Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / fe87fffd274980d967891b8dcf11d29141da1796 / . / docs / LangRef.html
blob: cac8fc79dad2650cc640fd909853f0c40e703235 [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>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]44 </ol>
45 </li>
46 <li><a href="#t_derived">Derived Types</a>
47 <ol>
Chris Lattner251ab812007-12-18 06:18:21 +0000[diff] [blame]48 <li><a href="#t_integer">Integer Type</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]49 <li><a href="#t_array">Array Type</a></li>
50 <li><a href="#t_function">Function Type</a></li>
51 <li><a href="#t_pointer">Pointer Type</a></li>
52 <li><a href="#t_struct">Structure Type</a></li>
53 <li><a href="#t_pstruct">Packed Structure Type</a></li>
54 <li><a href="#t_vector">Vector Type</a></li>
55 <li><a href="#t_opaque">Opaque Type</a></li>
56 </ol>
57 </li>
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]58 <li><a href="#t_uprefs">Type Up-references</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]59 </ol>
60 </li>
61 <li><a href="#constants">Constants</a>
62 <ol>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]63 <li><a href="#simpleconstants">Simple Constants</a></li>
Chris Lattner97063852009-02-28 18:32:25 +0000[diff] [blame]64 <li><a href="#complexconstants">Complex Constants</a></li>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]65 <li><a href="#globalconstants">Global Variable and Function Addresses</a></li>
66 <li><a href="#undefvalues">Undefined Values</a></li>
67 <li><a href="#constantexprs">Constant Expressions</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]68 </ol>
69 </li>
70 <li><a href="#othervalues">Other Values</a>
71 <ol>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]72 <li><a href="#inlineasm">Inline Assembler Expressions</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]73 </ol>
74 </li>
75 <li><a href="#instref">Instruction Reference</a>
76 <ol>
77 <li><a href="#terminators">Terminator Instructions</a>
78 <ol>
79 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
80 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
81 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
82 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
83 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
84 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
85 </ol>
86 </li>
87 <li><a href="#binaryops">Binary Operations</a>
88 <ol>
89 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
90 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
91 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
92 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
93 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
94 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
95 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
96 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
97 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
98 </ol>
99 </li>
100 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
101 <ol>
102 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
103 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
104 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
105 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
106 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
107 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
108 </ol>
109 </li>
110 <li><a href="#vectorops">Vector Operations</a>
111 <ol>
112 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
113 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
114 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
115 </ol>
116 </li>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]117 <li><a href="#aggregateops">Aggregate Operations</a>
118 <ol>
119 <li><a href="#i_extractvalue">'<tt>extractvalue</tt>' Instruction</a></li>
120 <li><a href="#i_insertvalue">'<tt>insertvalue</tt>' Instruction</a></li>
121 </ol>
122 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]123 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
124 <ol>
125 <li><a href="#i_malloc">'<tt>malloc</tt>' Instruction</a></li>
126 <li><a href="#i_free">'<tt>free</tt>' Instruction</a></li>
127 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
128 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
129 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
130 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
131 </ol>
132 </li>
133 <li><a href="#convertops">Conversion Operations</a>
134 <ol>
135 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
136 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
137 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
138 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
139 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
140 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
141 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
142 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
143 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
144 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
145 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
146 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
147 </ol>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]148 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]149 <li><a href="#otherops">Other Operations</a>
150 <ol>
151 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
152 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]153 <li><a href="#i_vicmp">'<tt>vicmp</tt>' Instruction</a></li>
154 <li><a href="#i_vfcmp">'<tt>vfcmp</tt>' Instruction</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]155 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
156 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
157 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
158 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
159 </ol>
160 </li>
161 </ol>
162 </li>
163 <li><a href="#intrinsics">Intrinsic Functions</a>
164 <ol>
165 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
166 <ol>
167 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
168 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
169 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
170 </ol>
171 </li>
172 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
173 <ol>
174 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
175 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
176 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
177 </ol>
178 </li>
179 <li><a href="#int_codegen">Code Generator Intrinsics</a>
180 <ol>
181 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
182 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
183 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
184 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
185 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
186 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
187 <li><a href="#int_readcyclecounter"><tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
188 </ol>
189 </li>
190 <li><a href="#int_libc">Standard C Library Intrinsics</a>
191 <ol>
192 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
193 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
194 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
195 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
196 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]197 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
198 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
199 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]200 </ol>
201 </li>
202 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
203 <ol>
204 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
205 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
206 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
207 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
208 <li><a href="#int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic </a></li>
209 <li><a href="#int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic </a></li>
210 </ol>
211 </li>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]212 <li><a href="#int_overflow">Arithmetic with Overflow Intrinsics</a>
213 <ol>
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]214 <li><a href="#int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt> Intrinsics</a></li>
215 <li><a href="#int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt> Intrinsics</a></li>
216 <li><a href="#int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt> Intrinsics</a></li>
217 <li><a href="#int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt> Intrinsics</a></li>
218 <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]219 <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]220 </ol>
221 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]222 <li><a href="#int_debugger">Debugger intrinsics</a></li>
223 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]224 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]225 <ol>
226 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]227 </ol>
228 </li>
Bill Wendling9127adb2008-11-18 22:10:53 +0000[diff] [blame]229 <li><a href="#int_atomics">Atomic intrinsics</a>
230 <ol>
231 <li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
232 <li><a href="#int_atomic_cmp_swap"><tt>llvm.atomic.cmp.swap</tt></a></li>
233 <li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
234 <li><a href="#int_atomic_load_add"><tt>llvm.atomic.load.add</tt></a></li>
235 <li><a href="#int_atomic_load_sub"><tt>llvm.atomic.load.sub</tt></a></li>
236 <li><a href="#int_atomic_load_and"><tt>llvm.atomic.load.and</tt></a></li>
237 <li><a href="#int_atomic_load_nand"><tt>llvm.atomic.load.nand</tt></a></li>
238 <li><a href="#int_atomic_load_or"><tt>llvm.atomic.load.or</tt></a></li>
239 <li><a href="#int_atomic_load_xor"><tt>llvm.atomic.load.xor</tt></a></li>
240 <li><a href="#int_atomic_load_max"><tt>llvm.atomic.load.max</tt></a></li>
241 <li><a href="#int_atomic_load_min"><tt>llvm.atomic.load.min</tt></a></li>
242 <li><a href="#int_atomic_load_umax"><tt>llvm.atomic.load.umax</tt></a></li>
243 <li><a href="#int_atomic_load_umin"><tt>llvm.atomic.load.umin</tt></a></li>
244 </ol>
245 </li>
Reid Spencerb043f672007-07-20 19:59:11 +0000[diff] [blame]246 <li><a href="#int_general">General intrinsics</a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]247 <ol>
Reid Spencerb043f672007-07-20 19:59:11 +0000[diff] [blame]248 <li><a href="#int_var_annotation">
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]249 '<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]250 <li><a href="#int_annotation">
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]251 '<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]252 <li><a href="#int_trap">
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]253 '<tt>llvm.trap</tt>' Intrinsic</a></li>
254 <li><a href="#int_stackprotector">
255 '<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]256 </ol>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]257 </li>
258 </ol>
259 </li>
260</ol>
261
262<div class="doc_author">
263 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
264 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
265</div>
266
267<!-- *********************************************************************** -->
268<div class="doc_section"> <a name="abstract">Abstract </a></div>
269<!-- *********************************************************************** -->
270
271<div class="doc_text">
272<p>This document is a reference manual for the LLVM assembly language.
Bill Wendlinge7846a52008-08-05 22:29:16 +0000[diff] [blame]273LLVM is a Static Single Assignment (SSA) based representation that provides
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]274type safety, low-level operations, flexibility, and the capability of
275representing 'all' high-level languages cleanly. It is the common code
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]276representation used throughout all phases of the LLVM compilation
277strategy.</p>
278</div>
279
280<!-- *********************************************************************** -->
281<div class="doc_section"> <a name="introduction">Introduction</a> </div>
282<!-- *********************************************************************** -->
283
284<div class="doc_text">
285
286<p>The LLVM code representation is designed to be used in three
287different forms: as an in-memory compiler IR, as an on-disk bitcode
288representation (suitable for fast loading by a Just-In-Time compiler),
289and as a human readable assembly language representation. This allows
290LLVM to provide a powerful intermediate representation for efficient
291compiler transformations and analysis, while providing a natural means
292to debug and visualize the transformations. The three different forms
293of LLVM are all equivalent. This document describes the human readable
294representation and notation.</p>
295
296<p>The LLVM representation aims to be light-weight and low-level
297while being expressive, typed, and extensible at the same time. It
298aims to be a "universal IR" of sorts, by being at a low enough level
299that high-level ideas may be cleanly mapped to it (similar to how
300microprocessors are "universal IR's", allowing many source languages to
301be mapped to them). By providing type information, LLVM can be used as
302the target of optimizations: for example, through pointer analysis, it
303can be proven that a C automatic variable is never accessed outside of
304the current function... allowing it to be promoted to a simple SSA
305value instead of a memory location.</p>
306
307</div>
308
309<!-- _______________________________________________________________________ -->
310<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
311
312<div class="doc_text">
313
314<p>It is important to note that this document describes 'well formed'
315LLVM assembly language. There is a difference between what the parser
316accepts and what is considered 'well formed'. For example, the
317following instruction is syntactically okay, but not well formed:</p>
318
319<div class="doc_code">
320<pre>
321%x = <a href="#i_add">add</a> i32 1, %x
322</pre>
323</div>
324
325<p>...because the definition of <tt>%x</tt> does not dominate all of
326its uses. The LLVM infrastructure provides a verification pass that may
327be used to verify that an LLVM module is well formed. This pass is
328automatically run by the parser after parsing input assembly and by
329the optimizer before it outputs bitcode. The violations pointed out
330by the verifier pass indicate bugs in transformation passes or input to
331the parser.</p>
332</div>
333
Chris Lattnera83fdc02007-10-03 17:34:29 +0000[diff] [blame]334<!-- Describe the typesetting conventions here. -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]335
336<!-- *********************************************************************** -->
337<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
338<!-- *********************************************************************** -->
339
340<div class="doc_text">
341
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]342 <p>LLVM identifiers come in two basic types: global and local. Global
343 identifiers (functions, global variables) begin with the @ character. Local
344 identifiers (register names, types) begin with the % character. Additionally,
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]345 there are three different formats for identifiers, for different purposes:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]346
347<ol>
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]348 <li>Named values are represented as a string of characters with their prefix.
349 For example, %foo, @DivisionByZero, %a.really.long.identifier. The actual
350 regular expression used is '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]351 Identifiers which require other characters in their names can be surrounded
Daniel Dunbar73f9a772008-10-14 23:51:43 +0000[diff] [blame]352 with quotes. Special characters may be escaped using "\xx" where xx is the
353 ASCII code for the character in hexadecimal. In this way, any character can
354 be used in a name value, even quotes themselves.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]355
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]356 <li>Unnamed values are represented as an unsigned numeric value with their
357 prefix. For example, %12, @2, %44.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]358
359 <li>Constants, which are described in a <a href="#constants">section about
360 constants</a>, below.</li>
361</ol>
362
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]363<p>LLVM requires that values start with a prefix for two reasons: Compilers
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]364don't need to worry about name clashes with reserved words, and the set of
365reserved words may be expanded in the future without penalty. Additionally,
366unnamed identifiers allow a compiler to quickly come up with a temporary
367variable without having to avoid symbol table conflicts.</p>
368
369<p>Reserved words in LLVM are very similar to reserved words in other
370languages. There are keywords for different opcodes
371('<tt><a href="#i_add">add</a></tt>',
372 '<tt><a href="#i_bitcast">bitcast</a></tt>',
373 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
374href="#t_void">void</a></tt>', '<tt><a href="#t_primitive">i32</a></tt>', etc...),
375and others. These reserved words cannot conflict with variable names, because
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]376none of them start with a prefix character ('%' or '@').</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]377
378<p>Here is an example of LLVM code to multiply the integer variable
379'<tt>%X</tt>' by 8:</p>
380
381<p>The easy way:</p>
382
383<div class="doc_code">
384<pre>
385%result = <a href="#i_mul">mul</a> i32 %X, 8
386</pre>
387</div>
388
389<p>After strength reduction:</p>
390
391<div class="doc_code">
392<pre>
393%result = <a href="#i_shl">shl</a> i32 %X, i8 3
394</pre>
395</div>
396
397<p>And the hard way:</p>
398
399<div class="doc_code">
400<pre>
401<a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
402<a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
403%result = <a href="#i_add">add</a> i32 %1, %1
404</pre>
405</div>
406
407<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
408important lexical features of LLVM:</p>
409
410<ol>
411
412 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
413 line.</li>
414
415 <li>Unnamed temporaries are created when the result of a computation is not
416 assigned to a named value.</li>
417
418 <li>Unnamed temporaries are numbered sequentially</li>
419
420</ol>
421
422<p>...and it also shows a convention that we follow in this document. When
423demonstrating instructions, we will follow an instruction with a comment that
424defines the type and name of value produced. Comments are shown in italic
425text.</p>
426
427</div>
428
429<!-- *********************************************************************** -->
430<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
431<!-- *********************************************************************** -->
432
433<!-- ======================================================================= -->
434<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
435</div>
436
437<div class="doc_text">
438
439<p>LLVM programs are composed of "Module"s, each of which is a
440translation unit of the input programs. Each module consists of
441functions, global variables, and symbol table entries. Modules may be
442combined together with the LLVM linker, which merges function (and
443global variable) definitions, resolves forward declarations, and merges
444symbol table entries. Here is an example of the "hello world" module:</p>
445
446<div class="doc_code">
447<pre><i>; Declare the string constant as a global constant...</i>
448<a href="#identifiers">@.LC0</a> = <a href="#linkage_internal">internal</a> <a
449 href="#globalvars">constant</a> <a href="#t_array">[13 x i8]</a> c"hello world\0A\00" <i>; [13 x i8]*</i>
450
451<i>; External declaration of the puts function</i>
452<a href="#functionstructure">declare</a> i32 @puts(i8 *) <i>; i32(i8 *)* </i>
453
454<i>; Definition of main function</i>
455define i32 @main() { <i>; i32()* </i>
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]456 <i>; Convert [13 x i8]* to i8 *...</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]457 %cast210 = <a
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]458 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]459
460 <i>; Call puts function to write out the string to stdout...</i>
461 <a
462 href="#i_call">call</a> i32 @puts(i8 * %cast210) <i>; i32</i>
463 <a
464 href="#i_ret">ret</a> i32 0<br>}<br>
465</pre>
466</div>
467
468<p>This example is made up of a <a href="#globalvars">global variable</a>
469named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
470function, and a <a href="#functionstructure">function definition</a>
471for "<tt>main</tt>".</p>
472
473<p>In general, a module is made up of a list of global values,
474where both functions and global variables are global values. Global values are
475represented by a pointer to a memory location (in this case, a pointer to an
476array of char, and a pointer to a function), and have one of the following <a
477href="#linkage">linkage types</a>.</p>
478
479</div>
480
481<!-- ======================================================================= -->
482<div class="doc_subsection">
483 <a name="linkage">Linkage Types</a>
484</div>
485
486<div class="doc_text">
487
488<p>
489All Global Variables and Functions have one of the following types of linkage:
490</p>
491
492<dl>
493
Rafael Espindolaa168fc92009-01-15 20:18:42 +0000[diff] [blame]494 <dt><tt><b><a name="linkage_private">private</a></b></tt>: </dt>
495
496 <dd>Global values with private linkage are only directly accessible by
497 objects in the current module. In particular, linking code into a module with
498 an private global value may cause the private to be renamed as necessary to
499 avoid collisions. Because the symbol is private to the module, all
500 references can be updated. This doesn't show up in any symbol table in the
501 object file.
502 </dd>
503
Dale Johannesen96e7e092008-05-23 23:13:41 +0000[diff] [blame]504 <dt><tt><b><a name="linkage_internal">internal</a></b></tt>: </dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]505
Duncan Sandsa75223a2009-01-16 09:29:46 +0000[diff] [blame]506 <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]507 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]508 '<tt>static</tt>' keyword in C.
509 </dd>
510
511 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
512
513 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
514 the same name when linkage occurs. This is typically used to implement
515 inline functions, templates, or other code which must be generated in each
516 translation unit that uses it. Unreferenced <tt>linkonce</tt> globals are
517 allowed to be discarded.
518 </dd>
519
Dale Johannesen96e7e092008-05-23 23:13:41 +0000[diff] [blame]520 <dt><tt><b><a name="linkage_common">common</a></b></tt>: </dt>
521
522 <dd>"<tt>common</tt>" linkage is exactly the same as <tt>linkonce</tt>
523 linkage, except that unreferenced <tt>common</tt> globals may not be
524 discarded. This is used for globals that may be emitted in multiple
525 translation units, but that are not guaranteed to be emitted into every
526 translation unit that uses them. One example of this is tentative
527 definitions in C, such as "<tt>int X;</tt>" at global scope.
528 </dd>
529
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]530 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
531
Dale Johannesen96e7e092008-05-23 23:13:41 +0000[diff] [blame]532 <dd>"<tt>weak</tt>" linkage is the same as <tt>common</tt> linkage, except
533 that some targets may choose to emit different assembly sequences for them
534 for target-dependent reasons. This is used for globals that are declared
535 "weak" in C source code.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]536 </dd>
537
538 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
539
540 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
541 pointer to array type. When two global variables with appending linkage are
542 linked together, the two global arrays are appended together. This is the
543 LLVM, typesafe, equivalent of having the system linker append together
544 "sections" with identical names when .o files are linked.
545 </dd>
546
547 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt>: </dt>
Duncan Sands19d161f2009-03-07 15:45:40 +0000[diff] [blame]548
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]549 <dd>The semantics of this linkage follow the ELF object file model: the
550 symbol is weak until linked, if not linked, the symbol becomes null instead
551 of being an undefined reference.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]552 </dd>
553
Duncan Sands19d161f2009-03-07 15:45:40 +0000[diff] [blame]554 <dt><tt><b><a name="linkage_linkonce">linkonce_odr</a></b></tt>: </dt>
Duncan Sands19d161f2009-03-07 15:45:40 +0000[diff] [blame]555 <dt><tt><b><a name="linkage_weak">weak_odr</a></b></tt>: </dt>
Duncan Sands19d161f2009-03-07 15:45:40 +0000[diff] [blame]556 <dd>Some languages allow inequivalent globals to be merged, such as two
557 functions with different semantics. Other languages, such as <tt>C++</tt>,
558 ensure that only equivalent globals are ever merged (the "one definition
Duncan Sandsb95df792009-03-11 20:14:15 +0000[diff] [blame]559 rule" - <tt>odr</tt>). Such languages can use the <tt>linkonce_odr</tt>
560 and <tt>weak_odr</tt> linkage types to indicate that the global will only
561 be merged with equivalent globals. These linkage types are otherwise the
562 same as their non-<tt>odr</tt> versions.
Duncan Sands19d161f2009-03-07 15:45:40 +0000[diff] [blame]563 </dd>
564
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]565 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
566
567 <dd>If none of the above identifiers are used, the global is externally
568 visible, meaning that it participates in linkage and can be used to resolve
569 external symbol references.
570 </dd>
571</dl>
572
573 <p>
574 The next two types of linkage are targeted for Microsoft Windows platform
575 only. They are designed to support importing (exporting) symbols from (to)
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]576 DLLs (Dynamic Link Libraries).
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]577 </p>
578
579 <dl>
580 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt>: </dt>
581
582 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
583 or variable via a global pointer to a pointer that is set up by the DLL
584 exporting the symbol. On Microsoft Windows targets, the pointer name is
Dan Gohman5ec99832009-01-12 21:35:55 +0000[diff] [blame]585 formed by combining <code>__imp_</code> and the function or variable name.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]586 </dd>
587
588 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt>: </dt>
589
590 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
591 pointer to a pointer in a DLL, so that it can be referenced with the
592 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
Dan Gohman5ec99832009-01-12 21:35:55 +0000[diff] [blame]593 name is formed by combining <code>__imp_</code> and the function or variable
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]594 name.
595 </dd>
596
597</dl>
598
Dan Gohman4dfac702008-11-24 17:18:39 +0000[diff] [blame]599<p>For example, since the "<tt>.LC0</tt>"
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]600variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
601variable and was linked with this one, one of the two would be renamed,
602preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
603external (i.e., lacking any linkage declarations), they are accessible
604outside of the current module.</p>
605<p>It is illegal for a function <i>declaration</i>
Duncan Sands565f65d2009-03-11 08:08:06 +0000[diff] [blame]606to have any linkage type other than "externally visible", <tt>dllimport</tt>
607or <tt>extern_weak</tt>.</p>
Duncan Sands19d161f2009-03-07 15:45:40 +0000[diff] [blame]608<p>Aliases can have only <tt>external</tt>, <tt>internal</tt>, <tt>weak</tt>
609or <tt>weak_odr</tt> linkages.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]610</div>
611
612<!-- ======================================================================= -->
613<div class="doc_subsection">
614 <a name="callingconv">Calling Conventions</a>
615</div>
616
617<div class="doc_text">
618
619<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
620and <a href="#i_invoke">invokes</a> can all have an optional calling convention
621specified for the call. The calling convention of any pair of dynamic
622caller/callee must match, or the behavior of the program is undefined. The
623following calling conventions are supported by LLVM, and more may be added in
624the future:</p>
625
626<dl>
627 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
628
629 <dd>This calling convention (the default if no other calling convention is
630 specified) matches the target C calling conventions. This calling convention
631 supports varargs function calls and tolerates some mismatch in the declared
632 prototype and implemented declaration of the function (as does normal C).
633 </dd>
634
635 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
636
637 <dd>This calling convention attempts to make calls as fast as possible
638 (e.g. by passing things in registers). This calling convention allows the
639 target to use whatever tricks it wants to produce fast code for the target,
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]640 without having to conform to an externally specified ABI (Application Binary
641 Interface). Implementations of this convention should allow arbitrary
Arnold Schwaighofer07444922008-05-14 09:17:12 +0000[diff] [blame]642 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> to be
643 supported. This calling convention does not support varargs and requires the
644 prototype of all callees to exactly match the prototype of the function
645 definition.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]646 </dd>
647
648 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
649
650 <dd>This calling convention attempts to make code in the caller as efficient
651 as possible under the assumption that the call is not commonly executed. As
652 such, these calls often preserve all registers so that the call does not break
653 any live ranges in the caller side. This calling convention does not support
654 varargs and requires the prototype of all callees to exactly match the
655 prototype of the function definition.
656 </dd>
657
658 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
659
660 <dd>Any calling convention may be specified by number, allowing
661 target-specific calling conventions to be used. Target specific calling
662 conventions start at 64.
663 </dd>
664</dl>
665
666<p>More calling conventions can be added/defined on an as-needed basis, to
667support pascal conventions or any other well-known target-independent
668convention.</p>
669
670</div>
671
672<!-- ======================================================================= -->
673<div class="doc_subsection">
674 <a name="visibility">Visibility Styles</a>
675</div>
676
677<div class="doc_text">
678
679<p>
680All Global Variables and Functions have one of the following visibility styles:
681</p>
682
683<dl>
684 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
685
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]686 <dd>On targets that use the ELF object file format, default visibility means
687 that the declaration is visible to other
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]688 modules and, in shared libraries, means that the declared entity may be
689 overridden. On Darwin, default visibility means that the declaration is
690 visible to other modules. Default visibility corresponds to "external
691 linkage" in the language.
692 </dd>
693
694 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
695
696 <dd>Two declarations of an object with hidden visibility refer to the same
697 object if they are in the same shared object. Usually, hidden visibility
698 indicates that the symbol will not be placed into the dynamic symbol table,
699 so no other module (executable or shared library) can reference it
700 directly.
701 </dd>
702
703 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
704
705 <dd>On ELF, protected visibility indicates that the symbol will be placed in
706 the dynamic symbol table, but that references within the defining module will
707 bind to the local symbol. That is, the symbol cannot be overridden by another
708 module.
709 </dd>
710</dl>
711
712</div>
713
714<!-- ======================================================================= -->
715<div class="doc_subsection">
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]716 <a name="namedtypes">Named Types</a>
717</div>
718
719<div class="doc_text">
720
721<p>LLVM IR allows you to specify name aliases for certain types. This can make
722it easier to read the IR and make the IR more condensed (particularly when
723recursive types are involved). An example of a name specification is:
724</p>
725
726<div class="doc_code">
727<pre>
728%mytype = type { %mytype*, i32 }
729</pre>
730</div>
731
732<p>You may give a name to any <a href="#typesystem">type</a> except "<a
733href="t_void">void</a>". Type name aliases may be used anywhere a type is
734expected with the syntax "%mytype".</p>
735
736<p>Note that type names are aliases for the structural type that they indicate,
737and that you can therefore specify multiple names for the same type. This often
738leads to confusing behavior when dumping out a .ll file. Since LLVM IR uses
739structural typing, the name is not part of the type. When printing out LLVM IR,
740the printer will pick <em>one name</em> to render all types of a particular
741shape. This means that if you have code where two different source types end up
742having the same LLVM type, that the dumper will sometimes print the "wrong" or
743unexpected type. This is an important design point and isn't going to
744change.</p>
745
746</div>
747
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]748<!-- ======================================================================= -->
749<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]750 <a name="globalvars">Global Variables</a>
751</div>
752
753<div class="doc_text">
754
755<p>Global variables define regions of memory allocated at compilation time
756instead of run-time. Global variables may optionally be initialized, may have
757an explicit section to be placed in, and may have an optional explicit alignment
758specified. A variable may be defined as "thread_local", which means that it
759will not be shared by threads (each thread will have a separated copy of the
760variable). A variable may be defined as a global "constant," which indicates
761that the contents of the variable will <b>never</b> be modified (enabling better
762optimization, allowing the global data to be placed in the read-only section of
763an executable, etc). Note that variables that need runtime initialization
764cannot be marked "constant" as there is a store to the variable.</p>
765
766<p>
767LLVM explicitly allows <em>declarations</em> of global variables to be marked
768constant, even if the final definition of the global is not. This capability
769can be used to enable slightly better optimization of the program, but requires
770the language definition to guarantee that optimizations based on the
771'constantness' are valid for the translation units that do not include the
772definition.
773</p>
774
775<p>As SSA values, global variables define pointer values that are in
776scope (i.e. they dominate) all basic blocks in the program. Global
777variables always define a pointer to their "content" type because they
778describe a region of memory, and all memory objects in LLVM are
779accessed through pointers.</p>
780
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]781<p>A global variable may be declared to reside in a target-specifc numbered
782address space. For targets that support them, address spaces may affect how
783optimizations are performed and/or what target instructions are used to access
Christopher Lamb20a39e92007-12-12 08:44:39 +0000[diff] [blame]784the variable. The default address space is zero. The address space qualifier
785must precede any other attributes.</p>
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]786
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]787<p>LLVM allows an explicit section to be specified for globals. If the target
788supports it, it will emit globals to the section specified.</p>
789
790<p>An explicit alignment may be specified for a global. If not present, or if
791the alignment is set to zero, the alignment of the global is set by the target
792to whatever it feels convenient. If an explicit alignment is specified, the
793global is forced to have at least that much alignment. All alignments must be
794a power of 2.</p>
795
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]796<p>For example, the following defines a global in a numbered address space with
797an initializer, section, and alignment:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]798
799<div class="doc_code">
800<pre>
Dan Gohman21ef02c2009-01-11 00:40:00 +0000[diff] [blame]801@G = addrspace(5) constant float 1.0, section "foo", align 4
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]802</pre>
803</div>
804
805</div>
806
807
808<!-- ======================================================================= -->
809<div class="doc_subsection">
810 <a name="functionstructure">Functions</a>
811</div>
812
813<div class="doc_text">
814
815<p>LLVM function definitions consist of the "<tt>define</tt>" keyord,
816an optional <a href="#linkage">linkage type</a>, an optional
817<a href="#visibility">visibility style</a>, an optional
818<a href="#callingconv">calling convention</a>, a return type, an optional
819<a href="#paramattrs">parameter attribute</a> for the return type, a function
820name, a (possibly empty) argument list (each with optional
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]821<a href="#paramattrs">parameter attributes</a>), optional
822<a href="#fnattrs">function attributes</a>, an optional section,
823an optional alignment, an optional <a href="#gc">garbage collector name</a>,
Chris Lattneree202542008-10-04 18:10:21 +0000[diff] [blame]824an opening curly brace, a list of basic blocks, and a closing curly brace.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]825
826LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
827optional <a href="#linkage">linkage type</a>, an optional
828<a href="#visibility">visibility style</a>, an optional
829<a href="#callingconv">calling convention</a>, a return type, an optional
830<a href="#paramattrs">parameter attribute</a> for the return type, a function
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]831name, a possibly empty list of arguments, an optional alignment, and an optional
Gordon Henriksend606f5b2007-12-10 03:30:21 +0000[diff] [blame]832<a href="#gc">garbage collector name</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]833
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]834<p>A function definition contains a list of basic blocks, forming the CFG
835(Control Flow Graph) for
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]836the function. Each basic block may optionally start with a label (giving the
837basic block a symbol table entry), contains a list of instructions, and ends
838with a <a href="#terminators">terminator</a> instruction (such as a branch or
839function return).</p>
840
841<p>The first basic block in a function is special in two ways: it is immediately
842executed on entrance to the function, and it is not allowed to have predecessor
843basic blocks (i.e. there can not be any branches to the entry block of a
844function). Because the block can have no predecessors, it also cannot have any
845<a href="#i_phi">PHI nodes</a>.</p>
846
847<p>LLVM allows an explicit section to be specified for functions. If the target
848supports it, it will emit functions to the section specified.</p>
849
850<p>An explicit alignment may be specified for a function. If not present, or if
851the alignment is set to zero, the alignment of the function is set by the target
852to whatever it feels convenient. If an explicit alignment is specified, the
853function is forced to have at least that much alignment. All alignments must be
854a power of 2.</p>
855
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]856 <h5>Syntax:</h5>
857
858<div class="doc_code">
Chris Lattner1e5c5cd02008-10-13 16:55:18 +0000[diff] [blame]859<tt>
860define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
861 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
862 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
863 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
864 [<a href="#gc">gc</a>] { ... }
865</tt>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]866</div>
867
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]868</div>
869
870
871<!-- ======================================================================= -->
872<div class="doc_subsection">
873 <a name="aliasstructure">Aliases</a>
874</div>
875<div class="doc_text">
876 <p>Aliases act as "second name" for the aliasee value (which can be either
Anton Korobeynikov96822812008-03-22 08:36:14 +0000[diff] [blame]877 function, global variable, another alias or bitcast of global value). Aliases
878 may have an optional <a href="#linkage">linkage type</a>, and an
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]879 optional <a href="#visibility">visibility style</a>.</p>
880
881 <h5>Syntax:</h5>
882
883<div class="doc_code">
884<pre>
Duncan Sandsd7bfabf2008-09-12 20:48:21 +0000[diff] [blame]885@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]886</pre>
887</div>
888
889</div>
890
891
892
893<!-- ======================================================================= -->
894<div class="doc_subsection"><a name="paramattrs">Parameter Attributes</a></div>
895<div class="doc_text">
896 <p>The return type and each parameter of a function type may have a set of
897 <i>parameter attributes</i> associated with them. Parameter attributes are
898 used to communicate additional information about the result or parameters of
Duncan Sandsf5588dc2007-11-27 13:23:08 +0000[diff] [blame]899 a function. Parameter attributes are considered to be part of the function,
900 not of the function type, so functions with different parameter attributes
901 can have the same function type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]902
903 <p>Parameter attributes are simple keywords that follow the type specified. If
904 multiple parameter attributes are needed, they are space separated. For
905 example:</p>
906
907<div class="doc_code">
908<pre>
Nick Lewycky3022a742009-02-15 23:06:14 +0000[diff] [blame]909declare i32 @printf(i8* noalias nocapture, ...)
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]910declare i32 @atoi(i8 zeroext)
911declare signext i8 @returns_signed_char()
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]912</pre>
913</div>
914
Duncan Sandsf5588dc2007-11-27 13:23:08 +0000[diff] [blame]915 <p>Note that any attributes for the function result (<tt>nounwind</tt>,
916 <tt>readonly</tt>) come immediately after the argument list.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]917
918 <p>Currently, only the following parameter attributes are defined:</p>
919 <dl>
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]920 <dt><tt>zeroext</tt></dt>
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]921 <dd>This indicates to the code generator that the parameter or return value
922 should be zero-extended to a 32-bit value by the caller (for a parameter)
923 or the callee (for a return value).</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]924
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]925 <dt><tt>signext</tt></dt>
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]926 <dd>This indicates to the code generator that the parameter or return value
927 should be sign-extended to a 32-bit value by the caller (for a parameter)
928 or the callee (for a return value).</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]929
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]930 <dt><tt>inreg</tt></dt>
Dale Johannesenc08a0e22008-09-25 20:47:45 +0000[diff] [blame]931 <dd>This indicates that this parameter or return value should be treated
932 in a special target-dependent fashion during while emitting code for a
933 function call or return (usually, by putting it in a register as opposed
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]934 to memory, though some targets use it to distinguish between two different
935 kinds of registers). Use of this attribute is target-specific.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]936
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]937 <dt><tt><a name="byval">byval</a></tt></dt>
Chris Lattner04c86182008-01-15 04:34:22 +0000[diff] [blame]938 <dd>This indicates that the pointer parameter should really be passed by
939 value to the function. The attribute implies that a hidden copy of the
940 pointee is made between the caller and the callee, so the callee is unable
Chris Lattner6a9f3c42008-08-05 18:21:08 +0000[diff] [blame]941 to modify the value in the callee. This attribute is only valid on LLVM
Chris Lattner04c86182008-01-15 04:34:22 +0000[diff] [blame]942 pointer arguments. It is generally used to pass structs and arrays by
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]943 value, but is also valid on pointers to scalars. The copy is considered to
944 belong to the caller not the callee (for example,
945 <tt><a href="#readonly">readonly</a></tt> functions should not write to
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]946 <tt>byval</tt> parameters). This is not a valid attribute for return
Chris Lattnere39c5bf2009-02-05 05:42:28 +0000[diff] [blame]947 values. The byval attribute also supports specifying an alignment with the
948 align attribute. This has a target-specific effect on the code generator
949 that usually indicates a desired alignment for the synthesized stack
950 slot.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]951
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]952 <dt><tt>sret</tt></dt>
Duncan Sands616cc032008-02-18 04:19:38 +0000[diff] [blame]953 <dd>This indicates that the pointer parameter specifies the address of a
954 structure that is the return value of the function in the source program.
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]955 This pointer must be guaranteed by the caller to be valid: loads and stores
956 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]957 be applied to the first parameter. This is not a valid attribute for
958 return values. </dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]959
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]960 <dt><tt>noalias</tt></dt>
Nick Lewyckyff384472008-11-24 03:41:24 +0000[diff] [blame]961 <dd>This indicates that the pointer does not alias any global or any other
962 parameter. The caller is responsible for ensuring that this is the
Nick Lewyckya604a942008-11-24 05:00:44 +0000[diff] [blame]963 case. On a function return value, <tt>noalias</tt> additionally indicates
964 that the pointer does not alias any other pointers visible to the
Nick Lewycky40fb6f22008-12-19 06:39:12 +0000[diff] [blame]965 caller. For further details, please see the discussion of the NoAlias
966 response in
967 <a href="http://llvm.org/docs/AliasAnalysis.html#MustMayNo">alias
968 analysis</a>.</dd>
969
970 <dt><tt>nocapture</tt></dt>
971 <dd>This indicates that the callee does not make any copies of the pointer
972 that outlive the callee itself. This is not a valid attribute for return
973 values.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]974
Duncan Sands4ee46812007-07-27 19:57:41 +0000[diff] [blame]975 <dt><tt>nest</tt></dt>
Duncan Sandsf1a7d4c2008-07-08 09:27:25 +0000[diff] [blame]976 <dd>This indicates that the pointer parameter can be excised using the
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]977 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
978 attribute for return values.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]979 </dl>
980
981</div>
982
983<!-- ======================================================================= -->
984<div class="doc_subsection">
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]985 <a name="gc">Garbage Collector Names</a>
986</div>
987
988<div class="doc_text">
989<p>Each function may specify a garbage collector name, which is simply a
990string.</p>
991
992<div class="doc_code"><pre
993>define void @f() gc "name" { ...</pre></div>
994
995<p>The compiler declares the supported values of <i>name</i>. Specifying a
996collector which will cause the compiler to alter its output in order to support
997the named garbage collection algorithm.</p>
998</div>
999
1000<!-- ======================================================================= -->
1001<div class="doc_subsection">
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1002 <a name="fnattrs">Function Attributes</a>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1003</div>
1004
1005<div class="doc_text">
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1006
1007<p>Function attributes are set to communicate additional information about
1008 a function. Function attributes are considered to be part of the function,
1009 not of the function type, so functions with different parameter attributes
1010 can have the same function type.</p>
1011
1012 <p>Function attributes are simple keywords that follow the type specified. If
1013 multiple attributes are needed, they are space separated. For
1014 example:</p>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1015
1016<div class="doc_code">
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1017<pre>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1018define void @f() noinline { ... }
1019define void @f() alwaysinline { ... }
1020define void @f() alwaysinline optsize { ... }
1021define void @f() optsize
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1022</pre>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1023</div>
1024
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1025<dl>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1026<dt><tt>alwaysinline</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]1027<dd>This attribute indicates that the inliner should attempt to inline this
1028function into callers whenever possible, ignoring any active inlining size
1029threshold for this caller.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1030
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1031<dt><tt>noinline</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]1032<dd>This attribute indicates that the inliner should never inline this function
Chris Lattner377f3ae2008-10-05 17:14:59 +0000[diff] [blame]1033in any situation. This attribute may not be used together with the
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]1034<tt>alwaysinline</tt> attribute.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1035
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1036<dt><tt>optsize</tt></dt>
Devang Patelc29099b2008-09-29 18:34:44 +0000[diff] [blame]1037<dd>This attribute suggests that optimization passes and code generator passes
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]1038make choices that keep the code size of this function low, and otherwise do
1039optimizations specifically to reduce code size.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1040
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1041<dt><tt>noreturn</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]1042<dd>This function attribute indicates that the function never returns normally.
1043This produces undefined behavior at runtime if the function ever does
1044dynamically return.</dd>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1045
1046<dt><tt>nounwind</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]1047<dd>This function attribute indicates that the function never returns with an
1048unwind or exceptional control flow. If the function does unwind, its runtime
1049behavior is undefined.</dd>
1050
1051<dt><tt>readnone</tt></dt>
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]1052<dd>This attribute indicates that the function computes its result (or the
1053exception it throws) based strictly on its arguments, without dereferencing any
1054pointer arguments or otherwise accessing any mutable state (e.g. memory, control
1055registers, etc) visible to caller functions. It does not write through any
1056pointer arguments (including <tt><a href="#byval">byval</a></tt> arguments) and
1057never changes any state visible to callers.</dd>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1058
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]1059<dt><tt><a name="readonly">readonly</a></tt></dt>
1060<dd>This attribute indicates that the function does not write through any
1061pointer arguments (including <tt><a href="#byval">byval</a></tt> arguments)
1062or otherwise modify any state (e.g. memory, control registers, etc) visible to
1063caller functions. It may dereference pointer arguments and read state that may
1064be set in the caller. A readonly function always returns the same value (or
1065throws the same exception) when called with the same set of arguments and global
1066state.</dd>
Bill Wendlingdfaabba2008-11-13 01:02:51 +0000[diff] [blame]1067
1068<dt><tt><a name="ssp">ssp</a></tt></dt>
Bill Wendling34f70d82008-11-26 19:19:05 +0000[diff] [blame]1069<dd>This attribute indicates that the function should emit a stack smashing
Bill Wendlingdfaabba2008-11-13 01:02:51 +0000[diff] [blame]1070protector. It is in the form of a "canary"&mdash;a random value placed on the
1071stack before the local variables that's checked upon return from the function to
1072see if it has been overwritten. A heuristic is used to determine if a function
Bill Wendling34f70d82008-11-26 19:19:05 +0000[diff] [blame]1073needs stack protectors or not.
Bill Wendlingdfaabba2008-11-13 01:02:51 +0000[diff] [blame]1074
Bill Wendlingbe9ec3f2008-11-26 19:07:40 +0000[diff] [blame]1075<p>If a function that has an <tt>ssp</tt> attribute is inlined into a function
1076that doesn't have an <tt>ssp</tt> attribute, then the resulting function will
1077have an <tt>ssp</tt> attribute.</p></dd>
1078
1079<dt><tt>sspreq</tt></dt>
Bill Wendling34f70d82008-11-26 19:19:05 +0000[diff] [blame]1080<dd>This attribute indicates that the function should <em>always</em> emit a
Bill Wendlingdfaabba2008-11-13 01:02:51 +0000[diff] [blame]1081stack smashing protector. This overrides the <tt><a href="#ssp">ssp</a></tt>
Bill Wendling34f70d82008-11-26 19:19:05 +0000[diff] [blame]1082function attribute.
Bill Wendlingbe9ec3f2008-11-26 19:07:40 +0000[diff] [blame]1083
1084<p>If a function that has an <tt>sspreq</tt> attribute is inlined into a
1085function that doesn't have an <tt>sspreq</tt> attribute or which has
1086an <tt>ssp</tt> attribute, then the resulting function will have
1087an <tt>sspreq</tt> attribute.</p></dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1088</dl>
1089
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1090</div>
1091
1092<!-- ======================================================================= -->
1093<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1094 <a name="moduleasm">Module-Level Inline Assembly</a>
1095</div>
1096
1097<div class="doc_text">
1098<p>
1099Modules may contain "module-level inline asm" blocks, which corresponds to the
1100GCC "file scope inline asm" blocks. These blocks are internally concatenated by
1101LLVM and treated as a single unit, but may be separated in the .ll file if
1102desired. The syntax is very simple:
1103</p>
1104
1105<div class="doc_code">
1106<pre>
1107module asm "inline asm code goes here"
1108module asm "more can go here"
1109</pre>
1110</div>
1111
1112<p>The strings can contain any character by escaping non-printable characters.
1113 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
1114 for the number.
1115</p>
1116
1117<p>
1118 The inline asm code is simply printed to the machine code .s file when
1119 assembly code is generated.
1120</p>
1121</div>
1122
1123<!-- ======================================================================= -->
1124<div class="doc_subsection">
1125 <a name="datalayout">Data Layout</a>
1126</div>
1127
1128<div class="doc_text">
1129<p>A module may specify a target specific data layout string that specifies how
1130data is to be laid out in memory. The syntax for the data layout is simply:</p>
1131<pre> target datalayout = "<i>layout specification</i>"</pre>
1132<p>The <i>layout specification</i> consists of a list of specifications
1133separated by the minus sign character ('-'). Each specification starts with a
1134letter and may include other information after the letter to define some
1135aspect of the data layout. The specifications accepted are as follows: </p>
1136<dl>
1137 <dt><tt>E</tt></dt>
1138 <dd>Specifies that the target lays out data in big-endian form. That is, the
1139 bits with the most significance have the lowest address location.</dd>
1140 <dt><tt>e</tt></dt>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1141 <dd>Specifies that the target lays out data in little-endian form. That is,
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1142 the bits with the least significance have the lowest address location.</dd>
1143 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1144 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
1145 <i>preferred</i> alignments. All sizes are in bits. Specifying the <i>pref</i>
1146 alignment is optional. If omitted, the preceding <tt>:</tt> should be omitted
1147 too.</dd>
1148 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1149 <dd>This specifies the alignment for an integer type of a given bit
1150 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1151 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1152 <dd>This specifies the alignment for a vector type of a given bit
1153 <i>size</i>.</dd>
1154 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1155 <dd>This specifies the alignment for a floating point type of a given bit
1156 <i>size</i>. The value of <i>size</i> must be either 32 (float) or 64
1157 (double).</dd>
1158 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1159 <dd>This specifies the alignment for an aggregate type of a given bit
1160 <i>size</i>.</dd>
1161</dl>
1162<p>When constructing the data layout for a given target, LLVM starts with a
1163default set of specifications which are then (possibly) overriden by the
1164specifications in the <tt>datalayout</tt> keyword. The default specifications
1165are given in this list:</p>
1166<ul>
1167 <li><tt>E</tt> - big endian</li>
1168 <li><tt>p:32:64:64</tt> - 32-bit pointers with 64-bit alignment</li>
1169 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1170 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1171 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1172 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1173 <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]1174 alignment of 64-bits</li>
1175 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1176 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1177 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1178 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1179 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
1180</ul>
Chris Lattner6a9f3c42008-08-05 18:21:08 +0000[diff] [blame]1181<p>When LLVM is determining the alignment for a given type, it uses the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]1182following rules:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1183<ol>
1184 <li>If the type sought is an exact match for one of the specifications, that
1185 specification is used.</li>
1186 <li>If no match is found, and the type sought is an integer type, then the
1187 smallest integer type that is larger than the bitwidth of the sought type is
1188 used. If none of the specifications are larger than the bitwidth then the the
1189 largest integer type is used. For example, given the default specifications
1190 above, the i7 type will use the alignment of i8 (next largest) while both
1191 i65 and i256 will use the alignment of i64 (largest specified).</li>
1192 <li>If no match is found, and the type sought is a vector type, then the
1193 largest vector type that is smaller than the sought vector type will be used
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]1194 as a fall back. This happens because &lt;128 x double&gt; can be implemented
1195 in terms of 64 &lt;2 x double&gt;, for example.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1196</ol>
1197</div>
1198
1199<!-- *********************************************************************** -->
1200<div class="doc_section"> <a name="typesystem">Type System</a> </div>
1201<!-- *********************************************************************** -->
1202
1203<div class="doc_text">
1204
1205<p>The LLVM type system is one of the most important features of the
1206intermediate representation. Being typed enables a number of
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1207optimizations to be performed on the intermediate representation directly,
1208without having to do
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1209extra analyses on the side before the transformation. A strong type
1210system makes it easier to read the generated code and enables novel
1211analyses and transformations that are not feasible to perform on normal
1212three address code representations.</p>
1213
1214</div>
1215
1216<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1217<div class="doc_subsection"> <a name="t_classifications">Type
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1218Classifications</a> </div>
1219<div class="doc_text">
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1220<p>The types fall into a few useful
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1221classifications:</p>
1222
1223<table border="1" cellspacing="0" cellpadding="4">
1224 <tbody>
1225 <tr><th>Classification</th><th>Types</th></tr>
1226 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1227 <td><a href="#t_integer">integer</a></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1228 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
1229 </tr>
1230 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1231 <td><a href="#t_floating">floating point</a></td>
1232 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1233 </tr>
1234 <tr>
1235 <td><a name="t_firstclass">first class</a></td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1236 <td><a href="#t_integer">integer</a>,
1237 <a href="#t_floating">floating point</a>,
1238 <a href="#t_pointer">pointer</a>,
Dan Gohmanf6237db2008-06-18 18:42:13 +0000[diff] [blame]1239 <a href="#t_vector">vector</a>,
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]1240 <a href="#t_struct">structure</a>,
1241 <a href="#t_array">array</a>,
Dan Gohmand13951e2008-05-23 22:50:26 +0000[diff] [blame]1242 <a href="#t_label">label</a>.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1243 </td>
1244 </tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1245 <tr>
1246 <td><a href="#t_primitive">primitive</a></td>
1247 <td><a href="#t_label">label</a>,
1248 <a href="#t_void">void</a>,
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1249 <a href="#t_floating">floating point</a>.</td>
1250 </tr>
1251 <tr>
1252 <td><a href="#t_derived">derived</a></td>
1253 <td><a href="#t_integer">integer</a>,
1254 <a href="#t_array">array</a>,
1255 <a href="#t_function">function</a>,
1256 <a href="#t_pointer">pointer</a>,
1257 <a href="#t_struct">structure</a>,
1258 <a href="#t_pstruct">packed structure</a>,
1259 <a href="#t_vector">vector</a>,
1260 <a href="#t_opaque">opaque</a>.
Dan Gohman032ba852008-10-14 16:32:04 +0000[diff] [blame]1261 </td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1262 </tr>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1263 </tbody>
1264</table>
1265
1266<p>The <a href="#t_firstclass">first class</a> types are perhaps the
1267most important. Values of these types are the only ones which can be
1268produced by instructions, passed as arguments, or used as operands to
Dan Gohman4f29e422008-05-23 21:53:15 +0000[diff] [blame]1269instructions.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1270</div>
1271
1272<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1273<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1274
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1275<div class="doc_text">
1276<p>The primitive types are the fundamental building blocks of the LLVM
1277system.</p>
1278
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1279</div>
1280
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1281<!-- _______________________________________________________________________ -->
1282<div class="doc_subsubsection"> <a name="t_floating">Floating Point Types</a> </div>
1283
1284<div class="doc_text">
1285 <table>
1286 <tbody>
1287 <tr><th>Type</th><th>Description</th></tr>
1288 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1289 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1290 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1291 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1292 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1293 </tbody>
1294 </table>
1295</div>
1296
1297<!-- _______________________________________________________________________ -->
1298<div class="doc_subsubsection"> <a name="t_void">Void Type</a> </div>
1299
1300<div class="doc_text">
1301<h5>Overview:</h5>
1302<p>The void type does not represent any value and has no size.</p>
1303
1304<h5>Syntax:</h5>
1305
1306<pre>
1307 void
1308</pre>
1309</div>
1310
1311<!-- _______________________________________________________________________ -->
1312<div class="doc_subsubsection"> <a name="t_label">Label Type</a> </div>
1313
1314<div class="doc_text">
1315<h5>Overview:</h5>
1316<p>The label type represents code labels.</p>
1317
1318<h5>Syntax:</h5>
1319
1320<pre>
1321 label
1322</pre>
1323</div>
1324
1325
1326<!-- ======================================================================= -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1327<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
1328
1329<div class="doc_text">
1330
1331<p>The real power in LLVM comes from the derived types in the system.
1332This is what allows a programmer to represent arrays, functions,
1333pointers, and other useful types. Note that these derived types may be
1334recursive: For example, it is possible to have a two dimensional array.</p>
1335
1336</div>
1337
1338<!-- _______________________________________________________________________ -->
1339<div class="doc_subsubsection"> <a name="t_integer">Integer Type</a> </div>
1340
1341<div class="doc_text">
1342
1343<h5>Overview:</h5>
1344<p>The integer type is a very simple derived type that simply specifies an
1345arbitrary bit width for the integer type desired. Any bit width from 1 bit to
13462^23-1 (about 8 million) can be specified.</p>
1347
1348<h5>Syntax:</h5>
1349
1350<pre>
1351 iN
1352</pre>
1353
1354<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1355value.</p>
1356
1357<h5>Examples:</h5>
1358<table class="layout">
Chris Lattner251ab812007-12-18 06:18:21 +0000[diff] [blame]1359 <tbody>
1360 <tr>
1361 <td><tt>i1</tt></td>
1362 <td>a single-bit integer.</td>
1363 </tr><tr>
1364 <td><tt>i32</tt></td>
1365 <td>a 32-bit integer.</td>
1366 </tr><tr>
1367 <td><tt>i1942652</tt></td>
1368 <td>a really big integer of over 1 million bits.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1369 </tr>
Chris Lattner251ab812007-12-18 06:18:21 +0000[diff] [blame]1370 </tbody>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1371</table>
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1372
1373<p>Note that the code generator does not yet support large integer types
1374to be used as function return types. The specific limit on how large a
1375return type the code generator can currently handle is target-dependent;
1376currently it's often 64 bits for 32-bit targets and 128 bits for 64-bit
1377targets.</p>
1378
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1379</div>
1380
1381<!-- _______________________________________________________________________ -->
1382<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
1383
1384<div class="doc_text">
1385
1386<h5>Overview:</h5>
1387
1388<p>The array type is a very simple derived type that arranges elements
1389sequentially in memory. The array type requires a size (number of
1390elements) and an underlying data type.</p>
1391
1392<h5>Syntax:</h5>
1393
1394<pre>
1395 [&lt;# elements&gt; x &lt;elementtype&gt;]
1396</pre>
1397
1398<p>The number of elements is a constant integer value; elementtype may
1399be any type with a size.</p>
1400
1401<h5>Examples:</h5>
1402<table class="layout">
1403 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1404 <td class="left"><tt>[40 x i32]</tt></td>
1405 <td class="left">Array of 40 32-bit integer values.</td>
1406 </tr>
1407 <tr class="layout">
1408 <td class="left"><tt>[41 x i32]</tt></td>
1409 <td class="left">Array of 41 32-bit integer values.</td>
1410 </tr>
1411 <tr class="layout">
1412 <td class="left"><tt>[4 x i8]</tt></td>
1413 <td class="left">Array of 4 8-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1414 </tr>
1415</table>
1416<p>Here are some examples of multidimensional arrays:</p>
1417<table class="layout">
1418 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1419 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1420 <td class="left">3x4 array of 32-bit integer values.</td>
1421 </tr>
1422 <tr class="layout">
1423 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1424 <td class="left">12x10 array of single precision floating point values.</td>
1425 </tr>
1426 <tr class="layout">
1427 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1428 <td class="left">2x3x4 array of 16-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1429 </tr>
1430</table>
1431
1432<p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
1433length array. Normally, accesses past the end of an array are undefined in
1434LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
1435As a special case, however, zero length arrays are recognized to be variable
1436length. This allows implementation of 'pascal style arrays' with the LLVM
1437type "{ i32, [0 x float]}", for example.</p>
1438
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1439<p>Note that the code generator does not yet support large aggregate types
1440to be used as function return types. The specific limit on how large an
1441aggregate return type the code generator can currently handle is
1442target-dependent, and also dependent on the aggregate element types.</p>
1443
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1444</div>
1445
1446<!-- _______________________________________________________________________ -->
1447<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
1448<div class="doc_text">
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1449
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1450<h5>Overview:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1451
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1452<p>The function type can be thought of as a function signature. It
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1453consists of a return type and a list of formal parameter types. The
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1454return 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]1455If 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]1456class types, and the struct must have at least one element.</p>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1457
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1458<h5>Syntax:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1459
1460<pre>
1461 &lt;returntype list&gt; (&lt;parameter list&gt;)
1462</pre>
1463
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1464<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
1465specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1466which indicates that the function takes a variable number of arguments.
1467Variable argument functions can access their arguments with the <a
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1468 href="#int_varargs">variable argument handling intrinsic</a> functions.
1469'<tt>&lt;returntype list&gt;</tt>' is a comma-separated list of
1470<a href="#t_firstclass">first class</a> type specifiers.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1471
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1472<h5>Examples:</h5>
1473<table class="layout">
1474 <tr class="layout">
1475 <td class="left"><tt>i32 (i32)</tt></td>
1476 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
1477 </td>
1478 </tr><tr class="layout">
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]1479 <td class="left"><tt>float&nbsp;(i16&nbsp;signext,&nbsp;i32&nbsp;*)&nbsp;*
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1480 </tt></td>
1481 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
1482 an <tt>i16</tt> that should be sign extended and a
1483 <a href="#t_pointer">pointer</a> to <tt>i32</tt>, returning
1484 <tt>float</tt>.
1485 </td>
1486 </tr><tr class="layout">
1487 <td class="left"><tt>i32 (i8*, ...)</tt></td>
1488 <td class="left">A vararg function that takes at least one
1489 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1490 which returns an integer. This is the signature for <tt>printf</tt> in
1491 LLVM.
1492 </td>
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1493 </tr><tr class="layout">
1494 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Misha Brukmanafc88b02008-11-27 06:41:20 +0000[diff] [blame]1495 <td class="left">A function taking an <tt>i32</tt>, returning two
1496 <tt>i32</tt> values as an aggregate of type <tt>{ i32, i32 }</tt>
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1497 </td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1498 </tr>
1499</table>
1500
1501</div>
1502<!-- _______________________________________________________________________ -->
1503<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
1504<div class="doc_text">
1505<h5>Overview:</h5>
1506<p>The structure type is used to represent a collection of data members
1507together in memory. The packing of the field types is defined to match
1508the ABI of the underlying processor. The elements of a structure may
1509be any type that has a size.</p>
1510<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1511and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1512field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1513instruction.</p>
1514<h5>Syntax:</h5>
1515<pre> { &lt;type list&gt; }<br></pre>
1516<h5>Examples:</h5>
1517<table class="layout">
1518 <tr class="layout">
1519 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1520 <td class="left">A triple of three <tt>i32</tt> values</td>
1521 </tr><tr class="layout">
1522 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1523 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1524 second element is a <a href="#t_pointer">pointer</a> to a
1525 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1526 an <tt>i32</tt>.</td>
1527 </tr>
1528</table>
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1529
1530<p>Note that the code generator does not yet support large aggregate types
1531to be used as function return types. The specific limit on how large an
1532aggregate return type the code generator can currently handle is
1533target-dependent, and also dependent on the aggregate element types.</p>
1534
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1535</div>
1536
1537<!-- _______________________________________________________________________ -->
1538<div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
1539</div>
1540<div class="doc_text">
1541<h5>Overview:</h5>
1542<p>The packed structure type is used to represent a collection of data members
1543together in memory. There is no padding between fields. Further, the alignment
1544of a packed structure is 1 byte. The elements of a packed structure may
1545be any type that has a size.</p>
1546<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1547and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1548field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1549instruction.</p>
1550<h5>Syntax:</h5>
1551<pre> &lt; { &lt;type list&gt; } &gt; <br></pre>
1552<h5>Examples:</h5>
1553<table class="layout">
1554 <tr class="layout">
1555 <td class="left"><tt>&lt; { i32, i32, i32 } &gt;</tt></td>
1556 <td class="left">A triple of three <tt>i32</tt> values</td>
1557 </tr><tr class="layout">
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1558 <td class="left">
1559<tt>&lt;&nbsp;{&nbsp;float,&nbsp;i32&nbsp;(i32)*&nbsp;}&nbsp;&gt;</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1560 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1561 second element is a <a href="#t_pointer">pointer</a> to a
1562 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1563 an <tt>i32</tt>.</td>
1564 </tr>
1565</table>
1566</div>
1567
1568<!-- _______________________________________________________________________ -->
1569<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
1570<div class="doc_text">
1571<h5>Overview:</h5>
1572<p>As in many languages, the pointer type represents a pointer or
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]1573reference to another object, which must live in memory. Pointer types may have
1574an optional address space attribute defining the target-specific numbered
1575address space where the pointed-to object resides. The default address space is
1576zero.</p>
Chris Lattner96edbd32009-02-08 19:53:29 +0000[diff] [blame]1577
1578<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]1579it permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</tt> instead.</p>
Chris Lattner96edbd32009-02-08 19:53:29 +0000[diff] [blame]1580
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1581<h5>Syntax:</h5>
1582<pre> &lt;type&gt; *<br></pre>
1583<h5>Examples:</h5>
1584<table class="layout">
1585 <tr class="layout">
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]1586 <td class="left"><tt>[4 x i32]*</tt></td>
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1587 <td class="left">A <a href="#t_pointer">pointer</a> to <a
1588 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
1589 </tr>
1590 <tr class="layout">
1591 <td class="left"><tt>i32 (i32 *) *</tt></td>
1592 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1593 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1594 <tt>i32</tt>.</td>
1595 </tr>
1596 <tr class="layout">
1597 <td class="left"><tt>i32 addrspace(5)*</tt></td>
1598 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
1599 that resides in address space #5.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1600 </tr>
1601</table>
1602</div>
1603
1604<!-- _______________________________________________________________________ -->
1605<div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
1606<div class="doc_text">
1607
1608<h5>Overview:</h5>
1609
1610<p>A vector type is a simple derived type that represents a vector
1611of elements. Vector types are used when multiple primitive data
1612are operated in parallel using a single instruction (SIMD).
1613A vector type requires a size (number of
1614elements) and an underlying primitive data type. Vectors must have a power
1615of two length (1, 2, 4, 8, 16 ...). Vector types are
1616considered <a href="#t_firstclass">first class</a>.</p>
1617
1618<h5>Syntax:</h5>
1619
1620<pre>
1621 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1622</pre>
1623
1624<p>The number of elements is a constant integer value; elementtype may
1625be any integer or floating point type.</p>
1626
1627<h5>Examples:</h5>
1628
1629<table class="layout">
1630 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1631 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
1632 <td class="left">Vector of 4 32-bit integer values.</td>
1633 </tr>
1634 <tr class="layout">
1635 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
1636 <td class="left">Vector of 8 32-bit floating-point values.</td>
1637 </tr>
1638 <tr class="layout">
1639 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
1640 <td class="left">Vector of 2 64-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1641 </tr>
1642</table>
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1643
1644<p>Note that the code generator does not yet support large vector types
1645to be used as function return types. The specific limit on how large a
1646vector return type codegen can currently handle is target-dependent;
1647currently it's often a few times longer than a hardware vector register.</p>
1648
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1649</div>
1650
1651<!-- _______________________________________________________________________ -->
1652<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1653<div class="doc_text">
1654
1655<h5>Overview:</h5>
1656
1657<p>Opaque types are used to represent unknown types in the system. This
Gordon Henriksenda0706e2007-10-14 00:34:53 +0000[diff] [blame]1658corresponds (for example) to the C notion of a forward declared structure type.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1659In LLVM, opaque types can eventually be resolved to any type (not just a
1660structure type).</p>
1661
1662<h5>Syntax:</h5>
1663
1664<pre>
1665 opaque
1666</pre>
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>opaque</tt></td>
1673 <td class="left">An opaque type.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1674 </tr>
1675</table>
1676</div>
1677
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]1678<!-- ======================================================================= -->
1679<div class="doc_subsection">
1680 <a name="t_uprefs">Type Up-references</a>
1681</div>
1682
1683<div class="doc_text">
1684<h5>Overview:</h5>
1685<p>
1686An "up reference" allows you to refer to a lexically enclosing type without
1687requiring it to have a name. For instance, a structure declaration may contain a
1688pointer to any of the types it is lexically a member of. Example of up
1689references (with their equivalent as named type declarations) include:</p>
1690
1691<pre>
Chris Lattner5ad632d2009-02-09 10:00:56 +0000[diff] [blame]1692 { \2 * } %x = type { %x* }
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]1693 { \2 }* %y = type { %y }*
1694 \1* %z = type %z*
1695</pre>
1696
1697<p>
1698An up reference is needed by the asmprinter for printing out cyclic types when
1699there is no declared name for a type in the cycle. Because the asmprinter does
1700not want to print out an infinite type string, it needs a syntax to handle
1701recursive types that have no names (all names are optional in llvm IR).
1702</p>
1703
1704<h5>Syntax:</h5>
1705<pre>
1706 \&lt;level&gt;
1707</pre>
1708
1709<p>
1710The level is the count of the lexical type that is being referred to.
1711</p>
1712
1713<h5>Examples:</h5>
1714
1715<table class="layout">
1716 <tr class="layout">
1717 <td class="left"><tt>\1*</tt></td>
1718 <td class="left">Self-referential pointer.</td>
1719 </tr>
1720 <tr class="layout">
1721 <td class="left"><tt>{ { \3*, i8 }, i32 }</tt></td>
1722 <td class="left">Recursive structure where the upref refers to the out-most
1723 structure.</td>
1724 </tr>
1725</table>
1726</div>
1727
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1728
1729<!-- *********************************************************************** -->
1730<div class="doc_section"> <a name="constants">Constants</a> </div>
1731<!-- *********************************************************************** -->
1732
1733<div class="doc_text">
1734
1735<p>LLVM has several different basic types of constants. This section describes
1736them all and their syntax.</p>
1737
1738</div>
1739
1740<!-- ======================================================================= -->
1741<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
1742
1743<div class="doc_text">
1744
1745<dl>
1746 <dt><b>Boolean constants</b></dt>
1747
1748 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
1749 constants of the <tt><a href="#t_primitive">i1</a></tt> type.
1750 </dd>
1751
1752 <dt><b>Integer constants</b></dt>
1753
1754 <dd>Standard integers (such as '4') are constants of the <a
1755 href="#t_integer">integer</a> type. Negative numbers may be used with
1756 integer types.
1757 </dd>
1758
1759 <dt><b>Floating point constants</b></dt>
1760
1761 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
1762 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]1763 notation (see below). The assembler requires the exact decimal value of
1764 a floating-point constant. For example, the assembler accepts 1.25 but
1765 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
1766 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1767
1768 <dt><b>Null pointer constants</b></dt>
1769
1770 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
1771 and must be of <a href="#t_pointer">pointer type</a>.</dd>
1772
1773</dl>
1774
Dale Johannesenf82a52f2009-02-11 22:14:51 +0000[diff] [blame]1775<p>The one non-intuitive notation for constants is the hexadecimal form
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1776of floating point constants. For example, the form '<tt>double
17770x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
17784.5e+15</tt>'. The only time hexadecimal floating point constants are required
1779(and the only time that they are generated by the disassembler) is when a
1780floating point constant must be emitted but it cannot be represented as a
Dale Johannesenf82a52f2009-02-11 22:14:51 +0000[diff] [blame]1781decimal floating point number in a reasonable number of digits. For example,
1782NaN's, infinities, and other
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1783special values are represented in their IEEE hexadecimal format so that
1784assembly and disassembly do not cause any bits to change in the constants.</p>
Dale Johannesenf82a52f2009-02-11 22:14:51 +0000[diff] [blame]1785<p>When using the hexadecimal form, constants of types float and double are
1786represented using the 16-digit form shown above (which matches the IEEE754
1787representation for double); float values must, however, be exactly representable
1788as IEE754 single precision.
1789Hexadecimal format is always used for long
1790double, and there are three forms of long double. The 80-bit
1791format used by x86 is represented as <tt>0xK</tt>
1792followed by 20 hexadecimal digits.
1793The 128-bit format used by PowerPC (two adjacent doubles) is represented
1794by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit
1795format is represented
1796by <tt>0xL</tt> followed by 32 hexadecimal digits; no currently supported
1797target uses this format. Long doubles will only work if they match
1798the long double format on your target. All hexadecimal formats are big-endian
1799(sign bit at the left).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1800</div>
1801
1802<!-- ======================================================================= -->
Chris Lattner97063852009-02-28 18:32:25 +0000[diff] [blame]1803<div class="doc_subsection">
1804<a name="aggregateconstants"> <!-- old anchor -->
1805<a name="complexconstants">Complex Constants</a></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1806</div>
1807
1808<div class="doc_text">
Chris Lattner97063852009-02-28 18:32:25 +0000[diff] [blame]1809<p>Complex constants are a (potentially recursive) combination of simple
1810constants and smaller complex constants.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1811
1812<dl>
1813 <dt><b>Structure constants</b></dt>
1814
1815 <dd>Structure constants are represented with notation similar to structure
1816 type definitions (a comma separated list of elements, surrounded by braces
Chris Lattner6c8de962007-12-25 20:34:52 +0000[diff] [blame]1817 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
1818 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]1819 must have <a href="#t_struct">structure type</a>, and the number and
1820 types of elements must match those specified by the type.
1821 </dd>
1822
1823 <dt><b>Array constants</b></dt>
1824
1825 <dd>Array constants are represented with notation similar to array type
1826 definitions (a comma separated list of elements, surrounded by square brackets
1827 (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74 ]</tt>". Array
1828 constants must have <a href="#t_array">array type</a>, and the number and
1829 types of elements must match those specified by the type.
1830 </dd>
1831
1832 <dt><b>Vector constants</b></dt>
1833
1834 <dd>Vector constants are represented with notation similar to vector type
1835 definitions (a comma separated list of elements, surrounded by
1836 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32 42,
1837 i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must have <a
1838 href="#t_vector">vector type</a>, and the number and types of elements must
1839 match those specified by the type.
1840 </dd>
1841
1842 <dt><b>Zero initialization</b></dt>
1843
1844 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1845 value to zero of <em>any</em> type, including scalar and aggregate types.
1846 This is often used to avoid having to print large zero initializers (e.g. for
1847 large arrays) and is always exactly equivalent to using explicit zero
1848 initializers.
1849 </dd>
1850</dl>
1851
1852</div>
1853
1854<!-- ======================================================================= -->
1855<div class="doc_subsection">
1856 <a name="globalconstants">Global Variable and Function Addresses</a>
1857</div>
1858
1859<div class="doc_text">
1860
1861<p>The addresses of <a href="#globalvars">global variables</a> and <a
1862href="#functionstructure">functions</a> are always implicitly valid (link-time)
1863constants. These constants are explicitly referenced when the <a
1864href="#identifiers">identifier for the global</a> is used and always have <a
1865href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1866file:</p>
1867
1868<div class="doc_code">
1869<pre>
1870@X = global i32 17
1871@Y = global i32 42
1872@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
1873</pre>
1874</div>
1875
1876</div>
1877
1878<!-- ======================================================================= -->
1879<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
1880<div class="doc_text">
1881 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
1882 no specific value. Undefined values may be of any type and be used anywhere
1883 a constant is permitted.</p>
1884
1885 <p>Undefined values indicate to the compiler that the program is well defined
1886 no matter what value is used, giving the compiler more freedom to optimize.
1887 </p>
1888</div>
1889
1890<!-- ======================================================================= -->
1891<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1892</div>
1893
1894<div class="doc_text">
1895
1896<p>Constant expressions are used to allow expressions involving other constants
1897to be used as constants. Constant expressions may be of any <a
1898href="#t_firstclass">first class</a> type and may involve any LLVM operation
1899that does not have side effects (e.g. load and call are not supported). The
1900following is the syntax for constant expressions:</p>
1901
1902<dl>
1903 <dt><b><tt>trunc ( CST to TYPE )</tt></b></dt>
1904 <dd>Truncate a constant to another type. The bit size of CST must be larger
1905 than the bit size of TYPE. Both types must be integers.</dd>
1906
1907 <dt><b><tt>zext ( CST to TYPE )</tt></b></dt>
1908 <dd>Zero extend a constant to another type. The bit size of CST must be
1909 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1910
1911 <dt><b><tt>sext ( CST to TYPE )</tt></b></dt>
1912 <dd>Sign extend a constant to another type. The bit size of CST must be
1913 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1914
1915 <dt><b><tt>fptrunc ( CST to TYPE )</tt></b></dt>
1916 <dd>Truncate a floating point constant to another floating point type. The
1917 size of CST must be larger than the size of TYPE. Both types must be
1918 floating point.</dd>
1919
1920 <dt><b><tt>fpext ( CST to TYPE )</tt></b></dt>
1921 <dd>Floating point extend a constant to another type. The size of CST must be
1922 smaller or equal to the size of TYPE. Both types must be floating point.</dd>
1923
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]1924 <dt><b><tt>fptoui ( CST to TYPE )</tt></b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1925 <dd>Convert a floating point constant to the corresponding unsigned integer
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1926 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1927 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1928 of the same number of elements. If the value won't fit in the integer type,
1929 the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1930
1931 <dt><b><tt>fptosi ( CST to TYPE )</tt></b></dt>
1932 <dd>Convert a floating point constant to the corresponding signed integer
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1933 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1934 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1935 of the same number of elements. If the value won't fit in the integer type,
1936 the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1937
1938 <dt><b><tt>uitofp ( CST to TYPE )</tt></b></dt>
1939 <dd>Convert an unsigned integer constant to the corresponding floating point
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1940 constant. TYPE must be a scalar or vector floating point type. CST must be of
1941 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1942 of the same number of elements. If the value won't fit in the floating point
1943 type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1944
1945 <dt><b><tt>sitofp ( CST to TYPE )</tt></b></dt>
1946 <dd>Convert a signed integer constant to the corresponding floating point
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1947 constant. TYPE must be a scalar or vector floating point type. CST must be of
1948 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1949 of the same number of elements. If the value won't fit in the floating point
1950 type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1951
1952 <dt><b><tt>ptrtoint ( CST to TYPE )</tt></b></dt>
1953 <dd>Convert a pointer typed constant to the corresponding integer constant
1954 TYPE must be an integer type. CST must be of pointer type. The CST value is
1955 zero extended, truncated, or unchanged to make it fit in TYPE.</dd>
1956
1957 <dt><b><tt>inttoptr ( CST to TYPE )</tt></b></dt>
1958 <dd>Convert a integer constant to a pointer constant. TYPE must be a
1959 pointer type. CST must be of integer type. The CST value is zero extended,
1960 truncated, or unchanged to make it fit in a pointer size. This one is
1961 <i>really</i> dangerous!</dd>
1962
1963 <dt><b><tt>bitcast ( CST to TYPE )</tt></b></dt>
Chris Lattner557bc5d2009-02-28 18:27:03 +0000[diff] [blame]1964 <dd>Convert a constant, CST, to another TYPE. The constraints of the operands
1965 are the same as those for the <a href="#i_bitcast">bitcast
1966 instruction</a>.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1967
1968 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
1969
1970 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
1971 constants. As with the <a href="#i_getelementptr">getelementptr</a>
1972 instruction, the index list may have zero or more indexes, which are required
1973 to make sense for the type of "CSTPTR".</dd>
1974
1975 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
1976
1977 <dd>Perform the <a href="#i_select">select operation</a> on
1978 constants.</dd>
1979
1980 <dt><b><tt>icmp COND ( VAL1, VAL2 )</tt></b></dt>
1981 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
1982
1983 <dt><b><tt>fcmp COND ( VAL1, VAL2 )</tt></b></dt>
1984 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
1985
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]1986 <dt><b><tt>vicmp COND ( VAL1, VAL2 )</tt></b></dt>
1987 <dd>Performs the <a href="#i_vicmp">vicmp operation</a> on constants.</dd>
1988
1989 <dt><b><tt>vfcmp COND ( VAL1, VAL2 )</tt></b></dt>
1990 <dd>Performs the <a href="#i_vfcmp">vfcmp operation</a> on constants.</dd>
1991
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1992 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
1993
1994 <dd>Perform the <a href="#i_extractelement">extractelement
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]1995 operation</a> on constants.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1996
1997 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
1998
1999 <dd>Perform the <a href="#i_insertelement">insertelement
2000 operation</a> on constants.</dd>
2001
2002
2003 <dt><b><tt>shufflevector ( VEC1, VEC2, IDXMASK )</tt></b></dt>
2004
2005 <dd>Perform the <a href="#i_shufflevector">shufflevector
2006 operation</a> on constants.</dd>
2007
2008 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
2009
2010 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
2011 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
2012 binary</a> operations. The constraints on operands are the same as those for
2013 the corresponding instruction (e.g. no bitwise operations on floating point
2014 values are allowed).</dd>
2015</dl>
2016</div>
2017
2018<!-- *********************************************************************** -->
2019<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
2020<!-- *********************************************************************** -->
2021
2022<!-- ======================================================================= -->
2023<div class="doc_subsection">
2024<a name="inlineasm">Inline Assembler Expressions</a>
2025</div>
2026
2027<div class="doc_text">
2028
2029<p>
2030LLVM supports inline assembler expressions (as opposed to <a href="#moduleasm">
2031Module-Level Inline Assembly</a>) through the use of a special value. This
2032value represents the inline assembler as a string (containing the instructions
2033to emit), a list of operand constraints (stored as a string), and a flag that
2034indicates whether or not the inline asm expression has side effects. An example
2035inline assembler expression is:
2036</p>
2037
2038<div class="doc_code">
2039<pre>
2040i32 (i32) asm "bswap $0", "=r,r"
2041</pre>
2042</div>
2043
2044<p>
2045Inline assembler expressions may <b>only</b> be used as the callee operand of
2046a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we have:
2047</p>
2048
2049<div class="doc_code">
2050<pre>
2051%X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
2052</pre>
2053</div>
2054
2055<p>
2056Inline asms with side effects not visible in the constraint list must be marked
2057as having side effects. This is done through the use of the
2058'<tt>sideeffect</tt>' keyword, like so:
2059</p>
2060
2061<div class="doc_code">
2062<pre>
2063call void asm sideeffect "eieio", ""()
2064</pre>
2065</div>
2066
2067<p>TODO: The format of the asm and constraints string still need to be
2068documented here. Constraints on what can be done (e.g. duplication, moving, etc
Chris Lattner1cdd64e2008-10-04 18:36:02 +0000[diff] [blame]2069need to be documented). This is probably best done by reference to another
2070document that covers inline asm from a holistic perspective.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2071</p>
2072
2073</div>
2074
2075<!-- *********************************************************************** -->
2076<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
2077<!-- *********************************************************************** -->
2078
2079<div class="doc_text">
2080
2081<p>The LLVM instruction set consists of several different
2082classifications of instructions: <a href="#terminators">terminator
2083instructions</a>, <a href="#binaryops">binary instructions</a>,
2084<a href="#bitwiseops">bitwise binary instructions</a>, <a
2085 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
2086instructions</a>.</p>
2087
2088</div>
2089
2090<!-- ======================================================================= -->
2091<div class="doc_subsection"> <a name="terminators">Terminator
2092Instructions</a> </div>
2093
2094<div class="doc_text">
2095
2096<p>As mentioned <a href="#functionstructure">previously</a>, every
2097basic block in a program ends with a "Terminator" instruction, which
2098indicates which block should be executed after the current block is
2099finished. These terminator instructions typically yield a '<tt>void</tt>'
2100value: they produce control flow, not values (the one exception being
2101the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
2102<p>There are six different terminator instructions: the '<a
2103 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
2104instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
2105the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
2106 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
2107 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
2108
2109</div>
2110
2111<!-- _______________________________________________________________________ -->
2112<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
2113Instruction</a> </div>
2114<div class="doc_text">
2115<h5>Syntax:</h5>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2116<pre>
2117 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]2118 ret void <i>; Return from void function</i>
2119</pre>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2120
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2121<h5>Overview:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2122
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2123<p>The '<tt>ret</tt>' instruction is used to return control flow (and
2124optionally a value) from a function back to the caller.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2125<p>There are two forms of the '<tt>ret</tt>' instruction: one that
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2126returns a value and then causes control flow, and one that just causes
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2127control flow to occur.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2128
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2129<h5>Arguments:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2130
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2131<p>The '<tt>ret</tt>' instruction optionally accepts a single argument,
2132the return value. The type of the return value must be a
2133'<a href="#t_firstclass">first class</a>' type.</p>
2134
2135<p>A function is not <a href="#wellformed">well formed</a> if
2136it it has a non-void return type and contains a '<tt>ret</tt>'
2137instruction with no return value or a return value with a type that
2138does not match its type, or if it has a void return type and contains
2139a '<tt>ret</tt>' instruction with a return value.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2140
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2141<h5>Semantics:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2142
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2143<p>When the '<tt>ret</tt>' instruction is executed, control flow
2144returns back to the calling function's context. If the caller is a "<a
2145 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
2146the instruction after the call. If the caller was an "<a
2147 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
2148at the beginning of the "normal" destination block. If the instruction
2149returns a value, that value shall set the call or invoke instruction's
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]2150return value.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2151
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2152<h5>Example:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2153
2154<pre>
2155 ret i32 5 <i>; Return an integer value of 5</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2156 ret void <i>; Return from a void function</i>
Bill Wendlingd163e2d2009-02-28 22:12:54 +0000[diff] [blame]2157 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]2158</pre>
Dan Gohman60967192009-01-12 23:12:39 +0000[diff] [blame]2159
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]2160<p>Note that the code generator does not yet fully support large
2161 return values. The specific sizes that are currently supported are
2162 dependent on the target. For integers, on 32-bit targets the limit
2163 is often 64 bits, and on 64-bit targets the limit is often 128 bits.
2164 For aggregate types, the current limits are dependent on the element
2165 types; for example targets are often limited to 2 total integer
2166 elements and 2 total floating-point elements.</p>
Dan Gohman60967192009-01-12 23:12:39 +0000[diff] [blame]2167
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2168</div>
2169<!-- _______________________________________________________________________ -->
2170<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
2171<div class="doc_text">
2172<h5>Syntax:</h5>
2173<pre> br i1 &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;<br> br label &lt;dest&gt; <i>; Unconditional branch</i>
2174</pre>
2175<h5>Overview:</h5>
2176<p>The '<tt>br</tt>' instruction is used to cause control flow to
2177transfer to a different basic block in the current function. There are
2178two forms of this instruction, corresponding to a conditional branch
2179and an unconditional branch.</p>
2180<h5>Arguments:</h5>
2181<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
2182single '<tt>i1</tt>' value and two '<tt>label</tt>' values. The
2183unconditional form of the '<tt>br</tt>' instruction takes a single
2184'<tt>label</tt>' value as a target.</p>
2185<h5>Semantics:</h5>
2186<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
2187argument is evaluated. If the value is <tt>true</tt>, control flows
2188to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
2189control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
2190<h5>Example:</h5>
2191<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
2192 href="#i_ret">ret</a> i32 1<br>IfUnequal:<br> <a href="#i_ret">ret</a> i32 0<br></pre>
2193</div>
2194<!-- _______________________________________________________________________ -->
2195<div class="doc_subsubsection">
2196 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
2197</div>
2198
2199<div class="doc_text">
2200<h5>Syntax:</h5>
2201
2202<pre>
2203 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
2204</pre>
2205
2206<h5>Overview:</h5>
2207
2208<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
2209several different places. It is a generalization of the '<tt>br</tt>'
2210instruction, allowing a branch to occur to one of many possible
2211destinations.</p>
2212
2213
2214<h5>Arguments:</h5>
2215
2216<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
2217comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
2218an array of pairs of comparison value constants and '<tt>label</tt>'s. The
2219table is not allowed to contain duplicate constant entries.</p>
2220
2221<h5>Semantics:</h5>
2222
2223<p>The <tt>switch</tt> instruction specifies a table of values and
2224destinations. When the '<tt>switch</tt>' instruction is executed, this
2225table is searched for the given value. If the value is found, control flow is
2226transfered to the corresponding destination; otherwise, control flow is
2227transfered to the default destination.</p>
2228
2229<h5>Implementation:</h5>
2230
2231<p>Depending on properties of the target machine and the particular
2232<tt>switch</tt> instruction, this instruction may be code generated in different
2233ways. For example, it could be generated as a series of chained conditional
2234branches or with a lookup table.</p>
2235
2236<h5>Example:</h5>
2237
2238<pre>
2239 <i>; Emulate a conditional br instruction</i>
2240 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]2241 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2242
2243 <i>; Emulate an unconditional br instruction</i>
2244 switch i32 0, label %dest [ ]
2245
2246 <i>; Implement a jump table:</i>
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]2247 switch i32 %val, label %otherwise [ i32 0, label %onzero
2248 i32 1, label %onone
2249 i32 2, label %ontwo ]
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2250</pre>
2251</div>
2252
2253<!-- _______________________________________________________________________ -->
2254<div class="doc_subsubsection">
2255 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
2256</div>
2257
2258<div class="doc_text">
2259
2260<h5>Syntax:</h5>
2261
2262<pre>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]2263 &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]2264 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
2265</pre>
2266
2267<h5>Overview:</h5>
2268
2269<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
2270function, with the possibility of control flow transfer to either the
2271'<tt>normal</tt>' label or the
2272'<tt>exception</tt>' label. If the callee function returns with the
2273"<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
2274"normal" label. If the callee (or any indirect callees) returns with the "<a
2275href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]2276continued at the dynamically nearest "exception" label.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2277
2278<h5>Arguments:</h5>
2279
2280<p>This instruction requires several arguments:</p>
2281
2282<ol>
2283 <li>
2284 The optional "cconv" marker indicates which <a href="#callingconv">calling
2285 convention</a> the call should use. If none is specified, the call defaults
2286 to using C calling conventions.
2287 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]2288
2289 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
2290 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>',
2291 and '<tt>inreg</tt>' attributes are valid here.</li>
2292
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2293 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
2294 function value being invoked. In most cases, this is a direct function
2295 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
2296 an arbitrary pointer to function value.
2297 </li>
2298
2299 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
2300 function to be invoked. </li>
2301
2302 <li>'<tt>function args</tt>': argument list whose types match the function
2303 signature argument types. If the function signature indicates the function
2304 accepts a variable number of arguments, the extra arguments can be
2305 specified. </li>
2306
2307 <li>'<tt>normal label</tt>': the label reached when the called function
2308 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
2309
2310 <li>'<tt>exception label</tt>': the label reached when a callee returns with
2311 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
2312
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]2313 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]2314 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
2315 '<tt>readnone</tt>' attributes are valid here.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2316</ol>
2317
2318<h5>Semantics:</h5>
2319
2320<p>This instruction is designed to operate as a standard '<tt><a
2321href="#i_call">call</a></tt>' instruction in most regards. The primary
2322difference is that it establishes an association with a label, which is used by
2323the runtime library to unwind the stack.</p>
2324
2325<p>This instruction is used in languages with destructors to ensure that proper
2326cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
2327exception. Additionally, this is important for implementation of
2328'<tt>catch</tt>' clauses in high-level languages that support them.</p>
2329
2330<h5>Example:</h5>
2331<pre>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2332 %retval = invoke i32 @Test(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2333 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2334 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2335 unwind label %TestCleanup <i>; {i32}:retval set</i>
2336</pre>
2337</div>
2338
2339
2340<!-- _______________________________________________________________________ -->
2341
2342<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
2343Instruction</a> </div>
2344
2345<div class="doc_text">
2346
2347<h5>Syntax:</h5>
2348<pre>
2349 unwind
2350</pre>
2351
2352<h5>Overview:</h5>
2353
2354<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
2355at the first callee in the dynamic call stack which used an <a
2356href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
2357primarily used to implement exception handling.</p>
2358
2359<h5>Semantics:</h5>
2360
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]2361<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2362immediately halt. The dynamic call stack is then searched for the first <a
2363href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
2364execution continues at the "exceptional" destination block specified by the
2365<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
2366dynamic call chain, undefined behavior results.</p>
2367</div>
2368
2369<!-- _______________________________________________________________________ -->
2370
2371<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
2372Instruction</a> </div>
2373
2374<div class="doc_text">
2375
2376<h5>Syntax:</h5>
2377<pre>
2378 unreachable
2379</pre>
2380
2381<h5>Overview:</h5>
2382
2383<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
2384instruction is used to inform the optimizer that a particular portion of the
2385code is not reachable. This can be used to indicate that the code after a
2386no-return function cannot be reached, and other facts.</p>
2387
2388<h5>Semantics:</h5>
2389
2390<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
2391</div>
2392
2393
2394
2395<!-- ======================================================================= -->
2396<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
2397<div class="doc_text">
2398<p>Binary operators are used to do most of the computation in a
Chris Lattnerab596d92008-04-01 18:47:32 +0000[diff] [blame]2399program. They require two operands of the same type, execute an operation on them, and
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2400produce a single value. The operands might represent
2401multiple 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]2402The result value has the same type as its operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2403<p>There are several different binary operators:</p>
2404</div>
2405<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2406<div class="doc_subsubsection">
2407 <a name="i_add">'<tt>add</tt>' Instruction</a>
2408</div>
2409
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2410<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2411
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2412<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2413
2414<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2415 &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]2416</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2417
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2418<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2419
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2420<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2421
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2422<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2423
2424<p>The two arguments to the '<tt>add</tt>' instruction must be <a
2425 href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>, or
2426 <a href="#t_vector">vector</a> values. Both arguments must have identical
2427 types.</p>
2428
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2429<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2430
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2431<p>The value produced is the integer or floating point sum of the two
2432operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2433
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2434<p>If an integer sum has unsigned overflow, the result returned is the
2435mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2436the result.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2437
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2438<p>Because LLVM integers use a two's complement representation, this
2439instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2440
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2441<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2442
2443<pre>
2444 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2445</pre>
2446</div>
2447<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2448<div class="doc_subsubsection">
2449 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
2450</div>
2451
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2452<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2453
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2454<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2455
2456<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2457 &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]2458</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2459
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2460<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2461
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2462<p>The '<tt>sub</tt>' instruction returns the difference of its two
2463operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2464
2465<p>Note that the '<tt>sub</tt>' instruction is used to represent the
2466'<tt>neg</tt>' instruction present in most other intermediate
2467representations.</p>
2468
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2469<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2470
2471<p>The two arguments to the '<tt>sub</tt>' instruction must be <a
2472 href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>,
2473 or <a href="#t_vector">vector</a> values. Both arguments must have identical
2474 types.</p>
2475
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2476<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2477
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2478<p>The value produced is the integer or floating point difference of
2479the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2480
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2481<p>If an integer difference has unsigned overflow, the result returned is the
2482mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2483the result.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2484
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2485<p>Because LLVM integers use a two's complement representation, this
2486instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2487
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2488<h5>Example:</h5>
2489<pre>
2490 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
2491 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
2492</pre>
2493</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2494
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2495<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2496<div class="doc_subsubsection">
2497 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
2498</div>
2499
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2500<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2501
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2502<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2503<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]2504</pre>
2505<h5>Overview:</h5>
2506<p>The '<tt>mul</tt>' instruction returns the product of its two
2507operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2508
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2509<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2510
2511<p>The two arguments to the '<tt>mul</tt>' instruction must be <a
2512href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>,
2513or <a href="#t_vector">vector</a> values. Both arguments must have identical
2514types.</p>
2515
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2516<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2517
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2518<p>The value produced is the integer or floating point product of the
2519two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2520
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2521<p>If the result of an integer multiplication has unsigned overflow,
2522the result returned is the mathematical result modulo
25232<sup>n</sup>, where n is the bit width of the result.</p>
2524<p>Because LLVM integers use a two's complement representation, and the
2525result is the same width as the operands, this instruction returns the
2526correct result for both signed and unsigned integers. If a full product
2527(e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands
2528should be sign-extended or zero-extended as appropriate to the
2529width of the full product.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2530<h5>Example:</h5>
2531<pre> &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
2532</pre>
2533</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2534
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2535<!-- _______________________________________________________________________ -->
2536<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
2537</a></div>
2538<div class="doc_text">
2539<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2540<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]2541</pre>
2542<h5>Overview:</h5>
2543<p>The '<tt>udiv</tt>' instruction returns the quotient of its two
2544operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2545
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2546<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2547
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2548<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2549<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2550values. Both arguments must have identical types.</p>
2551
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2552<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2553
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2554<p>The value produced is the unsigned integer quotient of the two operands.</p>
2555<p>Note that unsigned integer division and signed integer division are distinct
2556operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
2557<p>Division by zero leads to undefined behavior.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2558<h5>Example:</h5>
2559<pre> &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2560</pre>
2561</div>
2562<!-- _______________________________________________________________________ -->
2563<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
2564</a> </div>
2565<div class="doc_text">
2566<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2567<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2568 &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]2569</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2570
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2571<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2572
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2573<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two
2574operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2575
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2576<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2577
2578<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
2579<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2580values. Both arguments must have identical types.</p>
2581
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2582<h5>Semantics:</h5>
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2583<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]2584<p>Note that signed integer division and unsigned integer division are distinct
2585operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
2586<p>Division by zero leads to undefined behavior. Overflow also leads to
2587undefined behavior; this is a rare case, but can occur, for example,
2588by doing a 32-bit division of -2147483648 by -1.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2589<h5>Example:</h5>
2590<pre> &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2591</pre>
2592</div>
2593<!-- _______________________________________________________________________ -->
2594<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
2595Instruction</a> </div>
2596<div class="doc_text">
2597<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2598<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2599 &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]2600</pre>
2601<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2602
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2603<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two
2604operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2605
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2606<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2607
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2608<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2609<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2610of floating point values. Both arguments must have identical types.</p>
2611
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2612<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2613
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2614<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2615
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2616<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2617
2618<pre>
2619 &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]2620</pre>
2621</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2622
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2623<!-- _______________________________________________________________________ -->
2624<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
2625</div>
2626<div class="doc_text">
2627<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2628<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]2629</pre>
2630<h5>Overview:</h5>
2631<p>The '<tt>urem</tt>' instruction returns the remainder from the
2632unsigned division of its two arguments.</p>
2633<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2634<p>The two arguments to the '<tt>urem</tt>' instruction must be
2635<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2636values. Both arguments must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2637<h5>Semantics:</h5>
2638<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2639This instruction always performs an unsigned division to get the remainder.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2640<p>Note that unsigned integer remainder and signed integer remainder are
2641distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
2642<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2643<h5>Example:</h5>
2644<pre> &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2645</pre>
2646
2647</div>
2648<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2649<div class="doc_subsubsection">
2650 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
2651</div>
2652
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2653<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2654
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2655<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2656
2657<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2658 &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]2659</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2660
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2661<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2662
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2663<p>The '<tt>srem</tt>' instruction returns the remainder from the
Dan Gohman3e3fd8c2007-11-05 23:35:22 +0000[diff] [blame]2664signed division of its two operands. This instruction can also take
2665<a href="#t_vector">vector</a> versions of the values in which case
2666the elements must be integers.</p>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]2667
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2668<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2669
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2670<p>The two arguments to the '<tt>srem</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2671<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2672values. Both arguments must have identical types.</p>
2673
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2674<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2675
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2676<p>This instruction returns the <i>remainder</i> of a division (where the result
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2677has the same sign as the dividend, <tt>op1</tt>), not the <i>modulo</i>
2678operator (where the result has the same sign as the divisor, <tt>op2</tt>) of
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2679a value. For more information about the difference, see <a
2680 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
2681Math Forum</a>. For a table of how this is implemented in various languages,
2682please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
2683Wikipedia: modulo operation</a>.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2684<p>Note that signed integer remainder and unsigned integer remainder are
2685distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
2686<p>Taking the remainder of a division by zero leads to undefined behavior.
2687Overflow also leads to undefined behavior; this is a rare case, but can occur,
2688for example, by taking the remainder of a 32-bit division of -2147483648 by -1.
2689(The remainder doesn't actually overflow, but this rule lets srem be
2690implemented using instructions that return both the result of the division
2691and the remainder.)</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2692<h5>Example:</h5>
2693<pre> &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2694</pre>
2695
2696</div>
2697<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2698<div class="doc_subsubsection">
2699 <a name="i_frem">'<tt>frem</tt>' Instruction</a> </div>
2700
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2701<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2702
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2703<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2704<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]2705</pre>
2706<h5>Overview:</h5>
2707<p>The '<tt>frem</tt>' instruction returns the remainder from the
2708division of its two operands.</p>
2709<h5>Arguments:</h5>
2710<p>The two arguments to the '<tt>frem</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2711<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2712of floating point values. Both arguments must have identical types.</p>
2713
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2714<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2715
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2716<p>This instruction returns the <i>remainder</i> of a division.
2717The remainder has the same sign as the dividend.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2718
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2719<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2720
2721<pre>
2722 &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]2723</pre>
2724</div>
2725
2726<!-- ======================================================================= -->
2727<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
2728Operations</a> </div>
2729<div class="doc_text">
2730<p>Bitwise binary operators are used to do various forms of
2731bit-twiddling in a program. They are generally very efficient
2732instructions and can commonly be strength reduced from other
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2733instructions. They require two operands of the same type, execute an operation on them,
2734and 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]2735</div>
2736
2737<!-- _______________________________________________________________________ -->
2738<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
2739Instruction</a> </div>
2740<div class="doc_text">
2741<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2742<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]2743</pre>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2744
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2745<h5>Overview:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2746
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2747<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
2748the left a specified number of bits.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2749
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2750<h5>Arguments:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2751
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2752<p>Both arguments to the '<tt>shl</tt>' instruction must be the same <a
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]2753 href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2754type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2755
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2756<h5>Semantics:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2757
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2758<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod 2<sup>n</sup>,
2759where 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]2760equal to or larger than the number of bits in <tt>op1</tt>, the result is undefined.
2761If the arguments are vectors, each vector element of <tt>op1</tt> is shifted by the
2762corresponding shift amount in <tt>op2</tt>.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2763
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2764<h5>Example:</h5><pre>
2765 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
2766 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
2767 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2768 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Mon P Wang9901e732008-12-09 05:46:39 +0000[diff] [blame]2769 &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]2770</pre>
2771</div>
2772<!-- _______________________________________________________________________ -->
2773<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
2774Instruction</a> </div>
2775<div class="doc_text">
2776<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2777<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]2778</pre>
2779
2780<h5>Overview:</h5>
2781<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
2782operand shifted to the right a specified number of bits with zero fill.</p>
2783
2784<h5>Arguments:</h5>
2785<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]2786<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]2787type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2788
2789<h5>Semantics:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2790
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2791<p>This instruction always performs a logical shift right operation. The most
2792significant bits of the result will be filled with zero bits after the
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2793shift. If <tt>op2</tt> is (statically or dynamically) equal to or larger than
Mon P Wangb0f51822008-12-10 08:55:09 +0000[diff] [blame]2794the number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
2795vectors, each vector element of <tt>op1</tt> is shifted by the corresponding shift
2796amount in <tt>op2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2797
2798<h5>Example:</h5>
2799<pre>
2800 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
2801 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
2802 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
2803 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2804 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Mon P Wang9901e732008-12-09 05:46:39 +0000[diff] [blame]2805 &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]2806</pre>
2807</div>
2808
2809<!-- _______________________________________________________________________ -->
2810<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
2811Instruction</a> </div>
2812<div class="doc_text">
2813
2814<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2815<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]2816</pre>
2817
2818<h5>Overview:</h5>
2819<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
2820operand shifted to the right a specified number of bits with sign extension.</p>
2821
2822<h5>Arguments:</h5>
2823<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]2824<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]2825type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2826
2827<h5>Semantics:</h5>
2828<p>This instruction always performs an arithmetic shift right operation,
2829The most significant bits of the result will be filled with the sign bit
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2830of <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]2831larger than the number of bits in <tt>op1</tt>, the result is undefined. If the
2832arguments are vectors, each vector element of <tt>op1</tt> is shifted by the
2833corresponding shift amount in <tt>op2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2834
2835<h5>Example:</h5>
2836<pre>
2837 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
2838 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
2839 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
2840 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2841 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Mon P Wang9901e732008-12-09 05:46:39 +0000[diff] [blame]2842 &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]2843</pre>
2844</div>
2845
2846<!-- _______________________________________________________________________ -->
2847<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
2848Instruction</a> </div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2849
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2850<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2851
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2852<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2853
2854<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2855 &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]2856</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2857
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2858<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2859
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2860<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
2861its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2862
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2863<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2864
2865<p>The two arguments to the '<tt>and</tt>' instruction must be
2866<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2867values. Both arguments must have identical types.</p>
2868
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2869<h5>Semantics:</h5>
2870<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
2871<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]2872<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2873<table border="1" cellspacing="0" cellpadding="4">
2874 <tbody>
2875 <tr>
2876 <td>In0</td>
2877 <td>In1</td>
2878 <td>Out</td>
2879 </tr>
2880 <tr>
2881 <td>0</td>
2882 <td>0</td>
2883 <td>0</td>
2884 </tr>
2885 <tr>
2886 <td>0</td>
2887 <td>1</td>
2888 <td>0</td>
2889 </tr>
2890 <tr>
2891 <td>1</td>
2892 <td>0</td>
2893 <td>0</td>
2894 </tr>
2895 <tr>
2896 <td>1</td>
2897 <td>1</td>
2898 <td>1</td>
2899 </tr>
2900 </tbody>
2901</table>
2902</div>
2903<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2904<pre>
2905 &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]2906 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
2907 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
2908</pre>
2909</div>
2910<!-- _______________________________________________________________________ -->
2911<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
2912<div class="doc_text">
2913<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2914<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]2915</pre>
2916<h5>Overview:</h5>
2917<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
2918or of its two operands.</p>
2919<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2920
2921<p>The two arguments to the '<tt>or</tt>' instruction must be
2922<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2923values. Both arguments must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2924<h5>Semantics:</h5>
2925<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
2926<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]2927<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2928<table border="1" cellspacing="0" cellpadding="4">
2929 <tbody>
2930 <tr>
2931 <td>In0</td>
2932 <td>In1</td>
2933 <td>Out</td>
2934 </tr>
2935 <tr>
2936 <td>0</td>
2937 <td>0</td>
2938 <td>0</td>
2939 </tr>
2940 <tr>
2941 <td>0</td>
2942 <td>1</td>
2943 <td>1</td>
2944 </tr>
2945 <tr>
2946 <td>1</td>
2947 <td>0</td>
2948 <td>1</td>
2949 </tr>
2950 <tr>
2951 <td>1</td>
2952 <td>1</td>
2953 <td>1</td>
2954 </tr>
2955 </tbody>
2956</table>
2957</div>
2958<h5>Example:</h5>
2959<pre> &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
2960 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
2961 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
2962</pre>
2963</div>
2964<!-- _______________________________________________________________________ -->
2965<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
2966Instruction</a> </div>
2967<div class="doc_text">
2968<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2969<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]2970</pre>
2971<h5>Overview:</h5>
2972<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
2973or of its two operands. The <tt>xor</tt> is used to implement the
2974"one's complement" operation, which is the "~" operator in C.</p>
2975<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2976<p>The two arguments to the '<tt>xor</tt>' instruction must be
2977<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2978values. Both arguments must have identical types.</p>
2979
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2980<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2981
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2982<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
2983<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]2984<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2985<table border="1" cellspacing="0" cellpadding="4">
2986 <tbody>
2987 <tr>
2988 <td>In0</td>
2989 <td>In1</td>
2990 <td>Out</td>
2991 </tr>
2992 <tr>
2993 <td>0</td>
2994 <td>0</td>
2995 <td>0</td>
2996 </tr>
2997 <tr>
2998 <td>0</td>
2999 <td>1</td>
3000 <td>1</td>
3001 </tr>
3002 <tr>
3003 <td>1</td>
3004 <td>0</td>
3005 <td>1</td>
3006 </tr>
3007 <tr>
3008 <td>1</td>
3009 <td>1</td>
3010 <td>0</td>
3011 </tr>
3012 </tbody>
3013</table>
3014</div>
3015<p> </p>
3016<h5>Example:</h5>
3017<pre> &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
3018 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
3019 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
3020 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
3021</pre>
3022</div>
3023
3024<!-- ======================================================================= -->
3025<div class="doc_subsection">
3026 <a name="vectorops">Vector Operations</a>
3027</div>
3028
3029<div class="doc_text">
3030
3031<p>LLVM supports several instructions to represent vector operations in a
3032target-independent manner. These instructions cover the element-access and
3033vector-specific operations needed to process vectors effectively. While LLVM
3034does directly support these vector operations, many sophisticated algorithms
3035will want to use target-specific intrinsics to take full advantage of a specific
3036target.</p>
3037
3038</div>
3039
3040<!-- _______________________________________________________________________ -->
3041<div class="doc_subsubsection">
3042 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
3043</div>
3044
3045<div class="doc_text">
3046
3047<h5>Syntax:</h5>
3048
3049<pre>
3050 &lt;result&gt; = extractelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, i32 &lt;idx&gt; <i>; yields &lt;ty&gt;</i>
3051</pre>
3052
3053<h5>Overview:</h5>
3054
3055<p>
3056The '<tt>extractelement</tt>' instruction extracts a single scalar
3057element from a vector at a specified index.
3058</p>
3059
3060
3061<h5>Arguments:</h5>
3062
3063<p>
3064The first operand of an '<tt>extractelement</tt>' instruction is a
3065value of <a href="#t_vector">vector</a> type. The second operand is
3066an index indicating the position from which to extract the element.
3067The index may be a variable.</p>
3068
3069<h5>Semantics:</h5>
3070
3071<p>
3072The result is a scalar of the same type as the element type of
3073<tt>val</tt>. Its value is the value at position <tt>idx</tt> of
3074<tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
3075results are undefined.
3076</p>
3077
3078<h5>Example:</h5>
3079
3080<pre>
3081 %result = extractelement &lt;4 x i32&gt; %vec, i32 0 <i>; yields i32</i>
3082</pre>
3083</div>
3084
3085
3086<!-- _______________________________________________________________________ -->
3087<div class="doc_subsubsection">
3088 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
3089</div>
3090
3091<div class="doc_text">
3092
3093<h5>Syntax:</h5>
3094
3095<pre>
Dan Gohmanbcc3c502008-05-12 23:38:42 +0000[diff] [blame]3096 &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]3097</pre>
3098
3099<h5>Overview:</h5>
3100
3101<p>
3102The '<tt>insertelement</tt>' instruction inserts a scalar
3103element into a vector at a specified index.
3104</p>
3105
3106
3107<h5>Arguments:</h5>
3108
3109<p>
3110The first operand of an '<tt>insertelement</tt>' instruction is a
3111value of <a href="#t_vector">vector</a> type. The second operand is a
3112scalar value whose type must equal the element type of the first
3113operand. The third operand is an index indicating the position at
3114which to insert the value. The index may be a variable.</p>
3115
3116<h5>Semantics:</h5>
3117
3118<p>
3119The result is a vector of the same type as <tt>val</tt>. Its
3120element values are those of <tt>val</tt> except at position
3121<tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
3122exceeds the length of <tt>val</tt>, the results are undefined.
3123</p>
3124
3125<h5>Example:</h5>
3126
3127<pre>
3128 %result = insertelement &lt;4 x i32&gt; %vec, i32 1, i32 0 <i>; yields &lt;4 x i32&gt;</i>
3129</pre>
3130</div>
3131
3132<!-- _______________________________________________________________________ -->
3133<div class="doc_subsubsection">
3134 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
3135</div>
3136
3137<div class="doc_text">
3138
3139<h5>Syntax:</h5>
3140
3141<pre>
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3142 &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]3143</pre>
3144
3145<h5>Overview:</h5>
3146
3147<p>
3148The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3149from two input vectors, returning a vector with the same element type as
3150the input and length that is the same as the shuffle mask.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3151</p>
3152
3153<h5>Arguments:</h5>
3154
3155<p>
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3156The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
3157with types that match each other. The third argument is a shuffle mask whose
3158element type is always 'i32'. The result of the instruction is a vector whose
3159length is the same as the shuffle mask and whose element type is the same as
3160the element type of the first two operands.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3161</p>
3162
3163<p>
3164The shuffle mask operand is required to be a constant vector with either
3165constant integer or undef values.
3166</p>
3167
3168<h5>Semantics:</h5>
3169
3170<p>
3171The elements of the two input vectors are numbered from left to right across
3172both of the vectors. The shuffle mask operand specifies, for each element of
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3173the result vector, which element of the two input vectors the result element
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3174gets. The element selector may be undef (meaning "don't care") and the second
3175operand may be undef if performing a shuffle from only one vector.
3176</p>
3177
3178<h5>Example:</h5>
3179
3180<pre>
3181 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
3182 &lt;4 x i32&gt; &lt;i32 0, i32 4, i32 1, i32 5&gt; <i>; yields &lt;4 x i32&gt;</i>
3183 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; undef,
3184 &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]3185 %result = shufflevector &lt;8 x i32&gt; %v1, &lt;8 x i32&gt; undef,
3186 &lt;4 x i32&gt; &lt;i32 0, i32 1, i32 2, i32 3&gt; <i>; yields &lt;4 x i32&gt;</i>
3187 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
3188 &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]3189</pre>
3190</div>
3191
3192
3193<!-- ======================================================================= -->
3194<div class="doc_subsection">
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3195 <a name="aggregateops">Aggregate Operations</a>
3196</div>
3197
3198<div class="doc_text">
3199
3200<p>LLVM supports several instructions for working with aggregate values.
3201</p>
3202
3203</div>
3204
3205<!-- _______________________________________________________________________ -->
3206<div class="doc_subsubsection">
3207 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
3208</div>
3209
3210<div class="doc_text">
3211
3212<h5>Syntax:</h5>
3213
3214<pre>
3215 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
3216</pre>
3217
3218<h5>Overview:</h5>
3219
3220<p>
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3221The '<tt>extractvalue</tt>' instruction extracts the value of a struct field
3222or array element from an aggregate value.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3223</p>
3224
3225
3226<h5>Arguments:</h5>
3227
3228<p>
3229The first operand of an '<tt>extractvalue</tt>' instruction is a
3230value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a>
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3231type. The operands are constant indices to specify which value to extract
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3232in a similar manner as indices in a
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3233'<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
3234</p>
3235
3236<h5>Semantics:</h5>
3237
3238<p>
3239The result is the value at the position in the aggregate specified by
3240the index operands.
3241</p>
3242
3243<h5>Example:</h5>
3244
3245<pre>
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3246 %result = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3247</pre>
3248</div>
3249
3250
3251<!-- _______________________________________________________________________ -->
3252<div class="doc_subsubsection">
3253 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
3254</div>
3255
3256<div class="doc_text">
3257
3258<h5>Syntax:</h5>
3259
3260<pre>
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3261 &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]3262</pre>
3263
3264<h5>Overview:</h5>
3265
3266<p>
3267The '<tt>insertvalue</tt>' instruction inserts a value
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3268into a struct field or array element in an aggregate.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3269</p>
3270
3271
3272<h5>Arguments:</h5>
3273
3274<p>
3275The first operand of an '<tt>insertvalue</tt>' instruction is a
3276value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a> type.
3277The second operand is a first-class value to insert.
Dan Gohman4f29e422008-05-23 21:53:15 +0000[diff] [blame]3278The following operands are constant indices
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3279indicating the position at which to insert the value in a similar manner as
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3280indices in a
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3281'<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
3282The value to insert must have the same type as the value identified
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3283by the indices.
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3284</p>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3285
3286<h5>Semantics:</h5>
3287
3288<p>
3289The result is an aggregate of the same type as <tt>val</tt>. Its
3290value is that of <tt>val</tt> except that the value at the position
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3291specified by the indices is that of <tt>elt</tt>.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3292</p>
3293
3294<h5>Example:</h5>
3295
3296<pre>
Dan Gohmanb1aab4e2008-06-23 15:26:37 +0000[diff] [blame]3297 %result = insertvalue {i32, float} %agg, i32 1, 0 <i>; yields {i32, float}</i>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3298</pre>
3299</div>
3300
3301
3302<!-- ======================================================================= -->
3303<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3304 <a name="memoryops">Memory Access and Addressing Operations</a>
3305</div>
3306
3307<div class="doc_text">
3308
3309<p>A key design point of an SSA-based representation is how it
3310represents memory. In LLVM, no memory locations are in SSA form, which
3311makes things very simple. This section describes how to read, write,
3312allocate, and free memory in LLVM.</p>
3313
3314</div>
3315
3316<!-- _______________________________________________________________________ -->
3317<div class="doc_subsubsection">
3318 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
3319</div>
3320
3321<div class="doc_text">
3322
3323<h5>Syntax:</h5>
3324
3325<pre>
3326 &lt;result&gt; = malloc &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
3327</pre>
3328
3329<h5>Overview:</h5>
3330
3331<p>The '<tt>malloc</tt>' instruction allocates memory from the system
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]3332heap and returns a pointer to it. The object is always allocated in the generic
3333address space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3334
3335<h5>Arguments:</h5>
3336
3337<p>The '<tt>malloc</tt>' instruction allocates
3338<tt>sizeof(&lt;type&gt;)*NumElements</tt>
3339bytes of memory from the operating system and returns a pointer of the
3340appropriate type to the program. If "NumElements" is specified, it is the
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3341number of elements allocated, otherwise "NumElements" is defaulted to be one.
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3342If a constant alignment is specified, the value result of the allocation is guaranteed to
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3343be aligned to at least that boundary. If not specified, or if zero, the target can
3344choose to align the allocation on any convenient boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3345
3346<p>'<tt>type</tt>' must be a sized type.</p>
3347
3348<h5>Semantics:</h5>
3349
3350<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
Nick Lewyckyff384472008-11-24 03:41:24 +0000[diff] [blame]3351a pointer is returned. The result of a zero byte allocation is undefined. The
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]3352result is null if there is insufficient memory available.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3353
3354<h5>Example:</h5>
3355
3356<pre>
Dan Gohmanf54f50a2009-01-04 23:49:44 +0000[diff] [blame]3357 %array = malloc [4 x i8] <i>; yields {[%4 x i8]*}:array</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3358
3359 %size = <a href="#i_add">add</a> i32 2, 2 <i>; yields {i32}:size = i32 4</i>
3360 %array1 = malloc i8, i32 4 <i>; yields {i8*}:array1</i>
3361 %array2 = malloc [12 x i8], i32 %size <i>; yields {[12 x i8]*}:array2</i>
3362 %array3 = malloc i32, i32 4, align 1024 <i>; yields {i32*}:array3</i>
3363 %array4 = malloc i32, align 1024 <i>; yields {i32*}:array4</i>
3364</pre>
Dan Gohman60967192009-01-12 23:12:39 +0000[diff] [blame]3365
3366<p>Note that the code generator does not yet respect the
3367 alignment value.</p>
3368
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3369</div>
3370
3371<!-- _______________________________________________________________________ -->
3372<div class="doc_subsubsection">
3373 <a name="i_free">'<tt>free</tt>' Instruction</a>
3374</div>
3375
3376<div class="doc_text">
3377
3378<h5>Syntax:</h5>
3379
3380<pre>
Dan Gohmanf54f50a2009-01-04 23:49:44 +0000[diff] [blame]3381 free &lt;type&gt; &lt;value&gt; <i>; yields {void}</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3382</pre>
3383
3384<h5>Overview:</h5>
3385
3386<p>The '<tt>free</tt>' instruction returns memory back to the unused
3387memory heap to be reallocated in the future.</p>
3388
3389<h5>Arguments:</h5>
3390
3391<p>'<tt>value</tt>' shall be a pointer value that points to a value
3392that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
3393instruction.</p>
3394
3395<h5>Semantics:</h5>
3396
3397<p>Access to the memory pointed to by the pointer is no longer defined
Chris Lattner329d6532008-04-19 22:41:32 +0000[diff] [blame]3398after this instruction executes. If the pointer is null, the operation
3399is a noop.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3400
3401<h5>Example:</h5>
3402
3403<pre>
Dan Gohmanf54f50a2009-01-04 23:49:44 +0000[diff] [blame]3404 %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]3405 free [4 x i8]* %array
3406</pre>
3407</div>
3408
3409<!-- _______________________________________________________________________ -->
3410<div class="doc_subsubsection">
3411 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
3412</div>
3413
3414<div class="doc_text">
3415
3416<h5>Syntax:</h5>
3417
3418<pre>
3419 &lt;result&gt; = alloca &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
3420</pre>
3421
3422<h5>Overview:</h5>
3423
3424<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
3425currently executing function, to be automatically released when this function
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]3426returns to its caller. The object is always allocated in the generic address
3427space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3428
3429<h5>Arguments:</h5>
3430
3431<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
3432bytes of memory on the runtime stack, returning a pointer of the
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3433appropriate type to the program. If "NumElements" is specified, it is the
3434number of elements allocated, otherwise "NumElements" is defaulted to be one.
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3435If a constant alignment is specified, the value result of the allocation is guaranteed
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3436to be aligned to at least that boundary. If not specified, or if zero, the target
3437can choose to align the allocation on any convenient boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3438
3439<p>'<tt>type</tt>' may be any sized type.</p>
3440
3441<h5>Semantics:</h5>
3442
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]3443<p>Memory is allocated; a pointer is returned. The operation is undefiend if
3444there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3445memory is automatically released when the function returns. The '<tt>alloca</tt>'
3446instruction is commonly used to represent automatic variables that must
3447have an address available. When the function returns (either with the <tt><a
3448 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]3449instructions), the memory is reclaimed. Allocating zero bytes
3450is legal, but the result is undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3451
3452<h5>Example:</h5>
3453
3454<pre>
Dan Gohmanf54f50a2009-01-04 23:49:44 +0000[diff] [blame]3455 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
3456 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
3457 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
3458 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3459</pre>
3460</div>
3461
3462<!-- _______________________________________________________________________ -->
3463<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
3464Instruction</a> </div>
3465<div class="doc_text">
3466<h5>Syntax:</h5>
3467<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>
3468<h5>Overview:</h5>
3469<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
3470<h5>Arguments:</h5>
3471<p>The argument to the '<tt>load</tt>' instruction specifies the memory
3472address from which to load. The pointer must point to a <a
3473 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
3474marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
3475the number or order of execution of this <tt>load</tt> with other
3476volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
3477instructions. </p>
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3478<p>
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3479The optional constant "align" argument specifies the alignment of the operation
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3480(that is, the alignment of the memory address). A value of 0 or an
3481omitted "align" argument means that the operation has the preferential
3482alignment for the target. It is the responsibility of the code emitter
3483to ensure that the alignment information is correct. Overestimating
3484the alignment results in an undefined behavior. Underestimating the
3485alignment may produce less efficient code. An alignment of 1 is always
3486safe.
3487</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3488<h5>Semantics:</h5>
Duncan Sandsfe87fff2009-03-22 11:33:16 +0000[diff] [blame^]3489<p>The location of memory pointed to is loaded. If the value being loaded
3490is of scalar type then the number of bytes read does not exceed the minimum
3491number of bytes needed to hold all bits of the type. For example, loading an
3492<tt>i24</tt> reads at most three bytes. When loading a value of a type like
3493<tt>i20</tt> with a size that is not an integral number of bytes, the result
3494is undefined if the value was not originally written using a store of the
3495same type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3496<h5>Examples:</h5>
3497<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
3498 <a
3499 href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
3500 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
3501</pre>
3502</div>
3503<!-- _______________________________________________________________________ -->
3504<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
3505Instruction</a> </div>
3506<div class="doc_text">
3507<h5>Syntax:</h5>
3508<pre> store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
3509 volatile store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
3510</pre>
3511<h5>Overview:</h5>
3512<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
3513<h5>Arguments:</h5>
3514<p>There are two arguments to the '<tt>store</tt>' instruction: a value
3515to 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]3516operand must be a pointer to the <a href="#t_firstclass">first class</a> type
3517of the '<tt>&lt;value&gt;</tt>'
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3518operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
3519optimizer is not allowed to modify the number or order of execution of
3520this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
3521 href="#i_store">store</a></tt> instructions.</p>
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3522<p>
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3523The optional constant "align" argument specifies the alignment of the operation
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3524(that is, the alignment of the memory address). A value of 0 or an
3525omitted "align" argument means that the operation has the preferential
3526alignment for the target. It is the responsibility of the code emitter
3527to ensure that the alignment information is correct. Overestimating
3528the alignment results in an undefined behavior. Underestimating the
3529alignment may produce less efficient code. An alignment of 1 is always
3530safe.
3531</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3532<h5>Semantics:</h5>
3533<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>'
Duncan Sandsfe87fff2009-03-22 11:33:16 +0000[diff] [blame^]3534at the location specified by the '<tt>&lt;pointer&gt;</tt>' operand.
3535If '<tt>&lt;value&gt;</tt>' is of scalar type then the number of bytes
3536written does not exceed the minimum number of bytes needed to hold all
3537bits of the type. For example, storing an <tt>i24</tt> writes at most
3538three bytes. When writing a value of a type like <tt>i20</tt> with a
3539size that is not an integral number of bytes, it is unspecified what
3540happens to the extra bits that do not belong to the type, but they will
3541typically be overwritten.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3542<h5>Example:</h5>
3543<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling63ffa142007-10-22 05:10:05 +0000[diff] [blame]3544 store i32 3, i32* %ptr <i>; yields {void}</i>
3545 %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]3546</pre>
3547</div>
3548
3549<!-- _______________________________________________________________________ -->
3550<div class="doc_subsubsection">
3551 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
3552</div>
3553
3554<div class="doc_text">
3555<h5>Syntax:</h5>
3556<pre>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3557 &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]3558</pre>
3559
3560<h5>Overview:</h5>
3561
3562<p>
3563The '<tt>getelementptr</tt>' instruction is used to get the address of a
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3564subelement of an aggregate data structure. It performs address calculation only
3565and does not access memory.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3566
3567<h5>Arguments:</h5>
3568
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3569<p>The first argument is always a pointer, and forms the basis of the
3570calculation. The remaining arguments are indices, that indicate which of the
3571elements of the aggregate object are indexed. The interpretation of each index
3572is dependent on the type being indexed into. The first index always indexes the
3573pointer value given as the first argument, the second index indexes a value of
3574the type pointed to (not necessarily the value directly pointed to, since the
3575first index can be non-zero), etc. The first type indexed into must be a pointer
3576value, subsequent types can be arrays, vectors and structs. Note that subsequent
3577types being indexed into can never be pointers, since that would require loading
3578the pointer before continuing calculation.</p>
3579
3580<p>The type of each index argument depends on the type it is indexing into.
3581When indexing into a (packed) structure, only <tt>i32</tt> integer
3582<b>constants</b> are allowed. When indexing into an array, pointer or vector,
3583only integers of 32 or 64 bits are allowed (also non-constants). 32-bit values
3584will be sign extended to 64-bits if required.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3585
3586<p>For example, let's consider a C code fragment and how it gets
3587compiled to LLVM:</p>
3588
3589<div class="doc_code">
3590<pre>
3591struct RT {
3592 char A;
3593 int B[10][20];
3594 char C;
3595};
3596struct ST {
3597 int X;
3598 double Y;
3599 struct RT Z;
3600};
3601
3602int *foo(struct ST *s) {
3603 return &amp;s[1].Z.B[5][13];
3604}
3605</pre>
3606</div>
3607
3608<p>The LLVM code generated by the GCC frontend is:</p>
3609
3610<div class="doc_code">
3611<pre>
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]3612%RT = <a href="#namedtypes">type</a> { i8 , [10 x [20 x i32]], i8 }
3613%ST = <a href="#namedtypes">type</a> { i32, double, %RT }
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3614
3615define i32* %foo(%ST* %s) {
3616entry:
3617 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
3618 ret i32* %reg
3619}
3620</pre>
3621</div>
3622
3623<h5>Semantics:</h5>
3624
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3625<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
3626type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
3627}</tt>' type, a structure. The second index indexes into the third element of
3628the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
3629i8 }</tt>' type, another structure. The third index indexes into the second
3630element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
3631array. The two dimensions of the array are subscripted into, yielding an
3632'<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
3633to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
3634
3635<p>Note that it is perfectly legal to index partially through a
3636structure, returning a pointer to an inner element. Because of this,
3637the LLVM code for the given testcase is equivalent to:</p>
3638
3639<pre>
3640 define i32* %foo(%ST* %s) {
3641 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
3642 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
3643 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
3644 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
3645 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
3646 ret i32* %t5
3647 }
3648</pre>
3649
Chris Lattner50609942009-03-09 20:55:18 +0000[diff] [blame]3650<p>Note that it is undefined to access an array out of bounds: array
3651and pointer indexes must always be within the defined bounds of the
3652array type when accessed with an instruction that dereferences the
3653pointer (e.g. a load or store instruction). The one exception for
3654this rule is zero length arrays. These arrays are defined to be
3655accessible as variable length arrays, which requires access beyond the
3656zero'th element.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3657
3658<p>The getelementptr instruction is often confusing. For some more insight
3659into how it works, see <a href="GetElementPtr.html">the getelementptr
3660FAQ</a>.</p>
3661
3662<h5>Example:</h5>
3663
3664<pre>
3665 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3666 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
3667 <i>; yields i8*:vptr</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3668 %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]3669 <i>; yields i8*:eptr</i>
3670 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3671</pre>
3672</div>
3673
3674<!-- ======================================================================= -->
3675<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
3676</div>
3677<div class="doc_text">
3678<p>The instructions in this category are the conversion instructions (casting)
3679which all take a single operand and a type. They perform various bit conversions
3680on the operand.</p>
3681</div>
3682
3683<!-- _______________________________________________________________________ -->
3684<div class="doc_subsubsection">
3685 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
3686</div>
3687<div class="doc_text">
3688
3689<h5>Syntax:</h5>
3690<pre>
3691 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3692</pre>
3693
3694<h5>Overview:</h5>
3695<p>
3696The '<tt>trunc</tt>' instruction truncates its operand to the type <tt>ty2</tt>.
3697</p>
3698
3699<h5>Arguments:</h5>
3700<p>
3701The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
3702be an <a href="#t_integer">integer</a> type, and a type that specifies the size
3703and type of the result, which must be an <a href="#t_integer">integer</a>
3704type. The bit size of <tt>value</tt> must be larger than the bit size of
3705<tt>ty2</tt>. Equal sized types are not allowed.</p>
3706
3707<h5>Semantics:</h5>
3708<p>
3709The '<tt>trunc</tt>' instruction truncates the high order bits in <tt>value</tt>
3710and converts the remaining bits to <tt>ty2</tt>. Since the source size must be
3711larger than the destination size, <tt>trunc</tt> cannot be a <i>no-op cast</i>.
3712It will always truncate bits.</p>
3713
3714<h5>Example:</h5>
3715<pre>
3716 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
3717 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
3718 %Y = trunc i32 122 to i1 <i>; yields i1:false</i>
3719</pre>
3720</div>
3721
3722<!-- _______________________________________________________________________ -->
3723<div class="doc_subsubsection">
3724 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
3725</div>
3726<div class="doc_text">
3727
3728<h5>Syntax:</h5>
3729<pre>
3730 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3731</pre>
3732
3733<h5>Overview:</h5>
3734<p>The '<tt>zext</tt>' instruction zero extends its operand to type
3735<tt>ty2</tt>.</p>
3736
3737
3738<h5>Arguments:</h5>
3739<p>The '<tt>zext</tt>' instruction takes a value to cast, which must be of
3740<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3741also be of <a href="#t_integer">integer</a> type. The bit size of the
3742<tt>value</tt> must be smaller than the bit size of the destination type,
3743<tt>ty2</tt>.</p>
3744
3745<h5>Semantics:</h5>
3746<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
3747bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
3748
3749<p>When zero extending from i1, the result will always be either 0 or 1.</p>
3750
3751<h5>Example:</h5>
3752<pre>
3753 %X = zext i32 257 to i64 <i>; yields i64:257</i>
3754 %Y = zext i1 true to i32 <i>; yields i32:1</i>
3755</pre>
3756</div>
3757
3758<!-- _______________________________________________________________________ -->
3759<div class="doc_subsubsection">
3760 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
3761</div>
3762<div class="doc_text">
3763
3764<h5>Syntax:</h5>
3765<pre>
3766 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3767</pre>
3768
3769<h5>Overview:</h5>
3770<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
3771
3772<h5>Arguments:</h5>
3773<p>
3774The '<tt>sext</tt>' instruction takes a value to cast, which must be of
3775<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3776also be of <a href="#t_integer">integer</a> type. The bit size of the
3777<tt>value</tt> must be smaller than the bit size of the destination type,
3778<tt>ty2</tt>.</p>
3779
3780<h5>Semantics:</h5>
3781<p>
3782The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
3783bit (highest order bit) of the <tt>value</tt> until it reaches the bit size of
3784the type <tt>ty2</tt>.</p>
3785
3786<p>When sign extending from i1, the extension always results in -1 or 0.</p>
3787
3788<h5>Example:</h5>
3789<pre>
3790 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
3791 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
3792</pre>
3793</div>
3794
3795<!-- _______________________________________________________________________ -->
3796<div class="doc_subsubsection">
3797 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
3798</div>
3799
3800<div class="doc_text">
3801
3802<h5>Syntax:</h5>
3803
3804<pre>
3805 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3806</pre>
3807
3808<h5>Overview:</h5>
3809<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
3810<tt>ty2</tt>.</p>
3811
3812
3813<h5>Arguments:</h5>
3814<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
3815 point</a> value to cast and a <a href="#t_floating">floating point</a> type to
3816cast it to. The size of <tt>value</tt> must be larger than the size of
3817<tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
3818<i>no-op cast</i>.</p>
3819
3820<h5>Semantics:</h5>
3821<p> The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
3822<a href="#t_floating">floating point</a> type to a smaller
3823<a href="#t_floating">floating point</a> type. If the value cannot fit within
3824the destination type, <tt>ty2</tt>, then the results are undefined.</p>
3825
3826<h5>Example:</h5>
3827<pre>
3828 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
3829 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
3830</pre>
3831</div>
3832
3833<!-- _______________________________________________________________________ -->
3834<div class="doc_subsubsection">
3835 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
3836</div>
3837<div class="doc_text">
3838
3839<h5>Syntax:</h5>
3840<pre>
3841 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3842</pre>
3843
3844<h5>Overview:</h5>
3845<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
3846floating point value.</p>
3847
3848<h5>Arguments:</h5>
3849<p>The '<tt>fpext</tt>' instruction takes a
3850<a href="#t_floating">floating point</a> <tt>value</tt> to cast,
3851and a <a href="#t_floating">floating point</a> type to cast it to. The source
3852type must be smaller than the destination type.</p>
3853
3854<h5>Semantics:</h5>
3855<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
3856<a href="#t_floating">floating point</a> type to a larger
3857<a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
3858used to make a <i>no-op cast</i> because it always changes bits. Use
3859<tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
3860
3861<h5>Example:</h5>
3862<pre>
3863 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
3864 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
3865</pre>
3866</div>
3867
3868<!-- _______________________________________________________________________ -->
3869<div class="doc_subsubsection">
3870 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
3871</div>
3872<div class="doc_text">
3873
3874<h5>Syntax:</h5>
3875<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3876 &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]3877</pre>
3878
3879<h5>Overview:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3880<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3881unsigned integer equivalent of type <tt>ty2</tt>.
3882</p>
3883
3884<h5>Arguments:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3885<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]3886scalar or vector <a href="#t_floating">floating point</a> value, and a type
3887to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
3888type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
3889vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3890
3891<h5>Semantics:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3892<p> The '<tt>fptoui</tt>' instruction converts its
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3893<a href="#t_floating">floating point</a> operand into the nearest (rounding
3894towards zero) unsigned integer value. If the value cannot fit in <tt>ty2</tt>,
3895the results are undefined.</p>
3896
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3897<h5>Example:</h5>
3898<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3899 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]3900 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3901 %X = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3902</pre>
3903</div>
3904
3905<!-- _______________________________________________________________________ -->
3906<div class="doc_subsubsection">
3907 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
3908</div>
3909<div class="doc_text">
3910
3911<h5>Syntax:</h5>
3912<pre>
3913 &lt;result&gt; = fptosi &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3914</pre>
3915
3916<h5>Overview:</h5>
3917<p>The '<tt>fptosi</tt>' instruction converts
3918<a href="#t_floating">floating point</a> <tt>value</tt> to type <tt>ty2</tt>.
3919</p>
3920
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3921<h5>Arguments:</h5>
3922<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]3923scalar or vector <a href="#t_floating">floating point</a> value, and a type
3924to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
3925type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
3926vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3927
3928<h5>Semantics:</h5>
3929<p>The '<tt>fptosi</tt>' instruction converts its
3930<a href="#t_floating">floating point</a> operand into the nearest (rounding
3931towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
3932the results are undefined.</p>
3933
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3934<h5>Example:</h5>
3935<pre>
3936 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]3937 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3938 %X = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
3939</pre>
3940</div>
3941
3942<!-- _______________________________________________________________________ -->
3943<div class="doc_subsubsection">
3944 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
3945</div>
3946<div class="doc_text">
3947
3948<h5>Syntax:</h5>
3949<pre>
3950 &lt;result&gt; = uitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3951</pre>
3952
3953<h5>Overview:</h5>
3954<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
3955integer and converts that value to the <tt>ty2</tt> type.</p>
3956
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3957<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]3958<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be a
3959scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
3960to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
3961type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
3962floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3963
3964<h5>Semantics:</h5>
3965<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
3966integer quantity and converts it to the corresponding floating point value. If
3967the value cannot fit in the floating point value, the results are undefined.</p>
3968
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3969<h5>Example:</h5>
3970<pre>
3971 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3972 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3973</pre>
3974</div>
3975
3976<!-- _______________________________________________________________________ -->
3977<div class="doc_subsubsection">
3978 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
3979</div>
3980<div class="doc_text">
3981
3982<h5>Syntax:</h5>
3983<pre>
3984 &lt;result&gt; = sitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3985</pre>
3986
3987<h5>Overview:</h5>
3988<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed
3989integer and converts that value to the <tt>ty2</tt> type.</p>
3990
3991<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]3992<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be a
3993scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
3994to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
3995type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
3996floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3997
3998<h5>Semantics:</h5>
3999<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed
4000integer quantity and converts it to the corresponding floating point value. If
4001the value cannot fit in the floating point value, the results are undefined.</p>
4002
4003<h5>Example:</h5>
4004<pre>
4005 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4006 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4007</pre>
4008</div>
4009
4010<!-- _______________________________________________________________________ -->
4011<div class="doc_subsubsection">
4012 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
4013</div>
4014<div class="doc_text">
4015
4016<h5>Syntax:</h5>
4017<pre>
4018 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4019</pre>
4020
4021<h5>Overview:</h5>
4022<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
4023the integer type <tt>ty2</tt>.</p>
4024
4025<h5>Arguments:</h5>
4026<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
4027must 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]4028<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]4029
4030<h5>Semantics:</h5>
4031<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
4032<tt>ty2</tt> by interpreting the pointer value as an integer and either
4033truncating or zero extending that value to the size of the integer type. If
4034<tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
4035<tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
4036are the same size, then nothing is done (<i>no-op cast</i>) other than a type
4037change.</p>
4038
4039<h5>Example:</h5>
4040<pre>
4041 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
4042 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
4043</pre>
4044</div>
4045
4046<!-- _______________________________________________________________________ -->
4047<div class="doc_subsubsection">
4048 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
4049</div>
4050<div class="doc_text">
4051
4052<h5>Syntax:</h5>
4053<pre>
4054 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4055</pre>
4056
4057<h5>Overview:</h5>
4058<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to
4059a pointer type, <tt>ty2</tt>.</p>
4060
4061<h5>Arguments:</h5>
4062<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
4063value to cast, and a type to cast it to, which must be a
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4064<a href="#t_pointer">pointer</a> type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4065
4066<h5>Semantics:</h5>
4067<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
4068<tt>ty2</tt> by applying either a zero extension or a truncation depending on
4069the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
4070size of a pointer then a truncation is done. If <tt>value</tt> is smaller than
4071the size of a pointer then a zero extension is done. If they are the same size,
4072nothing is done (<i>no-op cast</i>).</p>
4073
4074<h5>Example:</h5>
4075<pre>
4076 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
4077 %X = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
4078 %Y = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
4079</pre>
4080</div>
4081
4082<!-- _______________________________________________________________________ -->
4083<div class="doc_subsubsection">
4084 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
4085</div>
4086<div class="doc_text">
4087
4088<h5>Syntax:</h5>
4089<pre>
4090 &lt;result&gt; = bitcast &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4091</pre>
4092
4093<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4094
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4095<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
4096<tt>ty2</tt> without changing any bits.</p>
4097
4098<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4099
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4100<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be
Dan Gohman7305fa02008-09-08 16:45:59 +0000[diff] [blame]4101a non-aggregate first class value, and a type to cast it to, which must also be
4102a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes of
4103<tt>value</tt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4104and the destination type, <tt>ty2</tt>, must be identical. If the source
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4105type is a pointer, the destination type must also be a pointer. This
4106instruction supports bitwise conversion of vectors to integers and to vectors
4107of other types (as long as they have the same size).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4108
4109<h5>Semantics:</h5>
4110<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
4111<tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
4112this conversion. The conversion is done as if the <tt>value</tt> had been
4113stored to memory and read back as type <tt>ty2</tt>. Pointer types may only be
4114converted to other pointer types with this instruction. To convert pointers to
4115other types, use the <a href="#i_inttoptr">inttoptr</a> or
4116<a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
4117
4118<h5>Example:</h5>
4119<pre>
4120 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
4121 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4122 %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]4123</pre>
4124</div>
4125
4126<!-- ======================================================================= -->
4127<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
4128<div class="doc_text">
4129<p>The instructions in this category are the "miscellaneous"
4130instructions, which defy better classification.</p>
4131</div>
4132
4133<!-- _______________________________________________________________________ -->
4134<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
4135</div>
4136<div class="doc_text">
4137<h5>Syntax:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4138<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]4139</pre>
4140<h5>Overview:</h5>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4141<p>The '<tt>icmp</tt>' instruction returns a boolean value or
4142a vector of boolean values based on comparison
4143of its two integer, integer vector, or pointer operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4144<h5>Arguments:</h5>
4145<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
4146the condition code indicating the kind of comparison to perform. It is not
4147a value, just a keyword. The possible condition code are:
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4148</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4149<ol>
4150 <li><tt>eq</tt>: equal</li>
4151 <li><tt>ne</tt>: not equal </li>
4152 <li><tt>ugt</tt>: unsigned greater than</li>
4153 <li><tt>uge</tt>: unsigned greater or equal</li>
4154 <li><tt>ult</tt>: unsigned less than</li>
4155 <li><tt>ule</tt>: unsigned less or equal</li>
4156 <li><tt>sgt</tt>: signed greater than</li>
4157 <li><tt>sge</tt>: signed greater or equal</li>
4158 <li><tt>slt</tt>: signed less than</li>
4159 <li><tt>sle</tt>: signed less or equal</li>
4160</ol>
4161<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4162<a href="#t_pointer">pointer</a>
4163or integer <a href="#t_vector">vector</a> typed.
4164They must also be identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4165<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4166<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]4167the condition code given as <tt>cond</tt>. The comparison performed always
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4168yields 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]4169</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4170<ol>
4171 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
4172 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
4173 </li>
4174 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4175 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4176 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4177 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4178 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4179 <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]4180 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4181 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4182 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4183 <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]4184 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4185 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4186 <li><tt>sge</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4187 <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]4188 <li><tt>slt</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4189 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4190 <li><tt>sle</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4191 <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]4192</ol>
4193<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
4194values are compared as if they were integers.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4195<p>If the operands are integer vectors, then they are compared
4196element by element. The result is an <tt>i1</tt> vector with
4197the same number of elements as the values being compared.
4198Otherwise, the result is an <tt>i1</tt>.
4199</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4200
4201<h5>Example:</h5>
4202<pre> &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
4203 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
4204 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
4205 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
4206 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
4207 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
4208</pre>
Dan Gohmana53eb382009-01-22 01:39:38 +0000[diff] [blame]4209
4210<p>Note that the code generator does not yet support vector types with
4211 the <tt>icmp</tt> instruction.</p>
4212
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4213</div>
4214
4215<!-- _______________________________________________________________________ -->
4216<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
4217</div>
4218<div class="doc_text">
4219<h5>Syntax:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4220<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]4221</pre>
4222<h5>Overview:</h5>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4223<p>The '<tt>fcmp</tt>' instruction returns a boolean value
4224or vector of boolean values based on comparison
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4225of its operands.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4226<p>
4227If the operands are floating point scalars, then the result
4228type is a boolean (<a href="#t_primitive"><tt>i1</tt></a>).
4229</p>
4230<p>If the operands are floating point vectors, then the result type
4231is a vector of boolean with the same number of elements as the
4232operands being compared.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4233<h5>Arguments:</h5>
4234<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
4235the condition code indicating the kind of comparison to perform. It is not
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4236a value, just a keyword. The possible condition code are:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4237<ol>
4238 <li><tt>false</tt>: no comparison, always returns false</li>
4239 <li><tt>oeq</tt>: ordered and equal</li>
4240 <li><tt>ogt</tt>: ordered and greater than </li>
4241 <li><tt>oge</tt>: ordered and greater than or equal</li>
4242 <li><tt>olt</tt>: ordered and less than </li>
4243 <li><tt>ole</tt>: ordered and less than or equal</li>
4244 <li><tt>one</tt>: ordered and not equal</li>
4245 <li><tt>ord</tt>: ordered (no nans)</li>
4246 <li><tt>ueq</tt>: unordered or equal</li>
4247 <li><tt>ugt</tt>: unordered or greater than </li>
4248 <li><tt>uge</tt>: unordered or greater than or equal</li>
4249 <li><tt>ult</tt>: unordered or less than </li>
4250 <li><tt>ule</tt>: unordered or less than or equal</li>
4251 <li><tt>une</tt>: unordered or not equal</li>
4252 <li><tt>uno</tt>: unordered (either nans)</li>
4253 <li><tt>true</tt>: no comparison, always returns true</li>
4254</ol>
4255<p><i>Ordered</i> means that neither operand is a QNAN while
4256<i>unordered</i> means that either operand may be a QNAN.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4257<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be
4258either a <a href="#t_floating">floating point</a> type
4259or a <a href="#t_vector">vector</a> of floating point type.
4260They must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4261<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4262<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4263according to the condition code given as <tt>cond</tt>.
4264If the operands are vectors, then the vectors are compared
4265element by element.
4266Each comparison performed
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4267always yields an <a href="#t_primitive">i1</a> result, as follows:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4268<ol>
4269 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
4270 <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]4271 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4272 <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]4273 <tt>op1</tt> is greather than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4274 <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]4275 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4276 <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]4277 <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4278 <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]4279 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4280 <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]4281 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4282 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
4283 <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]4284 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4285 <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]4286 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4287 <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]4288 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4289 <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]4290 <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4291 <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]4292 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4293 <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]4294 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4295 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
4296 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
4297</ol>
4298
4299<h5>Example:</h5>
4300<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]4301 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
4302 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
4303 &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]4304</pre>
Dan Gohmana53eb382009-01-22 01:39:38 +0000[diff] [blame]4305
4306<p>Note that the code generator does not yet support vector types with
4307 the <tt>fcmp</tt> instruction.</p>
4308
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4309</div>
4310
4311<!-- _______________________________________________________________________ -->
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4312<div class="doc_subsubsection">
4313 <a name="i_vicmp">'<tt>vicmp</tt>' Instruction</a>
4314</div>
4315<div class="doc_text">
4316<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4317<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]4318</pre>
4319<h5>Overview:</h5>
4320<p>The '<tt>vicmp</tt>' instruction returns an integer vector value based on
4321element-wise comparison of its two integer vector operands.</p>
4322<h5>Arguments:</h5>
4323<p>The '<tt>vicmp</tt>' instruction takes three operands. The first operand is
4324the condition code indicating the kind of comparison to perform. It is not
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4325a value, just a keyword. The possible condition code are:</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4326<ol>
4327 <li><tt>eq</tt>: equal</li>
4328 <li><tt>ne</tt>: not equal </li>
4329 <li><tt>ugt</tt>: unsigned greater than</li>
4330 <li><tt>uge</tt>: unsigned greater or equal</li>
4331 <li><tt>ult</tt>: unsigned less than</li>
4332 <li><tt>ule</tt>: unsigned less or equal</li>
4333 <li><tt>sgt</tt>: signed greater than</li>
4334 <li><tt>sge</tt>: signed greater or equal</li>
4335 <li><tt>slt</tt>: signed less than</li>
4336 <li><tt>sle</tt>: signed less or equal</li>
4337</ol>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4338<p>The remaining two arguments must be <a href="#t_vector">vector</a> or
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4339<a href="#t_integer">integer</a> typed. They must also be identical types.</p>
4340<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4341<p>The '<tt>vicmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4342according to the condition code given as <tt>cond</tt>. The comparison yields a
4343<a href="#t_vector">vector</a> of <a href="#t_integer">integer</a> result, of
4344identical type as the values being compared. The most significant bit in each
4345element is 1 if the element-wise comparison evaluates to true, and is 0
4346otherwise. All other bits of the result are undefined. The condition codes
4347are evaluated identically to the <a href="#i_icmp">'<tt>icmp</tt>'
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4348instruction</a>.</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4349
4350<h5>Example:</h5>
4351<pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4352 &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>
4353 &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]4354</pre>
4355</div>
4356
4357<!-- _______________________________________________________________________ -->
4358<div class="doc_subsubsection">
4359 <a name="i_vfcmp">'<tt>vfcmp</tt>' Instruction</a>
4360</div>
4361<div class="doc_text">
4362<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4363<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]4364<h5>Overview:</h5>
4365<p>The '<tt>vfcmp</tt>' instruction returns an integer vector value based on
4366element-wise comparison of its two floating point vector operands. The output
4367elements have the same width as the input elements.</p>
4368<h5>Arguments:</h5>
4369<p>The '<tt>vfcmp</tt>' instruction takes three operands. The first operand is
4370the condition code indicating the kind of comparison to perform. It is not
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4371a value, just a keyword. The possible condition code are:</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4372<ol>
4373 <li><tt>false</tt>: no comparison, always returns false</li>
4374 <li><tt>oeq</tt>: ordered and equal</li>
4375 <li><tt>ogt</tt>: ordered and greater than </li>
4376 <li><tt>oge</tt>: ordered and greater than or equal</li>
4377 <li><tt>olt</tt>: ordered and less than </li>
4378 <li><tt>ole</tt>: ordered and less than or equal</li>
4379 <li><tt>one</tt>: ordered and not equal</li>
4380 <li><tt>ord</tt>: ordered (no nans)</li>
4381 <li><tt>ueq</tt>: unordered or equal</li>
4382 <li><tt>ugt</tt>: unordered or greater than </li>
4383 <li><tt>uge</tt>: unordered or greater than or equal</li>
4384 <li><tt>ult</tt>: unordered or less than </li>
4385 <li><tt>ule</tt>: unordered or less than or equal</li>
4386 <li><tt>une</tt>: unordered or not equal</li>
4387 <li><tt>uno</tt>: unordered (either nans)</li>
4388 <li><tt>true</tt>: no comparison, always returns true</li>
4389</ol>
4390<p>The remaining two arguments must be <a href="#t_vector">vector</a> of
4391<a href="#t_floating">floating point</a> typed. They must also be identical
4392types.</p>
4393<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4394<p>The '<tt>vfcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4395according to the condition code given as <tt>cond</tt>. The comparison yields a
4396<a href="#t_vector">vector</a> of <a href="#t_integer">integer</a> result, with
4397an identical number of elements as the values being compared, and each element
4398having identical with to the width of the floating point elements. The most
4399significant bit in each element is 1 if the element-wise comparison evaluates to
4400true, and is 0 otherwise. All other bits of the result are undefined. The
4401condition codes are evaluated identically to the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4402<a href="#i_fcmp">'<tt>fcmp</tt>' instruction</a>.</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4403
4404<h5>Example:</h5>
4405<pre>
Chris Lattner1e5c5cd02008-10-13 16:55:18 +0000[diff] [blame]4406 <i>; yields: result=&lt;2 x i32&gt; &lt; i32 0, i32 -1 &gt;</i>
4407 &lt;result&gt; = vfcmp oeq &lt;2 x float&gt; &lt; float 4, float 0 &gt;, &lt; float 5, float 0 &gt;
4408
4409 <i>; yields: result=&lt;2 x i64&gt; &lt; i64 -1, i64 0 &gt;</i>
4410 &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]4411</pre>
4412</div>
4413
4414<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4415<div class="doc_subsubsection">
4416 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
4417</div>
4418
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4419<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4420
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4421<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4422
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4423<pre> &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...<br></pre>
4424<h5>Overview:</h5>
4425<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in
4426the SSA graph representing the function.</p>
4427<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4428
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4429<p>The type of the incoming values is specified with the first type
4430field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
4431as arguments, with one pair for each predecessor basic block of the
4432current block. Only values of <a href="#t_firstclass">first class</a>
4433type may be used as the value arguments to the PHI node. Only labels
4434may be used as the label arguments.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4435
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4436<p>There must be no non-phi instructions between the start of a basic
4437block and the PHI instructions: i.e. PHI instructions must be first in
4438a basic block.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4439
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4440<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4441
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4442<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
4443specified by the pair corresponding to the predecessor basic block that executed
4444just prior to the current block.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4445
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4446<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4447<pre>
4448Loop: ; Infinite loop that counts from 0 on up...
4449 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
4450 %nextindvar = add i32 %indvar, 1
4451 br label %Loop
4452</pre>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4453</div>
4454
4455<!-- _______________________________________________________________________ -->
4456<div class="doc_subsubsection">
4457 <a name="i_select">'<tt>select</tt>' Instruction</a>
4458</div>
4459
4460<div class="doc_text">
4461
4462<h5>Syntax:</h5>
4463
4464<pre>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4465 &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>
4466
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4467 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4468</pre>
4469
4470<h5>Overview:</h5>
4471
4472<p>
4473The '<tt>select</tt>' instruction is used to choose one value based on a
4474condition, without branching.
4475</p>
4476
4477
4478<h5>Arguments:</h5>
4479
4480<p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4481The '<tt>select</tt>' instruction requires an 'i1' value or
4482a vector of 'i1' values indicating the
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4483condition, and two values of the same <a href="#t_firstclass">first class</a>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4484type. If the val1/val2 are vectors and
4485the condition is a scalar, then entire vectors are selected, not
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4486individual elements.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4487</p>
4488
4489<h5>Semantics:</h5>
4490
4491<p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4492If 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]4493value argument; otherwise, it returns the second value argument.
4494</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4495<p>
4496If the condition is a vector of i1, then the value arguments must
4497be vectors of the same size, and the selection is done element
4498by element.
4499</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4500
4501<h5>Example:</h5>
4502
4503<pre>
4504 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
4505</pre>
Dan Gohmana53eb382009-01-22 01:39:38 +0000[diff] [blame]4506
4507<p>Note that the code generator does not yet support conditions
4508 with vector type.</p>
4509
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4510</div>
4511
4512
4513<!-- _______________________________________________________________________ -->
4514<div class="doc_subsubsection">
4515 <a name="i_call">'<tt>call</tt>' Instruction</a>
4516</div>
4517
4518<div class="doc_text">
4519
4520<h5>Syntax:</h5>
4521<pre>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]4522 &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]4523</pre>
4524
4525<h5>Overview:</h5>
4526
4527<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
4528
4529<h5>Arguments:</h5>
4530
4531<p>This instruction requires several arguments:</p>
4532
4533<ol>
4534 <li>
4535 <p>The optional "tail" marker indicates whether the callee function accesses
4536 any allocas or varargs in the caller. If the "tail" marker is present, the
4537 function call is eligible for tail call optimization. Note that calls may
4538 be marked "tail" even if they do not occur before a <a
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4539 href="#i_ret"><tt>ret</tt></a> instruction.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4540 </li>
4541 <li>
4542 <p>The optional "cconv" marker indicates which <a href="#callingconv">calling
4543 convention</a> the call should use. If none is specified, the call defaults
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4544 to using C calling conventions.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4545 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4546
4547 <li>
4548 <p>The optional <a href="#paramattrs">Parameter Attributes</a> list for
4549 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>',
4550 and '<tt>inreg</tt>' attributes are valid here.</p>
4551 </li>
4552
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4553 <li>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]4554 <p>'<tt>ty</tt>': the type of the call instruction itself which is also
4555 the type of the return value. Functions that return no value are marked
4556 <tt><a href="#t_void">void</a></tt>.</p>
4557 </li>
4558 <li>
4559 <p>'<tt>fnty</tt>': shall be the signature of the pointer to function
4560 value being invoked. The argument types must match the types implied by
4561 this signature. This type can be omitted if the function is not varargs
4562 and if the function type does not return a pointer to a function.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4563 </li>
4564 <li>
4565 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
4566 be invoked. In most cases, this is a direct function invocation, but
4567 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
4568 to function value.</p>
4569 </li>
4570 <li>
4571 <p>'<tt>function args</tt>': argument list whose types match the
4572 function signature argument types. All arguments must be of
4573 <a href="#t_firstclass">first class</a> type. If the function signature
4574 indicates the function accepts a variable number of arguments, the extra
4575 arguments can be specified.</p>
4576 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4577 <li>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]4578 <p>The optional <a href="#fnattrs">function attributes</a> list. Only
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4579 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
4580 '<tt>readnone</tt>' attributes are valid here.</p>
4581 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4582</ol>
4583
4584<h5>Semantics:</h5>
4585
4586<p>The '<tt>call</tt>' instruction is used to cause control flow to
4587transfer to a specified function, with its incoming arguments bound to
4588the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
4589instruction in the called function, control flow continues with the
4590instruction after the function call, and the return value of the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4591function is bound to the result argument.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4592
4593<h5>Example:</h5>
4594
4595<pre>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]4596 %retval = call i32 @test(i32 %argc)
Chris Lattner5e893ef2008-03-21 17:24:17 +0000[diff] [blame]4597 call i32 (i8 *, ...)* @printf(i8 * %msg, i32 12, i8 42) <i>; yields i32</i>
4598 %X = tail call i32 @foo() <i>; yields i32</i>
4599 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
4600 call void %foo(i8 97 signext)
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]4601
4602 %struct.A = type { i32, i8 }
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4603 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]4604 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
4605 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattnerac454b32008-10-08 06:26:11 +0000[diff] [blame]4606 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijman2c4e05a2008-10-07 10:03:45 +0000[diff] [blame]4607 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4608</pre>
4609
4610</div>
4611
4612<!-- _______________________________________________________________________ -->
4613<div class="doc_subsubsection">
4614 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
4615</div>
4616
4617<div class="doc_text">
4618
4619<h5>Syntax:</h5>
4620
4621<pre>
4622 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
4623</pre>
4624
4625<h5>Overview:</h5>
4626
4627<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
4628the "variable argument" area of a function call. It is used to implement the
4629<tt>va_arg</tt> macro in C.</p>
4630
4631<h5>Arguments:</h5>
4632
4633<p>This instruction takes a <tt>va_list*</tt> value and the type of
4634the argument. It returns a value of the specified argument type and
4635increments the <tt>va_list</tt> to point to the next argument. The
4636actual type of <tt>va_list</tt> is target specific.</p>
4637
4638<h5>Semantics:</h5>
4639
4640<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
4641type from the specified <tt>va_list</tt> and causes the
4642<tt>va_list</tt> to point to the next argument. For more information,
4643see the variable argument handling <a href="#int_varargs">Intrinsic
4644Functions</a>.</p>
4645
4646<p>It is legal for this instruction to be called in a function which does not
4647take a variable number of arguments, for example, the <tt>vfprintf</tt>
4648function.</p>
4649
4650<p><tt>va_arg</tt> is an LLVM instruction instead of an <a
4651href="#intrinsics">intrinsic function</a> because it takes a type as an
4652argument.</p>
4653
4654<h5>Example:</h5>
4655
4656<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
4657
Dan Gohman60967192009-01-12 23:12:39 +0000[diff] [blame]4658<p>Note that the code generator does not yet fully support va_arg
4659 on many targets. Also, it does not currently support va_arg with
4660 aggregate types on any target.</p>
4661
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4662</div>
4663
4664<!-- *********************************************************************** -->
4665<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
4666<!-- *********************************************************************** -->
4667
4668<div class="doc_text">
4669
4670<p>LLVM supports the notion of an "intrinsic function". These functions have
4671well known names and semantics and are required to follow certain restrictions.
4672Overall, these intrinsics represent an extension mechanism for the LLVM
4673language that does not require changing all of the transformations in LLVM when
4674adding to the language (or the bitcode reader/writer, the parser, etc...).</p>
4675
4676<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
4677prefix is reserved in LLVM for intrinsic names; thus, function names may not
4678begin with this prefix. Intrinsic functions must always be external functions:
4679you cannot define the body of intrinsic functions. Intrinsic functions may
4680only be used in call or invoke instructions: it is illegal to take the address
4681of an intrinsic function. Additionally, because intrinsic functions are part
4682of the LLVM language, it is required if any are added that they be documented
4683here.</p>
4684
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4685<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents
4686a family of functions that perform the same operation but on different data
4687types. Because LLVM can represent over 8 million different integer types,
4688overloading is used commonly to allow an intrinsic function to operate on any
4689integer type. One or more of the argument types or the result type can be
4690overloaded to accept any integer type. Argument types may also be defined as
4691exactly matching a previous argument's type or the result type. This allows an
4692intrinsic function which accepts multiple arguments, but needs all of them to
4693be of the same type, to only be overloaded with respect to a single argument or
4694the result.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4695
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4696<p>Overloaded intrinsics will have the names of its overloaded argument types
4697encoded into its function name, each preceded by a period. Only those types
4698which are overloaded result in a name suffix. Arguments whose type is matched
4699against another type do not. For example, the <tt>llvm.ctpop</tt> function can
4700take an integer of any width and returns an integer of exactly the same integer
4701width. This leads to a family of functions such as
4702<tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29 %val)</tt>.
4703Only one type, the return type, is overloaded, and only one type suffix is
4704required. Because the argument's type is matched against the return type, it
4705does not require its own name suffix.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4706
4707<p>To learn how to add an intrinsic function, please see the
4708<a href="ExtendingLLVM.html">Extending LLVM Guide</a>.
4709</p>
4710
4711</div>
4712
4713<!-- ======================================================================= -->
4714<div class="doc_subsection">
4715 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
4716</div>
4717
4718<div class="doc_text">
4719
4720<p>Variable argument support is defined in LLVM with the <a
4721 href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
4722intrinsic functions. These functions are related to the similarly
4723named macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
4724
4725<p>All of these functions operate on arguments that use a
4726target-specific value type "<tt>va_list</tt>". The LLVM assembly
4727language reference manual does not define what this type is, so all
4728transformations should be prepared to handle these functions regardless of
4729the type used.</p>
4730
4731<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
4732instruction and the variable argument handling intrinsic functions are
4733used.</p>
4734
4735<div class="doc_code">
4736<pre>
4737define i32 @test(i32 %X, ...) {
4738 ; Initialize variable argument processing
4739 %ap = alloca i8*
4740 %ap2 = bitcast i8** %ap to i8*
4741 call void @llvm.va_start(i8* %ap2)
4742
4743 ; Read a single integer argument
4744 %tmp = va_arg i8** %ap, i32
4745
4746 ; Demonstrate usage of llvm.va_copy and llvm.va_end
4747 %aq = alloca i8*
4748 %aq2 = bitcast i8** %aq to i8*
4749 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
4750 call void @llvm.va_end(i8* %aq2)
4751
4752 ; Stop processing of arguments.
4753 call void @llvm.va_end(i8* %ap2)
4754 ret i32 %tmp
4755}
4756
4757declare void @llvm.va_start(i8*)
4758declare void @llvm.va_copy(i8*, i8*)
4759declare void @llvm.va_end(i8*)
4760</pre>
4761</div>
4762
4763</div>
4764
4765<!-- _______________________________________________________________________ -->
4766<div class="doc_subsubsection">
4767 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
4768</div>
4769
4770
4771<div class="doc_text">
4772<h5>Syntax:</h5>
4773<pre> declare void %llvm.va_start(i8* &lt;arglist&gt;)<br></pre>
4774<h5>Overview:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4775<p>The '<tt>llvm.va_start</tt>' intrinsic initializes
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4776<tt>*&lt;arglist&gt;</tt> for subsequent use by <tt><a
4777href="#i_va_arg">va_arg</a></tt>.</p>
4778
4779<h5>Arguments:</h5>
4780
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4781<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]4782
4783<h5>Semantics:</h5>
4784
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4785<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]4786macro available in C. In a target-dependent way, it initializes the
4787<tt>va_list</tt> element to which the argument points, so that the next call to
4788<tt>va_arg</tt> will produce the first variable argument passed to the function.
4789Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
4790last argument of the function as the compiler can figure that out.</p>
4791
4792</div>
4793
4794<!-- _______________________________________________________________________ -->
4795<div class="doc_subsubsection">
4796 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
4797</div>
4798
4799<div class="doc_text">
4800<h5>Syntax:</h5>
4801<pre> declare void @llvm.va_end(i8* &lt;arglist&gt;)<br></pre>
4802<h5>Overview:</h5>
4803
4804<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
4805which has been initialized previously with <tt><a href="#int_va_start">llvm.va_start</a></tt>
4806or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
4807
4808<h5>Arguments:</h5>
4809
4810<p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
4811
4812<h5>Semantics:</h5>
4813
4814<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
4815macro available in C. In a target-dependent way, it destroys the
4816<tt>va_list</tt> element to which the argument points. Calls to <a
4817href="#int_va_start"><tt>llvm.va_start</tt></a> and <a href="#int_va_copy">
4818<tt>llvm.va_copy</tt></a> must be matched exactly with calls to
4819<tt>llvm.va_end</tt>.</p>
4820
4821</div>
4822
4823<!-- _______________________________________________________________________ -->
4824<div class="doc_subsubsection">
4825 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
4826</div>
4827
4828<div class="doc_text">
4829
4830<h5>Syntax:</h5>
4831
4832<pre>
4833 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
4834</pre>
4835
4836<h5>Overview:</h5>
4837
4838<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
4839from the source argument list to the destination argument list.</p>
4840
4841<h5>Arguments:</h5>
4842
4843<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
4844The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
4845
4846
4847<h5>Semantics:</h5>
4848
4849<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
4850macro available in C. In a target-dependent way, it copies the source
4851<tt>va_list</tt> element into the destination <tt>va_list</tt> element. This
4852intrinsic is necessary because the <tt><a href="#int_va_start">
4853llvm.va_start</a></tt> intrinsic may be arbitrarily complex and require, for
4854example, memory allocation.</p>
4855
4856</div>
4857
4858<!-- ======================================================================= -->
4859<div class="doc_subsection">
4860 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
4861</div>
4862
4863<div class="doc_text">
4864
4865<p>
4866LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]4867Collection</a> (GC) requires the implementation and generation of these
4868intrinsics.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4869These intrinsics allow identification of <a href="#int_gcroot">GC roots on the
4870stack</a>, as well as garbage collector implementations that require <a
4871href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a> barriers.
4872Front-ends for type-safe garbage collected languages should generate these
4873intrinsics to make use of the LLVM garbage collectors. For more details, see <a
4874href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
4875</p>
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]4876
4877<p>The garbage collection intrinsics only operate on objects in the generic
4878 address space (address space zero).</p>
4879
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4880</div>
4881
4882<!-- _______________________________________________________________________ -->
4883<div class="doc_subsubsection">
4884 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
4885</div>
4886
4887<div class="doc_text">
4888
4889<h5>Syntax:</h5>
4890
4891<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4892 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4893</pre>
4894
4895<h5>Overview:</h5>
4896
4897<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
4898the code generator, and allows some metadata to be associated with it.</p>
4899
4900<h5>Arguments:</h5>
4901
4902<p>The first argument specifies the address of a stack object that contains the
4903root pointer. The second pointer (which must be either a constant or a global
4904value address) contains the meta-data to be associated with the root.</p>
4905
4906<h5>Semantics:</h5>
4907
Chris Lattnera7d94ba2008-04-24 05:59:56 +0000[diff] [blame]4908<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]4909location. At compile-time, the code generator generates information to allow
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4910the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
4911intrinsic may only be used in a function which <a href="#gc">specifies a GC
4912algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4913
4914</div>
4915
4916
4917<!-- _______________________________________________________________________ -->
4918<div class="doc_subsubsection">
4919 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
4920</div>
4921
4922<div class="doc_text">
4923
4924<h5>Syntax:</h5>
4925
4926<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4927 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4928</pre>
4929
4930<h5>Overview:</h5>
4931
4932<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
4933locations, allowing garbage collector implementations that require read
4934barriers.</p>
4935
4936<h5>Arguments:</h5>
4937
4938<p>The second argument is the address to read from, which should be an address
4939allocated from the garbage collector. The first object is a pointer to the
4940start of the referenced object, if needed by the language runtime (otherwise
4941null).</p>
4942
4943<h5>Semantics:</h5>
4944
4945<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
4946instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4947garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
4948may only be used in a function which <a href="#gc">specifies a GC
4949algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4950
4951</div>
4952
4953
4954<!-- _______________________________________________________________________ -->
4955<div class="doc_subsubsection">
4956 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
4957</div>
4958
4959<div class="doc_text">
4960
4961<h5>Syntax:</h5>
4962
4963<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4964 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4965</pre>
4966
4967<h5>Overview:</h5>
4968
4969<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
4970locations, allowing garbage collector implementations that require write
4971barriers (such as generational or reference counting collectors).</p>
4972
4973<h5>Arguments:</h5>
4974
4975<p>The first argument is the reference to store, the second is the start of the
4976object to store it to, and the third is the address of the field of Obj to
4977store to. If the runtime does not require a pointer to the object, Obj may be
4978null.</p>
4979
4980<h5>Semantics:</h5>
4981
4982<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
4983instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4984garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
4985may only be used in a function which <a href="#gc">specifies a GC
4986algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4987
4988</div>
4989
4990
4991
4992<!-- ======================================================================= -->
4993<div class="doc_subsection">
4994 <a name="int_codegen">Code Generator Intrinsics</a>
4995</div>
4996
4997<div class="doc_text">
4998<p>
4999These intrinsics are provided by LLVM to expose special features that may only
5000be implemented with code generator support.
5001</p>
5002
5003</div>
5004
5005<!-- _______________________________________________________________________ -->
5006<div class="doc_subsubsection">
5007 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
5008</div>
5009
5010<div class="doc_text">
5011
5012<h5>Syntax:</h5>
5013<pre>
5014 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
5015</pre>
5016
5017<h5>Overview:</h5>
5018
5019<p>
5020The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
5021target-specific value indicating the return address of the current function
5022or one of its callers.
5023</p>
5024
5025<h5>Arguments:</h5>
5026
5027<p>
5028The argument to this intrinsic indicates which function to return the address
5029for. Zero indicates the calling function, one indicates its caller, etc. The
5030argument is <b>required</b> to be a constant integer value.
5031</p>
5032
5033<h5>Semantics:</h5>
5034
5035<p>
5036The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
5037the return address of the specified call frame, or zero if it cannot be
5038identified. The value returned by this intrinsic is likely to be incorrect or 0
5039for arguments other than zero, so it should only be used for debugging purposes.
5040</p>
5041
5042<p>
5043Note that calling this intrinsic does not prevent function inlining or other
5044aggressive transformations, so the value returned may not be that of the obvious
5045source-language caller.
5046</p>
5047</div>
5048
5049
5050<!-- _______________________________________________________________________ -->
5051<div class="doc_subsubsection">
5052 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
5053</div>
5054
5055<div class="doc_text">
5056
5057<h5>Syntax:</h5>
5058<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5059 declare i8 *@llvm.frameaddress(i32 &lt;level&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5060</pre>
5061
5062<h5>Overview:</h5>
5063
5064<p>
5065The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
5066target-specific frame pointer value for the specified stack frame.
5067</p>
5068
5069<h5>Arguments:</h5>
5070
5071<p>
5072The argument to this intrinsic indicates which function to return the frame
5073pointer for. Zero indicates the calling function, one indicates its caller,
5074etc. The argument is <b>required</b> to be a constant integer value.
5075</p>
5076
5077<h5>Semantics:</h5>
5078
5079<p>
5080The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
5081the frame address of the specified call frame, or zero if it cannot be
5082identified. The value returned by this intrinsic is likely to be incorrect or 0
5083for arguments other than zero, so it should only be used for debugging purposes.
5084</p>
5085
5086<p>
5087Note that calling this intrinsic does not prevent function inlining or other
5088aggressive transformations, so the value returned may not be that of the obvious
5089source-language caller.
5090</p>
5091</div>
5092
5093<!-- _______________________________________________________________________ -->
5094<div class="doc_subsubsection">
5095 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
5096</div>
5097
5098<div class="doc_text">
5099
5100<h5>Syntax:</h5>
5101<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5102 declare i8 *@llvm.stacksave()
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5103</pre>
5104
5105<h5>Overview:</h5>
5106
5107<p>
5108The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
5109the function stack, for use with <a href="#int_stackrestore">
5110<tt>llvm.stackrestore</tt></a>. This is useful for implementing language
5111features like scoped automatic variable sized arrays in C99.
5112</p>
5113
5114<h5>Semantics:</h5>
5115
5116<p>
5117This intrinsic returns a opaque pointer value that can be passed to <a
5118href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
5119<tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
5120<tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
5121state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
5122practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
5123that were allocated after the <tt>llvm.stacksave</tt> was executed.
5124</p>
5125
5126</div>
5127
5128<!-- _______________________________________________________________________ -->
5129<div class="doc_subsubsection">
5130 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
5131</div>
5132
5133<div class="doc_text">
5134
5135<h5>Syntax:</h5>
5136<pre>
5137 declare void @llvm.stackrestore(i8 * %ptr)
5138</pre>
5139
5140<h5>Overview:</h5>
5141
5142<p>
5143The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
5144the function stack to the state it was in when the corresponding <a
5145href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
5146useful for implementing language features like scoped automatic variable sized
5147arrays in C99.
5148</p>
5149
5150<h5>Semantics:</h5>
5151
5152<p>
5153See the description for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.
5154</p>
5155
5156</div>
5157
5158
5159<!-- _______________________________________________________________________ -->
5160<div class="doc_subsubsection">
5161 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
5162</div>
5163
5164<div class="doc_text">
5165
5166<h5>Syntax:</h5>
5167<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5168 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]5169</pre>
5170
5171<h5>Overview:</h5>
5172
5173
5174<p>
5175The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
5176a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
5177no
5178effect on the behavior of the program but can change its performance
5179characteristics.
5180</p>
5181
5182<h5>Arguments:</h5>
5183
5184<p>
5185<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
5186determining if the fetch should be for a read (0) or write (1), and
5187<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
5188locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
5189<tt>locality</tt> arguments must be constant integers.
5190</p>
5191
5192<h5>Semantics:</h5>
5193
5194<p>
5195This intrinsic does not modify the behavior of the program. In particular,
5196prefetches cannot trap and do not produce a value. On targets that support this
5197intrinsic, the prefetch can provide hints to the processor cache for better
5198performance.
5199</p>
5200
5201</div>
5202
5203<!-- _______________________________________________________________________ -->
5204<div class="doc_subsubsection">
5205 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
5206</div>
5207
5208<div class="doc_text">
5209
5210<h5>Syntax:</h5>
5211<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5212 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5213</pre>
5214
5215<h5>Overview:</h5>
5216
5217
5218<p>
5219The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]5220(PC) in a region of
5221code to simulators and other tools. The method is target specific, but it is
5222expected that the marker will use exported symbols to transmit the PC of the
5223marker.
5224The marker makes no guarantees that it will remain with any specific instruction
5225after optimizations. It is possible that the presence of a marker will inhibit
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5226optimizations. The intended use is to be inserted after optimizations to allow
5227correlations of simulation runs.
5228</p>
5229
5230<h5>Arguments:</h5>
5231
5232<p>
5233<tt>id</tt> is a numerical id identifying the marker.
5234</p>
5235
5236<h5>Semantics:</h5>
5237
5238<p>
5239This intrinsic does not modify the behavior of the program. Backends that do not
5240support this intrinisic may ignore it.
5241</p>
5242
5243</div>
5244
5245<!-- _______________________________________________________________________ -->
5246<div class="doc_subsubsection">
5247 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
5248</div>
5249
5250<div class="doc_text">
5251
5252<h5>Syntax:</h5>
5253<pre>
5254 declare i64 @llvm.readcyclecounter( )
5255</pre>
5256
5257<h5>Overview:</h5>
5258
5259
5260<p>
5261The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
5262counter register (or similar low latency, high accuracy clocks) on those targets
5263that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
5264As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
5265should only be used for small timings.
5266</p>
5267
5268<h5>Semantics:</h5>
5269
5270<p>
5271When directly supported, reading the cycle counter should not modify any memory.
5272Implementations are allowed to either return a application specific value or a
5273system wide value. On backends without support, this is lowered to a constant 0.
5274</p>
5275
5276</div>
5277
5278<!-- ======================================================================= -->
5279<div class="doc_subsection">
5280 <a name="int_libc">Standard C Library Intrinsics</a>
5281</div>
5282
5283<div class="doc_text">
5284<p>
5285LLVM provides intrinsics for a few important standard C library functions.
5286These intrinsics allow source-language front-ends to pass information about the
5287alignment of the pointer arguments to the code generator, providing opportunity
5288for more efficient code generation.
5289</p>
5290
5291</div>
5292
5293<!-- _______________________________________________________________________ -->
5294<div class="doc_subsubsection">
5295 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
5296</div>
5297
5298<div class="doc_text">
5299
5300<h5>Syntax:</h5>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5301<p>This is an overloaded intrinsic. You can use llvm.memcpy on any integer bit
5302width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5303<pre>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5304 declare void @llvm.memcpy.i8(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5305 i8 &lt;len&gt;, i32 &lt;align&gt;)
5306 declare void @llvm.memcpy.i16(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5307 i16 &lt;len&gt;, i32 &lt;align&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5308 declare void @llvm.memcpy.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5309 i32 &lt;len&gt;, i32 &lt;align&gt;)
5310 declare void @llvm.memcpy.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5311 i64 &lt;len&gt;, i32 &lt;align&gt;)
5312</pre>
5313
5314<h5>Overview:</h5>
5315
5316<p>
5317The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
5318location to the destination location.
5319</p>
5320
5321<p>
5322Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
5323intrinsics do not return a value, and takes an extra alignment argument.
5324</p>
5325
5326<h5>Arguments:</h5>
5327
5328<p>
5329The first argument is a pointer to the destination, the second is a pointer to
5330the source. The third argument is an integer argument
5331specifying the number of bytes to copy, and the fourth argument is the alignment
5332of the source and destination locations.
5333</p>
5334
5335<p>
5336If the call to this intrinisic has an alignment value that is not 0 or 1, then
5337the caller guarantees that both the source and destination pointers are aligned
5338to that boundary.
5339</p>
5340
5341<h5>Semantics:</h5>
5342
5343<p>
5344The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
5345location to the destination location, which are not allowed to overlap. It
5346copies "len" bytes of memory over. If the argument is known to be aligned to
5347some boundary, this can be specified as the fourth argument, otherwise it should
5348be set to 0 or 1.
5349</p>
5350</div>
5351
5352
5353<!-- _______________________________________________________________________ -->
5354<div class="doc_subsubsection">
5355 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
5356</div>
5357
5358<div class="doc_text">
5359
5360<h5>Syntax:</h5>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5361<p>This is an overloaded intrinsic. You can use llvm.memmove on any integer bit
5362width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5363<pre>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5364 declare void @llvm.memmove.i8(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5365 i8 &lt;len&gt;, i32 &lt;align&gt;)
5366 declare void @llvm.memmove.i16(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5367 i16 &lt;len&gt;, i32 &lt;align&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5368 declare void @llvm.memmove.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5369 i32 &lt;len&gt;, i32 &lt;align&gt;)
5370 declare void @llvm.memmove.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5371 i64 &lt;len&gt;, i32 &lt;align&gt;)
5372</pre>
5373
5374<h5>Overview:</h5>
5375
5376<p>
5377The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the source
5378location to the destination location. It is similar to the
Chris Lattnerdba16ea2008-01-06 19:51:52 +0000[diff] [blame]5379'<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to overlap.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5380</p>
5381
5382<p>
5383Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
5384intrinsics do not return a value, and takes an extra alignment argument.
5385</p>
5386
5387<h5>Arguments:</h5>
5388
5389<p>
5390The first argument is a pointer to the destination, the second is a pointer to
5391the source. The third argument is an integer argument
5392specifying the number of bytes to copy, and the fourth argument is the alignment
5393of the source and destination locations.
5394</p>
5395
5396<p>
5397If the call to this intrinisic has an alignment value that is not 0 or 1, then
5398the caller guarantees that the source and destination pointers are aligned to
5399that boundary.
5400</p>
5401
5402<h5>Semantics:</h5>
5403
5404<p>
5405The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the source
5406location to the destination location, which may overlap. It
5407copies "len" bytes of memory over. If the argument is known to be aligned to
5408some boundary, this can be specified as the fourth argument, otherwise it should
5409be set to 0 or 1.
5410</p>
5411</div>
5412
5413
5414<!-- _______________________________________________________________________ -->
5415<div class="doc_subsubsection">
5416 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
5417</div>
5418
5419<div class="doc_text">
5420
5421<h5>Syntax:</h5>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5422<p>This is an overloaded intrinsic. You can use llvm.memset on any integer bit
5423width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5424<pre>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5425 declare void @llvm.memset.i8(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5426 i8 &lt;len&gt;, i32 &lt;align&gt;)
5427 declare void @llvm.memset.i16(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5428 i16 &lt;len&gt;, i32 &lt;align&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5429 declare void @llvm.memset.i32(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5430 i32 &lt;len&gt;, i32 &lt;align&gt;)
5431 declare void @llvm.memset.i64(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5432 i64 &lt;len&gt;, i32 &lt;align&gt;)
5433</pre>
5434
5435<h5>Overview:</h5>
5436
5437<p>
5438The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a particular
5439byte value.
5440</p>
5441
5442<p>
5443Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
5444does not return a value, and takes an extra alignment argument.
5445</p>
5446
5447<h5>Arguments:</h5>
5448
5449<p>
5450The first argument is a pointer to the destination to fill, the second is the
5451byte value to fill it with, the third argument is an integer
5452argument specifying the number of bytes to fill, and the fourth argument is the
5453known alignment of destination location.
5454</p>
5455
5456<p>
5457If the call to this intrinisic has an alignment value that is not 0 or 1, then
5458the caller guarantees that the destination pointer is aligned to that boundary.
5459</p>
5460
5461<h5>Semantics:</h5>
5462
5463<p>
5464The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting at
5465the
5466destination location. If the argument is known to be aligned to some boundary,
5467this can be specified as the fourth argument, otherwise it should be set to 0 or
54681.
5469</p>
5470</div>
5471
5472
5473<!-- _______________________________________________________________________ -->
5474<div class="doc_subsubsection">
5475 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
5476</div>
5477
5478<div class="doc_text">
5479
5480<h5>Syntax:</h5>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5481<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]5482floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5483types however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5484<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5485 declare float @llvm.sqrt.f32(float %Val)
5486 declare double @llvm.sqrt.f64(double %Val)
5487 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
5488 declare fp128 @llvm.sqrt.f128(fp128 %Val)
5489 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5490</pre>
5491
5492<h5>Overview:</h5>
5493
5494<p>
5495The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5496returning the same value as the libm '<tt>sqrt</tt>' functions would. Unlike
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5497<tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
Chris Lattnerf914a002008-01-29 07:00:44 +0000[diff] [blame]5498negative numbers other than -0.0 (which allows for better optimization, because
5499there is no need to worry about errno being set). <tt>llvm.sqrt(-0.0)</tt> is
5500defined to return -0.0 like IEEE sqrt.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5501</p>
5502
5503<h5>Arguments:</h5>
5504
5505<p>
5506The argument and return value are floating point numbers of the same type.
5507</p>
5508
5509<h5>Semantics:</h5>
5510
5511<p>
5512This function returns the sqrt of the specified operand if it is a nonnegative
5513floating point number.
5514</p>
5515</div>
5516
5517<!-- _______________________________________________________________________ -->
5518<div class="doc_subsubsection">
5519 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
5520</div>
5521
5522<div class="doc_text">
5523
5524<h5>Syntax:</h5>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5525<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]5526floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5527types however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5528<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5529 declare float @llvm.powi.f32(float %Val, i32 %power)
5530 declare double @llvm.powi.f64(double %Val, i32 %power)
5531 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
5532 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
5533 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5534</pre>
5535
5536<h5>Overview:</h5>
5537
5538<p>
5539The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
5540specified (positive or negative) power. The order of evaluation of
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5541multiplications is not defined. When a vector of floating point type is
5542used, the second argument remains a scalar integer value.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5543</p>
5544
5545<h5>Arguments:</h5>
5546
5547<p>
5548The second argument is an integer power, and the first is a value to raise to
5549that power.
5550</p>
5551
5552<h5>Semantics:</h5>
5553
5554<p>
5555This function returns the first value raised to the second power with an
5556unspecified sequence of rounding operations.</p>
5557</div>
5558
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5559<!-- _______________________________________________________________________ -->
5560<div class="doc_subsubsection">
5561 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
5562</div>
5563
5564<div class="doc_text">
5565
5566<h5>Syntax:</h5>
5567<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
5568floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5569types however.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5570<pre>
5571 declare float @llvm.sin.f32(float %Val)
5572 declare double @llvm.sin.f64(double %Val)
5573 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
5574 declare fp128 @llvm.sin.f128(fp128 %Val)
5575 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
5576</pre>
5577
5578<h5>Overview:</h5>
5579
5580<p>
5581The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.
5582</p>
5583
5584<h5>Arguments:</h5>
5585
5586<p>
5587The argument and return value are floating point numbers of the same type.
5588</p>
5589
5590<h5>Semantics:</h5>
5591
5592<p>
5593This function returns the sine of the specified operand, returning the
5594same values as the libm <tt>sin</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5595conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5596</div>
5597
5598<!-- _______________________________________________________________________ -->
5599<div class="doc_subsubsection">
5600 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
5601</div>
5602
5603<div class="doc_text">
5604
5605<h5>Syntax:</h5>
5606<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
5607floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5608types however.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5609<pre>
5610 declare float @llvm.cos.f32(float %Val)
5611 declare double @llvm.cos.f64(double %Val)
5612 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
5613 declare fp128 @llvm.cos.f128(fp128 %Val)
5614 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
5615</pre>
5616
5617<h5>Overview:</h5>
5618
5619<p>
5620The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.
5621</p>
5622
5623<h5>Arguments:</h5>
5624
5625<p>
5626The argument and return value are floating point numbers of the same type.
5627</p>
5628
5629<h5>Semantics:</h5>
5630
5631<p>
5632This function returns the cosine of the specified operand, returning the
5633same values as the libm <tt>cos</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5634conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5635</div>
5636
5637<!-- _______________________________________________________________________ -->
5638<div class="doc_subsubsection">
5639 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
5640</div>
5641
5642<div class="doc_text">
5643
5644<h5>Syntax:</h5>
5645<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
5646floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5647types however.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5648<pre>
5649 declare float @llvm.pow.f32(float %Val, float %Power)
5650 declare double @llvm.pow.f64(double %Val, double %Power)
5651 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
5652 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
5653 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
5654</pre>
5655
5656<h5>Overview:</h5>
5657
5658<p>
5659The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
5660specified (positive or negative) power.
5661</p>
5662
5663<h5>Arguments:</h5>
5664
5665<p>
5666The second argument is a floating point power, and the first is a value to
5667raise to that power.
5668</p>
5669
5670<h5>Semantics:</h5>
5671
5672<p>
5673This function returns the first value raised to the second power,
5674returning the
5675same values as the libm <tt>pow</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5676conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5677</div>
5678
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5679
5680<!-- ======================================================================= -->
5681<div class="doc_subsection">
5682 <a name="int_manip">Bit Manipulation Intrinsics</a>
5683</div>
5684
5685<div class="doc_text">
5686<p>
5687LLVM provides intrinsics for a few important bit manipulation operations.
5688These allow efficient code generation for some algorithms.
5689</p>
5690
5691</div>
5692
5693<!-- _______________________________________________________________________ -->
5694<div class="doc_subsubsection">
5695 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
5696</div>
5697
5698<div class="doc_text">
5699
5700<h5>Syntax:</h5>
5701<p>This is an overloaded intrinsic function. You can use bswap on any integer
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5702type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5703<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5704 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
5705 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
5706 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5707</pre>
5708
5709<h5>Overview:</h5>
5710
5711<p>
5712The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
5713values with an even number of bytes (positive multiple of 16 bits). These are
5714useful for performing operations on data that is not in the target's native
5715byte order.
5716</p>
5717
5718<h5>Semantics:</h5>
5719
5720<p>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5721The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5722and low byte of the input i16 swapped. Similarly, the <tt>llvm.bswap.i32</tt>
5723intrinsic returns an i32 value that has the four bytes of the input i32
5724swapped, 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]5725i32 will have its bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i48</tt>,
5726<tt>llvm.bswap.i64</tt> and other intrinsics extend this concept to
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5727additional even-byte lengths (6 bytes, 8 bytes and more, respectively).
5728</p>
5729
5730</div>
5731
5732<!-- _______________________________________________________________________ -->
5733<div class="doc_subsubsection">
5734 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
5735</div>
5736
5737<div class="doc_text">
5738
5739<h5>Syntax:</h5>
5740<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]5741width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5742<pre>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5743 declare i8 @llvm.ctpop.i8(i8 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5744 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5745 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5746 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
5747 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5748</pre>
5749
5750<h5>Overview:</h5>
5751
5752<p>
5753The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
5754value.
5755</p>
5756
5757<h5>Arguments:</h5>
5758
5759<p>
5760The only argument is the value to be counted. The argument may be of any
5761integer type. The return type must match the argument type.
5762</p>
5763
5764<h5>Semantics:</h5>
5765
5766<p>
5767The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
5768</p>
5769</div>
5770
5771<!-- _______________________________________________________________________ -->
5772<div class="doc_subsubsection">
5773 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
5774</div>
5775
5776<div class="doc_text">
5777
5778<h5>Syntax:</h5>
5779<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]5780integer bit width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5781<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5782 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
5783 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5784 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5785 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
5786 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5787</pre>
5788
5789<h5>Overview:</h5>
5790
5791<p>
5792The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
5793leading zeros in a variable.
5794</p>
5795
5796<h5>Arguments:</h5>
5797
5798<p>
5799The only argument is the value to be counted. The argument may be of any
5800integer type. The return type must match the argument type.
5801</p>
5802
5803<h5>Semantics:</h5>
5804
5805<p>
5806The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
5807in a variable. If the src == 0 then the result is the size in bits of the type
5808of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.
5809</p>
5810</div>
5811
5812
5813
5814<!-- _______________________________________________________________________ -->
5815<div class="doc_subsubsection">
5816 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
5817</div>
5818
5819<div class="doc_text">
5820
5821<h5>Syntax:</h5>
5822<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]5823integer bit width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5824<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5825 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
5826 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5827 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5828 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
5829 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5830</pre>
5831
5832<h5>Overview:</h5>
5833
5834<p>
5835The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
5836trailing zeros.
5837</p>
5838
5839<h5>Arguments:</h5>
5840
5841<p>
5842The only argument is the value to be counted. The argument may be of any
5843integer type. The return type must match the argument type.
5844</p>
5845
5846<h5>Semantics:</h5>
5847
5848<p>
5849The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
5850in a variable. If the src == 0 then the result is the size in bits of the type
5851of src. For example, <tt>llvm.cttz(2) = 1</tt>.
5852</p>
5853</div>
5854
5855<!-- _______________________________________________________________________ -->
5856<div class="doc_subsubsection">
5857 <a name="int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic</a>
5858</div>
5859
5860<div class="doc_text">
5861
5862<h5>Syntax:</h5>
5863<p>This is an overloaded intrinsic. You can use <tt>llvm.part.select</tt>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5864on any integer bit width.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5865<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5866 declare i17 @llvm.part.select.i17 (i17 %val, i32 %loBit, i32 %hiBit)
5867 declare i29 @llvm.part.select.i29 (i29 %val, i32 %loBit, i32 %hiBit)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5868</pre>
5869
5870<h5>Overview:</h5>
5871<p>The '<tt>llvm.part.select</tt>' family of intrinsic functions selects a
5872range of bits from an integer value and returns them in the same bit width as
5873the original value.</p>
5874
5875<h5>Arguments:</h5>
5876<p>The first argument, <tt>%val</tt> and the result may be integer types of
5877any bit width but they must have the same bit width. The second and third
5878arguments must be <tt>i32</tt> type since they specify only a bit index.</p>
5879
5880<h5>Semantics:</h5>
5881<p>The operation of the '<tt>llvm.part.select</tt>' intrinsic has two modes
5882of operation: forwards and reverse. If <tt>%loBit</tt> is greater than
5883<tt>%hiBits</tt> then the intrinsic operates in reverse mode. Otherwise it
5884operates in forward mode.</p>
5885<p>In forward mode, this intrinsic is the equivalent of shifting <tt>%val</tt>
5886right by <tt>%loBit</tt> bits and then ANDing it with a mask with
5887only the <tt>%hiBit - %loBit</tt> bits set, as follows:</p>
5888<ol>
5889 <li>The <tt>%val</tt> is shifted right (LSHR) by the number of bits specified
5890 by <tt>%loBits</tt>. This normalizes the value to the low order bits.</li>
5891 <li>The <tt>%loBits</tt> value is subtracted from the <tt>%hiBits</tt> value
5892 to determine the number of bits to retain.</li>
5893 <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]5894 <li>The mask is ANDed with <tt>%val</tt> to produce the result.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5895</ol>
5896<p>In reverse mode, a similar computation is made except that the bits are
5897returned in the reverse order. So, for example, if <tt>X</tt> has the value
5898<tt>i16 0x0ACF (101011001111)</tt> and we apply
5899<tt>part.select(i16 X, 8, 3)</tt> to it, we get back the value
5900<tt>i16 0x0026 (000000100110)</tt>.</p>
5901</div>
5902
5903<div class="doc_subsubsection">
5904 <a name="int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic</a>
5905</div>
5906
5907<div class="doc_text">
5908
5909<h5>Syntax:</h5>
5910<p>This is an overloaded intrinsic. You can use <tt>llvm.part.set</tt>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5911on any integer bit width.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5912<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5913 declare i17 @llvm.part.set.i17.i9 (i17 %val, i9 %repl, i32 %lo, i32 %hi)
5914 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]5915</pre>
5916
5917<h5>Overview:</h5>
5918<p>The '<tt>llvm.part.set</tt>' family of intrinsic functions replaces a range
5919of bits in an integer value with another integer value. It returns the integer
5920with the replaced bits.</p>
5921
5922<h5>Arguments:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5923<p>The first argument, <tt>%val</tt>, and the result may be integer types of
5924any 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]5925whose bits will be replaced. The second argument, <tt>%repl</tt> may be an
5926integer of any bit width. The third and fourth arguments must be <tt>i32</tt>
5927type since they specify only a bit index.</p>
5928
5929<h5>Semantics:</h5>
5930<p>The operation of the '<tt>llvm.part.set</tt>' intrinsic has two modes
5931of operation: forwards and reverse. If <tt>%lo</tt> is greater than
5932<tt>%hi</tt> then the intrinsic operates in reverse mode. Otherwise it
5933operates in forward mode.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5934
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5935<p>For both modes, the <tt>%repl</tt> value is prepared for use by either
5936truncating it down to the size of the replacement area or zero extending it
5937up to that size.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5938
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5939<p>In forward mode, the bits between <tt>%lo</tt> and <tt>%hi</tt> (inclusive)
5940are replaced with corresponding bits from <tt>%repl</tt>. That is the 0th bit
5941in <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]5942to the <tt>%hi</tt>th bit.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5943
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5944<p>In reverse mode, a similar computation is made except that the bits are
5945reversed. That is, the <tt>0</tt>th bit in <tt>%repl</tt> replaces the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5946<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]5947
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5948<h5>Examples:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5949
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5950<pre>
5951 llvm.part.set(0xFFFF, 0, 4, 7) -&gt; 0xFF0F
5952 llvm.part.set(0xFFFF, 0, 7, 4) -&gt; 0xFF0F
5953 llvm.part.set(0xFFFF, 1, 7, 4) -&gt; 0xFF8F
5954 llvm.part.set(0xFFFF, F, 8, 3) -&gt; 0xFFE7
5955 llvm.part.set(0xFFFF, 0, 3, 8) -&gt; 0xFE07
5956</pre>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5957
5958</div>
5959
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]5960<!-- ======================================================================= -->
5961<div class="doc_subsection">
5962 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
5963</div>
5964
5965<div class="doc_text">
5966<p>
5967LLVM provides intrinsics for some arithmetic with overflow operations.
5968</p>
5969
5970</div>
5971
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5972<!-- _______________________________________________________________________ -->
5973<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]5974 <a name="int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5975</div>
5976
5977<div class="doc_text">
5978
5979<h5>Syntax:</h5>
5980
5981<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]5982on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5983
5984<pre>
5985 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
5986 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
5987 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
5988</pre>
5989
5990<h5>Overview:</h5>
5991
5992<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
5993a signed addition of the two arguments, and indicate whether an overflow
5994occurred during the signed summation.</p>
5995
5996<h5>Arguments:</h5>
5997
5998<p>The arguments (%a and %b) and the first element of the result structure may
5999be of integer types of any bit width, but they must have the same bit width. The
6000second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6001and <tt>%b</tt> are the two values that will undergo signed addition.</p>
6002
6003<h5>Semantics:</h5>
6004
6005<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
6006a signed addition of the two variables. They return a structure &mdash; the
6007first element of which is the signed summation, and the second element of which
6008is a bit specifying if the signed summation resulted in an overflow.</p>
6009
6010<h5>Examples:</h5>
6011<pre>
6012 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6013 %sum = extractvalue {i32, i1} %res, 0
6014 %obit = extractvalue {i32, i1} %res, 1
6015 br i1 %obit, label %overflow, label %normal
6016</pre>
6017
6018</div>
6019
6020<!-- _______________________________________________________________________ -->
6021<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6022 <a name="int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6023</div>
6024
6025<div class="doc_text">
6026
6027<h5>Syntax:</h5>
6028
6029<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]6030on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6031
6032<pre>
6033 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
6034 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6035 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
6036</pre>
6037
6038<h5>Overview:</h5>
6039
6040<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
6041an unsigned addition of the two arguments, and indicate whether a carry occurred
6042during the unsigned summation.</p>
6043
6044<h5>Arguments:</h5>
6045
6046<p>The arguments (%a and %b) and the first element of the result structure may
6047be of integer types of any bit width, but they must have the same bit width. The
6048second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6049and <tt>%b</tt> are the two values that will undergo unsigned addition.</p>
6050
6051<h5>Semantics:</h5>
6052
6053<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
6054an unsigned addition of the two arguments. They return a structure &mdash; the
6055first element of which is the sum, and the second element of which is a bit
6056specifying if the unsigned summation resulted in a carry.</p>
6057
6058<h5>Examples:</h5>
6059<pre>
6060 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6061 %sum = extractvalue {i32, i1} %res, 0
6062 %obit = extractvalue {i32, i1} %res, 1
6063 br i1 %obit, label %carry, label %normal
6064</pre>
6065
6066</div>
6067
6068<!-- _______________________________________________________________________ -->
6069<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6070 <a name="int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6071</div>
6072
6073<div class="doc_text">
6074
6075<h5>Syntax:</h5>
6076
6077<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]6078on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6079
6080<pre>
6081 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
6082 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6083 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
6084</pre>
6085
6086<h5>Overview:</h5>
6087
6088<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
6089a signed subtraction of the two arguments, and indicate whether an overflow
6090occurred during the signed subtraction.</p>
6091
6092<h5>Arguments:</h5>
6093
6094<p>The arguments (%a and %b) and the first element of the result structure may
6095be of integer types of any bit width, but they must have the same bit width. The
6096second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6097and <tt>%b</tt> are the two values that will undergo signed subtraction.</p>
6098
6099<h5>Semantics:</h5>
6100
6101<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
6102a signed subtraction of the two arguments. They return a structure &mdash; the
6103first element of which is the subtraction, and the second element of which is a bit
6104specifying if the signed subtraction resulted in an overflow.</p>
6105
6106<h5>Examples:</h5>
6107<pre>
6108 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6109 %sum = extractvalue {i32, i1} %res, 0
6110 %obit = extractvalue {i32, i1} %res, 1
6111 br i1 %obit, label %overflow, label %normal
6112</pre>
6113
6114</div>
6115
6116<!-- _______________________________________________________________________ -->
6117<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6118 <a name="int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6119</div>
6120
6121<div class="doc_text">
6122
6123<h5>Syntax:</h5>
6124
6125<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]6126on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6127
6128<pre>
6129 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
6130 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6131 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
6132</pre>
6133
6134<h5>Overview:</h5>
6135
6136<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
6137an unsigned subtraction of the two arguments, and indicate whether an overflow
6138occurred during the unsigned subtraction.</p>
6139
6140<h5>Arguments:</h5>
6141
6142<p>The arguments (%a and %b) and the first element of the result structure may
6143be of integer types of any bit width, but they must have the same bit width. The
6144second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6145and <tt>%b</tt> are the two values that will undergo unsigned subtraction.</p>
6146
6147<h5>Semantics:</h5>
6148
6149<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
6150an unsigned subtraction of the two arguments. They return a structure &mdash; the
6151first element of which is the subtraction, and the second element of which is a bit
6152specifying if the unsigned subtraction resulted in an overflow.</p>
6153
6154<h5>Examples:</h5>
6155<pre>
6156 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6157 %sum = extractvalue {i32, i1} %res, 0
6158 %obit = extractvalue {i32, i1} %res, 1
6159 br i1 %obit, label %overflow, label %normal
6160</pre>
6161
6162</div>
6163
6164<!-- _______________________________________________________________________ -->
6165<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6166 <a name="int_smul_overflow">'<tt>llvm.smul.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6167</div>
6168
6169<div class="doc_text">
6170
6171<h5>Syntax:</h5>
6172
6173<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]6174on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6175
6176<pre>
6177 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
6178 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6179 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
6180</pre>
6181
6182<h5>Overview:</h5>
6183
6184<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
6185a signed multiplication of the two arguments, and indicate whether an overflow
6186occurred during the signed multiplication.</p>
6187
6188<h5>Arguments:</h5>
6189
6190<p>The arguments (%a and %b) and the first element of the result structure may
6191be of integer types of any bit width, but they must have the same bit width. The
6192second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6193and <tt>%b</tt> are the two values that will undergo signed multiplication.</p>
6194
6195<h5>Semantics:</h5>
6196
6197<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
6198a signed multiplication of the two arguments. They return a structure &mdash;
6199the first element of which is the multiplication, and the second element of
6200which is a bit specifying if the signed multiplication resulted in an
6201overflow.</p>
6202
6203<h5>Examples:</h5>
6204<pre>
6205 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6206 %sum = extractvalue {i32, i1} %res, 0
6207 %obit = extractvalue {i32, i1} %res, 1
6208 br i1 %obit, label %overflow, label %normal
6209</pre>
6210
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6211</div>
6212
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]6213<!-- _______________________________________________________________________ -->
6214<div class="doc_subsubsection">
6215 <a name="int_umul_overflow">'<tt>llvm.umul.with.overflow.*</tt>' Intrinsics</a>
6216</div>
6217
6218<div class="doc_text">
6219
6220<h5>Syntax:</h5>
6221
6222<p>This is an overloaded intrinsic. You can use <tt>llvm.umul.with.overflow</tt>
6223on any integer bit width.</p>
6224
6225<pre>
6226 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
6227 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6228 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
6229</pre>
6230
6231<h5>Overview:</h5>
6232
6233<p><i><b>Warning:</b> '<tt>llvm.umul.with.overflow</tt>' is badly broken. It is
6234actively being fixed, but it should not currently be used!</i></p>
6235
6236<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
6237a unsigned multiplication of the two arguments, and indicate whether an overflow
6238occurred during the unsigned multiplication.</p>
6239
6240<h5>Arguments:</h5>
6241
6242<p>The arguments (%a and %b) and the first element of the result structure may
6243be of integer types of any bit width, but they must have the same bit width. The
6244second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6245and <tt>%b</tt> are the two values that will undergo unsigned
6246multiplication.</p>
6247
6248<h5>Semantics:</h5>
6249
6250<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
6251an unsigned multiplication of the two arguments. They return a structure &mdash;
6252the first element of which is the multiplication, and the second element of
6253which is a bit specifying if the unsigned multiplication resulted in an
6254overflow.</p>
6255
6256<h5>Examples:</h5>
6257<pre>
6258 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6259 %sum = extractvalue {i32, i1} %res, 0
6260 %obit = extractvalue {i32, i1} %res, 1
6261 br i1 %obit, label %overflow, label %normal
6262</pre>
6263
6264</div>
6265
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6266<!-- ======================================================================= -->
6267<div class="doc_subsection">
6268 <a name="int_debugger">Debugger Intrinsics</a>
6269</div>
6270
6271<div class="doc_text">
6272<p>
6273The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
6274are described in the <a
6275href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
6276Debugging</a> document.
6277</p>
6278</div>
6279
6280
6281<!-- ======================================================================= -->
6282<div class="doc_subsection">
6283 <a name="int_eh">Exception Handling Intrinsics</a>
6284</div>
6285
6286<div class="doc_text">
6287<p> The LLVM exception handling intrinsics (which all start with
6288<tt>llvm.eh.</tt> prefix), are described in the <a
6289href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
6290Handling</a> document. </p>
6291</div>
6292
6293<!-- ======================================================================= -->
6294<div class="doc_subsection">
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6295 <a name="int_trampoline">Trampoline Intrinsic</a>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6296</div>
6297
6298<div class="doc_text">
6299<p>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6300 This intrinsic makes it possible to excise one parameter, marked with
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6301 the <tt>nest</tt> attribute, from a function. The result is a callable
6302 function pointer lacking the nest parameter - the caller does not need
6303 to provide a value for it. Instead, the value to use is stored in
6304 advance in a "trampoline", a block of memory usually allocated
6305 on the stack, which also contains code to splice the nest value into the
6306 argument list. This is used to implement the GCC nested function address
6307 extension.
6308</p>
6309<p>
6310 For example, if the function is
6311 <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]6312 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]6313<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6314 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
6315 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
6316 %p = call i8* @llvm.init.trampoline( i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8*), i8* %nval )
6317 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6318</pre>
Bill Wendlinge262b4c2007-09-22 09:23:55 +0000[diff] [blame]6319 <p>The call <tt>%val = call i32 %fp( i32 %x, i32 %y )</tt> is then equivalent
6320 to <tt>%val = call i32 %f( i8* %nval, i32 %x, i32 %y )</tt>.</p>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6321</div>
6322
6323<!-- _______________________________________________________________________ -->
6324<div class="doc_subsubsection">
6325 <a name="int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a>
6326</div>
6327<div class="doc_text">
6328<h5>Syntax:</h5>
6329<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6330declare 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]6331</pre>
6332<h5>Overview:</h5>
6333<p>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6334 This fills the memory pointed to by <tt>tramp</tt> with code
6335 and returns a function pointer suitable for executing it.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6336</p>
6337<h5>Arguments:</h5>
6338<p>
6339 The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
6340 pointers. The <tt>tramp</tt> argument must point to a sufficiently large
6341 and sufficiently aligned block of memory; this memory is written to by the
Duncan Sands35012212007-08-22 23:39:54 +0000[diff] [blame]6342 intrinsic. Note that the size and the alignment are target-specific - LLVM
6343 currently provides no portable way of determining them, so a front-end that
6344 generates this intrinsic needs to have some target-specific knowledge.
6345 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]6346</p>
6347<h5>Semantics:</h5>
6348<p>
6349 The block of memory pointed to by <tt>tramp</tt> is filled with target
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6350 dependent code, turning it into a function. A pointer to this function is
6351 returned, but needs to be bitcast to an
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6352 <a href="#int_trampoline">appropriate function pointer type</a>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6353 before being called. The new function's signature is the same as that of
6354 <tt>func</tt> with any arguments marked with the <tt>nest</tt> attribute
6355 removed. At most one such <tt>nest</tt> argument is allowed, and it must be
6356 of pointer type. Calling the new function is equivalent to calling
6357 <tt>func</tt> with the same argument list, but with <tt>nval</tt> used for the
6358 missing <tt>nest</tt> argument. If, after calling
6359 <tt>llvm.init.trampoline</tt>, the memory pointed to by <tt>tramp</tt> is
6360 modified, then the effect of any later call to the returned function pointer is
6361 undefined.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6362</p>
6363</div>
6364
6365<!-- ======================================================================= -->
6366<div class="doc_subsection">
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6367 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
6368</div>
6369
6370<div class="doc_text">
6371<p>
6372 These intrinsic functions expand the "universal IR" of LLVM to represent
6373 hardware constructs for atomic operations and memory synchronization. This
6374 provides an interface to the hardware, not an interface to the programmer. It
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]6375 is aimed at a low enough level to allow any programming models or APIs
6376 (Application Programming Interfaces) which
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6377 need atomic behaviors to map cleanly onto it. It is also modeled primarily on
6378 hardware behavior. Just as hardware provides a "universal IR" for source
6379 languages, it also provides a starting point for developing a "universal"
6380 atomic operation and synchronization IR.
6381</p>
6382<p>
6383 These do <em>not</em> form an API such as high-level threading libraries,
6384 software transaction memory systems, atomic primitives, and intrinsic
6385 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
6386 application libraries. The hardware interface provided by LLVM should allow
6387 a clean implementation of all of these APIs and parallel programming models.
6388 No one model or paradigm should be selected above others unless the hardware
6389 itself ubiquitously does so.
6390
6391</p>
6392</div>
6393
6394<!-- _______________________________________________________________________ -->
6395<div class="doc_subsubsection">
6396 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
6397</div>
6398<div class="doc_text">
6399<h5>Syntax:</h5>
6400<pre>
6401declare void @llvm.memory.barrier( i1 &lt;ll&gt;, i1 &lt;ls&gt;, i1 &lt;sl&gt;, i1 &lt;ss&gt;,
6402i1 &lt;device&gt; )
6403
6404</pre>
6405<h5>Overview:</h5>
6406<p>
6407 The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
6408 specific pairs of memory access types.
6409</p>
6410<h5>Arguments:</h5>
6411<p>
6412 The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
6413 The first four arguments enables a specific barrier as listed below. The fith
6414 argument specifies that the barrier applies to io or device or uncached memory.
6415
6416</p>
6417 <ul>
6418 <li><tt>ll</tt>: load-load barrier</li>
6419 <li><tt>ls</tt>: load-store barrier</li>
6420 <li><tt>sl</tt>: store-load barrier</li>
6421 <li><tt>ss</tt>: store-store barrier</li>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]6422 <li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6423 </ul>
6424<h5>Semantics:</h5>
6425<p>
6426 This intrinsic causes the system to enforce some ordering constraints upon
6427 the loads and stores of the program. This barrier does not indicate
6428 <em>when</em> any events will occur, it only enforces an <em>order</em> in
6429 which they occur. For any of the specified pairs of load and store operations
6430 (f.ex. load-load, or store-load), all of the first operations preceding the
6431 barrier will complete before any of the second operations succeeding the
6432 barrier begin. Specifically the semantics for each pairing is as follows:
6433</p>
6434 <ul>
6435 <li><tt>ll</tt>: All loads before the barrier must complete before any load
6436 after the barrier begins.</li>
6437
6438 <li><tt>ls</tt>: All loads before the barrier must complete before any
6439 store after the barrier begins.</li>
6440 <li><tt>ss</tt>: All stores before the barrier must complete before any
6441 store after the barrier begins.</li>
6442 <li><tt>sl</tt>: All stores before the barrier must complete before any
6443 load after the barrier begins.</li>
6444 </ul>
6445<p>
6446 These semantics are applied with a logical "and" behavior when more than one
6447 is enabled in a single memory barrier intrinsic.
6448</p>
6449<p>
6450 Backends may implement stronger barriers than those requested when they do not
6451 support as fine grained a barrier as requested. Some architectures do not
6452 need all types of barriers and on such architectures, these become noops.
6453</p>
6454<h5>Example:</h5>
6455<pre>
6456%ptr = malloc i32
6457 store i32 4, %ptr
6458
6459%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
6460 call void @llvm.memory.barrier( i1 false, i1 true, i1 false, i1 false )
6461 <i>; guarantee the above finishes</i>
6462 store i32 8, %ptr <i>; before this begins</i>
6463</pre>
6464</div>
6465
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6466<!-- _______________________________________________________________________ -->
6467<div class="doc_subsubsection">
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6468 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6469</div>
6470<div class="doc_text">
6471<h5>Syntax:</h5>
6472<p>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6473 This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
6474 any integer bit width and for different address spaces. Not all targets
6475 support all bit widths however.</p>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6476
6477<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6478declare i8 @llvm.atomic.cmp.swap.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt; )
6479declare i16 @llvm.atomic.cmp.swap.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt; )
6480declare i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt; )
6481declare 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]6482
6483</pre>
6484<h5>Overview:</h5>
6485<p>
6486 This loads a value in memory and compares it to a given value. If they are
6487 equal, it stores a new value into the memory.
6488</p>
6489<h5>Arguments:</h5>
6490<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6491 The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result as
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6492 well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
6493 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
6494 this integer type. While any bit width integer may be used, targets may only
6495 lower representations they support in hardware.
6496
6497</p>
6498<h5>Semantics:</h5>
6499<p>
6500 This entire intrinsic must be executed atomically. It first loads the value
6501 in memory pointed to by <tt>ptr</tt> and compares it with the value
6502 <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the memory. The
6503 loaded value is yielded in all cases. This provides the equivalent of an
6504 atomic compare-and-swap operation within the SSA framework.
6505</p>
6506<h5>Examples:</h5>
6507
6508<pre>
6509%ptr = malloc i32
6510 store i32 4, %ptr
6511
6512%val1 = add i32 4, 4
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6513%result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 4, %val1 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6514 <i>; yields {i32}:result1 = 4</i>
6515%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
6516%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
6517
6518%val2 = add i32 1, 1
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6519%result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 5, %val2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6520 <i>; yields {i32}:result2 = 8</i>
6521%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
6522
6523%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
6524</pre>
6525</div>
6526
6527<!-- _______________________________________________________________________ -->
6528<div class="doc_subsubsection">
6529 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
6530</div>
6531<div class="doc_text">
6532<h5>Syntax:</h5>
6533
6534<p>
6535 This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
6536 integer bit width. Not all targets support all bit widths however.</p>
6537<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6538declare i8 @llvm.atomic.swap.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;val&gt; )
6539declare i16 @llvm.atomic.swap.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;val&gt; )
6540declare i32 @llvm.atomic.swap.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;val&gt; )
6541declare i64 @llvm.atomic.swap.i64.p0i64( i64* &lt;ptr&gt;, i64 &lt;val&gt; )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6542
6543</pre>
6544<h5>Overview:</h5>
6545<p>
6546 This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
6547 the value from memory. It then stores the value in <tt>val</tt> in the memory
6548 at <tt>ptr</tt>.
6549</p>
6550<h5>Arguments:</h5>
6551
6552<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6553 The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both the
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6554 <tt>val</tt> argument and the result must be integers of the same bit width.
6555 The first argument, <tt>ptr</tt>, must be a pointer to a value of this
6556 integer type. The targets may only lower integer representations they
6557 support.
6558</p>
6559<h5>Semantics:</h5>
6560<p>
6561 This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
6562 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
6563 equivalent of an atomic swap operation within the SSA framework.
6564
6565</p>
6566<h5>Examples:</h5>
6567<pre>
6568%ptr = malloc i32
6569 store i32 4, %ptr
6570
6571%val1 = add i32 4, 4
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6572%result1 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val1 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6573 <i>; yields {i32}:result1 = 4</i>
6574%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
6575%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
6576
6577%val2 = add i32 1, 1
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6578%result2 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6579 <i>; yields {i32}:result2 = 8</i>
6580
6581%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
6582%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
6583</pre>
6584</div>
6585
6586<!-- _______________________________________________________________________ -->
6587<div class="doc_subsubsection">
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6588 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6589
6590</div>
6591<div class="doc_text">
6592<h5>Syntax:</h5>
6593<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6594 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]6595 integer bit width. Not all targets support all bit widths however.</p>
6596<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6597declare i8 @llvm.atomic.load.add.i8..p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6598declare i16 @llvm.atomic.load.add.i16..p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6599declare i32 @llvm.atomic.load.add.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6600declare 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]6601
6602</pre>
6603<h5>Overview:</h5>
6604<p>
6605 This intrinsic adds <tt>delta</tt> to the value stored in memory at
6606 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
6607</p>
6608<h5>Arguments:</h5>
6609<p>
6610
6611 The intrinsic takes two arguments, the first a pointer to an integer value
6612 and the second an integer value. The result is also an integer value. These
6613 integer types can have any bit width, but they must all have the same bit
6614 width. The targets may only lower integer representations they support.
6615</p>
6616<h5>Semantics:</h5>
6617<p>
6618 This intrinsic does a series of operations atomically. It first loads the
6619 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
6620 to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
6621</p>
6622
6623<h5>Examples:</h5>
6624<pre>
6625%ptr = malloc i32
6626 store i32 4, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6627%result1 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 4 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6628 <i>; yields {i32}:result1 = 4</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6629%result2 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6630 <i>; yields {i32}:result2 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6631%result3 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 5 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6632 <i>; yields {i32}:result3 = 10</i>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6633%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6634</pre>
6635</div>
6636
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6637<!-- _______________________________________________________________________ -->
6638<div class="doc_subsubsection">
6639 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
6640
6641</div>
6642<div class="doc_text">
6643<h5>Syntax:</h5>
6644<p>
6645 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]6646 any integer bit width and for different address spaces. Not all targets
6647 support all bit widths however.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6648<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6649declare i8 @llvm.atomic.load.sub.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6650declare i16 @llvm.atomic.load.sub.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6651declare i32 @llvm.atomic.load.sub.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6652declare 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]6653
6654</pre>
6655<h5>Overview:</h5>
6656<p>
6657 This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
6658 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
6659</p>
6660<h5>Arguments:</h5>
6661<p>
6662
6663 The intrinsic takes two arguments, the first a pointer to an integer value
6664 and the second an integer value. The result is also an integer value. These
6665 integer types can have any bit width, but they must all have the same bit
6666 width. The targets may only lower integer representations they support.
6667</p>
6668<h5>Semantics:</h5>
6669<p>
6670 This intrinsic does a series of operations atomically. It first loads the
6671 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
6672 result to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
6673</p>
6674
6675<h5>Examples:</h5>
6676<pre>
6677%ptr = malloc i32
6678 store i32 8, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6679%result1 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 4 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6680 <i>; yields {i32}:result1 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6681%result2 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 2 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6682 <i>; yields {i32}:result2 = 4</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6683%result3 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 5 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6684 <i>; yields {i32}:result3 = 2</i>
6685%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
6686</pre>
6687</div>
6688
6689<!-- _______________________________________________________________________ -->
6690<div class="doc_subsubsection">
6691 <a name="int_atomic_load_and">'<tt>llvm.atomic.load.and.*</tt>' Intrinsic</a><br>
6692 <a name="int_atomic_load_nand">'<tt>llvm.atomic.load.nand.*</tt>' Intrinsic</a><br>
6693 <a name="int_atomic_load_or">'<tt>llvm.atomic.load.or.*</tt>' Intrinsic</a><br>
6694 <a name="int_atomic_load_xor">'<tt>llvm.atomic.load.xor.*</tt>' Intrinsic</a><br>
6695
6696</div>
6697<div class="doc_text">
6698<h5>Syntax:</h5>
6699<p>
6700 These are overloaded intrinsics. You can use <tt>llvm.atomic.load_and</tt>,
6701 <tt>llvm.atomic.load_nand</tt>, <tt>llvm.atomic.load_or</tt>, and
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6702 <tt>llvm.atomic.load_xor</tt> on any integer bit width and for different
6703 address spaces. Not all targets support all bit widths however.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6704<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6705declare i8 @llvm.atomic.load.and.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6706declare i16 @llvm.atomic.load.and.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6707declare i32 @llvm.atomic.load.and.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6708declare 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]6709
6710</pre>
6711
6712<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6713declare i8 @llvm.atomic.load.or.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6714declare i16 @llvm.atomic.load.or.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6715declare i32 @llvm.atomic.load.or.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6716declare 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]6717
6718</pre>
6719
6720<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6721declare i8 @llvm.atomic.load.nand.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6722declare i16 @llvm.atomic.load.nand.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6723declare i32 @llvm.atomic.load.nand.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6724declare 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]6725
6726</pre>
6727
6728<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6729declare i8 @llvm.atomic.load.xor.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6730declare i16 @llvm.atomic.load.xor.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6731declare i32 @llvm.atomic.load.xor.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6732declare 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]6733
6734</pre>
6735<h5>Overview:</h5>
6736<p>
6737 These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
6738 the value stored in memory at <tt>ptr</tt>. It yields the original value
6739 at <tt>ptr</tt>.
6740</p>
6741<h5>Arguments:</h5>
6742<p>
6743
6744 These intrinsics take two arguments, the first a pointer to an integer value
6745 and the second an integer value. The result is also an integer value. These
6746 integer types can have any bit width, but they must all have the same bit
6747 width. The targets may only lower integer representations they support.
6748</p>
6749<h5>Semantics:</h5>
6750<p>
6751 These intrinsics does a series of operations atomically. They first load the
6752 value stored at <tt>ptr</tt>. They then do the bitwise operation
6753 <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the original
6754 value stored at <tt>ptr</tt>.
6755</p>
6756
6757<h5>Examples:</h5>
6758<pre>
6759%ptr = malloc i32
6760 store i32 0x0F0F, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6761%result0 = call i32 @llvm.atomic.load.nand.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6762 <i>; yields {i32}:result0 = 0x0F0F</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6763%result1 = call i32 @llvm.atomic.load.and.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6764 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6765%result2 = call i32 @llvm.atomic.load.or.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6766 <i>; yields {i32}:result2 = 0xF0</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6767%result3 = call i32 @llvm.atomic.load.xor.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6768 <i>; yields {i32}:result3 = FF</i>
6769%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
6770</pre>
6771</div>
6772
6773
6774<!-- _______________________________________________________________________ -->
6775<div class="doc_subsubsection">
6776 <a name="int_atomic_load_max">'<tt>llvm.atomic.load.max.*</tt>' Intrinsic</a><br>
6777 <a name="int_atomic_load_min">'<tt>llvm.atomic.load.min.*</tt>' Intrinsic</a><br>
6778 <a name="int_atomic_load_umax">'<tt>llvm.atomic.load.umax.*</tt>' Intrinsic</a><br>
6779 <a name="int_atomic_load_umin">'<tt>llvm.atomic.load.umin.*</tt>' Intrinsic</a><br>
6780
6781</div>
6782<div class="doc_text">
6783<h5>Syntax:</h5>
6784<p>
6785 These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
6786 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6787 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
6788 address spaces. Not all targets
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6789 support all bit widths however.</p>
6790<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6791declare i8 @llvm.atomic.load.max.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6792declare i16 @llvm.atomic.load.max.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6793declare i32 @llvm.atomic.load.max.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6794declare 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]6795
6796</pre>
6797
6798<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6799declare i8 @llvm.atomic.load.min.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6800declare i16 @llvm.atomic.load.min.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6801declare i32 @llvm.atomic.load.min.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6802declare 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]6803
6804</pre>
6805
6806<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6807declare i8 @llvm.atomic.load.umax.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6808declare i16 @llvm.atomic.load.umax.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6809declare i32 @llvm.atomic.load.umax.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6810declare 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]6811
6812</pre>
6813
6814<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6815declare i8 @llvm.atomic.load.umin.i8..p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6816declare i16 @llvm.atomic.load.umin.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6817declare i32 @llvm.atomic.load.umin.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6818declare 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]6819
6820</pre>
6821<h5>Overview:</h5>
6822<p>
6823 These intrinsics takes the signed or unsigned minimum or maximum of
6824 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
6825 original value at <tt>ptr</tt>.
6826</p>
6827<h5>Arguments:</h5>
6828<p>
6829
6830 These intrinsics take two arguments, the first a pointer to an integer value
6831 and the second an integer value. The result is also an integer value. These
6832 integer types can have any bit width, but they must all have the same bit
6833 width. The targets may only lower integer representations they support.
6834</p>
6835<h5>Semantics:</h5>
6836<p>
6837 These intrinsics does a series of operations atomically. They first load the
6838 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or max
6839 <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They yield
6840 the original value stored at <tt>ptr</tt>.
6841</p>
6842
6843<h5>Examples:</h5>
6844<pre>
6845%ptr = malloc i32
6846 store i32 7, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6847%result0 = call i32 @llvm.atomic.load.min.i32.p0i32( i32* %ptr, i32 -2 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6848 <i>; yields {i32}:result0 = 7</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6849%result1 = call i32 @llvm.atomic.load.max.i32.p0i32( i32* %ptr, i32 8 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6850 <i>; yields {i32}:result1 = -2</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6851%result2 = call i32 @llvm.atomic.load.umin.i32.p0i32( i32* %ptr, i32 10 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6852 <i>; yields {i32}:result2 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6853%result3 = call i32 @llvm.atomic.load.umax.i32.p0i32( i32* %ptr, i32 30 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6854 <i>; yields {i32}:result3 = 8</i>
6855%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
6856</pre>
6857</div>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6858
6859<!-- ======================================================================= -->
6860<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6861 <a name="int_general">General Intrinsics</a>
6862</div>
6863
6864<div class="doc_text">
6865<p> This class of intrinsics is designed to be generic and has
6866no specific purpose. </p>
6867</div>
6868
6869<!-- _______________________________________________________________________ -->
6870<div class="doc_subsubsection">
6871 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
6872</div>
6873
6874<div class="doc_text">
6875
6876<h5>Syntax:</h5>
6877<pre>
6878 declare void @llvm.var.annotation(i8* &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6879</pre>
6880
6881<h5>Overview:</h5>
6882
6883<p>
6884The '<tt>llvm.var.annotation</tt>' intrinsic
6885</p>
6886
6887<h5>Arguments:</h5>
6888
6889<p>
6890The first argument is a pointer to a value, the second is a pointer to a
6891global string, the third is a pointer to a global string which is the source
6892file name, and the last argument is the line number.
6893</p>
6894
6895<h5>Semantics:</h5>
6896
6897<p>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]6898This intrinsic allows annotation of local variables with arbitrary strings.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6899This can be useful for special purpose optimizations that want to look for these
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]6900annotations. These have no other defined use, they are ignored by code
6901generation and optimization.
6902</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6903</div>
6904
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6905<!-- _______________________________________________________________________ -->
6906<div class="doc_subsubsection">
Tanya Lattnerc9869b12007-09-21 23:57:59 +0000[diff] [blame]6907 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6908</div>
6909
6910<div class="doc_text">
6911
6912<h5>Syntax:</h5>
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]6913<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
6914any integer bit width.
6915</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6916<pre>
Tanya Lattner09161fe2007-09-22 00:03:01 +0000[diff] [blame]6917 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6918 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6919 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6920 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6921 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]6922</pre>
6923
6924<h5>Overview:</h5>
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]6925
6926<p>
6927The '<tt>llvm.annotation</tt>' intrinsic.
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6928</p>
6929
6930<h5>Arguments:</h5>
6931
6932<p>
6933The first argument is an integer value (result of some expression),
6934the second is a pointer to a global string, the third is a pointer to a global
6935string which is the source file name, and the last argument is the line number.
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]6936It returns the value of the first argument.
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6937</p>
6938
6939<h5>Semantics:</h5>
6940
6941<p>
6942This intrinsic allows annotations to be put on arbitrary expressions
6943with arbitrary strings. This can be useful for special purpose optimizations
6944that want to look for these annotations. These have no other defined use, they
6945are ignored by code generation and optimization.
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]6946</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6947</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6948
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]6949<!-- _______________________________________________________________________ -->
6950<div class="doc_subsubsection">
6951 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
6952</div>
6953
6954<div class="doc_text">
6955
6956<h5>Syntax:</h5>
6957<pre>
6958 declare void @llvm.trap()
6959</pre>
6960
6961<h5>Overview:</h5>
6962
6963<p>
6964The '<tt>llvm.trap</tt>' intrinsic
6965</p>
6966
6967<h5>Arguments:</h5>
6968
6969<p>
6970None
6971</p>
6972
6973<h5>Semantics:</h5>
6974
6975<p>
6976This intrinsics is lowered to the target dependent trap instruction. If the
6977target does not have a trap instruction, this intrinsic will be lowered to the
6978call of the abort() function.
6979</p>
6980</div>
6981
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]6982<!-- _______________________________________________________________________ -->
6983<div class="doc_subsubsection">
Misha Brukman5dd7f4d2008-11-22 23:55:29 +0000[diff] [blame]6984 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]6985</div>
6986<div class="doc_text">
6987<h5>Syntax:</h5>
6988<pre>
6989declare void @llvm.stackprotector( i8* &lt;guard&gt;, i8** &lt;slot&gt; )
6990
6991</pre>
6992<h5>Overview:</h5>
6993<p>
6994 The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and stores
6995 it onto the stack at <tt>slot</tt>. The stack slot is adjusted to ensure that
6996 it is placed on the stack before local variables.
6997</p>
6998<h5>Arguments:</h5>
6999<p>
7000 The <tt>llvm.stackprotector</tt> intrinsic requires two pointer arguments. The
7001 first argument is the value loaded from the stack guard
7002 <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt> that
7003 has enough space to hold the value of the guard.
7004</p>
7005<h5>Semantics:</h5>
7006<p>
7007 This intrinsic causes the prologue/epilogue inserter to force the position of
7008 the <tt>AllocaInst</tt> stack slot to be before local variables on the
7009 stack. This is to ensure that if a local variable on the stack is overwritten,
7010 it will destroy the value of the guard. When the function exits, the guard on
7011 the stack is checked against the original guard. If they're different, then
7012 the program aborts by calling the <tt>__stack_chk_fail()</tt> function.
7013</p>
7014</div>
7015
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7016<!-- *********************************************************************** -->
7017<hr>
7018<address>
7019 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman947321d2008-12-11 17:34:48 +0000[diff] [blame]7020 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7021 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman947321d2008-12-11 17:34:48 +0000[diff] [blame]7022 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]7023
7024 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
7025 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
7026 Last modified: $Date$
7027</address>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]7028
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7029</body>
7030</html>