
3647<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
3648Instruction</a> </div>
3649<div class="doc_text">
3650<h5>Syntax:</h5>
3651<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>
3652<h5>Overview:</h5>
3653<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
3654<h5>Arguments:</h5>
3655<p>The argument to the '<tt>load</tt>' instruction specifies the memory
3656address from which to load. The pointer must point to a <a
3657 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
3658marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
3659the number or order of execution of this <tt>load</tt> with other
3660volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
3661instructions. </p>
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3662<p>
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3663The optional constant "align" argument specifies the alignment of the operation
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3664(that is, the alignment of the memory address). A value of 0 or an
3665omitted "align" argument means that the operation has the preferential
3666alignment for the target. It is the responsibility of the code emitter
3667to ensure that the alignment information is correct. Overestimating
3668the alignment results in an undefined behavior. Underestimating the
3669alignment may produce less efficient code. An alignment of 1 is always
3670safe.
3671</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3672<h5>Semantics:</h5>
Duncan Sandsfe87fff2009-03-22 11:33:16 +0000[diff] [blame]3673<p>The location of memory pointed to is loaded. If the value being loaded
3674is of scalar type then the number of bytes read does not exceed the minimum
3675number of bytes needed to hold all bits of the type. For example, loading an
3676<tt>i24</tt> reads at most three bytes. When loading a value of a type like
3677<tt>i20</tt> with a size that is not an integral number of bytes, the result
3678is undefined if the value was not originally written using a store of the
3679same type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3680<h5>Examples:</h5>
3681<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
3682 <a
3683 href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
3684 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
3685</pre>
3686</div>
3687<!-- _______________________________________________________________________ -->
3688<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
3689Instruction</a> </div>
3690<div class="doc_text">
3691<h5>Syntax:</h5>
3692<pre> store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
3693 volatile store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
3694</pre>
3695<h5>Overview:</h5>
3696<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
3697<h5>Arguments:</h5>
3698<p>There are two arguments to the '<tt>store</tt>' instruction: a value
3699to 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]3700operand must be a pointer to the <a href="#t_firstclass">first class</a> type
3701of the '<tt>&lt;value&gt;</tt>'
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3702operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
3703optimizer is not allowed to modify the number or order of execution of
3704this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
3705 href="#i_store">store</a></tt> instructions.</p>
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3706<p>
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3707The optional constant "align" argument specifies the alignment of the operation
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3708(that is, the alignment of the memory address). A value of 0 or an
3709omitted "align" argument means that the operation has the preferential
3710alignment for the target. It is the responsibility of the code emitter
3711to ensure that the alignment information is correct. Overestimating
3712the alignment results in an undefined behavior. Underestimating the
3713alignment may produce less efficient code. An alignment of 1 is always
3714safe.
3715</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3716<h5>Semantics:</h5>
3717<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>'
Duncan Sandsfe87fff2009-03-22 11:33:16 +0000[diff] [blame]3718at the location specified by the '<tt>&lt;pointer&gt;</tt>' operand.
3719If '<tt>&lt;value&gt;</tt>' is of scalar type then the number of bytes
3720written does not exceed the minimum number of bytes needed to hold all
3721bits of the type. For example, storing an <tt>i24</tt> writes at most
3722three bytes. When writing a value of a type like <tt>i20</tt> with a
3723size that is not an integral number of bytes, it is unspecified what
3724happens to the extra bits that do not belong to the type, but they will
3725typically be overwritten.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3726<h5>Example:</h5>
3727<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling63ffa142007-10-22 05:10:05 +0000[diff] [blame]3728 store i32 3, i32* %ptr <i>; yields {void}</i>
3729 %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]3730</pre>
3731</div>
3732
3733<!-- _______________________________________________________________________ -->
3734<div class="doc_subsubsection">
3735 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
3736</div>
3737
3738<div class="doc_text">
3739<h5>Syntax:</h5>
3740<pre>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3741 &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]3742</pre>
3743
3744<h5>Overview:</h5>
3745
3746<p>
3747The '<tt>getelementptr</tt>' instruction is used to get the address of a
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3748subelement of an aggregate data structure. It performs address calculation only
3749and does not access memory.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3750
3751<h5>Arguments:</h5>
3752
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3753<p>The first argument is always a pointer, and forms the basis of the
3754calculation. The remaining arguments are indices, that indicate which of the
3755elements of the aggregate object are indexed. The interpretation of each index
3756is dependent on the type being indexed into. The first index always indexes the
3757pointer value given as the first argument, the second index indexes a value of
3758the type pointed to (not necessarily the value directly pointed to, since the
3759first index can be non-zero), etc. The first type indexed into must be a pointer
3760value, subsequent types can be arrays, vectors and structs. Note that subsequent
3761types being indexed into can never be pointers, since that would require loading
3762the pointer before continuing calculation.</p>
3763
3764<p>The type of each index argument depends on the type it is indexing into.
3765When indexing into a (packed) structure, only <tt>i32</tt> integer
3766<b>constants</b> are allowed. When indexing into an array, pointer or vector,
Sanjiv Guptaa2e72d92009-04-27 03:21:00 +0000[diff] [blame]3767integers of any width are allowed (also non-constants).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3768
3769<p>For example, let's consider a C code fragment and how it gets
3770compiled to LLVM:</p>
3771
3772<div class="doc_code">
3773<pre>
3774struct RT {
3775 char A;
3776 int B[10][20];
3777 char C;
3778};
3779struct ST {
3780 int X;
3781 double Y;
3782 struct RT Z;
3783};
3784
3785int *foo(struct ST *s) {
3786 return &amp;s[1].Z.B[5][13];
3787}
3788</pre>
3789</div>
3790
3791<p>The LLVM code generated by the GCC frontend is:</p>
3792
3793<div class="doc_code">
3794<pre>
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]3795%RT = <a href="#namedtypes">type</a> { i8 , [10 x [20 x i32]], i8 }
3796%ST = <a href="#namedtypes">type</a> { i32, double, %RT }
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3797
3798define i32* %foo(%ST* %s) {
3799entry:
3800 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
3801 ret i32* %reg
3802}
3803</pre>
3804</div>
3805
3806<h5>Semantics:</h5>
3807
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3808<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
3809type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
3810}</tt>' type, a structure. The second index indexes into the third element of
3811the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
3812i8 }</tt>' type, another structure. The third index indexes into the second
3813element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
3814array. The two dimensions of the array are subscripted into, yielding an
3815'<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
3816to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
3817
3818<p>Note that it is perfectly legal to index partially through a
3819structure, returning a pointer to an inner element. Because of this,
3820the LLVM code for the given testcase is equivalent to:</p>
3821
3822<pre>
3823 define i32* %foo(%ST* %s) {
3824 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
3825 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
3826 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
3827 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
3828 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
3829 ret i32* %t5
3830 }
3831</pre>
3832
Chris Lattner50609942009-03-09 20:55:18 +0000[diff] [blame]3833<p>Note that it is undefined to access an array out of bounds: array
3834and pointer indexes must always be within the defined bounds of the
3835array type when accessed with an instruction that dereferences the
3836pointer (e.g. a load or store instruction). The one exception for
3837this rule is zero length arrays. These arrays are defined to be
3838accessible as variable length arrays, which requires access beyond the
3839zero'th element.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3840
3841<p>The getelementptr instruction is often confusing. For some more insight
3842into how it works, see <a href="GetElementPtr.html">the getelementptr
3843FAQ</a>.</p>
3844
3845<h5>Example:</h5>
3846
3847<pre>
3848 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3849 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
3850 <i>; yields i8*:vptr</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3851 %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]3852 <i>; yields i8*:eptr</i>
3853 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Sanjiv Gupta4f9a0dc2009-04-25 07:27:44 +0000[diff] [blame]3854 <i>; yields i32*:iptr</i>
Sanjiv Gupta1e46c582009-04-24 16:38:13 +0000[diff] [blame]3855 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3856</pre>
3857</div>
3858
3859<!-- ======================================================================= -->
3860<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
3861</div>
3862<div class="doc_text">
3863<p>The instructions in this category are the conversion instructions (casting)
3864which all take a single operand and a type. They perform various bit conversions
3865on the operand.</p>
3866</div>
3867
3868<!-- _______________________________________________________________________ -->
3869<div class="doc_subsubsection">
3870 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
3871</div>
3872<div class="doc_text">
3873
3874<h5>Syntax:</h5>
3875<pre>
3876 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3877</pre>
3878
3879<h5>Overview:</h5>
3880<p>
3881The '<tt>trunc</tt>' instruction truncates its operand to the type <tt>ty2</tt>.
3882</p>
3883
3884<h5>Arguments:</h5>
3885<p>
3886The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
3887be an <a href="#t_integer">integer</a> type, and a type that specifies the size
3888and type of the result, which must be an <a href="#t_integer">integer</a>
3889type. The bit size of <tt>value</tt> must be larger than the bit size of
3890<tt>ty2</tt>. Equal sized types are not allowed.</p>
3891
3892<h5>Semantics:</h5>
3893<p>
3894The '<tt>trunc</tt>' instruction truncates the high order bits in <tt>value</tt>
3895and converts the remaining bits to <tt>ty2</tt>. Since the source size must be
3896larger than the destination size, <tt>trunc</tt> cannot be a <i>no-op cast</i>.
3897It will always truncate bits.</p>
3898
3899<h5>Example:</h5>
3900<pre>
3901 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
3902 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
3903 %Y = trunc i32 122 to i1 <i>; yields i1:false</i>
3904</pre>
3905</div>
3906
3907<!-- _______________________________________________________________________ -->
3908<div class="doc_subsubsection">
3909 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
3910</div>
3911<div class="doc_text">
3912
3913<h5>Syntax:</h5>
3914<pre>
3915 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3916</pre>
3917
3918<h5>Overview:</h5>
3919<p>The '<tt>zext</tt>' instruction zero extends its operand to type
3920<tt>ty2</tt>.</p>
3921
3922
3923<h5>Arguments:</h5>
3924<p>The '<tt>zext</tt>' instruction takes a value to cast, which must be of
3925<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3926also be of <a href="#t_integer">integer</a> type. The bit size of the
3927<tt>value</tt> must be smaller than the bit size of the destination type,
3928<tt>ty2</tt>.</p>
3929
3930<h5>Semantics:</h5>
3931<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
3932bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
3933
3934<p>When zero extending from i1, the result will always be either 0 or 1.</p>
3935
3936<h5>Example:</h5>
3937<pre>
3938 %X = zext i32 257 to i64 <i>; yields i64:257</i>
3939 %Y = zext i1 true to i32 <i>; yields i32:1</i>
3940</pre>
3941</div>
3942
3943<!-- _______________________________________________________________________ -->
3944<div class="doc_subsubsection">
3945 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
3946</div>
3947<div class="doc_text">
3948
3949<h5>Syntax:</h5>
3950<pre>
3951 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3952</pre>
3953
3954<h5>Overview:</h5>
3955<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
3956
3957<h5>Arguments:</h5>
3958<p>
3959The '<tt>sext</tt>' instruction takes a value to cast, which must be of
3960<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3961also be of <a href="#t_integer">integer</a> type. The bit size of the
3962<tt>value</tt> must be smaller than the bit size of the destination type,
3963<tt>ty2</tt>.</p>
3964
3965<h5>Semantics:</h5>
3966<p>
3967The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
3968bit (highest order bit) of the <tt>value</tt> until it reaches the bit size of
3969the type <tt>ty2</tt>.</p>
3970
3971<p>When sign extending from i1, the extension always results in -1 or 0.</p>
3972
3973<h5>Example:</h5>
3974<pre>
3975 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
3976 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
3977</pre>
3978</div>
3979
3980<!-- _______________________________________________________________________ -->
3981<div class="doc_subsubsection">
3982 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
3983</div>
3984
3985<div class="doc_text">
3986
3987<h5>Syntax:</h5>
3988
3989<pre>
3990 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3991</pre>
3992
3993<h5>Overview:</h5>
3994<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
3995<tt>ty2</tt>.</p>
3996
3997
3998<h5>Arguments:</h5>
3999<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
4000 point</a> value to cast and a <a href="#t_floating">floating point</a> type to
4001cast it to. The size of <tt>value</tt> must be larger than the size of
4002<tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
4003<i>no-op cast</i>.</p>
4004
4005<h5>Semantics:</h5>
4006<p> The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
4007<a href="#t_floating">floating point</a> type to a smaller
4008<a href="#t_floating">floating point</a> type. If the value cannot fit within
4009the destination type, <tt>ty2</tt>, then the results are undefined.</p>
4010
4011<h5>Example:</h5>
4012<pre>
4013 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
4014 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
4015</pre>
4016</div>
4017
4018<!-- _______________________________________________________________________ -->
4019<div class="doc_subsubsection">
4020 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
4021</div>
4022<div class="doc_text">
4023
4024<h5>Syntax:</h5>
4025<pre>
4026 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4027</pre>
4028
4029<h5>Overview:</h5>
4030<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
4031floating point value.</p>
4032
4033<h5>Arguments:</h5>
4034<p>The '<tt>fpext</tt>' instruction takes a
4035<a href="#t_floating">floating point</a> <tt>value</tt> to cast,
4036and a <a href="#t_floating">floating point</a> type to cast it to. The source
4037type must be smaller than the destination type.</p>
4038
4039<h5>Semantics:</h5>
4040<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
4041<a href="#t_floating">floating point</a> type to a larger
4042<a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
4043used to make a <i>no-op cast</i> because it always changes bits. Use
4044<tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
4045
4046<h5>Example:</h5>
4047<pre>
4048 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
4049 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
4050</pre>
4051</div>
4052
4053<!-- _______________________________________________________________________ -->
4054<div class="doc_subsubsection">
4055 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
4056</div>
4057<div class="doc_text">
4058
4059<h5>Syntax:</h5>
4060<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]4061 &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]4062</pre>
4063
4064<h5>Overview:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]4065<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4066unsigned integer equivalent of type <tt>ty2</tt>.
4067</p>
4068
4069<h5>Arguments:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]4070<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]4071scalar or vector <a href="#t_floating">floating point</a> value, and a type
4072to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4073type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4074vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4075
4076<h5>Semantics:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]4077<p> The '<tt>fptoui</tt>' instruction converts its
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4078<a href="#t_floating">floating point</a> operand into the nearest (rounding
4079towards zero) unsigned integer value. If the value cannot fit in <tt>ty2</tt>,
4080the results are undefined.</p>
4081
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4082<h5>Example:</h5>
4083<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]4084 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]4085 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]4086 %X = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4087</pre>
4088</div>
4089
4090<!-- _______________________________________________________________________ -->
4091<div class="doc_subsubsection">
4092 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
4093</div>
4094<div class="doc_text">
4095
4096<h5>Syntax:</h5>
4097<pre>
4098 &lt;result&gt; = fptosi &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4099</pre>
4100
4101<h5>Overview:</h5>
4102<p>The '<tt>fptosi</tt>' instruction converts
4103<a href="#t_floating">floating point</a> <tt>value</tt> to type <tt>ty2</tt>.
4104</p>
4105
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4106<h5>Arguments:</h5>
4107<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]4108scalar or vector <a href="#t_floating">floating point</a> value, and a type
4109to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4110type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4111vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4112
4113<h5>Semantics:</h5>
4114<p>The '<tt>fptosi</tt>' instruction converts its
4115<a href="#t_floating">floating point</a> operand into the nearest (rounding
4116towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
4117the results are undefined.</p>
4118
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4119<h5>Example:</h5>
4120<pre>
4121 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]4122 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4123 %X = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
4124</pre>
4125</div>
4126
4127<!-- _______________________________________________________________________ -->
4128<div class="doc_subsubsection">
4129 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
4130</div>
4131<div class="doc_text">
4132
4133<h5>Syntax:</h5>
4134<pre>
4135 &lt;result&gt; = uitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4136</pre>
4137
4138<h5>Overview:</h5>
4139<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
4140integer and converts that value to the <tt>ty2</tt> type.</p>
4141
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4142<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]4143<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be a
4144scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
4145to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4146type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4147floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4148
4149<h5>Semantics:</h5>
4150<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
4151integer quantity and converts it to the corresponding floating point value. If
4152the value cannot fit in the floating point value, the results are undefined.</p>
4153
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4154<h5>Example:</h5>
4155<pre>
4156 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4157 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4158</pre>
4159</div>
4160
4161<!-- _______________________________________________________________________ -->
4162<div class="doc_subsubsection">
4163 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
4164</div>
4165<div class="doc_text">
4166
4167<h5>Syntax:</h5>
4168<pre>
4169 &lt;result&gt; = sitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4170</pre>
4171
4172<h5>Overview:</h5>
4173<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed
4174integer and converts that value to the <tt>ty2</tt> type.</p>
4175
4176<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]4177<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be a
4178scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
4179to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4180type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4181floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4182
4183<h5>Semantics:</h5>
4184<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed
4185integer quantity and converts it to the corresponding floating point value. If
4186the value cannot fit in the floating point value, the results are undefined.</p>
4187
4188<h5>Example:</h5>
4189<pre>
4190 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4191 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4192</pre>
4193</div>
4194
4195<!-- _______________________________________________________________________ -->
4196<div class="doc_subsubsection">
4197 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
4198</div>
4199<div class="doc_text">
4200
4201<h5>Syntax:</h5>
4202<pre>
4203 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4204</pre>
4205
4206<h5>Overview:</h5>
4207<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
4208the integer type <tt>ty2</tt>.</p>
4209
4210<h5>Arguments:</h5>
4211<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
4212must 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]4213<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]4214
4215<h5>Semantics:</h5>
4216<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
4217<tt>ty2</tt> by interpreting the pointer value as an integer and either
4218truncating or zero extending that value to the size of the integer type. If
4219<tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
4220<tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
4221are the same size, then nothing is done (<i>no-op cast</i>) other than a type
4222change.</p>
4223
4224<h5>Example:</h5>
4225<pre>
4226 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
4227 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
4228</pre>
4229</div>
4230
4231<!-- _______________________________________________________________________ -->
4232<div class="doc_subsubsection">
4233 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
4234</div>
4235<div class="doc_text">
4236
4237<h5>Syntax:</h5>
4238<pre>
4239 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4240</pre>
4241
4242<h5>Overview:</h5>
4243<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to
4244a pointer type, <tt>ty2</tt>.</p>
4245
4246<h5>Arguments:</h5>
4247<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
4248value to cast, and a type to cast it to, which must be a
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4249<a href="#t_pointer">pointer</a> type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4250
4251<h5>Semantics:</h5>
4252<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
4253<tt>ty2</tt> by applying either a zero extension or a truncation depending on
4254the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
4255size of a pointer then a truncation is done. If <tt>value</tt> is smaller than
4256the size of a pointer then a zero extension is done. If they are the same size,
4257nothing is done (<i>no-op cast</i>).</p>
4258
4259<h5>Example:</h5>
4260<pre>
4261 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
4262 %X = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
4263 %Y = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
4264</pre>
4265</div>
4266
4267<!-- _______________________________________________________________________ -->
4268<div class="doc_subsubsection">
4269 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
4270</div>
4271<div class="doc_text">
4272
4273<h5>Syntax:</h5>
4274<pre>
4275 &lt;result&gt; = bitcast &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4276</pre>
4277
4278<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4279
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4280<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
4281<tt>ty2</tt> without changing any bits.</p>
4282
4283<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4284
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4285<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be
Dan Gohman7305fa02008-09-08 16:45:59 +0000[diff] [blame]4286a non-aggregate first class value, and a type to cast it to, which must also be
4287a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes of
4288<tt>value</tt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4289and the destination type, <tt>ty2</tt>, must be identical. If the source
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4290type is a pointer, the destination type must also be a pointer. This
4291instruction supports bitwise conversion of vectors to integers and to vectors
4292of other types (as long as they have the same size).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4293
4294<h5>Semantics:</h5>
4295<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
4296<tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
4297this conversion. The conversion is done as if the <tt>value</tt> had been
4298stored to memory and read back as type <tt>ty2</tt>. Pointer types may only be
4299converted to other pointer types with this instruction. To convert pointers to
4300other types, use the <a href="#i_inttoptr">inttoptr</a> or
4301<a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
4302
4303<h5>Example:</h5>
4304<pre>
4305 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
4306 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4307 %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]4308</pre>
4309</div>
4310
4311<!-- ======================================================================= -->
4312<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
4313<div class="doc_text">
4314<p>The instructions in this category are the "miscellaneous"
4315instructions, which defy better classification.</p>
4316</div>
4317
4318<!-- _______________________________________________________________________ -->
4319<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
4320</div>
4321<div class="doc_text">
4322<h5>Syntax:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4323<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]4324</pre>
4325<h5>Overview:</h5>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4326<p>The '<tt>icmp</tt>' instruction returns a boolean value or
4327a vector of boolean values based on comparison
4328of its two integer, integer vector, or pointer operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4329<h5>Arguments:</h5>
4330<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
4331the condition code indicating the kind of comparison to perform. It is not
4332a value, just a keyword. The possible condition code are:
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4333</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4334<ol>
4335 <li><tt>eq</tt>: equal</li>
4336 <li><tt>ne</tt>: not equal </li>
4337 <li><tt>ugt</tt>: unsigned greater than</li>
4338 <li><tt>uge</tt>: unsigned greater or equal</li>
4339 <li><tt>ult</tt>: unsigned less than</li>
4340 <li><tt>ule</tt>: unsigned less or equal</li>
4341 <li><tt>sgt</tt>: signed greater than</li>
4342 <li><tt>sge</tt>: signed greater or equal</li>
4343 <li><tt>slt</tt>: signed less than</li>
4344 <li><tt>sle</tt>: signed less or equal</li>
4345</ol>
4346<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4347<a href="#t_pointer">pointer</a>
4348or integer <a href="#t_vector">vector</a> typed.
4349They must also be identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4350<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4351<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]4352the condition code given as <tt>cond</tt>. The comparison performed always
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4353yields 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]4354</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4355<ol>
4356 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
4357 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
4358 </li>
4359 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4360 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4361 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4362 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4363 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4364 <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]4365 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4366 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4367 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4368 <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]4369 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4370 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4371 <li><tt>sge</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4372 <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]4373 <li><tt>slt</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4374 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4375 <li><tt>sle</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4376 <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]4377</ol>
4378<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
4379values are compared as if they were integers.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4380<p>If the operands are integer vectors, then they are compared
4381element by element. The result is an <tt>i1</tt> vector with
4382the same number of elements as the values being compared.
4383Otherwise, the result is an <tt>i1</tt>.
4384</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4385
4386<h5>Example:</h5>
4387<pre> &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
4388 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
4389 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
4390 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
4391 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
4392 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
4393</pre>
Dan Gohmana53eb382009-01-22 01:39:38 +0000[diff] [blame]4394
4395<p>Note that the code generator does not yet support vector types with
4396 the <tt>icmp</tt> instruction.</p>
4397
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4398</div>
4399
4400<!-- _______________________________________________________________________ -->
4401<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
4402</div>
4403<div class="doc_text">
4404<h5>Syntax:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4405<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]4406</pre>
4407<h5>Overview:</h5>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4408<p>The '<tt>fcmp</tt>' instruction returns a boolean value
4409or vector of boolean values based on comparison
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4410of its operands.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4411<p>
4412If the operands are floating point scalars, then the result
4413type is a boolean (<a href="#t_primitive"><tt>i1</tt></a>).
4414</p>
4415<p>If the operands are floating point vectors, then the result type
4416is a vector of boolean with the same number of elements as the
4417operands being compared.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4418<h5>Arguments:</h5>
4419<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
4420the condition code indicating the kind of comparison to perform. It is not
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4421a value, just a keyword. The possible condition code are:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4422<ol>
4423 <li><tt>false</tt>: no comparison, always returns false</li>
4424 <li><tt>oeq</tt>: ordered and equal</li>
4425 <li><tt>ogt</tt>: ordered and greater than </li>
4426 <li><tt>oge</tt>: ordered and greater than or equal</li>
4427 <li><tt>olt</tt>: ordered and less than </li>
4428 <li><tt>ole</tt>: ordered and less than or equal</li>
4429 <li><tt>one</tt>: ordered and not equal</li>
4430 <li><tt>ord</tt>: ordered (no nans)</li>
4431 <li><tt>ueq</tt>: unordered or equal</li>
4432 <li><tt>ugt</tt>: unordered or greater than </li>
4433 <li><tt>uge</tt>: unordered or greater than or equal</li>
4434 <li><tt>ult</tt>: unordered or less than </li>
4435 <li><tt>ule</tt>: unordered or less than or equal</li>
4436 <li><tt>une</tt>: unordered or not equal</li>
4437 <li><tt>uno</tt>: unordered (either nans)</li>
4438 <li><tt>true</tt>: no comparison, always returns true</li>
4439</ol>
4440<p><i>Ordered</i> means that neither operand is a QNAN while
4441<i>unordered</i> means that either operand may be a QNAN.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4442<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be
4443either a <a href="#t_floating">floating point</a> type
4444or a <a href="#t_vector">vector</a> of floating point type.
4445They must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4446<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4447<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4448according to the condition code given as <tt>cond</tt>.
4449If the operands are vectors, then the vectors are compared
4450element by element.
4451Each comparison performed
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4452always yields an <a href="#t_primitive">i1</a> result, as follows:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4453<ol>
4454 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
4455 <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]4456 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4457 <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]4458 <tt>op1</tt> is greather than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4459 <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]4460 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4461 <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]4462 <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4463 <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]4464 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4465 <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]4466 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4467 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
4468 <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]4469 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4470 <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]4471 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4472 <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]4473 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4474 <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]4475 <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4476 <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]4477 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4478 <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]4479 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4480 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
4481 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
4482</ol>
4483
4484<h5>Example:</h5>
4485<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]4486 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
4487 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
4488 &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]4489</pre>
Dan Gohmana53eb382009-01-22 01:39:38 +0000[diff] [blame]4490
4491<p>Note that the code generator does not yet support vector types with
4492 the <tt>fcmp</tt> instruction.</p>
4493
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4494</div>
4495
4496<!-- _______________________________________________________________________ -->
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4497<div class="doc_subsubsection">
4498 <a name="i_vicmp">'<tt>vicmp</tt>' Instruction</a>
4499</div>
4500<div class="doc_text">
4501<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4502<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]4503</pre>
4504<h5>Overview:</h5>
4505<p>The '<tt>vicmp</tt>' instruction returns an integer vector value based on
4506element-wise comparison of its two integer vector operands.</p>
4507<h5>Arguments:</h5>
4508<p>The '<tt>vicmp</tt>' instruction takes three operands. The first operand is
4509the condition code indicating the kind of comparison to perform. It is not
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4510a value, just a keyword. The possible condition code are:</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4511<ol>
4512 <li><tt>eq</tt>: equal</li>
4513 <li><tt>ne</tt>: not equal </li>
4514 <li><tt>ugt</tt>: unsigned greater than</li>
4515 <li><tt>uge</tt>: unsigned greater or equal</li>
4516 <li><tt>ult</tt>: unsigned less than</li>
4517 <li><tt>ule</tt>: unsigned less or equal</li>
4518 <li><tt>sgt</tt>: signed greater than</li>
4519 <li><tt>sge</tt>: signed greater or equal</li>
4520 <li><tt>slt</tt>: signed less than</li>
4521 <li><tt>sle</tt>: signed less or equal</li>
4522</ol>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4523<p>The remaining two arguments must be <a href="#t_vector">vector</a> or
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4524<a href="#t_integer">integer</a> typed. They must also be identical types.</p>
4525<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4526<p>The '<tt>vicmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4527according to the condition code given as <tt>cond</tt>. The comparison yields a
4528<a href="#t_vector">vector</a> of <a href="#t_integer">integer</a> result, of
4529identical type as the values being compared. The most significant bit in each
4530element is 1 if the element-wise comparison evaluates to true, and is 0
4531otherwise. All other bits of the result are undefined. The condition codes
4532are evaluated identically to the <a href="#i_icmp">'<tt>icmp</tt>'
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4533instruction</a>.</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4534
4535<h5>Example:</h5>
4536<pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4537 &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>
4538 &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]4539</pre>
4540</div>
4541
4542<!-- _______________________________________________________________________ -->
4543<div class="doc_subsubsection">
4544 <a name="i_vfcmp">'<tt>vfcmp</tt>' Instruction</a>
4545</div>
4546<div class="doc_text">
4547<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4548<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]4549<h5>Overview:</h5>
4550<p>The '<tt>vfcmp</tt>' instruction returns an integer vector value based on
4551element-wise comparison of its two floating point vector operands. The output
4552elements have the same width as the input elements.</p>
4553<h5>Arguments:</h5>
4554<p>The '<tt>vfcmp</tt>' instruction takes three operands. The first operand is
4555the condition code indicating the kind of comparison to perform. It is not
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4556a value, just a keyword. The possible condition code are:</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4557<ol>
4558 <li><tt>false</tt>: no comparison, always returns false</li>
4559 <li><tt>oeq</tt>: ordered and equal</li>
4560 <li><tt>ogt</tt>: ordered and greater than </li>
4561 <li><tt>oge</tt>: ordered and greater than or equal</li>
4562 <li><tt>olt</tt>: ordered and less than </li>
4563 <li><tt>ole</tt>: ordered and less than or equal</li>
4564 <li><tt>one</tt>: ordered and not equal</li>
4565 <li><tt>ord</tt>: ordered (no nans)</li>
4566 <li><tt>ueq</tt>: unordered or equal</li>
4567 <li><tt>ugt</tt>: unordered or greater than </li>
4568 <li><tt>uge</tt>: unordered or greater than or equal</li>
4569 <li><tt>ult</tt>: unordered or less than </li>
4570 <li><tt>ule</tt>: unordered or less than or equal</li>
4571 <li><tt>une</tt>: unordered or not equal</li>
4572 <li><tt>uno</tt>: unordered (either nans)</li>
4573 <li><tt>true</tt>: no comparison, always returns true</li>
4574</ol>
4575<p>The remaining two arguments must be <a href="#t_vector">vector</a> of
4576<a href="#t_floating">floating point</a> typed. They must also be identical
4577types.</p>
4578<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4579<p>The '<tt>vfcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4580according to the condition code given as <tt>cond</tt>. The comparison yields a
4581<a href="#t_vector">vector</a> of <a href="#t_integer">integer</a> result, with
4582an identical number of elements as the values being compared, and each element
4583having identical with to the width of the floating point elements. The most
4584significant bit in each element is 1 if the element-wise comparison evaluates to
4585true, and is 0 otherwise. All other bits of the result are undefined. The
4586condition codes are evaluated identically to the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4587<a href="#i_fcmp">'<tt>fcmp</tt>' instruction</a>.</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4588
4589<h5>Example:</h5>
4590<pre>
Chris Lattner1e5c5cd02008-10-13 16:55:18 +0000[diff] [blame]4591 <i>; yields: result=&lt;2 x i32&gt; &lt; i32 0, i32 -1 &gt;</i>
4592 &lt;result&gt; = vfcmp oeq &lt;2 x float&gt; &lt; float 4, float 0 &gt;, &lt; float 5, float 0 &gt;
4593
4594 <i>; yields: result=&lt;2 x i64&gt; &lt; i64 -1, i64 0 &gt;</i>
4595 &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]4596</pre>
4597</div>
4598
4599<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4600<div class="doc_subsubsection">
4601 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
4602</div>
4603
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4604<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4605
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4606<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4607
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4608<pre> &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...<br></pre>
4609<h5>Overview:</h5>
4610<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in
4611the SSA graph representing the function.</p>
4612<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4613
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4614<p>The type of the incoming values is specified with the first type
4615field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
4616as arguments, with one pair for each predecessor basic block of the
4617current block. Only values of <a href="#t_firstclass">first class</a>
4618type may be used as the value arguments to the PHI node. Only labels
4619may be used as the label arguments.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4620
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4621<p>There must be no non-phi instructions between the start of a basic
4622block and the PHI instructions: i.e. PHI instructions must be first in
4623a basic block.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4624
Jay Foad8e2fd2c2009-06-03 10:20:10 +0000[diff] [blame]4625<p>For the purposes of the SSA form, the use of each incoming value is
4626deemed to occur on the edge from the corresponding predecessor block
4627to the current block (but after any definition of an '<tt>invoke</tt>'
4628instruction's return value on the same edge).</p>
4629
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4630<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4631
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4632<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
4633specified by the pair corresponding to the predecessor basic block that executed
4634just prior to the current block.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4635
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4636<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4637<pre>
4638Loop: ; Infinite loop that counts from 0 on up...
4639 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
4640 %nextindvar = add i32 %indvar, 1
4641 br label %Loop
4642</pre>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4643</div>
4644
4645<!-- _______________________________________________________________________ -->
4646<div class="doc_subsubsection">
4647 <a name="i_select">'<tt>select</tt>' Instruction</a>
4648</div>
4649
4650<div class="doc_text">
4651
4652<h5>Syntax:</h5>
4653
4654<pre>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4655 &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>
4656
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4657 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4658</pre>
4659
4660<h5>Overview:</h5>
4661
4662<p>
4663The '<tt>select</tt>' instruction is used to choose one value based on a
4664condition, without branching.
4665</p>
4666
4667
4668<h5>Arguments:</h5>
4669
4670<p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4671The '<tt>select</tt>' instruction requires an 'i1' value or
4672a vector of 'i1' values indicating the
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4673condition, and two values of the same <a href="#t_firstclass">first class</a>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4674type. If the val1/val2 are vectors and
4675the condition is a scalar, then entire vectors are selected, not
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4676individual elements.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4677</p>
4678
4679<h5>Semantics:</h5>
4680
4681<p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4682If 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]4683value argument; otherwise, it returns the second value argument.
4684</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4685<p>
4686If the condition is a vector of i1, then the value arguments must
4687be vectors of the same size, and the selection is done element
4688by element.
4689</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4690
4691<h5>Example:</h5>
4692
4693<pre>
4694 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
4695</pre>
Dan Gohmana53eb382009-01-22 01:39:38 +0000[diff] [blame]4696
4697<p>Note that the code generator does not yet support conditions
4698 with vector type.</p>
4699
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4700</div>
4701
4702
4703<!-- _______________________________________________________________________ -->
4704<div class="doc_subsubsection">
4705 <a name="i_call">'<tt>call</tt>' Instruction</a>
4706</div>
4707
4708<div class="doc_text">
4709
4710<h5>Syntax:</h5>
4711<pre>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]4712 &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]4713</pre>
4714
4715<h5>Overview:</h5>
4716
4717<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
4718
4719<h5>Arguments:</h5>
4720
4721<p>This instruction requires several arguments:</p>
4722
4723<ol>
4724 <li>
4725 <p>The optional "tail" marker indicates whether the callee function accesses
4726 any allocas or varargs in the caller. If the "tail" marker is present, the
4727 function call is eligible for tail call optimization. Note that calls may
4728 be marked "tail" even if they do not occur before a <a
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4729 href="#i_ret"><tt>ret</tt></a> instruction.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4730 </li>
4731 <li>
4732 <p>The optional "cconv" marker indicates which <a href="#callingconv">calling
4733 convention</a> the call should use. If none is specified, the call defaults
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4734 to using C calling conventions.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4735 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4736
4737 <li>
4738 <p>The optional <a href="#paramattrs">Parameter Attributes</a> list for
4739 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>',
4740 and '<tt>inreg</tt>' attributes are valid here.</p>
4741 </li>
4742
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4743 <li>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]4744 <p>'<tt>ty</tt>': the type of the call instruction itself which is also
4745 the type of the return value. Functions that return no value are marked
4746 <tt><a href="#t_void">void</a></tt>.</p>
4747 </li>
4748 <li>
4749 <p>'<tt>fnty</tt>': shall be the signature of the pointer to function
4750 value being invoked. The argument types must match the types implied by
4751 this signature. This type can be omitted if the function is not varargs
4752 and if the function type does not return a pointer to a function.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4753 </li>
4754 <li>
4755 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
4756 be invoked. In most cases, this is a direct function invocation, but
4757 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
4758 to function value.</p>
4759 </li>
4760 <li>
4761 <p>'<tt>function args</tt>': argument list whose types match the
4762 function signature argument types. All arguments must be of
4763 <a href="#t_firstclass">first class</a> type. If the function signature
4764 indicates the function accepts a variable number of arguments, the extra
4765 arguments can be specified.</p>
4766 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4767 <li>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]4768 <p>The optional <a href="#fnattrs">function attributes</a> list. Only
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4769 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
4770 '<tt>readnone</tt>' attributes are valid here.</p>
4771 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4772</ol>
4773
4774<h5>Semantics:</h5>
4775
4776<p>The '<tt>call</tt>' instruction is used to cause control flow to
4777transfer to a specified function, with its incoming arguments bound to
4778the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
4779instruction in the called function, control flow continues with the
4780instruction after the function call, and the return value of the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4781function is bound to the result argument.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4782
4783<h5>Example:</h5>
4784
4785<pre>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]4786 %retval = call i32 @test(i32 %argc)
Chris Lattner5e893ef2008-03-21 17:24:17 +0000[diff] [blame]4787 call i32 (i8 *, ...)* @printf(i8 * %msg, i32 12, i8 42) <i>; yields i32</i>
4788 %X = tail call i32 @foo() <i>; yields i32</i>
4789 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
4790 call void %foo(i8 97 signext)
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]4791
4792 %struct.A = type { i32, i8 }
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4793 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]4794 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
4795 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattnerac454b32008-10-08 06:26:11 +0000[diff] [blame]4796 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijman2c4e05a2008-10-07 10:03:45 +0000[diff] [blame]4797 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4798</pre>
4799
4800</div>
4801
4802<!-- _______________________________________________________________________ -->
4803<div class="doc_subsubsection">
4804 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
4805</div>
4806
4807<div class="doc_text">
4808
4809<h5>Syntax:</h5>
4810
4811<pre>
4812 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
4813</pre>
4814
4815<h5>Overview:</h5>
4816
4817<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
4818the "variable argument" area of a function call. It is used to implement the
4819<tt>va_arg</tt> macro in C.</p>
4820
4821<h5>Arguments:</h5>
4822
4823<p>This instruction takes a <tt>va_list*</tt> value and the type of
4824the argument. It returns a value of the specified argument type and
4825increments the <tt>va_list</tt> to point to the next argument. The
4826actual type of <tt>va_list</tt> is target specific.</p>
4827
4828<h5>Semantics:</h5>
4829
4830<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
4831type from the specified <tt>va_list</tt> and causes the
4832<tt>va_list</tt> to point to the next argument. For more information,
4833see the variable argument handling <a href="#int_varargs">Intrinsic
4834Functions</a>.</p>
4835
4836<p>It is legal for this instruction to be called in a function which does not
4837take a variable number of arguments, for example, the <tt>vfprintf</tt>
4838function.</p>
4839
4840<p><tt>va_arg</tt> is an LLVM instruction instead of an <a
4841href="#intrinsics">intrinsic function</a> because it takes a type as an
4842argument.</p>
4843
4844<h5>Example:</h5>
4845
4846<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
4847
Dan Gohman60967192009-01-12 23:12:39 +0000[diff] [blame]4848<p>Note that the code generator does not yet fully support va_arg
4849 on many targets. Also, it does not currently support va_arg with
4850 aggregate types on any target.</p>
4851
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4852</div>
4853
4854<!-- *********************************************************************** -->
4855<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
4856<!-- *********************************************************************** -->
4857
4858<div class="doc_text">
4859
4860<p>LLVM supports the notion of an "intrinsic function". These functions have
4861well known names and semantics and are required to follow certain restrictions.
4862Overall, these intrinsics represent an extension mechanism for the LLVM
4863language that does not require changing all of the transformations in LLVM when
4864adding to the language (or the bitcode reader/writer, the parser, etc...).</p>
4865
4866<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
4867prefix is reserved in LLVM for intrinsic names; thus, function names may not
4868begin with this prefix. Intrinsic functions must always be external functions:
4869you cannot define the body of intrinsic functions. Intrinsic functions may
4870only be used in call or invoke instructions: it is illegal to take the address
4871of an intrinsic function. Additionally, because intrinsic functions are part
4872of the LLVM language, it is required if any are added that they be documented
4873here.</p>
4874
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4875<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents
4876a family of functions that perform the same operation but on different data
4877types. Because LLVM can represent over 8 million different integer types,
4878overloading is used commonly to allow an intrinsic function to operate on any
4879integer type. One or more of the argument types or the result type can be
4880overloaded to accept any integer type. Argument types may also be defined as
4881exactly matching a previous argument's type or the result type. This allows an
4882intrinsic function which accepts multiple arguments, but needs all of them to
4883be of the same type, to only be overloaded with respect to a single argument or
4884the result.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4885
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4886<p>Overloaded intrinsics will have the names of its overloaded argument types
4887encoded into its function name, each preceded by a period. Only those types
4888which are overloaded result in a name suffix. Arguments whose type is matched
4889against another type do not. For example, the <tt>llvm.ctpop</tt> function can
4890take an integer of any width and returns an integer of exactly the same integer
4891width. This leads to a family of functions such as
4892<tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29 %val)</tt>.
4893Only one type, the return type, is overloaded, and only one type suffix is
4894required. Because the argument's type is matched against the return type, it
4895does not require its own name suffix.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4896
4897<p>To learn how to add an intrinsic function, please see the
4898<a href="ExtendingLLVM.html">Extending LLVM Guide</a>.
4899</p>
4900
4901</div>
4902
4903<!-- ======================================================================= -->
4904<div class="doc_subsection">
4905 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
4906</div>
4907
4908<div class="doc_text">
4909
4910<p>Variable argument support is defined in LLVM with the <a
4911 href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
4912intrinsic functions. These functions are related to the similarly
4913named macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
4914
4915<p>All of these functions operate on arguments that use a
4916target-specific value type "<tt>va_list</tt>". The LLVM assembly
4917language reference manual does not define what this type is, so all
4918transformations should be prepared to handle these functions regardless of
4919the type used.</p>
4920
4921<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
4922instruction and the variable argument handling intrinsic functions are
4923used.</p>
4924
4925<div class="doc_code">
4926<pre>
4927define i32 @test(i32 %X, ...) {
4928 ; Initialize variable argument processing
4929 %ap = alloca i8*
4930 %ap2 = bitcast i8** %ap to i8*
4931 call void @llvm.va_start(i8* %ap2)
4932
4933 ; Read a single integer argument
4934 %tmp = va_arg i8** %ap, i32
4935
4936 ; Demonstrate usage of llvm.va_copy and llvm.va_end
4937 %aq = alloca i8*
4938 %aq2 = bitcast i8** %aq to i8*
4939 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
4940 call void @llvm.va_end(i8* %aq2)
4941
4942 ; Stop processing of arguments.
4943 call void @llvm.va_end(i8* %ap2)
4944 ret i32 %tmp
4945}
4946
4947declare void @llvm.va_start(i8*)
4948declare void @llvm.va_copy(i8*, i8*)
4949declare void @llvm.va_end(i8*)
4950</pre>
4951</div>
4952
4953</div>
4954
4955<!-- _______________________________________________________________________ -->
4956<div class="doc_subsubsection">
4957 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
4958</div>
4959
4960
4961<div class="doc_text">
4962<h5>Syntax:</h5>
4963<pre> declare void %llvm.va_start(i8* &lt;arglist&gt;)<br></pre>
4964<h5>Overview:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4965<p>The '<tt>llvm.va_start</tt>' intrinsic initializes
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4966<tt>*&lt;arglist&gt;</tt> for subsequent use by <tt><a
4967href="#i_va_arg">va_arg</a></tt>.</p>
4968
4969<h5>Arguments:</h5>
4970
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4971<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]4972
4973<h5>Semantics:</h5>
4974
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4975<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]4976macro available in C. In a target-dependent way, it initializes the
4977<tt>va_list</tt> element to which the argument points, so that the next call to
4978<tt>va_arg</tt> will produce the first variable argument passed to the function.
4979Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
4980last argument of the function as the compiler can figure that out.</p>
4981
4982</div>
4983
4984<!-- _______________________________________________________________________ -->
4985<div class="doc_subsubsection">
4986 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
4987</div>
4988
4989<div class="doc_text">
4990<h5>Syntax:</h5>
4991<pre> declare void @llvm.va_end(i8* &lt;arglist&gt;)<br></pre>
4992<h5>Overview:</h5>
4993
4994<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
4995which has been initialized previously with <tt><a href="#int_va_start">llvm.va_start</a></tt>
4996or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
4997
4998<h5>Arguments:</h5>
4999
5000<p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
5001
5002<h5>Semantics:</h5>
5003
5004<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
5005macro available in C. In a target-dependent way, it destroys the
5006<tt>va_list</tt> element to which the argument points. Calls to <a
5007href="#int_va_start"><tt>llvm.va_start</tt></a> and <a href="#int_va_copy">
5008<tt>llvm.va_copy</tt></a> must be matched exactly with calls to
5009<tt>llvm.va_end</tt>.</p>
5010
5011</div>
5012
5013<!-- _______________________________________________________________________ -->
5014<div class="doc_subsubsection">
5015 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
5016</div>
5017
5018<div class="doc_text">
5019
5020<h5>Syntax:</h5>
5021
5022<pre>
5023 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
5024</pre>
5025
5026<h5>Overview:</h5>
5027
5028<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
5029from the source argument list to the destination argument list.</p>
5030
5031<h5>Arguments:</h5>
5032
5033<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
5034The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
5035
5036
5037<h5>Semantics:</h5>
5038
5039<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
5040macro available in C. In a target-dependent way, it copies the source
5041<tt>va_list</tt> element into the destination <tt>va_list</tt> element. This
5042intrinsic is necessary because the <tt><a href="#int_va_start">
5043llvm.va_start</a></tt> intrinsic may be arbitrarily complex and require, for
5044example, memory allocation.</p>
5045
5046</div>
5047
5048<!-- ======================================================================= -->
5049<div class="doc_subsection">
5050 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
5051</div>
5052
5053<div class="doc_text">
5054
5055<p>
5056LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]5057Collection</a> (GC) requires the implementation and generation of these
5058intrinsics.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5059These intrinsics allow identification of <a href="#int_gcroot">GC roots on the
5060stack</a>, as well as garbage collector implementations that require <a
5061href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a> barriers.
5062Front-ends for type-safe garbage collected languages should generate these
5063intrinsics to make use of the LLVM garbage collectors. For more details, see <a
5064href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
5065</p>
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]5066
5067<p>The garbage collection intrinsics only operate on objects in the generic
5068 address space (address space zero).</p>
5069
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5070</div>
5071
5072<!-- _______________________________________________________________________ -->
5073<div class="doc_subsubsection">
5074 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
5075</div>
5076
5077<div class="doc_text">
5078
5079<h5>Syntax:</h5>
5080
5081<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5082 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5083</pre>
5084
5085<h5>Overview:</h5>
5086
5087<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
5088the code generator, and allows some metadata to be associated with it.</p>
5089
5090<h5>Arguments:</h5>
5091
5092<p>The first argument specifies the address of a stack object that contains the
5093root pointer. The second pointer (which must be either a constant or a global
5094value address) contains the meta-data to be associated with the root.</p>
5095
5096<h5>Semantics:</h5>
5097
Chris Lattnera7d94ba2008-04-24 05:59:56 +0000[diff] [blame]5098<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]5099location. At compile-time, the code generator generates information to allow
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]5100the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
5101intrinsic may only be used in a function which <a href="#gc">specifies a GC
5102algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5103
5104</div>
5105
5106
5107<!-- _______________________________________________________________________ -->
5108<div class="doc_subsubsection">
5109 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
5110</div>
5111
5112<div class="doc_text">
5113
5114<h5>Syntax:</h5>
5115
5116<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5117 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5118</pre>
5119
5120<h5>Overview:</h5>
5121
5122<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
5123locations, allowing garbage collector implementations that require read
5124barriers.</p>
5125
5126<h5>Arguments:</h5>
5127
5128<p>The second argument is the address to read from, which should be an address
5129allocated from the garbage collector. The first object is a pointer to the
5130start of the referenced object, if needed by the language runtime (otherwise
5131null).</p>
5132
5133<h5>Semantics:</h5>
5134
5135<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
5136instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]5137garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
5138may only be used in a function which <a href="#gc">specifies a GC
5139algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5140
5141</div>
5142
5143
5144<!-- _______________________________________________________________________ -->
5145<div class="doc_subsubsection">
5146 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
5147</div>
5148
5149<div class="doc_text">
5150
5151<h5>Syntax:</h5>
5152
5153<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5154 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5155</pre>
5156
5157<h5>Overview:</h5>
5158
5159<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
5160locations, allowing garbage collector implementations that require write
5161barriers (such as generational or reference counting collectors).</p>
5162
5163<h5>Arguments:</h5>
5164
5165<p>The first argument is the reference to store, the second is the start of the
5166object to store it to, and the third is the address of the field of Obj to
5167store to. If the runtime does not require a pointer to the object, Obj may be
5168null.</p>
5169
5170<h5>Semantics:</h5>
5171
5172<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
5173instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]5174garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
5175may only be used in a function which <a href="#gc">specifies a GC
5176algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5177
5178</div>
5179
5180
5181
5182<!-- ======================================================================= -->
5183<div class="doc_subsection">
5184 <a name="int_codegen">Code Generator Intrinsics</a>
5185</div>
5186
5187<div class="doc_text">
5188<p>
5189These intrinsics are provided by LLVM to expose special features that may only
5190be implemented with code generator support.
5191</p>
5192
5193</div>
5194
5195<!-- _______________________________________________________________________ -->
5196<div class="doc_subsubsection">
5197 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
5198</div>
5199
5200<div class="doc_text">
5201
5202<h5>Syntax:</h5>
5203<pre>
5204 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
5205</pre>
5206
5207<h5>Overview:</h5>
5208
5209<p>
5210The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
5211target-specific value indicating the return address of the current function
5212or one of its callers.
5213</p>
5214
5215<h5>Arguments:</h5>
5216
5217<p>
5218The argument to this intrinsic indicates which function to return the address
5219for. Zero indicates the calling function, one indicates its caller, etc. The
5220argument is <b>required</b> to be a constant integer value.
5221</p>
5222
5223<h5>Semantics:</h5>
5224
5225<p>
5226The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
5227the return address of the specified call frame, or zero if it cannot be
5228identified. The value returned by this intrinsic is likely to be incorrect or 0
5229for arguments other than zero, so it should only be used for debugging purposes.
5230</p>
5231
5232<p>
5233Note that calling this intrinsic does not prevent function inlining or other
5234aggressive transformations, so the value returned may not be that of the obvious
5235source-language caller.
5236</p>
5237</div>
5238
5239
5240<!-- _______________________________________________________________________ -->
5241<div class="doc_subsubsection">
5242 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
5243</div>
5244
5245<div class="doc_text">
5246
5247<h5>Syntax:</h5>
5248<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5249 declare i8 *@llvm.frameaddress(i32 &lt;level&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5250</pre>
5251
5252<h5>Overview:</h5>
5253
5254<p>
5255The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
5256target-specific frame pointer value for the specified stack frame.
5257</p>
5258
5259<h5>Arguments:</h5>
5260
5261<p>
5262The argument to this intrinsic indicates which function to return the frame
5263pointer for. Zero indicates the calling function, one indicates its caller,
5264etc. The argument is <b>required</b> to be a constant integer value.
5265</p>
5266
5267<h5>Semantics:</h5>
5268
5269<p>
5270The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
5271the frame address of the specified call frame, or zero if it cannot be
5272identified. The value returned by this intrinsic is likely to be incorrect or 0
5273for arguments other than zero, so it should only be used for debugging purposes.
5274</p>
5275
5276<p>
5277Note that calling this intrinsic does not prevent function inlining or other
5278aggressive transformations, so the value returned may not be that of the obvious
5279source-language caller.
5280</p>
5281</div>
5282
5283<!-- _______________________________________________________________________ -->
5284<div class="doc_subsubsection">
5285 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
5286</div>
5287
5288<div class="doc_text">
5289
5290<h5>Syntax:</h5>
5291<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5292 declare i8 *@llvm.stacksave()
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5293</pre>
5294
5295<h5>Overview:</h5>
5296
5297<p>
5298The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
5299the function stack, for use with <a href="#int_stackrestore">
5300<tt>llvm.stackrestore</tt></a>. This is useful for implementing language
5301features like scoped automatic variable sized arrays in C99.
5302</p>
5303
5304<h5>Semantics:</h5>
5305
5306<p>
5307This intrinsic returns a opaque pointer value that can be passed to <a
5308href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
5309<tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
5310<tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
5311state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
5312practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
5313that were allocated after the <tt>llvm.stacksave</tt> was executed.
5314</p>
5315
5316</div>
5317
5318<!-- _______________________________________________________________________ -->
5319<div class="doc_subsubsection">
5320 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
5321</div>
5322
5323<div class="doc_text">
5324
5325<h5>Syntax:</h5>
5326<pre>
5327 declare void @llvm.stackrestore(i8 * %ptr)
5328</pre>
5329
5330<h5>Overview:</h5>
5331
5332<p>
5333The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
5334the function stack to the state it was in when the corresponding <a
5335href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
5336useful for implementing language features like scoped automatic variable sized
5337arrays in C99.
5338</p>
5339
5340<h5>Semantics:</h5>
5341
5342<p>
5343See the description for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.
5344</p>
5345
5346</div>
5347
5348
5349<!-- _______________________________________________________________________ -->
5350<div class="doc_subsubsection">
5351 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
5352</div>
5353
5354<div class="doc_text">
5355
5356<h5>Syntax:</h5>
5357<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5358 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]5359</pre>
5360
5361<h5>Overview:</h5>
5362
5363
5364<p>
5365The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
5366a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
5367no
5368effect on the behavior of the program but can change its performance
5369characteristics.
5370</p>
5371
5372<h5>Arguments:</h5>
5373
5374<p>
5375<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
5376determining if the fetch should be for a read (0) or write (1), and
5377<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
5378locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
5379<tt>locality</tt> arguments must be constant integers.
5380</p>
5381
5382<h5>Semantics:</h5>
5383
5384<p>
5385This intrinsic does not modify the behavior of the program. In particular,
5386prefetches cannot trap and do not produce a value. On targets that support this
5387intrinsic, the prefetch can provide hints to the processor cache for better
5388performance.
5389</p>
5390
5391</div>
5392
5393<!-- _______________________________________________________________________ -->
5394<div class="doc_subsubsection">
5395 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
5396</div>
5397
5398<div class="doc_text">
5399
5400<h5>Syntax:</h5>
5401<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5402 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5403</pre>
5404
5405<h5>Overview:</h5>
5406
5407
5408<p>
5409The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]5410(PC) in a region of
5411code to simulators and other tools. The method is target specific, but it is
5412expected that the marker will use exported symbols to transmit the PC of the
5413marker.
5414The marker makes no guarantees that it will remain with any specific instruction
5415after optimizations. It is possible that the presence of a marker will inhibit
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5416optimizations. The intended use is to be inserted after optimizations to allow
5417correlations of simulation runs.
5418</p>
5419
5420<h5>Arguments:</h5>
5421
5422<p>
5423<tt>id</tt> is a numerical id identifying the marker.
5424</p>
5425
5426<h5>Semantics:</h5>
5427
5428<p>
5429This intrinsic does not modify the behavior of the program. Backends that do not
5430support this intrinisic may ignore it.
5431</p>
5432
5433</div>
5434
5435<!-- _______________________________________________________________________ -->
5436<div class="doc_subsubsection">
5437 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
5438</div>
5439
5440<div class="doc_text">
5441
5442<h5>Syntax:</h5>
5443<pre>
5444 declare i64 @llvm.readcyclecounter( )
5445</pre>
5446
5447<h5>Overview:</h5>
5448
5449
5450<p>
5451The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
5452counter register (or similar low latency, high accuracy clocks) on those targets
5453that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
5454As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
5455should only be used for small timings.
5456</p>
5457
5458<h5>Semantics:</h5>
5459
5460<p>
5461When directly supported, reading the cycle counter should not modify any memory.
5462Implementations are allowed to either return a application specific value or a
5463system wide value. On backends without support, this is lowered to a constant 0.
5464</p>
5465
5466</div>
5467
5468<!-- ======================================================================= -->
5469<div class="doc_subsection">
5470 <a name="int_libc">Standard C Library Intrinsics</a>
5471</div>
5472
5473<div class="doc_text">
5474<p>
5475LLVM provides intrinsics for a few important standard C library functions.
5476These intrinsics allow source-language front-ends to pass information about the
5477alignment of the pointer arguments to the code generator, providing opportunity
5478for more efficient code generation.
5479</p>
5480
5481</div>
5482
5483<!-- _______________________________________________________________________ -->
5484<div class="doc_subsubsection">
5485 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
5486</div>
5487
5488<div class="doc_text">
5489
5490<h5>Syntax:</h5>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5491<p>This is an overloaded intrinsic. You can use llvm.memcpy on any integer bit
5492width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5493<pre>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5494 declare void @llvm.memcpy.i8(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5495 i8 &lt;len&gt;, i32 &lt;align&gt;)
5496 declare void @llvm.memcpy.i16(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5497 i16 &lt;len&gt;, i32 &lt;align&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5498 declare void @llvm.memcpy.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5499 i32 &lt;len&gt;, i32 &lt;align&gt;)
5500 declare void @llvm.memcpy.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5501 i64 &lt;len&gt;, i32 &lt;align&gt;)
5502</pre>
5503
5504<h5>Overview:</h5>
5505
5506<p>
5507The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
5508location to the destination location.
5509</p>
5510
5511<p>
5512Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
5513intrinsics do not return a value, and takes an extra alignment argument.
5514</p>
5515
5516<h5>Arguments:</h5>
5517
5518<p>
5519The first argument is a pointer to the destination, the second is a pointer to
5520the source. The third argument is an integer argument
5521specifying the number of bytes to copy, and the fourth argument is the alignment
5522of the source and destination locations.
5523</p>
5524
5525<p>
5526If the call to this intrinisic has an alignment value that is not 0 or 1, then
5527the caller guarantees that both the source and destination pointers are aligned
5528to that boundary.
5529</p>
5530
5531<h5>Semantics:</h5>
5532
5533<p>
5534The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
5535location to the destination location, which are not allowed to overlap. It
5536copies "len" bytes of memory over. If the argument is known to be aligned to
5537some boundary, this can be specified as the fourth argument, otherwise it should
5538be set to 0 or 1.
5539</p>
5540</div>
5541
5542
5543<!-- _______________________________________________________________________ -->
5544<div class="doc_subsubsection">
5545 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
5546</div>
5547
5548<div class="doc_text">
5549
5550<h5>Syntax:</h5>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5551<p>This is an overloaded intrinsic. You can use llvm.memmove on any integer bit
5552width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5553<pre>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5554 declare void @llvm.memmove.i8(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5555 i8 &lt;len&gt;, i32 &lt;align&gt;)
5556 declare void @llvm.memmove.i16(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5557 i16 &lt;len&gt;, i32 &lt;align&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5558 declare void @llvm.memmove.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5559 i32 &lt;len&gt;, i32 &lt;align&gt;)
5560 declare void @llvm.memmove.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5561 i64 &lt;len&gt;, i32 &lt;align&gt;)
5562</pre>
5563
5564<h5>Overview:</h5>
5565
5566<p>
5567The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the source
5568location to the destination location. It is similar to the
Chris Lattnerdba16ea2008-01-06 19:51:52 +0000[diff] [blame]5569'<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to overlap.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5570</p>
5571
5572<p>
5573Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
5574intrinsics do not return a value, and takes an extra alignment argument.
5575</p>
5576
5577<h5>Arguments:</h5>
5578
5579<p>
5580The first argument is a pointer to the destination, the second is a pointer to
5581the source. The third argument is an integer argument
5582specifying the number of bytes to copy, and the fourth argument is the alignment
5583of the source and destination locations.
5584</p>
5585
5586<p>
5587If the call to this intrinisic has an alignment value that is not 0 or 1, then
5588the caller guarantees that the source and destination pointers are aligned to
5589that boundary.
5590</p>
5591
5592<h5>Semantics:</h5>
5593
5594<p>
5595The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the source
5596location to the destination location, which may overlap. It
5597copies "len" bytes of memory over. If the argument is known to be aligned to
5598some boundary, this can be specified as the fourth argument, otherwise it should
5599be set to 0 or 1.
5600</p>
5601</div>
5602
5603
5604<!-- _______________________________________________________________________ -->
5605<div class="doc_subsubsection">
5606 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
5607</div>
5608
5609<div class="doc_text">
5610
5611<h5>Syntax:</h5>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5612<p>This is an overloaded intrinsic. You can use llvm.memset on any integer bit
5613width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5614<pre>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5615 declare void @llvm.memset.i8(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5616 i8 &lt;len&gt;, i32 &lt;align&gt;)
5617 declare void @llvm.memset.i16(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5618 i16 &lt;len&gt;, i32 &lt;align&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5619 declare void @llvm.memset.i32(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5620 i32 &lt;len&gt;, i32 &lt;align&gt;)
5621 declare void @llvm.memset.i64(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5622 i64 &lt;len&gt;, i32 &lt;align&gt;)
5623</pre>
5624
5625<h5>Overview:</h5>
5626
5627<p>
5628The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a particular
5629byte value.
5630</p>
5631
5632<p>
5633Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
5634does not return a value, and takes an extra alignment argument.
5635</p>
5636
5637<h5>Arguments:</h5>
5638
5639<p>
5640The first argument is a pointer to the destination to fill, the second is the
5641byte value to fill it with, the third argument is an integer
5642argument specifying the number of bytes to fill, and the fourth argument is the
5643known alignment of destination location.
5644</p>
5645
5646<p>
5647If the call to this intrinisic has an alignment value that is not 0 or 1, then
5648the caller guarantees that the destination pointer is aligned to that boundary.
5649</p>
5650
5651<h5>Semantics:</h5>
5652
5653<p>
5654The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting at
5655the
5656destination location. If the argument is known to be aligned to some boundary,
5657this can be specified as the fourth argument, otherwise it should be set to 0 or
56581.
5659</p>
5660</div>
5661
5662
5663<!-- _______________________________________________________________________ -->
5664<div class="doc_subsubsection">
5665 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
5666</div>
5667
5668<div class="doc_text">
5669
5670<h5>Syntax:</h5>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5671<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]5672floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5673types however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5674<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5675 declare float @llvm.sqrt.f32(float %Val)
5676 declare double @llvm.sqrt.f64(double %Val)
5677 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
5678 declare fp128 @llvm.sqrt.f128(fp128 %Val)
5679 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5680</pre>
5681
5682<h5>Overview:</h5>
5683
5684<p>
5685The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5686returning the same value as the libm '<tt>sqrt</tt>' functions would. Unlike
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5687<tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
Chris Lattnerf914a002008-01-29 07:00:44 +0000[diff] [blame]5688negative numbers other than -0.0 (which allows for better optimization, because
5689there is no need to worry about errno being set). <tt>llvm.sqrt(-0.0)</tt> is
5690defined to return -0.0 like IEEE sqrt.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5691</p>
5692
5693<h5>Arguments:</h5>
5694
5695<p>
5696The argument and return value are floating point numbers of the same type.
5697</p>
5698
5699<h5>Semantics:</h5>
5700
5701<p>
5702This function returns the sqrt of the specified operand if it is a nonnegative
5703floating point number.
5704</p>
5705</div>
5706
5707<!-- _______________________________________________________________________ -->
5708<div class="doc_subsubsection">
5709 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
5710</div>
5711
5712<div class="doc_text">
5713
5714<h5>Syntax:</h5>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5715<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]5716floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5717types however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5718<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5719 declare float @llvm.powi.f32(float %Val, i32 %power)
5720 declare double @llvm.powi.f64(double %Val, i32 %power)
5721 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
5722 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
5723 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5724</pre>
5725
5726<h5>Overview:</h5>
5727
5728<p>
5729The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
5730specified (positive or negative) power. The order of evaluation of
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5731multiplications is not defined. When a vector of floating point type is
5732used, the second argument remains a scalar integer value.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5733</p>
5734
5735<h5>Arguments:</h5>
5736
5737<p>
5738The second argument is an integer power, and the first is a value to raise to
5739that power.
5740</p>
5741
5742<h5>Semantics:</h5>
5743
5744<p>
5745This function returns the first value raised to the second power with an
5746unspecified sequence of rounding operations.</p>
5747</div>
5748
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5749<!-- _______________________________________________________________________ -->
5750<div class="doc_subsubsection">
5751 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
5752</div>
5753
5754<div class="doc_text">
5755
5756<h5>Syntax:</h5>
5757<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
5758floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5759types however.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5760<pre>
5761 declare float @llvm.sin.f32(float %Val)
5762 declare double @llvm.sin.f64(double %Val)
5763 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
5764 declare fp128 @llvm.sin.f128(fp128 %Val)
5765 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
5766</pre>
5767
5768<h5>Overview:</h5>
5769
5770<p>
5771The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.
5772</p>
5773
5774<h5>Arguments:</h5>
5775
5776<p>
5777The argument and return value are floating point numbers of the same type.
5778</p>
5779
5780<h5>Semantics:</h5>
5781
5782<p>
5783This function returns the sine of the specified operand, returning the
5784same values as the libm <tt>sin</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5785conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5786</div>
5787
5788<!-- _______________________________________________________________________ -->
5789<div class="doc_subsubsection">
5790 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
5791</div>
5792
5793<div class="doc_text">
5794
5795<h5>Syntax:</h5>
5796<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
5797floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5798types however.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5799<pre>
5800 declare float @llvm.cos.f32(float %Val)
5801 declare double @llvm.cos.f64(double %Val)
5802 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
5803 declare fp128 @llvm.cos.f128(fp128 %Val)
5804 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
5805</pre>
5806
5807<h5>Overview:</h5>
5808
5809<p>
5810The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.
5811</p>
5812
5813<h5>Arguments:</h5>
5814
5815<p>
5816The argument and return value are floating point numbers of the same type.
5817</p>
5818
5819<h5>Semantics:</h5>
5820
5821<p>
5822This function returns the cosine of the specified operand, returning the
5823same values as the libm <tt>cos</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5824conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5825</div>
5826
5827<!-- _______________________________________________________________________ -->
5828<div class="doc_subsubsection">
5829 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
5830</div>
5831
5832<div class="doc_text">
5833
5834<h5>Syntax:</h5>
5835<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
5836floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5837types however.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5838<pre>
5839 declare float @llvm.pow.f32(float %Val, float %Power)
5840 declare double @llvm.pow.f64(double %Val, double %Power)
5841 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
5842 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
5843 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
5844</pre>
5845
5846<h5>Overview:</h5>
5847
5848<p>
5849The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
5850specified (positive or negative) power.
5851</p>
5852
5853<h5>Arguments:</h5>
5854
5855<p>
5856The second argument is a floating point power, and the first is a value to
5857raise to that power.
5858</p>
5859
5860<h5>Semantics:</h5>
5861
5862<p>
5863This function returns the first value raised to the second power,
5864returning the
5865same values as the libm <tt>pow</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5866conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5867</div>
5868
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5869
5870<!-- ======================================================================= -->
5871<div class="doc_subsection">
5872 <a name="int_manip">Bit Manipulation Intrinsics</a>
5873</div>
5874
5875<div class="doc_text">
5876<p>
5877LLVM provides intrinsics for a few important bit manipulation operations.
5878These allow efficient code generation for some algorithms.
5879</p>
5880
5881</div>
5882
5883<!-- _______________________________________________________________________ -->
5884<div class="doc_subsubsection">
5885 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
5886</div>
5887
5888<div class="doc_text">
5889
5890<h5>Syntax:</h5>
5891<p>This is an overloaded intrinsic function. You can use bswap on any integer
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5892type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5893<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5894 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
5895 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
5896 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5897</pre>
5898
5899<h5>Overview:</h5>
5900
5901<p>
5902The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
5903values with an even number of bytes (positive multiple of 16 bits). These are
5904useful for performing operations on data that is not in the target's native
5905byte order.
5906</p>
5907
5908<h5>Semantics:</h5>
5909
5910<p>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5911The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5912and low byte of the input i16 swapped. Similarly, the <tt>llvm.bswap.i32</tt>
5913intrinsic returns an i32 value that has the four bytes of the input i32
5914swapped, 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]5915i32 will have its bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i48</tt>,
5916<tt>llvm.bswap.i64</tt> and other intrinsics extend this concept to
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5917additional even-byte lengths (6 bytes, 8 bytes and more, respectively).
5918</p>
5919
5920</div>
5921
5922<!-- _______________________________________________________________________ -->
5923<div class="doc_subsubsection">
5924 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
5925</div>
5926
5927<div class="doc_text">
5928
5929<h5>Syntax:</h5>
5930<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]5931width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5932<pre>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5933 declare i8 @llvm.ctpop.i8(i8 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5934 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5935 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5936 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
5937 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5938</pre>
5939
5940<h5>Overview:</h5>
5941
5942<p>
5943The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
5944value.
5945</p>
5946
5947<h5>Arguments:</h5>
5948
5949<p>
5950The only argument is the value to be counted. The argument may be of any
5951integer type. The return type must match the argument type.
5952</p>
5953
5954<h5>Semantics:</h5>
5955
5956<p>
5957The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
5958</p>
5959</div>
5960
5961<!-- _______________________________________________________________________ -->
5962<div class="doc_subsubsection">
5963 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
5964</div>
5965
5966<div class="doc_text">
5967
5968<h5>Syntax:</h5>
5969<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]5970integer bit width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5971<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5972 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
5973 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5974 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5975 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
5976 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5977</pre>
5978
5979<h5>Overview:</h5>
5980
5981<p>
5982The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
5983leading zeros in a variable.
5984</p>
5985
5986<h5>Arguments:</h5>
5987
5988<p>
5989The only argument is the value to be counted. The argument may be of any
5990integer type. The return type must match the argument type.
5991</p>
5992
5993<h5>Semantics:</h5>
5994
5995<p>
5996The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
5997in a variable. If the src == 0 then the result is the size in bits of the type
5998of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.
5999</p>
6000</div>
6001
6002
6003
6004<!-- _______________________________________________________________________ -->
6005<div class="doc_subsubsection">
6006 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
6007</div>
6008
6009<div class="doc_text">
6010
6011<h5>Syntax:</h5>
6012<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]6013integer bit width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6014<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]6015 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
6016 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6017 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]6018 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
6019 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6020</pre>
6021
6022<h5>Overview:</h5>
6023
6024<p>
6025The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
6026trailing zeros.
6027</p>
6028
6029<h5>Arguments:</h5>
6030
6031<p>
6032The only argument is the value to be counted. The argument may be of any
6033integer type. The return type must match the argument type.
6034</p>
6035
6036<h5>Semantics:</h5>
6037
6038<p>
6039The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
6040in a variable. If the src == 0 then the result is the size in bits of the type
6041of src. For example, <tt>llvm.cttz(2) = 1</tt>.
6042</p>
6043</div>
6044
6045<!-- _______________________________________________________________________ -->
6046<div class="doc_subsubsection">
6047 <a name="int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic</a>
6048</div>
6049
6050<div class="doc_text">
6051
6052<h5>Syntax:</h5>
6053<p>This is an overloaded intrinsic. You can use <tt>llvm.part.select</tt>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]6054on any integer bit width.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6055<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]6056 declare i17 @llvm.part.select.i17 (i17 %val, i32 %loBit, i32 %hiBit)
6057 declare i29 @llvm.part.select.i29 (i29 %val, i32 %loBit, i32 %hiBit)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6058</pre>
6059
6060<h5>Overview:</h5>
6061<p>The '<tt>llvm.part.select</tt>' family of intrinsic functions selects a
6062range of bits from an integer value and returns them in the same bit width as
6063the original value.</p>
6064
6065<h5>Arguments:</h5>
6066<p>The first argument, <tt>%val</tt> and the result may be integer types of
6067any bit width but they must have the same bit width. The second and third
6068arguments must be <tt>i32</tt> type since they specify only a bit index.</p>
6069
6070<h5>Semantics:</h5>
6071<p>The operation of the '<tt>llvm.part.select</tt>' intrinsic has two modes
6072of operation: forwards and reverse. If <tt>%loBit</tt> is greater than
6073<tt>%hiBits</tt> then the intrinsic operates in reverse mode. Otherwise it
6074operates in forward mode.</p>
6075<p>In forward mode, this intrinsic is the equivalent of shifting <tt>%val</tt>
6076right by <tt>%loBit</tt> bits and then ANDing it with a mask with
6077only the <tt>%hiBit - %loBit</tt> bits set, as follows:</p>
6078<ol>
6079 <li>The <tt>%val</tt> is shifted right (LSHR) by the number of bits specified
6080 by <tt>%loBits</tt>. This normalizes the value to the low order bits.</li>
6081 <li>The <tt>%loBits</tt> value is subtracted from the <tt>%hiBits</tt> value
6082 to determine the number of bits to retain.</li>
6083 <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]6084 <li>The mask is ANDed with <tt>%val</tt> to produce the result.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6085</ol>
6086<p>In reverse mode, a similar computation is made except that the bits are
6087returned in the reverse order. So, for example, if <tt>X</tt> has the value
6088<tt>i16 0x0ACF (101011001111)</tt> and we apply
6089<tt>part.select(i16 X, 8, 3)</tt> to it, we get back the value
6090<tt>i16 0x0026 (000000100110)</tt>.</p>
6091</div>
6092
6093<div class="doc_subsubsection">
6094 <a name="int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic</a>
6095</div>
6096
6097<div class="doc_text">
6098
6099<h5>Syntax:</h5>
6100<p>This is an overloaded intrinsic. You can use <tt>llvm.part.set</tt>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]6101on any integer bit width.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6102<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]6103 declare i17 @llvm.part.set.i17.i9 (i17 %val, i9 %repl, i32 %lo, i32 %hi)
6104 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]6105</pre>
6106
6107<h5>Overview:</h5>
6108<p>The '<tt>llvm.part.set</tt>' family of intrinsic functions replaces a range
6109of bits in an integer value with another integer value. It returns the integer
6110with the replaced bits.</p>
6111
6112<h5>Arguments:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6113<p>The first argument, <tt>%val</tt>, and the result may be integer types of
6114any 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]6115whose bits will be replaced. The second argument, <tt>%repl</tt> may be an
6116integer of any bit width. The third and fourth arguments must be <tt>i32</tt>
6117type since they specify only a bit index.</p>
6118
6119<h5>Semantics:</h5>
6120<p>The operation of the '<tt>llvm.part.set</tt>' intrinsic has two modes
6121of operation: forwards and reverse. If <tt>%lo</tt> is greater than
6122<tt>%hi</tt> then the intrinsic operates in reverse mode. Otherwise it
6123operates in forward mode.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6124
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6125<p>For both modes, the <tt>%repl</tt> value is prepared for use by either
6126truncating it down to the size of the replacement area or zero extending it
6127up to that size.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6128
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6129<p>In forward mode, the bits between <tt>%lo</tt> and <tt>%hi</tt> (inclusive)
6130are replaced with corresponding bits from <tt>%repl</tt>. That is the 0th bit
6131in <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]6132to the <tt>%hi</tt>th bit.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6133
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6134<p>In reverse mode, a similar computation is made except that the bits are
6135reversed. That is, the <tt>0</tt>th bit in <tt>%repl</tt> replaces the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]6136<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]6137
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6138<h5>Examples:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6139
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6140<pre>
6141 llvm.part.set(0xFFFF, 0, 4, 7) -&gt; 0xFF0F
6142 llvm.part.set(0xFFFF, 0, 7, 4) -&gt; 0xFF0F
6143 llvm.part.set(0xFFFF, 1, 7, 4) -&gt; 0xFF8F
6144 llvm.part.set(0xFFFF, F, 8, 3) -&gt; 0xFFE7
6145 llvm.part.set(0xFFFF, 0, 3, 8) -&gt; 0xFE07
6146</pre>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6147
6148</div>
6149
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6150<!-- ======================================================================= -->
6151<div class="doc_subsection">
6152 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
6153</div>
6154
6155<div class="doc_text">
6156<p>
6157LLVM provides intrinsics for some arithmetic with overflow operations.
6158</p>
6159
6160</div>
6161
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6162<!-- _______________________________________________________________________ -->
6163<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6164 <a name="int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6165</div>
6166
6167<div class="doc_text">
6168
6169<h5>Syntax:</h5>
6170
6171<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]6172on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6173
6174<pre>
6175 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
6176 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6177 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
6178</pre>
6179
6180<h5>Overview:</h5>
6181
6182<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
6183a signed addition of the two arguments, and indicate whether an overflow
6184occurred during the signed summation.</p>
6185
6186<h5>Arguments:</h5>
6187
6188<p>The arguments (%a and %b) and the first element of the result structure may
6189be of integer types of any bit width, but they must have the same bit width. The
6190second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6191and <tt>%b</tt> are the two values that will undergo signed addition.</p>
6192
6193<h5>Semantics:</h5>
6194
6195<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
6196a signed addition of the two variables. They return a structure &mdash; the
6197first element of which is the signed summation, and the second element of which
6198is a bit specifying if the signed summation resulted in an overflow.</p>
6199
6200<h5>Examples:</h5>
6201<pre>
6202 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6203 %sum = extractvalue {i32, i1} %res, 0
6204 %obit = extractvalue {i32, i1} %res, 1
6205 br i1 %obit, label %overflow, label %normal
6206</pre>
6207
6208</div>
6209
6210<!-- _______________________________________________________________________ -->
6211<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6212 <a name="int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6213</div>
6214
6215<div class="doc_text">
6216
6217<h5>Syntax:</h5>
6218
6219<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]6220on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6221
6222<pre>
6223 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
6224 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6225 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
6226</pre>
6227
6228<h5>Overview:</h5>
6229
6230<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
6231an unsigned addition of the two arguments, and indicate whether a carry occurred
6232during the unsigned summation.</p>
6233
6234<h5>Arguments:</h5>
6235
6236<p>The arguments (%a and %b) and the first element of the result structure may
6237be of integer types of any bit width, but they must have the same bit width. The
6238second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6239and <tt>%b</tt> are the two values that will undergo unsigned addition.</p>
6240
6241<h5>Semantics:</h5>
6242
6243<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
6244an unsigned addition of the two arguments. They return a structure &mdash; the
6245first element of which is the sum, and the second element of which is a bit
6246specifying if the unsigned summation resulted in a carry.</p>
6247
6248<h5>Examples:</h5>
6249<pre>
6250 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6251 %sum = extractvalue {i32, i1} %res, 0
6252 %obit = extractvalue {i32, i1} %res, 1
6253 br i1 %obit, label %carry, label %normal
6254</pre>
6255
6256</div>
6257
6258<!-- _______________________________________________________________________ -->
6259<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6260 <a name="int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6261</div>
6262
6263<div class="doc_text">
6264
6265<h5>Syntax:</h5>
6266
6267<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]6268on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6269
6270<pre>
6271 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
6272 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6273 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
6274</pre>
6275
6276<h5>Overview:</h5>
6277
6278<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
6279a signed subtraction of the two arguments, and indicate whether an overflow
6280occurred during the signed subtraction.</p>
6281
6282<h5>Arguments:</h5>
6283
6284<p>The arguments (%a and %b) and the first element of the result structure may
6285be of integer types of any bit width, but they must have the same bit width. The
6286second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6287and <tt>%b</tt> are the two values that will undergo signed subtraction.</p>
6288
6289<h5>Semantics:</h5>
6290
6291<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
6292a signed subtraction of the two arguments. They return a structure &mdash; the
6293first element of which is the subtraction, and the second element of which is a bit
6294specifying if the signed subtraction resulted in an overflow.</p>
6295
6296<h5>Examples:</h5>
6297<pre>
6298 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6299 %sum = extractvalue {i32, i1} %res, 0
6300 %obit = extractvalue {i32, i1} %res, 1
6301 br i1 %obit, label %overflow, label %normal
6302</pre>
6303
6304</div>
6305
6306<!-- _______________________________________________________________________ -->
6307<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6308 <a name="int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6309</div>
6310
6311<div class="doc_text">
6312
6313<h5>Syntax:</h5>
6314
6315<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]6316on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6317
6318<pre>
6319 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
6320 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6321 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
6322</pre>
6323
6324<h5>Overview:</h5>
6325
6326<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
6327an unsigned subtraction of the two arguments, and indicate whether an overflow
6328occurred during the unsigned subtraction.</p>
6329
6330<h5>Arguments:</h5>
6331
6332<p>The arguments (%a and %b) and the first element of the result structure may
6333be of integer types of any bit width, but they must have the same bit width. The
6334second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6335and <tt>%b</tt> are the two values that will undergo unsigned subtraction.</p>
6336
6337<h5>Semantics:</h5>
6338
6339<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
6340an unsigned subtraction of the two arguments. They return a structure &mdash; the
6341first element of which is the subtraction, and the second element of which is a bit
6342specifying if the unsigned subtraction resulted in an overflow.</p>
6343
6344<h5>Examples:</h5>
6345<pre>
6346 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6347 %sum = extractvalue {i32, i1} %res, 0
6348 %obit = extractvalue {i32, i1} %res, 1
6349 br i1 %obit, label %overflow, label %normal
6350</pre>
6351
6352</div>
6353
6354<!-- _______________________________________________________________________ -->
6355<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6356 <a name="int_smul_overflow">'<tt>llvm.smul.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6357</div>
6358
6359<div class="doc_text">
6360
6361<h5>Syntax:</h5>
6362
6363<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]6364on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6365
6366<pre>
6367 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
6368 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6369 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
6370</pre>
6371
6372<h5>Overview:</h5>
6373
6374<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
6375a signed multiplication of the two arguments, and indicate whether an overflow
6376occurred during the signed multiplication.</p>
6377
6378<h5>Arguments:</h5>
6379
6380<p>The arguments (%a and %b) and the first element of the result structure may
6381be of integer types of any bit width, but they must have the same bit width. The
6382second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6383and <tt>%b</tt> are the two values that will undergo signed multiplication.</p>
6384
6385<h5>Semantics:</h5>
6386
6387<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
6388a signed multiplication of the two arguments. They return a structure &mdash;
6389the first element of which is the multiplication, and the second element of
6390which is a bit specifying if the signed multiplication resulted in an
6391overflow.</p>
6392
6393<h5>Examples:</h5>
6394<pre>
6395 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6396 %sum = extractvalue {i32, i1} %res, 0
6397 %obit = extractvalue {i32, i1} %res, 1
6398 br i1 %obit, label %overflow, label %normal
6399</pre>
6400
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6401</div>
6402
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]6403<!-- _______________________________________________________________________ -->
6404<div class="doc_subsubsection">
6405 <a name="int_umul_overflow">'<tt>llvm.umul.with.overflow.*</tt>' Intrinsics</a>
6406</div>
6407
6408<div class="doc_text">
6409
6410<h5>Syntax:</h5>
6411
6412<p>This is an overloaded intrinsic. You can use <tt>llvm.umul.with.overflow</tt>
6413on any integer bit width.</p>
6414
6415<pre>
6416 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
6417 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6418 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
6419</pre>
6420
6421<h5>Overview:</h5>
6422
6423<p><i><b>Warning:</b> '<tt>llvm.umul.with.overflow</tt>' is badly broken. It is
6424actively being fixed, but it should not currently be used!</i></p>
6425
6426<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
6427a unsigned multiplication of the two arguments, and indicate whether an overflow
6428occurred during the unsigned multiplication.</p>
6429
6430<h5>Arguments:</h5>
6431
6432<p>The arguments (%a and %b) and the first element of the result structure may
6433be of integer types of any bit width, but they must have the same bit width. The
6434second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6435and <tt>%b</tt> are the two values that will undergo unsigned
6436multiplication.</p>
6437
6438<h5>Semantics:</h5>
6439
6440<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
6441an unsigned multiplication of the two arguments. They return a structure &mdash;
6442the first element of which is the multiplication, and the second element of
6443which is a bit specifying if the unsigned multiplication resulted in an
6444overflow.</p>
6445
6446<h5>Examples:</h5>
6447<pre>
6448 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6449 %sum = extractvalue {i32, i1} %res, 0
6450 %obit = extractvalue {i32, i1} %res, 1
6451 br i1 %obit, label %overflow, label %normal
6452</pre>
6453
6454</div>
6455
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6456<!-- ======================================================================= -->
6457<div class="doc_subsection">
6458 <a name="int_debugger">Debugger Intrinsics</a>
6459</div>
6460
6461<div class="doc_text">
6462<p>
6463The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
6464are described in the <a
6465href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
6466Debugging</a> document.
6467</p>
6468</div>
6469
6470
6471<!-- ======================================================================= -->
6472<div class="doc_subsection">
6473 <a name="int_eh">Exception Handling Intrinsics</a>
6474</div>
6475
6476<div class="doc_text">
6477<p> The LLVM exception handling intrinsics (which all start with
6478<tt>llvm.eh.</tt> prefix), are described in the <a
6479href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
6480Handling</a> document. </p>
6481</div>
6482
6483<!-- ======================================================================= -->
6484<div class="doc_subsection">
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6485 <a name="int_trampoline">Trampoline Intrinsic</a>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6486</div>
6487
6488<div class="doc_text">
6489<p>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6490 This intrinsic makes it possible to excise one parameter, marked with
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6491 the <tt>nest</tt> attribute, from a function. The result is a callable
6492 function pointer lacking the nest parameter - the caller does not need
6493 to provide a value for it. Instead, the value to use is stored in
6494 advance in a "trampoline", a block of memory usually allocated
6495 on the stack, which also contains code to splice the nest value into the
6496 argument list. This is used to implement the GCC nested function address
6497 extension.
6498</p>
6499<p>
6500 For example, if the function is
6501 <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]6502 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]6503<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6504 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
6505 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
6506 %p = call i8* @llvm.init.trampoline( i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8*), i8* %nval )
6507 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6508</pre>
Bill Wendlinge262b4c2007-09-22 09:23:55 +0000[diff] [blame]6509 <p>The call <tt>%val = call i32 %fp( i32 %x, i32 %y )</tt> is then equivalent
6510 to <tt>%val = call i32 %f( i8* %nval, i32 %x, i32 %y )</tt>.</p>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6511</div>
6512
6513<!-- _______________________________________________________________________ -->
6514<div class="doc_subsubsection">
6515 <a name="int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a>
6516</div>
6517<div class="doc_text">
6518<h5>Syntax:</h5>
6519<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6520declare 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]6521</pre>
6522<h5>Overview:</h5>
6523<p>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6524 This fills the memory pointed to by <tt>tramp</tt> with code
6525 and returns a function pointer suitable for executing it.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6526</p>
6527<h5>Arguments:</h5>
6528<p>
6529 The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
6530 pointers. The <tt>tramp</tt> argument must point to a sufficiently large
6531 and sufficiently aligned block of memory; this memory is written to by the
Duncan Sands35012212007-08-22 23:39:54 +0000[diff] [blame]6532 intrinsic. Note that the size and the alignment are target-specific - LLVM
6533 currently provides no portable way of determining them, so a front-end that
6534 generates this intrinsic needs to have some target-specific knowledge.
6535 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]6536</p>
6537<h5>Semantics:</h5>
6538<p>
6539 The block of memory pointed to by <tt>tramp</tt> is filled with target
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6540 dependent code, turning it into a function. A pointer to this function is
6541 returned, but needs to be bitcast to an
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6542 <a href="#int_trampoline">appropriate function pointer type</a>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6543 before being called. The new function's signature is the same as that of
6544 <tt>func</tt> with any arguments marked with the <tt>nest</tt> attribute
6545 removed. At most one such <tt>nest</tt> argument is allowed, and it must be
6546 of pointer type. Calling the new function is equivalent to calling
6547 <tt>func</tt> with the same argument list, but with <tt>nval</tt> used for the
6548 missing <tt>nest</tt> argument. If, after calling
6549 <tt>llvm.init.trampoline</tt>, the memory pointed to by <tt>tramp</tt> is
6550 modified, then the effect of any later call to the returned function pointer is
6551 undefined.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6552</p>
6553</div>
6554
6555<!-- ======================================================================= -->
6556<div class="doc_subsection">
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6557 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
6558</div>
6559
6560<div class="doc_text">
6561<p>
6562 These intrinsic functions expand the "universal IR" of LLVM to represent
6563 hardware constructs for atomic operations and memory synchronization. This
6564 provides an interface to the hardware, not an interface to the programmer. It
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]6565 is aimed at a low enough level to allow any programming models or APIs
6566 (Application Programming Interfaces) which
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6567 need atomic behaviors to map cleanly onto it. It is also modeled primarily on
6568 hardware behavior. Just as hardware provides a "universal IR" for source
6569 languages, it also provides a starting point for developing a "universal"
6570 atomic operation and synchronization IR.
6571</p>
6572<p>
6573 These do <em>not</em> form an API such as high-level threading libraries,
6574 software transaction memory systems, atomic primitives, and intrinsic
6575 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
6576 application libraries. The hardware interface provided by LLVM should allow
6577 a clean implementation of all of these APIs and parallel programming models.
6578 No one model or paradigm should be selected above others unless the hardware
6579 itself ubiquitously does so.
6580
6581</p>
6582</div>
6583
6584<!-- _______________________________________________________________________ -->
6585<div class="doc_subsubsection">
6586 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
6587</div>
6588<div class="doc_text">
6589<h5>Syntax:</h5>
6590<pre>
6591declare void @llvm.memory.barrier( i1 &lt;ll&gt;, i1 &lt;ls&gt;, i1 &lt;sl&gt;, i1 &lt;ss&gt;,
6592i1 &lt;device&gt; )
6593
6594</pre>
6595<h5>Overview:</h5>
6596<p>
6597 The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
6598 specific pairs of memory access types.
6599</p>
6600<h5>Arguments:</h5>
6601<p>
6602 The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
6603 The first four arguments enables a specific barrier as listed below. The fith
6604 argument specifies that the barrier applies to io or device or uncached memory.
6605
6606</p>
6607 <ul>
6608 <li><tt>ll</tt>: load-load barrier</li>
6609 <li><tt>ls</tt>: load-store barrier</li>
6610 <li><tt>sl</tt>: store-load barrier</li>
6611 <li><tt>ss</tt>: store-store barrier</li>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]6612 <li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6613 </ul>
6614<h5>Semantics:</h5>
6615<p>
6616 This intrinsic causes the system to enforce some ordering constraints upon
6617 the loads and stores of the program. This barrier does not indicate
6618 <em>when</em> any events will occur, it only enforces an <em>order</em> in
6619 which they occur. For any of the specified pairs of load and store operations
6620 (f.ex. load-load, or store-load), all of the first operations preceding the
6621 barrier will complete before any of the second operations succeeding the
6622 barrier begin. Specifically the semantics for each pairing is as follows:
6623</p>
6624 <ul>
6625 <li><tt>ll</tt>: All loads before the barrier must complete before any load
6626 after the barrier begins.</li>
6627
6628 <li><tt>ls</tt>: All loads before the barrier must complete before any
6629 store after the barrier begins.</li>
6630 <li><tt>ss</tt>: All stores before the barrier must complete before any
6631 store after the barrier begins.</li>
6632 <li><tt>sl</tt>: All stores before the barrier must complete before any
6633 load after the barrier begins.</li>
6634 </ul>
6635<p>
6636 These semantics are applied with a logical "and" behavior when more than one
6637 is enabled in a single memory barrier intrinsic.
6638</p>
6639<p>
6640 Backends may implement stronger barriers than those requested when they do not
6641 support as fine grained a barrier as requested. Some architectures do not
6642 need all types of barriers and on such architectures, these become noops.
6643</p>
6644<h5>Example:</h5>
6645<pre>
6646%ptr = malloc i32
6647 store i32 4, %ptr
6648
6649%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
6650 call void @llvm.memory.barrier( i1 false, i1 true, i1 false, i1 false )
6651 <i>; guarantee the above finishes</i>
6652 store i32 8, %ptr <i>; before this begins</i>
6653</pre>
6654</div>
6655
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6656<!-- _______________________________________________________________________ -->
6657<div class="doc_subsubsection">
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6658 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6659</div>
6660<div class="doc_text">
6661<h5>Syntax:</h5>
6662<p>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6663 This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
6664 any integer bit width and for different address spaces. Not all targets
6665 support all bit widths however.</p>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6666
6667<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6668declare i8 @llvm.atomic.cmp.swap.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt; )
6669declare i16 @llvm.atomic.cmp.swap.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt; )
6670declare i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt; )
6671declare 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]6672
6673</pre>
6674<h5>Overview:</h5>
6675<p>
6676 This loads a value in memory and compares it to a given value. If they are
6677 equal, it stores a new value into the memory.
6678</p>
6679<h5>Arguments:</h5>
6680<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6681 The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result as
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6682 well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
6683 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
6684 this integer type. While any bit width integer may be used, targets may only
6685 lower representations they support in hardware.
6686
6687</p>
6688<h5>Semantics:</h5>
6689<p>
6690 This entire intrinsic must be executed atomically. It first loads the value
6691 in memory pointed to by <tt>ptr</tt> and compares it with the value
6692 <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the memory. The
6693 loaded value is yielded in all cases. This provides the equivalent of an
6694 atomic compare-and-swap operation within the SSA framework.
6695</p>
6696<h5>Examples:</h5>
6697
6698<pre>
6699%ptr = malloc i32
6700 store i32 4, %ptr
6701
6702%val1 = add i32 4, 4
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6703%result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 4, %val1 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6704 <i>; yields {i32}:result1 = 4</i>
6705%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
6706%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
6707
6708%val2 = add i32 1, 1
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6709%result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 5, %val2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6710 <i>; yields {i32}:result2 = 8</i>
6711%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
6712
6713%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
6714</pre>
6715</div>
6716
6717<!-- _______________________________________________________________________ -->
6718<div class="doc_subsubsection">
6719 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
6720</div>
6721<div class="doc_text">
6722<h5>Syntax:</h5>
6723
6724<p>
6725 This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
6726 integer bit width. Not all targets support all bit widths however.</p>
6727<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6728declare i8 @llvm.atomic.swap.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;val&gt; )
6729declare i16 @llvm.atomic.swap.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;val&gt; )
6730declare i32 @llvm.atomic.swap.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;val&gt; )
6731declare i64 @llvm.atomic.swap.i64.p0i64( i64* &lt;ptr&gt;, i64 &lt;val&gt; )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6732
6733</pre>
6734<h5>Overview:</h5>
6735<p>
6736 This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
6737 the value from memory. It then stores the value in <tt>val</tt> in the memory
6738 at <tt>ptr</tt>.
6739</p>
6740<h5>Arguments:</h5>
6741
6742<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6743 The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both the
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6744 <tt>val</tt> argument and the result must be integers of the same bit width.
6745 The first argument, <tt>ptr</tt>, must be a pointer to a value of this
6746 integer type. The targets may only lower integer representations they
6747 support.
6748</p>
6749<h5>Semantics:</h5>
6750<p>
6751 This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
6752 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
6753 equivalent of an atomic swap operation within the SSA framework.
6754
6755</p>
6756<h5>Examples:</h5>
6757<pre>
6758%ptr = malloc i32
6759 store i32 4, %ptr
6760
6761%val1 = add i32 4, 4
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6762%result1 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val1 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6763 <i>; yields {i32}:result1 = 4</i>
6764%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
6765%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
6766
6767%val2 = add i32 1, 1
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6768%result2 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6769 <i>; yields {i32}:result2 = 8</i>
6770
6771%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
6772%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
6773</pre>
6774</div>
6775
6776<!-- _______________________________________________________________________ -->
6777<div class="doc_subsubsection">
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6778 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6779
6780</div>
6781<div class="doc_text">
6782<h5>Syntax:</h5>
6783<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6784 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]6785 integer bit width. Not all targets support all bit widths however.</p>
6786<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6787declare i8 @llvm.atomic.load.add.i8..p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6788declare i16 @llvm.atomic.load.add.i16..p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6789declare i32 @llvm.atomic.load.add.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6790declare 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]6791
6792</pre>
6793<h5>Overview:</h5>
6794<p>
6795 This intrinsic adds <tt>delta</tt> to the value stored in memory at
6796 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
6797</p>
6798<h5>Arguments:</h5>
6799<p>
6800
6801 The intrinsic takes two arguments, the first a pointer to an integer value
6802 and the second an integer value. The result is also an integer value. These
6803 integer types can have any bit width, but they must all have the same bit
6804 width. The targets may only lower integer representations they support.
6805</p>
6806<h5>Semantics:</h5>
6807<p>
6808 This intrinsic does a series of operations atomically. It first loads the
6809 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
6810 to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
6811</p>
6812
6813<h5>Examples:</h5>
6814<pre>
6815%ptr = malloc i32
6816 store i32 4, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6817%result1 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 4 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6818 <i>; yields {i32}:result1 = 4</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6819%result2 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6820 <i>; yields {i32}:result2 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6821%result3 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 5 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6822 <i>; yields {i32}:result3 = 10</i>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6823%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6824</pre>
6825</div>
6826
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6827<!-- _______________________________________________________________________ -->
6828<div class="doc_subsubsection">
6829 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
6830
6831</div>
6832<div class="doc_text">
6833<h5>Syntax:</h5>
6834<p>
6835 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]6836 any integer bit width and for different address spaces. Not all targets
6837 support all bit widths however.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6838<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6839declare i8 @llvm.atomic.load.sub.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6840declare i16 @llvm.atomic.load.sub.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6841declare i32 @llvm.atomic.load.sub.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6842declare 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]6843
6844</pre>
6845<h5>Overview:</h5>
6846<p>
6847 This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
6848 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
6849</p>
6850<h5>Arguments:</h5>
6851<p>
6852
6853 The intrinsic takes two arguments, the first a pointer to an integer value
6854 and the second an integer value. The result is also an integer value. These
6855 integer types can have any bit width, but they must all have the same bit
6856 width. The targets may only lower integer representations they support.
6857</p>
6858<h5>Semantics:</h5>
6859<p>
6860 This intrinsic does a series of operations atomically. It first loads the
6861 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
6862 result to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
6863</p>
6864
6865<h5>Examples:</h5>
6866<pre>
6867%ptr = malloc i32
6868 store i32 8, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6869%result1 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 4 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6870 <i>; yields {i32}:result1 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6871%result2 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 2 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6872 <i>; yields {i32}:result2 = 4</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6873%result3 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 5 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6874 <i>; yields {i32}:result3 = 2</i>
6875%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
6876</pre>
6877</div>
6878
6879<!-- _______________________________________________________________________ -->
6880<div class="doc_subsubsection">
6881 <a name="int_atomic_load_and">'<tt>llvm.atomic.load.and.*</tt>' Intrinsic</a><br>
6882 <a name="int_atomic_load_nand">'<tt>llvm.atomic.load.nand.*</tt>' Intrinsic</a><br>
6883 <a name="int_atomic_load_or">'<tt>llvm.atomic.load.or.*</tt>' Intrinsic</a><br>
6884 <a name="int_atomic_load_xor">'<tt>llvm.atomic.load.xor.*</tt>' Intrinsic</a><br>
6885
6886</div>
6887<div class="doc_text">
6888<h5>Syntax:</h5>
6889<p>
6890 These are overloaded intrinsics. You can use <tt>llvm.atomic.load_and</tt>,
6891 <tt>llvm.atomic.load_nand</tt>, <tt>llvm.atomic.load_or</tt>, and
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6892 <tt>llvm.atomic.load_xor</tt> on any integer bit width and for different
6893 address spaces. Not all targets support all bit widths however.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6894<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6895declare i8 @llvm.atomic.load.and.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6896declare i16 @llvm.atomic.load.and.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6897declare i32 @llvm.atomic.load.and.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6898declare 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]6899
6900</pre>
6901
6902<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6903declare i8 @llvm.atomic.load.or.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6904declare i16 @llvm.atomic.load.or.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6905declare i32 @llvm.atomic.load.or.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6906declare 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]6907
6908</pre>
6909
6910<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6911declare i8 @llvm.atomic.load.nand.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6912declare i16 @llvm.atomic.load.nand.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6913declare i32 @llvm.atomic.load.nand.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6914declare 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]6915
6916</pre>
6917
6918<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6919declare i8 @llvm.atomic.load.xor.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6920declare i16 @llvm.atomic.load.xor.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6921declare i32 @llvm.atomic.load.xor.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6922declare 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]6923
6924</pre>
6925<h5>Overview:</h5>
6926<p>
6927 These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
6928 the value stored in memory at <tt>ptr</tt>. It yields the original value
6929 at <tt>ptr</tt>.
6930</p>
6931<h5>Arguments:</h5>
6932<p>
6933
6934 These intrinsics take two arguments, the first a pointer to an integer value
6935 and the second an integer value. The result is also an integer value. These
6936 integer types can have any bit width, but they must all have the same bit
6937 width. The targets may only lower integer representations they support.
6938</p>
6939<h5>Semantics:</h5>
6940<p>
6941 These intrinsics does a series of operations atomically. They first load the
6942 value stored at <tt>ptr</tt>. They then do the bitwise operation
6943 <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the original
6944 value stored at <tt>ptr</tt>.
6945</p>
6946
6947<h5>Examples:</h5>
6948<pre>
6949%ptr = malloc i32
6950 store i32 0x0F0F, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6951%result0 = call i32 @llvm.atomic.load.nand.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6952 <i>; yields {i32}:result0 = 0x0F0F</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6953%result1 = call i32 @llvm.atomic.load.and.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6954 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6955%result2 = call i32 @llvm.atomic.load.or.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6956 <i>; yields {i32}:result2 = 0xF0</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6957%result3 = call i32 @llvm.atomic.load.xor.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6958 <i>; yields {i32}:result3 = FF</i>
6959%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
6960</pre>
6961</div>
6962
6963
6964<!-- _______________________________________________________________________ -->
6965<div class="doc_subsubsection">
6966 <a name="int_atomic_load_max">'<tt>llvm.atomic.load.max.*</tt>' Intrinsic</a><br>
6967 <a name="int_atomic_load_min">'<tt>llvm.atomic.load.min.*</tt>' Intrinsic</a><br>
6968 <a name="int_atomic_load_umax">'<tt>llvm.atomic.load.umax.*</tt>' Intrinsic</a><br>
6969 <a name="int_atomic_load_umin">'<tt>llvm.atomic.load.umin.*</tt>' Intrinsic</a><br>
6970
6971</div>
6972<div class="doc_text">
6973<h5>Syntax:</h5>
6974<p>
6975 These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
6976 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6977 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
6978 address spaces. Not all targets
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6979 support all bit widths however.</p>
6980<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6981declare i8 @llvm.atomic.load.max.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6982declare i16 @llvm.atomic.load.max.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6983declare i32 @llvm.atomic.load.max.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6984declare 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]6985
6986</pre>
6987
6988<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6989declare i8 @llvm.atomic.load.min.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6990declare i16 @llvm.atomic.load.min.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6991declare i32 @llvm.atomic.load.min.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6992declare 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]6993
6994</pre>
6995
6996<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6997declare i8 @llvm.atomic.load.umax.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6998declare i16 @llvm.atomic.load.umax.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6999declare i32 @llvm.atomic.load.umax.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
7000declare 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]7001
7002</pre>
7003
7004<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7005declare i8 @llvm.atomic.load.umin.i8..p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
7006declare i16 @llvm.atomic.load.umin.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
7007declare i32 @llvm.atomic.load.umin.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
7008declare 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]7009
7010</pre>
7011<h5>Overview:</h5>
7012<p>
7013 These intrinsics takes the signed or unsigned minimum or maximum of
7014 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
7015 original value at <tt>ptr</tt>.
7016</p>
7017<h5>Arguments:</h5>
7018<p>
7019
7020 These intrinsics take two arguments, the first a pointer to an integer value
7021 and the second an integer value. The result is also an integer value. These
7022 integer types can have any bit width, but they must all have the same bit
7023 width. The targets may only lower integer representations they support.
7024</p>
7025<h5>Semantics:</h5>
7026<p>
7027 These intrinsics does a series of operations atomically. They first load the
7028 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or max
7029 <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They yield
7030 the original value stored at <tt>ptr</tt>.
7031</p>
7032
7033<h5>Examples:</h5>
7034<pre>
7035%ptr = malloc i32
7036 store i32 7, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7037%result0 = call i32 @llvm.atomic.load.min.i32.p0i32( i32* %ptr, i32 -2 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7038 <i>; yields {i32}:result0 = 7</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7039%result1 = call i32 @llvm.atomic.load.max.i32.p0i32( i32* %ptr, i32 8 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7040 <i>; yields {i32}:result1 = -2</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7041%result2 = call i32 @llvm.atomic.load.umin.i32.p0i32( i32* %ptr, i32 10 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7042 <i>; yields {i32}:result2 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]7043%result3 = call i32 @llvm.atomic.load.umax.i32.p0i32( i32* %ptr, i32 30 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]7044 <i>; yields {i32}:result3 = 8</i>
7045%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
7046</pre>
7047</div>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]7048
7049<!-- ======================================================================= -->
7050<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7051 <a name="int_general">General Intrinsics</a>
7052</div>
7053
7054<div class="doc_text">
7055<p> This class of intrinsics is designed to be generic and has
7056no specific purpose. </p>
7057</div>
7058
7059<!-- _______________________________________________________________________ -->
7060<div class="doc_subsubsection">
7061 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
7062</div>
7063
7064<div class="doc_text">
7065
7066<h5>Syntax:</h5>
7067<pre>
7068 declare void @llvm.var.annotation(i8* &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
7069</pre>
7070
7071<h5>Overview:</h5>
7072
7073<p>
7074The '<tt>llvm.var.annotation</tt>' intrinsic
7075</p>
7076
7077<h5>Arguments:</h5>
7078
7079<p>
7080The first argument is a pointer to a value, the second is a pointer to a
7081global string, the third is a pointer to a global string which is the source
7082file name, and the last argument is the line number.
7083</p>
7084
7085<h5>Semantics:</h5>
7086
7087<p>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]7088This intrinsic allows annotation of local variables with arbitrary strings.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7089This can be useful for special purpose optimizations that want to look for these
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]7090annotations. These have no other defined use, they are ignored by code
7091generation and optimization.
7092</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7093</div>
7094
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7095<!-- _______________________________________________________________________ -->
7096<div class="doc_subsubsection">
Tanya Lattnerc9869b12007-09-21 23:57:59 +0000[diff] [blame]7097 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7098</div>
7099
7100<div class="doc_text">
7101
7102<h5>Syntax:</h5>
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]7103<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
7104any integer bit width.
7105</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7106<pre>
Tanya Lattner09161fe2007-09-22 00:03:01 +0000[diff] [blame]7107 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
7108 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
7109 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
7110 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
7111 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]7112</pre>
7113
7114<h5>Overview:</h5>
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]7115
7116<p>
7117The '<tt>llvm.annotation</tt>' intrinsic.
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7118</p>
7119
7120<h5>Arguments:</h5>
7121
7122<p>
7123The first argument is an integer value (result of some expression),
7124the second is a pointer to a global string, the third is a pointer to a global
7125string which is the source file name, and the last argument is the line number.
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]7126It returns the value of the first argument.
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7127</p>
7128
7129<h5>Semantics:</h5>
7130
7131<p>
7132This intrinsic allows annotations to be put on arbitrary expressions
7133with arbitrary strings. This can be useful for special purpose optimizations
7134that want to look for these annotations. These have no other defined use, they
7135are ignored by code generation and optimization.
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]7136</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]7137</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7138
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]7139<!-- _______________________________________________________________________ -->
7140<div class="doc_subsubsection">
7141 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
7142</div>
7143
7144<div class="doc_text">
7145
7146<h5>Syntax:</h5>
7147<pre>
7148 declare void @llvm.trap()
7149</pre>
7150
7151<h5>Overview:</h5>
7152
7153<p>
7154The '<tt>llvm.trap</tt>' intrinsic
7155</p>
7156
7157<h5>Arguments:</h5>
7158
7159<p>
7160None
7161</p>
7162
7163<h5>Semantics:</h5>
7164
7165<p>
7166This intrinsics is lowered to the target dependent trap instruction. If the
7167target does not have a trap instruction, this intrinsic will be lowered to the
7168call of the abort() function.
7169</p>
7170</div>
7171
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]7172<!-- _______________________________________________________________________ -->
7173<div class="doc_subsubsection">
Misha Brukman5dd7f4d2008-11-22 23:55:29 +0000[diff] [blame]7174 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]7175</div>
7176<div class="doc_text">
7177<h5>Syntax:</h5>
7178<pre>
7179declare void @llvm.stackprotector( i8* &lt;guard&gt;, i8** &lt;slot&gt; )
7180
7181</pre>
7182<h5>Overview:</h5>
7183<p>
7184 The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and stores
7185 it onto the stack at <tt>slot</tt>. The stack slot is adjusted to ensure that
7186 it is placed on the stack before local variables.
7187</p>
7188<h5>Arguments:</h5>
7189<p>
7190 The <tt>llvm.stackprotector</tt> intrinsic requires two pointer arguments. The
7191 first argument is the value loaded from the stack guard
7192 <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt> that
7193 has enough space to hold the value of the guard.
7194</p>
7195<h5>Semantics:</h5>
7196<p>
7197 This intrinsic causes the prologue/epilogue inserter to force the position of
7198 the <tt>AllocaInst</tt> stack slot to be before local variables on the
7199 stack. This is to ensure that if a local variable on the stack is overwritten,
7200 it will destroy the value of the guard. When the function exits, the guard on
7201 the stack is checked against the original guard. If they're different, then
7202 the program aborts by calling the <tt>__stack_chk_fail()</tt> function.
7203</p>
7204</div>
7205
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7206<!-- *********************************************************************** -->
7207<hr>
7208<address>
7209 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman947321d2008-12-11 17:34:48 +0000[diff] [blame]7210 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7211 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman947321d2008-12-11 17:34:48 +0000[diff] [blame]7212 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]7213
7214 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
7215 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
7216 Last modified: $Date$
7217</address>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]7218
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7219</body>
7220</html>