Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / d901b6640e0f1d1757a2db9fb57a140ad828ea75 / . / docs / LangRef.html
blob: 137ed9e8dae3060282bac52ebab0a96aad465641 [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>
25 <li><a href="#globalvars">Global Variables</a></li>
26 <li><a href="#functionstructure">Functions</a></li>
27 <li><a href="#aliasstructure">Aliases</a>
28 <li><a href="#paramattrs">Parameter Attributes</a></li>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]29 <li><a href="#fnattrs">Function Attributes</a></li>
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]30 <li><a href="#gc">Garbage Collector Names</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]31 <li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
32 <li><a href="#datalayout">Data Layout</a></li>
33 </ol>
34 </li>
35 <li><a href="#typesystem">Type System</a>
36 <ol>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]37 <li><a href="#t_classifications">Type Classifications</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]38 <li><a href="#t_primitive">Primitive Types</a>
39 <ol>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]40 <li><a href="#t_floating">Floating Point Types</a></li>
41 <li><a href="#t_void">Void Type</a></li>
42 <li><a href="#t_label">Label Type</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]43 </ol>
44 </li>
45 <li><a href="#t_derived">Derived Types</a>
46 <ol>
Chris Lattner251ab812007-12-18 06:18:21 +0000[diff] [blame]47 <li><a href="#t_integer">Integer Type</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]48 <li><a href="#t_array">Array Type</a></li>
49 <li><a href="#t_function">Function Type</a></li>
50 <li><a href="#t_pointer">Pointer Type</a></li>
51 <li><a href="#t_struct">Structure Type</a></li>
52 <li><a href="#t_pstruct">Packed Structure Type</a></li>
53 <li><a href="#t_vector">Vector Type</a></li>
54 <li><a href="#t_opaque">Opaque Type</a></li>
55 </ol>
56 </li>
57 </ol>
58 </li>
59 <li><a href="#constants">Constants</a>
60 <ol>
61 <li><a href="#simpleconstants">Simple Constants</a>
62 <li><a href="#aggregateconstants">Aggregate Constants</a>
63 <li><a href="#globalconstants">Global Variable and Function Addresses</a>
64 <li><a href="#undefvalues">Undefined Values</a>
65 <li><a href="#constantexprs">Constant Expressions</a>
66 </ol>
67 </li>
68 <li><a href="#othervalues">Other Values</a>
69 <ol>
70 <li><a href="#inlineasm">Inline Assembler Expressions</a>
71 </ol>
72 </li>
73 <li><a href="#instref">Instruction Reference</a>
74 <ol>
75 <li><a href="#terminators">Terminator Instructions</a>
76 <ol>
77 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
78 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
79 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
80 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
81 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
82 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
83 </ol>
84 </li>
85 <li><a href="#binaryops">Binary Operations</a>
86 <ol>
87 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
88 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
89 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
90 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
91 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
92 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
93 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
94 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
95 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
96 </ol>
97 </li>
98 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
99 <ol>
100 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
101 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
102 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
103 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
104 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
105 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
106 </ol>
107 </li>
108 <li><a href="#vectorops">Vector Operations</a>
109 <ol>
110 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
111 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
112 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
113 </ol>
114 </li>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]115 <li><a href="#aggregateops">Aggregate Operations</a>
116 <ol>
117 <li><a href="#i_extractvalue">'<tt>extractvalue</tt>' Instruction</a></li>
118 <li><a href="#i_insertvalue">'<tt>insertvalue</tt>' Instruction</a></li>
119 </ol>
120 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]121 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
122 <ol>
123 <li><a href="#i_malloc">'<tt>malloc</tt>' Instruction</a></li>
124 <li><a href="#i_free">'<tt>free</tt>' Instruction</a></li>
125 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
126 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
127 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
128 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
129 </ol>
130 </li>
131 <li><a href="#convertops">Conversion Operations</a>
132 <ol>
133 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
134 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
135 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
136 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
137 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
138 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
139 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
140 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
141 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
142 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
143 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
144 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
145 </ol>
146 <li><a href="#otherops">Other Operations</a>
147 <ol>
148 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
149 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]150 <li><a href="#i_vicmp">'<tt>vicmp</tt>' Instruction</a></li>
151 <li><a href="#i_vfcmp">'<tt>vfcmp</tt>' Instruction</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]152 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
153 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
154 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
155 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
156 </ol>
157 </li>
158 </ol>
159 </li>
160 <li><a href="#intrinsics">Intrinsic Functions</a>
161 <ol>
162 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
163 <ol>
164 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
165 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
166 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
167 </ol>
168 </li>
169 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
170 <ol>
171 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
172 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
173 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
174 </ol>
175 </li>
176 <li><a href="#int_codegen">Code Generator Intrinsics</a>
177 <ol>
178 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
179 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
180 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
181 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
182 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
183 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
184 <li><a href="#int_readcyclecounter"><tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
185 </ol>
186 </li>
187 <li><a href="#int_libc">Standard C Library Intrinsics</a>
188 <ol>
189 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
190 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
191 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
192 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
193 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]194 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
195 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
196 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]197 </ol>
198 </li>
199 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
200 <ol>
201 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
202 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
203 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
204 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
205 <li><a href="#int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic </a></li>
206 <li><a href="#int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic </a></li>
207 </ol>
208 </li>
209 <li><a href="#int_debugger">Debugger intrinsics</a></li>
210 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]211 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]212 <ol>
213 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]214 </ol>
215 </li>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]216 <li><a href="#int_atomics">Atomic intrinsics</a>
217 <ol>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]218 <li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]219 <li><a href="#int_atomic_cmp_swap"><tt>llvm.atomic.cmp.swap</tt></a></li>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]220 <li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]221 <li><a href="#int_atomic_load_add"><tt>llvm.atomic.load.add</tt></a></li>
222 <li><a href="#int_atomic_load_sub"><tt>llvm.atomic.load.sub</tt></a></li>
223 <li><a href="#int_atomic_load_and"><tt>llvm.atomic.load.and</tt></a></li>
224 <li><a href="#int_atomic_load_nand"><tt>llvm.atomic.load.nand</tt></a></li>
225 <li><a href="#int_atomic_load_or"><tt>llvm.atomic.load.or</tt></a></li>
226 <li><a href="#int_atomic_load_xor"><tt>llvm.atomic.load.xor</tt></a></li>
227 <li><a href="#int_atomic_load_max"><tt>llvm.atomic.load.max</tt></a></li>
228 <li><a href="#int_atomic_load_min"><tt>llvm.atomic.load.min</tt></a></li>
229 <li><a href="#int_atomic_load_umax"><tt>llvm.atomic.load.umax</tt></a></li>
230 <li><a href="#int_atomic_load_umin"><tt>llvm.atomic.load.umin</tt></a></li>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]231 </ol>
232 </li>
Reid Spencerb043f672007-07-20 19:59:11 +0000[diff] [blame]233 <li><a href="#int_general">General intrinsics</a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]234 <ol>
Reid Spencerb043f672007-07-20 19:59:11 +0000[diff] [blame]235 <li><a href="#int_var_annotation">
Tanya Lattner51369f32007-09-22 00:01:26 +0000[diff] [blame]236 <tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]237 <li><a href="#int_annotation">
Tanya Lattner51369f32007-09-22 00:01:26 +0000[diff] [blame]238 <tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]239 <li><a href="#int_trap">
240 <tt>llvm.trap</tt>' Intrinsic</a></li>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]241 </ol>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]242 </li>
243 </ol>
244 </li>
245</ol>
246
247<div class="doc_author">
248 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
249 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
250</div>
251
252<!-- *********************************************************************** -->
253<div class="doc_section"> <a name="abstract">Abstract </a></div>
254<!-- *********************************************************************** -->
255
256<div class="doc_text">
257<p>This document is a reference manual for the LLVM assembly language.
Bill Wendlinge7846a52008-08-05 22:29:16 +0000[diff] [blame]258LLVM is a Static Single Assignment (SSA) based representation that provides
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]259type safety, low-level operations, flexibility, and the capability of
260representing 'all' high-level languages cleanly. It is the common code
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]261representation used throughout all phases of the LLVM compilation
262strategy.</p>
263</div>
264
265<!-- *********************************************************************** -->
266<div class="doc_section"> <a name="introduction">Introduction</a> </div>
267<!-- *********************************************************************** -->
268
269<div class="doc_text">
270
271<p>The LLVM code representation is designed to be used in three
272different forms: as an in-memory compiler IR, as an on-disk bitcode
273representation (suitable for fast loading by a Just-In-Time compiler),
274and as a human readable assembly language representation. This allows
275LLVM to provide a powerful intermediate representation for efficient
276compiler transformations and analysis, while providing a natural means
277to debug and visualize the transformations. The three different forms
278of LLVM are all equivalent. This document describes the human readable
279representation and notation.</p>
280
281<p>The LLVM representation aims to be light-weight and low-level
282while being expressive, typed, and extensible at the same time. It
283aims to be a "universal IR" of sorts, by being at a low enough level
284that high-level ideas may be cleanly mapped to it (similar to how
285microprocessors are "universal IR's", allowing many source languages to
286be mapped to them). By providing type information, LLVM can be used as
287the target of optimizations: for example, through pointer analysis, it
288can be proven that a C automatic variable is never accessed outside of
289the current function... allowing it to be promoted to a simple SSA
290value instead of a memory location.</p>
291
292</div>
293
294<!-- _______________________________________________________________________ -->
295<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
296
297<div class="doc_text">
298
299<p>It is important to note that this document describes 'well formed'
300LLVM assembly language. There is a difference between what the parser
301accepts and what is considered 'well formed'. For example, the
302following instruction is syntactically okay, but not well formed:</p>
303
304<div class="doc_code">
305<pre>
306%x = <a href="#i_add">add</a> i32 1, %x
307</pre>
308</div>
309
310<p>...because the definition of <tt>%x</tt> does not dominate all of
311its uses. The LLVM infrastructure provides a verification pass that may
312be used to verify that an LLVM module is well formed. This pass is
313automatically run by the parser after parsing input assembly and by
314the optimizer before it outputs bitcode. The violations pointed out
315by the verifier pass indicate bugs in transformation passes or input to
316the parser.</p>
317</div>
318
Chris Lattnera83fdc02007-10-03 17:34:29 +0000[diff] [blame]319<!-- Describe the typesetting conventions here. -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]320
321<!-- *********************************************************************** -->
322<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
323<!-- *********************************************************************** -->
324
325<div class="doc_text">
326
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]327 <p>LLVM identifiers come in two basic types: global and local. Global
328 identifiers (functions, global variables) begin with the @ character. Local
329 identifiers (register names, types) begin with the % character. Additionally,
330 there are three different formats for identifiers, for different purposes:
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]331
332<ol>
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]333 <li>Named values are represented as a string of characters with their prefix.
334 For example, %foo, @DivisionByZero, %a.really.long.identifier. The actual
335 regular expression used is '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]336 Identifiers which require other characters in their names can be surrounded
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]337 with quotes. In this way, anything except a <tt>&quot;</tt> character can
338 be used in a named value.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]339
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]340 <li>Unnamed values are represented as an unsigned numeric value with their
341 prefix. For example, %12, @2, %44.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]342
343 <li>Constants, which are described in a <a href="#constants">section about
344 constants</a>, below.</li>
345</ol>
346
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]347<p>LLVM requires that values start with a prefix for two reasons: Compilers
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]348don't need to worry about name clashes with reserved words, and the set of
349reserved words may be expanded in the future without penalty. Additionally,
350unnamed identifiers allow a compiler to quickly come up with a temporary
351variable without having to avoid symbol table conflicts.</p>
352
353<p>Reserved words in LLVM are very similar to reserved words in other
354languages. There are keywords for different opcodes
355('<tt><a href="#i_add">add</a></tt>',
356 '<tt><a href="#i_bitcast">bitcast</a></tt>',
357 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
358href="#t_void">void</a></tt>', '<tt><a href="#t_primitive">i32</a></tt>', etc...),
359and others. These reserved words cannot conflict with variable names, because
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]360none of them start with a prefix character ('%' or '@').</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]361
362<p>Here is an example of LLVM code to multiply the integer variable
363'<tt>%X</tt>' by 8:</p>
364
365<p>The easy way:</p>
366
367<div class="doc_code">
368<pre>
369%result = <a href="#i_mul">mul</a> i32 %X, 8
370</pre>
371</div>
372
373<p>After strength reduction:</p>
374
375<div class="doc_code">
376<pre>
377%result = <a href="#i_shl">shl</a> i32 %X, i8 3
378</pre>
379</div>
380
381<p>And the hard way:</p>
382
383<div class="doc_code">
384<pre>
385<a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
386<a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
387%result = <a href="#i_add">add</a> i32 %1, %1
388</pre>
389</div>
390
391<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
392important lexical features of LLVM:</p>
393
394<ol>
395
396 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
397 line.</li>
398
399 <li>Unnamed temporaries are created when the result of a computation is not
400 assigned to a named value.</li>
401
402 <li>Unnamed temporaries are numbered sequentially</li>
403
404</ol>
405
406<p>...and it also shows a convention that we follow in this document. When
407demonstrating instructions, we will follow an instruction with a comment that
408defines the type and name of value produced. Comments are shown in italic
409text.</p>
410
411</div>
412
413<!-- *********************************************************************** -->
414<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
415<!-- *********************************************************************** -->
416
417<!-- ======================================================================= -->
418<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
419</div>
420
421<div class="doc_text">
422
423<p>LLVM programs are composed of "Module"s, each of which is a
424translation unit of the input programs. Each module consists of
425functions, global variables, and symbol table entries. Modules may be
426combined together with the LLVM linker, which merges function (and
427global variable) definitions, resolves forward declarations, and merges
428symbol table entries. Here is an example of the "hello world" module:</p>
429
430<div class="doc_code">
431<pre><i>; Declare the string constant as a global constant...</i>
432<a href="#identifiers">@.LC0</a> = <a href="#linkage_internal">internal</a> <a
433 href="#globalvars">constant</a> <a href="#t_array">[13 x i8]</a> c"hello world\0A\00" <i>; [13 x i8]*</i>
434
435<i>; External declaration of the puts function</i>
436<a href="#functionstructure">declare</a> i32 @puts(i8 *) <i>; i32(i8 *)* </i>
437
438<i>; Definition of main function</i>
439define i32 @main() { <i>; i32()* </i>
440 <i>; Convert [13x i8 ]* to i8 *...</i>
441 %cast210 = <a
442 href="#i_getelementptr">getelementptr</a> [13 x i8 ]* @.LC0, i64 0, i64 0 <i>; i8 *</i>
443
444 <i>; Call puts function to write out the string to stdout...</i>
445 <a
446 href="#i_call">call</a> i32 @puts(i8 * %cast210) <i>; i32</i>
447 <a
448 href="#i_ret">ret</a> i32 0<br>}<br>
449</pre>
450</div>
451
452<p>This example is made up of a <a href="#globalvars">global variable</a>
453named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
454function, and a <a href="#functionstructure">function definition</a>
455for "<tt>main</tt>".</p>
456
457<p>In general, a module is made up of a list of global values,
458where both functions and global variables are global values. Global values are
459represented by a pointer to a memory location (in this case, a pointer to an
460array of char, and a pointer to a function), and have one of the following <a
461href="#linkage">linkage types</a>.</p>
462
463</div>
464
465<!-- ======================================================================= -->
466<div class="doc_subsection">
467 <a name="linkage">Linkage Types</a>
468</div>
469
470<div class="doc_text">
471
472<p>
473All Global Variables and Functions have one of the following types of linkage:
474</p>
475
476<dl>
477
Dale Johannesen96e7e092008-05-23 23:13:41 +0000[diff] [blame]478 <dt><tt><b><a name="linkage_internal">internal</a></b></tt>: </dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]479
480 <dd>Global values with internal linkage are only directly accessible by
481 objects in the current module. In particular, linking code into a module with
482 an internal global value may cause the internal to be renamed as necessary to
483 avoid collisions. Because the symbol is internal to the module, all
484 references can be updated. This corresponds to the notion of the
485 '<tt>static</tt>' keyword in C.
486 </dd>
487
488 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
489
490 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
491 the same name when linkage occurs. This is typically used to implement
492 inline functions, templates, or other code which must be generated in each
493 translation unit that uses it. Unreferenced <tt>linkonce</tt> globals are
494 allowed to be discarded.
495 </dd>
496
Dale Johannesen96e7e092008-05-23 23:13:41 +0000[diff] [blame]497 <dt><tt><b><a name="linkage_common">common</a></b></tt>: </dt>
498
499 <dd>"<tt>common</tt>" linkage is exactly the same as <tt>linkonce</tt>
500 linkage, except that unreferenced <tt>common</tt> globals may not be
501 discarded. This is used for globals that may be emitted in multiple
502 translation units, but that are not guaranteed to be emitted into every
503 translation unit that uses them. One example of this is tentative
504 definitions in C, such as "<tt>int X;</tt>" at global scope.
505 </dd>
506
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]507 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
508
Dale Johannesen96e7e092008-05-23 23:13:41 +0000[diff] [blame]509 <dd>"<tt>weak</tt>" linkage is the same as <tt>common</tt> linkage, except
510 that some targets may choose to emit different assembly sequences for them
511 for target-dependent reasons. This is used for globals that are declared
512 "weak" in C source code.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]513 </dd>
514
515 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
516
517 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
518 pointer to array type. When two global variables with appending linkage are
519 linked together, the two global arrays are appended together. This is the
520 LLVM, typesafe, equivalent of having the system linker append together
521 "sections" with identical names when .o files are linked.
522 </dd>
523
524 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt>: </dt>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]525 <dd>The semantics of this linkage follow the ELF object file model: the
526 symbol is weak until linked, if not linked, the symbol becomes null instead
527 of being an undefined reference.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]528 </dd>
529
530 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
531
532 <dd>If none of the above identifiers are used, the global is externally
533 visible, meaning that it participates in linkage and can be used to resolve
534 external symbol references.
535 </dd>
536</dl>
537
538 <p>
539 The next two types of linkage are targeted for Microsoft Windows platform
540 only. They are designed to support importing (exporting) symbols from (to)
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]541 DLLs (Dynamic Link Libraries).
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]542 </p>
543
544 <dl>
545 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt>: </dt>
546
547 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
548 or variable via a global pointer to a pointer that is set up by the DLL
549 exporting the symbol. On Microsoft Windows targets, the pointer name is
550 formed by combining <code>_imp__</code> and the function or variable name.
551 </dd>
552
553 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt>: </dt>
554
555 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
556 pointer to a pointer in a DLL, so that it can be referenced with the
557 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
558 name is formed by combining <code>_imp__</code> and the function or variable
559 name.
560 </dd>
561
562</dl>
563
564<p><a name="linkage_external"></a>For example, since the "<tt>.LC0</tt>"
565variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
566variable and was linked with this one, one of the two would be renamed,
567preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
568external (i.e., lacking any linkage declarations), they are accessible
569outside of the current module.</p>
570<p>It is illegal for a function <i>declaration</i>
571to have any linkage type other than "externally visible", <tt>dllimport</tt>,
572or <tt>extern_weak</tt>.</p>
573<p>Aliases can have only <tt>external</tt>, <tt>internal</tt> and <tt>weak</tt>
574linkages.
575</div>
576
577<!-- ======================================================================= -->
578<div class="doc_subsection">
579 <a name="callingconv">Calling Conventions</a>
580</div>
581
582<div class="doc_text">
583
584<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
585and <a href="#i_invoke">invokes</a> can all have an optional calling convention
586specified for the call. The calling convention of any pair of dynamic
587caller/callee must match, or the behavior of the program is undefined. The
588following calling conventions are supported by LLVM, and more may be added in
589the future:</p>
590
591<dl>
592 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
593
594 <dd>This calling convention (the default if no other calling convention is
595 specified) matches the target C calling conventions. This calling convention
596 supports varargs function calls and tolerates some mismatch in the declared
597 prototype and implemented declaration of the function (as does normal C).
598 </dd>
599
600 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
601
602 <dd>This calling convention attempts to make calls as fast as possible
603 (e.g. by passing things in registers). This calling convention allows the
604 target to use whatever tricks it wants to produce fast code for the target,
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]605 without having to conform to an externally specified ABI (Application Binary
606 Interface). Implementations of this convention should allow arbitrary
Arnold Schwaighofer07444922008-05-14 09:17:12 +0000[diff] [blame]607 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> to be
608 supported. This calling convention does not support varargs and requires the
609 prototype of all callees to exactly match the prototype of the function
610 definition.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]611 </dd>
612
613 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
614
615 <dd>This calling convention attempts to make code in the caller as efficient
616 as possible under the assumption that the call is not commonly executed. As
617 such, these calls often preserve all registers so that the call does not break
618 any live ranges in the caller side. This calling convention does not support
619 varargs and requires the prototype of all callees to exactly match the
620 prototype of the function definition.
621 </dd>
622
623 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
624
625 <dd>Any calling convention may be specified by number, allowing
626 target-specific calling conventions to be used. Target specific calling
627 conventions start at 64.
628 </dd>
629</dl>
630
631<p>More calling conventions can be added/defined on an as-needed basis, to
632support pascal conventions or any other well-known target-independent
633convention.</p>
634
635</div>
636
637<!-- ======================================================================= -->
638<div class="doc_subsection">
639 <a name="visibility">Visibility Styles</a>
640</div>
641
642<div class="doc_text">
643
644<p>
645All Global Variables and Functions have one of the following visibility styles:
646</p>
647
648<dl>
649 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
650
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]651 <dd>On targets that use the ELF object file format, default visibility means
652 that the declaration is visible to other
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]653 modules and, in shared libraries, means that the declared entity may be
654 overridden. On Darwin, default visibility means that the declaration is
655 visible to other modules. Default visibility corresponds to "external
656 linkage" in the language.
657 </dd>
658
659 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
660
661 <dd>Two declarations of an object with hidden visibility refer to the same
662 object if they are in the same shared object. Usually, hidden visibility
663 indicates that the symbol will not be placed into the dynamic symbol table,
664 so no other module (executable or shared library) can reference it
665 directly.
666 </dd>
667
668 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
669
670 <dd>On ELF, protected visibility indicates that the symbol will be placed in
671 the dynamic symbol table, but that references within the defining module will
672 bind to the local symbol. That is, the symbol cannot be overridden by another
673 module.
674 </dd>
675</dl>
676
677</div>
678
679<!-- ======================================================================= -->
680<div class="doc_subsection">
681 <a name="globalvars">Global Variables</a>
682</div>
683
684<div class="doc_text">
685
686<p>Global variables define regions of memory allocated at compilation time
687instead of run-time. Global variables may optionally be initialized, may have
688an explicit section to be placed in, and may have an optional explicit alignment
689specified. A variable may be defined as "thread_local", which means that it
690will not be shared by threads (each thread will have a separated copy of the
691variable). A variable may be defined as a global "constant," which indicates
692that the contents of the variable will <b>never</b> be modified (enabling better
693optimization, allowing the global data to be placed in the read-only section of
694an executable, etc). Note that variables that need runtime initialization
695cannot be marked "constant" as there is a store to the variable.</p>
696
697<p>
698LLVM explicitly allows <em>declarations</em> of global variables to be marked
699constant, even if the final definition of the global is not. This capability
700can be used to enable slightly better optimization of the program, but requires
701the language definition to guarantee that optimizations based on the
702'constantness' are valid for the translation units that do not include the
703definition.
704</p>
705
706<p>As SSA values, global variables define pointer values that are in
707scope (i.e. they dominate) all basic blocks in the program. Global
708variables always define a pointer to their "content" type because they
709describe a region of memory, and all memory objects in LLVM are
710accessed through pointers.</p>
711
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]712<p>A global variable may be declared to reside in a target-specifc numbered
713address space. For targets that support them, address spaces may affect how
714optimizations are performed and/or what target instructions are used to access
Christopher Lamb20a39e92007-12-12 08:44:39 +0000[diff] [blame]715the variable. The default address space is zero. The address space qualifier
716must precede any other attributes.</p>
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]717
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]718<p>LLVM allows an explicit section to be specified for globals. If the target
719supports it, it will emit globals to the section specified.</p>
720
721<p>An explicit alignment may be specified for a global. If not present, or if
722the alignment is set to zero, the alignment of the global is set by the target
723to whatever it feels convenient. If an explicit alignment is specified, the
724global is forced to have at least that much alignment. All alignments must be
725a power of 2.</p>
726
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]727<p>For example, the following defines a global in a numbered address space with
728an initializer, section, and alignment:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]729
730<div class="doc_code">
731<pre>
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]732@G = constant float 1.0 addrspace(5), section "foo", align 4
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]733</pre>
734</div>
735
736</div>
737
738
739<!-- ======================================================================= -->
740<div class="doc_subsection">
741 <a name="functionstructure">Functions</a>
742</div>
743
744<div class="doc_text">
745
746<p>LLVM function definitions consist of the "<tt>define</tt>" keyord,
747an optional <a href="#linkage">linkage type</a>, an optional
748<a href="#visibility">visibility style</a>, an optional
749<a href="#callingconv">calling convention</a>, a return type, an optional
750<a href="#paramattrs">parameter attribute</a> for the return type, a function
751name, a (possibly empty) argument list (each with optional
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]752<a href="#paramattrs">parameter attributes</a>), optional
753<a href="#fnattrs">function attributes</a>, an optional section,
754an optional alignment, an optional <a href="#gc">garbage collector name</a>,
Chris Lattneree202542008-10-04 18:10:21 +0000[diff] [blame]755an opening curly brace, a list of basic blocks, and a closing curly brace.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]756
757LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
758optional <a href="#linkage">linkage type</a>, an optional
759<a href="#visibility">visibility style</a>, an optional
760<a href="#callingconv">calling convention</a>, a return type, an optional
761<a href="#paramattrs">parameter attribute</a> for the return type, a function
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]762name, a possibly empty list of arguments, an optional alignment, and an optional
Gordon Henriksend606f5b2007-12-10 03:30:21 +0000[diff] [blame]763<a href="#gc">garbage collector name</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]764
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]765<p>A function definition contains a list of basic blocks, forming the CFG
766(Control Flow Graph) for
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]767the function. Each basic block may optionally start with a label (giving the
768basic block a symbol table entry), contains a list of instructions, and ends
769with a <a href="#terminators">terminator</a> instruction (such as a branch or
770function return).</p>
771
772<p>The first basic block in a function is special in two ways: it is immediately
773executed on entrance to the function, and it is not allowed to have predecessor
774basic blocks (i.e. there can not be any branches to the entry block of a
775function). Because the block can have no predecessors, it also cannot have any
776<a href="#i_phi">PHI nodes</a>.</p>
777
778<p>LLVM allows an explicit section to be specified for functions. If the target
779supports it, it will emit functions to the section specified.</p>
780
781<p>An explicit alignment may be specified for a function. If not present, or if
782the alignment is set to zero, the alignment of the function is set by the target
783to whatever it feels convenient. If an explicit alignment is specified, the
784function is forced to have at least that much alignment. All alignments must be
785a power of 2.</p>
786
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]787 <h5>Syntax:</h5>
788
789<div class="doc_code">
Chris Lattner1e5c5cd02008-10-13 16:55:18 +0000[diff] [blame]790<tt>
791define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
792 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
793 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
794 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
795 [<a href="#gc">gc</a>] { ... }
796</tt>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]797</div>
798
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]799</div>
800
801
802<!-- ======================================================================= -->
803<div class="doc_subsection">
804 <a name="aliasstructure">Aliases</a>
805</div>
806<div class="doc_text">
807 <p>Aliases act as "second name" for the aliasee value (which can be either
Anton Korobeynikov96822812008-03-22 08:36:14 +0000[diff] [blame]808 function, global variable, another alias or bitcast of global value). Aliases
809 may have an optional <a href="#linkage">linkage type</a>, and an
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]810 optional <a href="#visibility">visibility style</a>.</p>
811
812 <h5>Syntax:</h5>
813
814<div class="doc_code">
815<pre>
Duncan Sandsd7bfabf2008-09-12 20:48:21 +0000[diff] [blame]816@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]817</pre>
818</div>
819
820</div>
821
822
823
824<!-- ======================================================================= -->
825<div class="doc_subsection"><a name="paramattrs">Parameter Attributes</a></div>
826<div class="doc_text">
827 <p>The return type and each parameter of a function type may have a set of
828 <i>parameter attributes</i> associated with them. Parameter attributes are
829 used to communicate additional information about the result or parameters of
Duncan Sandsf5588dc2007-11-27 13:23:08 +0000[diff] [blame]830 a function. Parameter attributes are considered to be part of the function,
831 not of the function type, so functions with different parameter attributes
832 can have the same function type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]833
834 <p>Parameter attributes are simple keywords that follow the type specified. If
835 multiple parameter attributes are needed, they are space separated. For
836 example:</p>
837
838<div class="doc_code">
839<pre>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]840declare i32 @printf(i8* noalias , ...)
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]841declare i32 @atoi(i8 zeroext)
842declare signext i8 @returns_signed_char()
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]843</pre>
844</div>
845
Duncan Sandsf5588dc2007-11-27 13:23:08 +0000[diff] [blame]846 <p>Note that any attributes for the function result (<tt>nounwind</tt>,
847 <tt>readonly</tt>) come immediately after the argument list.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]848
849 <p>Currently, only the following parameter attributes are defined:</p>
850 <dl>
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]851 <dt><tt>zeroext</tt></dt>
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]852 <dd>This indicates to the code generator that the parameter or return value
853 should be zero-extended to a 32-bit value by the caller (for a parameter)
854 or the callee (for a return value).</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]855
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]856 <dt><tt>signext</tt></dt>
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]857 <dd>This indicates to the code generator that the parameter or return value
858 should be sign-extended to a 32-bit value by the caller (for a parameter)
859 or the callee (for a return value).</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]860
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]861 <dt><tt>inreg</tt></dt>
Dale Johannesenc08a0e22008-09-25 20:47:45 +0000[diff] [blame]862 <dd>This indicates that this parameter or return value should be treated
863 in a special target-dependent fashion during while emitting code for a
864 function call or return (usually, by putting it in a register as opposed
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]865 to memory, though some targets use it to distinguish between two different
866 kinds of registers). Use of this attribute is target-specific.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]867
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]868 <dt><tt><a name="byval">byval</a></tt></dt>
Chris Lattner04c86182008-01-15 04:34:22 +0000[diff] [blame]869 <dd>This indicates that the pointer parameter should really be passed by
870 value to the function. The attribute implies that a hidden copy of the
871 pointee is made between the caller and the callee, so the callee is unable
Chris Lattner6a9f3c42008-08-05 18:21:08 +0000[diff] [blame]872 to modify the value in the callee. This attribute is only valid on LLVM
Chris Lattner04c86182008-01-15 04:34:22 +0000[diff] [blame]873 pointer arguments. It is generally used to pass structs and arrays by
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]874 value, but is also valid on pointers to scalars. The copy is considered to
875 belong to the caller not the callee (for example,
876 <tt><a href="#readonly">readonly</a></tt> functions should not write to
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]877 <tt>byval</tt> parameters). This is not a valid attribute for return
878 values. </dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]879
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]880 <dt><tt>sret</tt></dt>
Duncan Sands616cc032008-02-18 04:19:38 +0000[diff] [blame]881 <dd>This indicates that the pointer parameter specifies the address of a
882 structure that is the return value of the function in the source program.
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]883 This pointer must be guaranteed by the caller to be valid: loads and stores
884 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]885 be applied to the first parameter. This is not a valid attribute for
886 return values. </dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]887
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]888 <dt><tt>noalias</tt></dt>
Owen Andersonc4fc4cd2008-02-18 04:09:01 +0000[diff] [blame]889 <dd>This indicates that the parameter does not alias any global or any other
890 parameter. The caller is responsible for ensuring that this is the case,
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]891 usually by placing the value in a stack allocation. This is not a valid
892 attribute for return values.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]893
Duncan Sands4ee46812007-07-27 19:57:41 +0000[diff] [blame]894 <dt><tt>nest</tt></dt>
Duncan Sandsf1a7d4c2008-07-08 09:27:25 +0000[diff] [blame]895 <dd>This indicates that the pointer parameter can be excised using the
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]896 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
897 attribute for return values.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]898 </dl>
899
900</div>
901
902<!-- ======================================================================= -->
903<div class="doc_subsection">
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]904 <a name="gc">Garbage Collector Names</a>
905</div>
906
907<div class="doc_text">
908<p>Each function may specify a garbage collector name, which is simply a
909string.</p>
910
911<div class="doc_code"><pre
912>define void @f() gc "name" { ...</pre></div>
913
914<p>The compiler declares the supported values of <i>name</i>. Specifying a
915collector which will cause the compiler to alter its output in order to support
916the named garbage collection algorithm.</p>
917</div>
918
919<!-- ======================================================================= -->
920<div class="doc_subsection">
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]921 <a name="fnattrs">Function Attributes</a>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]922</div>
923
924<div class="doc_text">
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]925
926<p>Function attributes are set to communicate additional information about
927 a function. Function attributes are considered to be part of the function,
928 not of the function type, so functions with different parameter attributes
929 can have the same function type.</p>
930
931 <p>Function attributes are simple keywords that follow the type specified. If
932 multiple attributes are needed, they are space separated. For
933 example:</p>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]934
935<div class="doc_code">
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]936<pre>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]937define void @f() noinline { ... }
938define void @f() alwaysinline { ... }
939define void @f() alwaysinline optsize { ... }
940define void @f() optsize
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]941</pre>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]942</div>
943
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]944<dl>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]945<dt><tt>alwaysinline</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]946<dd>This attribute indicates that the inliner should attempt to inline this
947function into callers whenever possible, ignoring any active inlining size
948threshold for this caller.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]949
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]950<dt><tt>noinline</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]951<dd>This attribute indicates that the inliner should never inline this function
Chris Lattner377f3ae2008-10-05 17:14:59 +0000[diff] [blame]952in any situation. This attribute may not be used together with the
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]953<tt>alwaysinline</tt> attribute.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]954
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]955<dt><tt>optsize</tt></dt>
Devang Patelc29099b2008-09-29 18:34:44 +0000[diff] [blame]956<dd>This attribute suggests that optimization passes and code generator passes
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]957make choices that keep the code size of this function low, and otherwise do
958optimizations specifically to reduce code size.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]959
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]960<dt><tt>noreturn</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]961<dd>This function attribute indicates that the function never returns normally.
962This produces undefined behavior at runtime if the function ever does
963dynamically return.</dd>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]964
965<dt><tt>nounwind</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]966<dd>This function attribute indicates that the function never returns with an
967unwind or exceptional control flow. If the function does unwind, its runtime
968behavior is undefined.</dd>
969
970<dt><tt>readnone</tt></dt>
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]971<dd>This attribute indicates that the function computes its result (or the
972exception it throws) based strictly on its arguments, without dereferencing any
973pointer arguments or otherwise accessing any mutable state (e.g. memory, control
974registers, etc) visible to caller functions. It does not write through any
975pointer arguments (including <tt><a href="#byval">byval</a></tt> arguments) and
976never changes any state visible to callers.</dd>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]977
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]978<dt><tt><a name="readonly">readonly</a></tt></dt>
979<dd>This attribute indicates that the function does not write through any
980pointer arguments (including <tt><a href="#byval">byval</a></tt> arguments)
981or otherwise modify any state (e.g. memory, control registers, etc) visible to
982caller functions. It may dereference pointer arguments and read state that may
983be set in the caller. A readonly function always returns the same value (or
984throws the same exception) when called with the same set of arguments and global
985state.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]986</dl>
987
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]988</div>
989
990<!-- ======================================================================= -->
991<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]992 <a name="moduleasm">Module-Level Inline Assembly</a>
993</div>
994
995<div class="doc_text">
996<p>
997Modules may contain "module-level inline asm" blocks, which corresponds to the
998GCC "file scope inline asm" blocks. These blocks are internally concatenated by
999LLVM and treated as a single unit, but may be separated in the .ll file if
1000desired. The syntax is very simple:
1001</p>
1002
1003<div class="doc_code">
1004<pre>
1005module asm "inline asm code goes here"
1006module asm "more can go here"
1007</pre>
1008</div>
1009
1010<p>The strings can contain any character by escaping non-printable characters.
1011 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
1012 for the number.
1013</p>
1014
1015<p>
1016 The inline asm code is simply printed to the machine code .s file when
1017 assembly code is generated.
1018</p>
1019</div>
1020
1021<!-- ======================================================================= -->
1022<div class="doc_subsection">
1023 <a name="datalayout">Data Layout</a>
1024</div>
1025
1026<div class="doc_text">
1027<p>A module may specify a target specific data layout string that specifies how
1028data is to be laid out in memory. The syntax for the data layout is simply:</p>
1029<pre> target datalayout = "<i>layout specification</i>"</pre>
1030<p>The <i>layout specification</i> consists of a list of specifications
1031separated by the minus sign character ('-'). Each specification starts with a
1032letter and may include other information after the letter to define some
1033aspect of the data layout. The specifications accepted are as follows: </p>
1034<dl>
1035 <dt><tt>E</tt></dt>
1036 <dd>Specifies that the target lays out data in big-endian form. That is, the
1037 bits with the most significance have the lowest address location.</dd>
1038 <dt><tt>e</tt></dt>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1039 <dd>Specifies that the target lays out data in little-endian form. That is,
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1040 the bits with the least significance have the lowest address location.</dd>
1041 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1042 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
1043 <i>preferred</i> alignments. All sizes are in bits. Specifying the <i>pref</i>
1044 alignment is optional. If omitted, the preceding <tt>:</tt> should be omitted
1045 too.</dd>
1046 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1047 <dd>This specifies the alignment for an integer type of a given bit
1048 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1049 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1050 <dd>This specifies the alignment for a vector type of a given bit
1051 <i>size</i>.</dd>
1052 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1053 <dd>This specifies the alignment for a floating point type of a given bit
1054 <i>size</i>. The value of <i>size</i> must be either 32 (float) or 64
1055 (double).</dd>
1056 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1057 <dd>This specifies the alignment for an aggregate type of a given bit
1058 <i>size</i>.</dd>
1059</dl>
1060<p>When constructing the data layout for a given target, LLVM starts with a
1061default set of specifications which are then (possibly) overriden by the
1062specifications in the <tt>datalayout</tt> keyword. The default specifications
1063are given in this list:</p>
1064<ul>
1065 <li><tt>E</tt> - big endian</li>
1066 <li><tt>p:32:64:64</tt> - 32-bit pointers with 64-bit alignment</li>
1067 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1068 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1069 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1070 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1071 <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]1072 alignment of 64-bits</li>
1073 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1074 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1075 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1076 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1077 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
1078</ul>
Chris Lattner6a9f3c42008-08-05 18:21:08 +0000[diff] [blame]1079<p>When LLVM is determining the alignment for a given type, it uses the
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1080following rules:
1081<ol>
1082 <li>If the type sought is an exact match for one of the specifications, that
1083 specification is used.</li>
1084 <li>If no match is found, and the type sought is an integer type, then the
1085 smallest integer type that is larger than the bitwidth of the sought type is
1086 used. If none of the specifications are larger than the bitwidth then the the
1087 largest integer type is used. For example, given the default specifications
1088 above, the i7 type will use the alignment of i8 (next largest) while both
1089 i65 and i256 will use the alignment of i64 (largest specified).</li>
1090 <li>If no match is found, and the type sought is a vector type, then the
1091 largest vector type that is smaller than the sought vector type will be used
1092 as a fall back. This happens because <128 x double> can be implemented in
1093 terms of 64 <2 x double>, for example.</li>
1094</ol>
1095</div>
1096
1097<!-- *********************************************************************** -->
1098<div class="doc_section"> <a name="typesystem">Type System</a> </div>
1099<!-- *********************************************************************** -->
1100
1101<div class="doc_text">
1102
1103<p>The LLVM type system is one of the most important features of the
1104intermediate representation. Being typed enables a number of
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1105optimizations to be performed on the intermediate representation directly,
1106without having to do
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1107extra analyses on the side before the transformation. A strong type
1108system makes it easier to read the generated code and enables novel
1109analyses and transformations that are not feasible to perform on normal
1110three address code representations.</p>
1111
1112</div>
1113
1114<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1115<div class="doc_subsection"> <a name="t_classifications">Type
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1116Classifications</a> </div>
1117<div class="doc_text">
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1118<p>The types fall into a few useful
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1119classifications:</p>
1120
1121<table border="1" cellspacing="0" cellpadding="4">
1122 <tbody>
1123 <tr><th>Classification</th><th>Types</th></tr>
1124 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1125 <td><a href="#t_integer">integer</a></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1126 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
1127 </tr>
1128 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1129 <td><a href="#t_floating">floating point</a></td>
1130 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1131 </tr>
1132 <tr>
1133 <td><a name="t_firstclass">first class</a></td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1134 <td><a href="#t_integer">integer</a>,
1135 <a href="#t_floating">floating point</a>,
1136 <a href="#t_pointer">pointer</a>,
Dan Gohmanf6237db2008-06-18 18:42:13 +0000[diff] [blame]1137 <a href="#t_vector">vector</a>,
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]1138 <a href="#t_struct">structure</a>,
1139 <a href="#t_array">array</a>,
Dan Gohmand13951e2008-05-23 22:50:26 +0000[diff] [blame]1140 <a href="#t_label">label</a>.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1141 </td>
1142 </tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1143 <tr>
1144 <td><a href="#t_primitive">primitive</a></td>
1145 <td><a href="#t_label">label</a>,
1146 <a href="#t_void">void</a>,
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1147 <a href="#t_floating">floating point</a>.</td>
1148 </tr>
1149 <tr>
1150 <td><a href="#t_derived">derived</a></td>
1151 <td><a href="#t_integer">integer</a>,
1152 <a href="#t_array">array</a>,
1153 <a href="#t_function">function</a>,
1154 <a href="#t_pointer">pointer</a>,
1155 <a href="#t_struct">structure</a>,
1156 <a href="#t_pstruct">packed structure</a>,
1157 <a href="#t_vector">vector</a>,
1158 <a href="#t_opaque">opaque</a>.
1159 </tr>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1160 </tbody>
1161</table>
1162
1163<p>The <a href="#t_firstclass">first class</a> types are perhaps the
1164most important. Values of these types are the only ones which can be
1165produced by instructions, passed as arguments, or used as operands to
Dan Gohman4f29e422008-05-23 21:53:15 +0000[diff] [blame]1166instructions.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1167</div>
1168
1169<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1170<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1171
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1172<div class="doc_text">
1173<p>The primitive types are the fundamental building blocks of the LLVM
1174system.</p>
1175
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1176</div>
1177
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1178<!-- _______________________________________________________________________ -->
1179<div class="doc_subsubsection"> <a name="t_floating">Floating Point Types</a> </div>
1180
1181<div class="doc_text">
1182 <table>
1183 <tbody>
1184 <tr><th>Type</th><th>Description</th></tr>
1185 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1186 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1187 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1188 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1189 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1190 </tbody>
1191 </table>
1192</div>
1193
1194<!-- _______________________________________________________________________ -->
1195<div class="doc_subsubsection"> <a name="t_void">Void Type</a> </div>
1196
1197<div class="doc_text">
1198<h5>Overview:</h5>
1199<p>The void type does not represent any value and has no size.</p>
1200
1201<h5>Syntax:</h5>
1202
1203<pre>
1204 void
1205</pre>
1206</div>
1207
1208<!-- _______________________________________________________________________ -->
1209<div class="doc_subsubsection"> <a name="t_label">Label Type</a> </div>
1210
1211<div class="doc_text">
1212<h5>Overview:</h5>
1213<p>The label type represents code labels.</p>
1214
1215<h5>Syntax:</h5>
1216
1217<pre>
1218 label
1219</pre>
1220</div>
1221
1222
1223<!-- ======================================================================= -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1224<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
1225
1226<div class="doc_text">
1227
1228<p>The real power in LLVM comes from the derived types in the system.
1229This is what allows a programmer to represent arrays, functions,
1230pointers, and other useful types. Note that these derived types may be
1231recursive: For example, it is possible to have a two dimensional array.</p>
1232
1233</div>
1234
1235<!-- _______________________________________________________________________ -->
1236<div class="doc_subsubsection"> <a name="t_integer">Integer Type</a> </div>
1237
1238<div class="doc_text">
1239
1240<h5>Overview:</h5>
1241<p>The integer type is a very simple derived type that simply specifies an
1242arbitrary bit width for the integer type desired. Any bit width from 1 bit to
12432^23-1 (about 8 million) can be specified.</p>
1244
1245<h5>Syntax:</h5>
1246
1247<pre>
1248 iN
1249</pre>
1250
1251<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1252value.</p>
1253
1254<h5>Examples:</h5>
1255<table class="layout">
Chris Lattner251ab812007-12-18 06:18:21 +0000[diff] [blame]1256 <tbody>
1257 <tr>
1258 <td><tt>i1</tt></td>
1259 <td>a single-bit integer.</td>
1260 </tr><tr>
1261 <td><tt>i32</tt></td>
1262 <td>a 32-bit integer.</td>
1263 </tr><tr>
1264 <td><tt>i1942652</tt></td>
1265 <td>a really big integer of over 1 million bits.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1266 </tr>
Chris Lattner251ab812007-12-18 06:18:21 +0000[diff] [blame]1267 </tbody>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1268</table>
1269</div>
1270
1271<!-- _______________________________________________________________________ -->
1272<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
1273
1274<div class="doc_text">
1275
1276<h5>Overview:</h5>
1277
1278<p>The array type is a very simple derived type that arranges elements
1279sequentially in memory. The array type requires a size (number of
1280elements) and an underlying data type.</p>
1281
1282<h5>Syntax:</h5>
1283
1284<pre>
1285 [&lt;# elements&gt; x &lt;elementtype&gt;]
1286</pre>
1287
1288<p>The number of elements is a constant integer value; elementtype may
1289be any type with a size.</p>
1290
1291<h5>Examples:</h5>
1292<table class="layout">
1293 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1294 <td class="left"><tt>[40 x i32]</tt></td>
1295 <td class="left">Array of 40 32-bit integer values.</td>
1296 </tr>
1297 <tr class="layout">
1298 <td class="left"><tt>[41 x i32]</tt></td>
1299 <td class="left">Array of 41 32-bit integer values.</td>
1300 </tr>
1301 <tr class="layout">
1302 <td class="left"><tt>[4 x i8]</tt></td>
1303 <td class="left">Array of 4 8-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1304 </tr>
1305</table>
1306<p>Here are some examples of multidimensional arrays:</p>
1307<table class="layout">
1308 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1309 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1310 <td class="left">3x4 array of 32-bit integer values.</td>
1311 </tr>
1312 <tr class="layout">
1313 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1314 <td class="left">12x10 array of single precision floating point values.</td>
1315 </tr>
1316 <tr class="layout">
1317 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1318 <td class="left">2x3x4 array of 16-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1319 </tr>
1320</table>
1321
1322<p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
1323length array. Normally, accesses past the end of an array are undefined in
1324LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
1325As a special case, however, zero length arrays are recognized to be variable
1326length. This allows implementation of 'pascal style arrays' with the LLVM
1327type "{ i32, [0 x float]}", for example.</p>
1328
1329</div>
1330
1331<!-- _______________________________________________________________________ -->
1332<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
1333<div class="doc_text">
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1334
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1335<h5>Overview:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1336
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1337<p>The function type can be thought of as a function signature. It
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1338consists of a return type and a list of formal parameter types. The
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1339return 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]1340If 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]1341class types, and the struct must have at least one element.</p>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1342
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1343<h5>Syntax:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1344
1345<pre>
1346 &lt;returntype list&gt; (&lt;parameter list&gt;)
1347</pre>
1348
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1349<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
1350specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1351which indicates that the function takes a variable number of arguments.
1352Variable argument functions can access their arguments with the <a
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1353 href="#int_varargs">variable argument handling intrinsic</a> functions.
1354'<tt>&lt;returntype list&gt;</tt>' is a comma-separated list of
1355<a href="#t_firstclass">first class</a> type specifiers.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1356
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1357<h5>Examples:</h5>
1358<table class="layout">
1359 <tr class="layout">
1360 <td class="left"><tt>i32 (i32)</tt></td>
1361 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
1362 </td>
1363 </tr><tr class="layout">
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]1364 <td class="left"><tt>float&nbsp;(i16&nbsp;signext,&nbsp;i32&nbsp;*)&nbsp;*
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1365 </tt></td>
1366 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
1367 an <tt>i16</tt> that should be sign extended and a
1368 <a href="#t_pointer">pointer</a> to <tt>i32</tt>, returning
1369 <tt>float</tt>.
1370 </td>
1371 </tr><tr class="layout">
1372 <td class="left"><tt>i32 (i8*, ...)</tt></td>
1373 <td class="left">A vararg function that takes at least one
1374 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1375 which returns an integer. This is the signature for <tt>printf</tt> in
1376 LLVM.
1377 </td>
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1378 </tr><tr class="layout">
1379 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Devang Patel6dddba22008-03-24 18:10:52 +0000[diff] [blame]1380 <td class="left">A function taking an <tt>i32></tt>, returning two
1381 <tt> i32 </tt> values as an aggregate of type <tt>{ i32, i32 }</tt>
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1382 </td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1383 </tr>
1384</table>
1385
1386</div>
1387<!-- _______________________________________________________________________ -->
1388<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
1389<div class="doc_text">
1390<h5>Overview:</h5>
1391<p>The structure type is used to represent a collection of data members
1392together in memory. The packing of the field types is defined to match
1393the ABI of the underlying processor. The elements of a structure may
1394be any type that has a size.</p>
1395<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1396and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1397field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1398instruction.</p>
1399<h5>Syntax:</h5>
1400<pre> { &lt;type list&gt; }<br></pre>
1401<h5>Examples:</h5>
1402<table class="layout">
1403 <tr class="layout">
1404 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1405 <td class="left">A triple of three <tt>i32</tt> values</td>
1406 </tr><tr class="layout">
1407 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1408 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1409 second element is a <a href="#t_pointer">pointer</a> to a
1410 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1411 an <tt>i32</tt>.</td>
1412 </tr>
1413</table>
1414</div>
1415
1416<!-- _______________________________________________________________________ -->
1417<div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
1418</div>
1419<div class="doc_text">
1420<h5>Overview:</h5>
1421<p>The packed structure type is used to represent a collection of data members
1422together in memory. There is no padding between fields. Further, the alignment
1423of a packed structure is 1 byte. The elements of a packed structure may
1424be any type that has a size.</p>
1425<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1426and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1427field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1428instruction.</p>
1429<h5>Syntax:</h5>
1430<pre> &lt; { &lt;type list&gt; } &gt; <br></pre>
1431<h5>Examples:</h5>
1432<table class="layout">
1433 <tr class="layout">
1434 <td class="left"><tt>&lt; { i32, i32, i32 } &gt;</tt></td>
1435 <td class="left">A triple of three <tt>i32</tt> values</td>
1436 </tr><tr class="layout">
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1437 <td class="left">
1438<tt>&lt;&nbsp;{&nbsp;float,&nbsp;i32&nbsp;(i32)*&nbsp;}&nbsp;&gt;</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1439 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1440 second element is a <a href="#t_pointer">pointer</a> to a
1441 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1442 an <tt>i32</tt>.</td>
1443 </tr>
1444</table>
1445</div>
1446
1447<!-- _______________________________________________________________________ -->
1448<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
1449<div class="doc_text">
1450<h5>Overview:</h5>
1451<p>As in many languages, the pointer type represents a pointer or
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]1452reference to another object, which must live in memory. Pointer types may have
1453an optional address space attribute defining the target-specific numbered
1454address space where the pointed-to object resides. The default address space is
1455zero.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1456<h5>Syntax:</h5>
1457<pre> &lt;type&gt; *<br></pre>
1458<h5>Examples:</h5>
1459<table class="layout">
1460 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1461 <td class="left"><tt>[4x i32]*</tt></td>
1462 <td class="left">A <a href="#t_pointer">pointer</a> to <a
1463 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
1464 </tr>
1465 <tr class="layout">
1466 <td class="left"><tt>i32 (i32 *) *</tt></td>
1467 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1468 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1469 <tt>i32</tt>.</td>
1470 </tr>
1471 <tr class="layout">
1472 <td class="left"><tt>i32 addrspace(5)*</tt></td>
1473 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
1474 that resides in address space #5.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1475 </tr>
1476</table>
1477</div>
1478
1479<!-- _______________________________________________________________________ -->
1480<div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
1481<div class="doc_text">
1482
1483<h5>Overview:</h5>
1484
1485<p>A vector type is a simple derived type that represents a vector
1486of elements. Vector types are used when multiple primitive data
1487are operated in parallel using a single instruction (SIMD).
1488A vector type requires a size (number of
1489elements) and an underlying primitive data type. Vectors must have a power
1490of two length (1, 2, 4, 8, 16 ...). Vector types are
1491considered <a href="#t_firstclass">first class</a>.</p>
1492
1493<h5>Syntax:</h5>
1494
1495<pre>
1496 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1497</pre>
1498
1499<p>The number of elements is a constant integer value; elementtype may
1500be any integer or floating point type.</p>
1501
1502<h5>Examples:</h5>
1503
1504<table class="layout">
1505 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1506 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
1507 <td class="left">Vector of 4 32-bit integer values.</td>
1508 </tr>
1509 <tr class="layout">
1510 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
1511 <td class="left">Vector of 8 32-bit floating-point values.</td>
1512 </tr>
1513 <tr class="layout">
1514 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
1515 <td class="left">Vector of 2 64-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1516 </tr>
1517</table>
1518</div>
1519
1520<!-- _______________________________________________________________________ -->
1521<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1522<div class="doc_text">
1523
1524<h5>Overview:</h5>
1525
1526<p>Opaque types are used to represent unknown types in the system. This
Gordon Henriksenda0706e2007-10-14 00:34:53 +0000[diff] [blame]1527corresponds (for example) to the C notion of a forward declared structure type.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1528In LLVM, opaque types can eventually be resolved to any type (not just a
1529structure type).</p>
1530
1531<h5>Syntax:</h5>
1532
1533<pre>
1534 opaque
1535</pre>
1536
1537<h5>Examples:</h5>
1538
1539<table class="layout">
1540 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1541 <td class="left"><tt>opaque</tt></td>
1542 <td class="left">An opaque type.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1543 </tr>
1544</table>
1545</div>
1546
1547
1548<!-- *********************************************************************** -->
1549<div class="doc_section"> <a name="constants">Constants</a> </div>
1550<!-- *********************************************************************** -->
1551
1552<div class="doc_text">
1553
1554<p>LLVM has several different basic types of constants. This section describes
1555them all and their syntax.</p>
1556
1557</div>
1558
1559<!-- ======================================================================= -->
1560<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
1561
1562<div class="doc_text">
1563
1564<dl>
1565 <dt><b>Boolean constants</b></dt>
1566
1567 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
1568 constants of the <tt><a href="#t_primitive">i1</a></tt> type.
1569 </dd>
1570
1571 <dt><b>Integer constants</b></dt>
1572
1573 <dd>Standard integers (such as '4') are constants of the <a
1574 href="#t_integer">integer</a> type. Negative numbers may be used with
1575 integer types.
1576 </dd>
1577
1578 <dt><b>Floating point constants</b></dt>
1579
1580 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
1581 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]1582 notation (see below). The assembler requires the exact decimal value of
1583 a floating-point constant. For example, the assembler accepts 1.25 but
1584 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
1585 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1586
1587 <dt><b>Null pointer constants</b></dt>
1588
1589 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
1590 and must be of <a href="#t_pointer">pointer type</a>.</dd>
1591
1592</dl>
1593
1594<p>The one non-intuitive notation for constants is the optional hexadecimal form
1595of floating point constants. For example, the form '<tt>double
15960x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
15974.5e+15</tt>'. The only time hexadecimal floating point constants are required
1598(and the only time that they are generated by the disassembler) is when a
1599floating point constant must be emitted but it cannot be represented as a
1600decimal floating point number. For example, NaN's, infinities, and other
1601special values are represented in their IEEE hexadecimal format so that
1602assembly and disassembly do not cause any bits to change in the constants.</p>
1603
1604</div>
1605
1606<!-- ======================================================================= -->
1607<div class="doc_subsection"><a name="aggregateconstants">Aggregate Constants</a>
1608</div>
1609
1610<div class="doc_text">
1611<p>Aggregate constants arise from aggregation of simple constants
1612and smaller aggregate constants.</p>
1613
1614<dl>
1615 <dt><b>Structure constants</b></dt>
1616
1617 <dd>Structure constants are represented with notation similar to structure
1618 type definitions (a comma separated list of elements, surrounded by braces
Chris Lattner6c8de962007-12-25 20:34:52 +0000[diff] [blame]1619 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
1620 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]1621 must have <a href="#t_struct">structure type</a>, and the number and
1622 types of elements must match those specified by the type.
1623 </dd>
1624
1625 <dt><b>Array constants</b></dt>
1626
1627 <dd>Array constants are represented with notation similar to array type
1628 definitions (a comma separated list of elements, surrounded by square brackets
1629 (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74 ]</tt>". Array
1630 constants must have <a href="#t_array">array type</a>, and the number and
1631 types of elements must match those specified by the type.
1632 </dd>
1633
1634 <dt><b>Vector constants</b></dt>
1635
1636 <dd>Vector constants are represented with notation similar to vector type
1637 definitions (a comma separated list of elements, surrounded by
1638 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32 42,
1639 i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must have <a
1640 href="#t_vector">vector type</a>, and the number and types of elements must
1641 match those specified by the type.
1642 </dd>
1643
1644 <dt><b>Zero initialization</b></dt>
1645
1646 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1647 value to zero of <em>any</em> type, including scalar and aggregate types.
1648 This is often used to avoid having to print large zero initializers (e.g. for
1649 large arrays) and is always exactly equivalent to using explicit zero
1650 initializers.
1651 </dd>
1652</dl>
1653
1654</div>
1655
1656<!-- ======================================================================= -->
1657<div class="doc_subsection">
1658 <a name="globalconstants">Global Variable and Function Addresses</a>
1659</div>
1660
1661<div class="doc_text">
1662
1663<p>The addresses of <a href="#globalvars">global variables</a> and <a
1664href="#functionstructure">functions</a> are always implicitly valid (link-time)
1665constants. These constants are explicitly referenced when the <a
1666href="#identifiers">identifier for the global</a> is used and always have <a
1667href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1668file:</p>
1669
1670<div class="doc_code">
1671<pre>
1672@X = global i32 17
1673@Y = global i32 42
1674@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
1675</pre>
1676</div>
1677
1678</div>
1679
1680<!-- ======================================================================= -->
1681<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
1682<div class="doc_text">
1683 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
1684 no specific value. Undefined values may be of any type and be used anywhere
1685 a constant is permitted.</p>
1686
1687 <p>Undefined values indicate to the compiler that the program is well defined
1688 no matter what value is used, giving the compiler more freedom to optimize.
1689 </p>
1690</div>
1691
1692<!-- ======================================================================= -->
1693<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1694</div>
1695
1696<div class="doc_text">
1697
1698<p>Constant expressions are used to allow expressions involving other constants
1699to be used as constants. Constant expressions may be of any <a
1700href="#t_firstclass">first class</a> type and may involve any LLVM operation
1701that does not have side effects (e.g. load and call are not supported). The
1702following is the syntax for constant expressions:</p>
1703
1704<dl>
1705 <dt><b><tt>trunc ( CST to TYPE )</tt></b></dt>
1706 <dd>Truncate a constant to another type. The bit size of CST must be larger
1707 than the bit size of TYPE. Both types must be integers.</dd>
1708
1709 <dt><b><tt>zext ( CST to TYPE )</tt></b></dt>
1710 <dd>Zero extend a constant to another type. The bit size of CST must be
1711 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1712
1713 <dt><b><tt>sext ( CST to TYPE )</tt></b></dt>
1714 <dd>Sign extend a constant to another type. The bit size of CST must be
1715 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1716
1717 <dt><b><tt>fptrunc ( CST to TYPE )</tt></b></dt>
1718 <dd>Truncate a floating point constant to another floating point type. The
1719 size of CST must be larger than the size of TYPE. Both types must be
1720 floating point.</dd>
1721
1722 <dt><b><tt>fpext ( CST to TYPE )</tt></b></dt>
1723 <dd>Floating point extend a constant to another type. The size of CST must be
1724 smaller or equal to the size of TYPE. Both types must be floating point.</dd>
1725
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]1726 <dt><b><tt>fptoui ( CST to TYPE )</tt></b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1727 <dd>Convert a floating point constant to the corresponding unsigned integer
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1728 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1729 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1730 of the same number of elements. If the value won't fit in the integer type,
1731 the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1732
1733 <dt><b><tt>fptosi ( CST to TYPE )</tt></b></dt>
1734 <dd>Convert a floating point constant to the corresponding signed integer
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1735 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1736 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1737 of the same number of elements. If the value won't fit in the integer type,
1738 the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1739
1740 <dt><b><tt>uitofp ( CST to TYPE )</tt></b></dt>
1741 <dd>Convert an unsigned integer constant to the corresponding floating point
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1742 constant. TYPE must be a scalar or vector floating point type. CST must be of
1743 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1744 of the same number of elements. If the value won't fit in the floating point
1745 type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1746
1747 <dt><b><tt>sitofp ( CST to TYPE )</tt></b></dt>
1748 <dd>Convert a signed integer constant to the corresponding floating point
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1749 constant. TYPE must be a scalar or vector floating point type. CST must be of
1750 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1751 of the same number of elements. If the value won't fit in the floating point
1752 type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1753
1754 <dt><b><tt>ptrtoint ( CST to TYPE )</tt></b></dt>
1755 <dd>Convert a pointer typed constant to the corresponding integer constant
1756 TYPE must be an integer type. CST must be of pointer type. The CST value is
1757 zero extended, truncated, or unchanged to make it fit in TYPE.</dd>
1758
1759 <dt><b><tt>inttoptr ( CST to TYPE )</tt></b></dt>
1760 <dd>Convert a integer constant to a pointer constant. TYPE must be a
1761 pointer type. CST must be of integer type. The CST value is zero extended,
1762 truncated, or unchanged to make it fit in a pointer size. This one is
1763 <i>really</i> dangerous!</dd>
1764
1765 <dt><b><tt>bitcast ( CST to TYPE )</tt></b></dt>
1766 <dd>Convert a constant, CST, to another TYPE. The size of CST and TYPE must be
1767 identical (same number of bits). The conversion is done as if the CST value
1768 was stored to memory and read back as TYPE. In other words, no bits change
1769 with this operator, just the type. This can be used for conversion of
1770 vector types to any other type, as long as they have the same bit width. For
Dan Gohman7305fa02008-09-08 16:45:59 +0000[diff] [blame]1771 pointers it is only valid to cast to another pointer type. It is not valid
1772 to bitcast to or from an aggregate type.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1773 </dd>
1774
1775 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
1776
1777 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
1778 constants. As with the <a href="#i_getelementptr">getelementptr</a>
1779 instruction, the index list may have zero or more indexes, which are required
1780 to make sense for the type of "CSTPTR".</dd>
1781
1782 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
1783
1784 <dd>Perform the <a href="#i_select">select operation</a> on
1785 constants.</dd>
1786
1787 <dt><b><tt>icmp COND ( VAL1, VAL2 )</tt></b></dt>
1788 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
1789
1790 <dt><b><tt>fcmp COND ( VAL1, VAL2 )</tt></b></dt>
1791 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
1792
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]1793 <dt><b><tt>vicmp COND ( VAL1, VAL2 )</tt></b></dt>
1794 <dd>Performs the <a href="#i_vicmp">vicmp operation</a> on constants.</dd>
1795
1796 <dt><b><tt>vfcmp COND ( VAL1, VAL2 )</tt></b></dt>
1797 <dd>Performs the <a href="#i_vfcmp">vfcmp operation</a> on constants.</dd>
1798
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1799 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
1800
1801 <dd>Perform the <a href="#i_extractelement">extractelement
1802 operation</a> on constants.
1803
1804 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
1805
1806 <dd>Perform the <a href="#i_insertelement">insertelement
1807 operation</a> on constants.</dd>
1808
1809
1810 <dt><b><tt>shufflevector ( VEC1, VEC2, IDXMASK )</tt></b></dt>
1811
1812 <dd>Perform the <a href="#i_shufflevector">shufflevector
1813 operation</a> on constants.</dd>
1814
1815 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
1816
1817 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
1818 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
1819 binary</a> operations. The constraints on operands are the same as those for
1820 the corresponding instruction (e.g. no bitwise operations on floating point
1821 values are allowed).</dd>
1822</dl>
1823</div>
1824
1825<!-- *********************************************************************** -->
1826<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
1827<!-- *********************************************************************** -->
1828
1829<!-- ======================================================================= -->
1830<div class="doc_subsection">
1831<a name="inlineasm">Inline Assembler Expressions</a>
1832</div>
1833
1834<div class="doc_text">
1835
1836<p>
1837LLVM supports inline assembler expressions (as opposed to <a href="#moduleasm">
1838Module-Level Inline Assembly</a>) through the use of a special value. This
1839value represents the inline assembler as a string (containing the instructions
1840to emit), a list of operand constraints (stored as a string), and a flag that
1841indicates whether or not the inline asm expression has side effects. An example
1842inline assembler expression is:
1843</p>
1844
1845<div class="doc_code">
1846<pre>
1847i32 (i32) asm "bswap $0", "=r,r"
1848</pre>
1849</div>
1850
1851<p>
1852Inline assembler expressions may <b>only</b> be used as the callee operand of
1853a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we have:
1854</p>
1855
1856<div class="doc_code">
1857<pre>
1858%X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
1859</pre>
1860</div>
1861
1862<p>
1863Inline asms with side effects not visible in the constraint list must be marked
1864as having side effects. This is done through the use of the
1865'<tt>sideeffect</tt>' keyword, like so:
1866</p>
1867
1868<div class="doc_code">
1869<pre>
1870call void asm sideeffect "eieio", ""()
1871</pre>
1872</div>
1873
1874<p>TODO: The format of the asm and constraints string still need to be
1875documented here. Constraints on what can be done (e.g. duplication, moving, etc
Chris Lattner1cdd64e2008-10-04 18:36:02 +0000[diff] [blame]1876need to be documented). This is probably best done by reference to another
1877document that covers inline asm from a holistic perspective.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1878</p>
1879
1880</div>
1881
1882<!-- *********************************************************************** -->
1883<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
1884<!-- *********************************************************************** -->
1885
1886<div class="doc_text">
1887
1888<p>The LLVM instruction set consists of several different
1889classifications of instructions: <a href="#terminators">terminator
1890instructions</a>, <a href="#binaryops">binary instructions</a>,
1891<a href="#bitwiseops">bitwise binary instructions</a>, <a
1892 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
1893instructions</a>.</p>
1894
1895</div>
1896
1897<!-- ======================================================================= -->
1898<div class="doc_subsection"> <a name="terminators">Terminator
1899Instructions</a> </div>
1900
1901<div class="doc_text">
1902
1903<p>As mentioned <a href="#functionstructure">previously</a>, every
1904basic block in a program ends with a "Terminator" instruction, which
1905indicates which block should be executed after the current block is
1906finished. These terminator instructions typically yield a '<tt>void</tt>'
1907value: they produce control flow, not values (the one exception being
1908the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
1909<p>There are six different terminator instructions: the '<a
1910 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
1911instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
1912the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
1913 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
1914 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
1915
1916</div>
1917
1918<!-- _______________________________________________________________________ -->
1919<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
1920Instruction</a> </div>
1921<div class="doc_text">
1922<h5>Syntax:</h5>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]1923<pre>
1924 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]1925 ret void <i>; Return from void function</i>
1926</pre>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1927
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1928<h5>Overview:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1929
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]1930<p>The '<tt>ret</tt>' instruction is used to return control flow (and
1931optionally a value) from a function back to the caller.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1932<p>There are two forms of the '<tt>ret</tt>' instruction: one that
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]1933returns a value and then causes control flow, and one that just causes
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1934control flow to occur.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1935
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1936<h5>Arguments:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1937
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]1938<p>The '<tt>ret</tt>' instruction optionally accepts a single argument,
1939the return value. The type of the return value must be a
1940'<a href="#t_firstclass">first class</a>' type.</p>
1941
1942<p>A function is not <a href="#wellformed">well formed</a> if
1943it it has a non-void return type and contains a '<tt>ret</tt>'
1944instruction with no return value or a return value with a type that
1945does not match its type, or if it has a void return type and contains
1946a '<tt>ret</tt>' instruction with a return value.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1947
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1948<h5>Semantics:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1949
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1950<p>When the '<tt>ret</tt>' instruction is executed, control flow
1951returns back to the calling function's context. If the caller is a "<a
1952 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
1953the instruction after the call. If the caller was an "<a
1954 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
1955at the beginning of the "normal" destination block. If the instruction
1956returns a value, that value shall set the call or invoke instruction's
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]1957return value.
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1958
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1959<h5>Example:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1960
1961<pre>
1962 ret i32 5 <i>; Return an integer value of 5</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1963 ret void <i>; Return from a void function</i>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]1964 ret { i32, i8 } { i32 4, i8 2 } <i>; Return an aggregate of values 4 and 2</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1965</pre>
1966</div>
1967<!-- _______________________________________________________________________ -->
1968<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
1969<div class="doc_text">
1970<h5>Syntax:</h5>
1971<pre> br i1 &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;<br> br label &lt;dest&gt; <i>; Unconditional branch</i>
1972</pre>
1973<h5>Overview:</h5>
1974<p>The '<tt>br</tt>' instruction is used to cause control flow to
1975transfer to a different basic block in the current function. There are
1976two forms of this instruction, corresponding to a conditional branch
1977and an unconditional branch.</p>
1978<h5>Arguments:</h5>
1979<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
1980single '<tt>i1</tt>' value and two '<tt>label</tt>' values. The
1981unconditional form of the '<tt>br</tt>' instruction takes a single
1982'<tt>label</tt>' value as a target.</p>
1983<h5>Semantics:</h5>
1984<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
1985argument is evaluated. If the value is <tt>true</tt>, control flows
1986to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
1987control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
1988<h5>Example:</h5>
1989<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
1990 href="#i_ret">ret</a> i32 1<br>IfUnequal:<br> <a href="#i_ret">ret</a> i32 0<br></pre>
1991</div>
1992<!-- _______________________________________________________________________ -->
1993<div class="doc_subsubsection">
1994 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
1995</div>
1996
1997<div class="doc_text">
1998<h5>Syntax:</h5>
1999
2000<pre>
2001 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
2002</pre>
2003
2004<h5>Overview:</h5>
2005
2006<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
2007several different places. It is a generalization of the '<tt>br</tt>'
2008instruction, allowing a branch to occur to one of many possible
2009destinations.</p>
2010
2011
2012<h5>Arguments:</h5>
2013
2014<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
2015comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
2016an array of pairs of comparison value constants and '<tt>label</tt>'s. The
2017table is not allowed to contain duplicate constant entries.</p>
2018
2019<h5>Semantics:</h5>
2020
2021<p>The <tt>switch</tt> instruction specifies a table of values and
2022destinations. When the '<tt>switch</tt>' instruction is executed, this
2023table is searched for the given value. If the value is found, control flow is
2024transfered to the corresponding destination; otherwise, control flow is
2025transfered to the default destination.</p>
2026
2027<h5>Implementation:</h5>
2028
2029<p>Depending on properties of the target machine and the particular
2030<tt>switch</tt> instruction, this instruction may be code generated in different
2031ways. For example, it could be generated as a series of chained conditional
2032branches or with a lookup table.</p>
2033
2034<h5>Example:</h5>
2035
2036<pre>
2037 <i>; Emulate a conditional br instruction</i>
2038 %Val = <a href="#i_zext">zext</a> i1 %value to i32
2039 switch i32 %Val, label %truedest [i32 0, label %falsedest ]
2040
2041 <i>; Emulate an unconditional br instruction</i>
2042 switch i32 0, label %dest [ ]
2043
2044 <i>; Implement a jump table:</i>
2045 switch i32 %val, label %otherwise [ i32 0, label %onzero
2046 i32 1, label %onone
2047 i32 2, label %ontwo ]
2048</pre>
2049</div>
2050
2051<!-- _______________________________________________________________________ -->
2052<div class="doc_subsubsection">
2053 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
2054</div>
2055
2056<div class="doc_text">
2057
2058<h5>Syntax:</h5>
2059
2060<pre>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]2061 &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]2062 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
2063</pre>
2064
2065<h5>Overview:</h5>
2066
2067<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
2068function, with the possibility of control flow transfer to either the
2069'<tt>normal</tt>' label or the
2070'<tt>exception</tt>' label. If the callee function returns with the
2071"<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
2072"normal" label. If the callee (or any indirect callees) returns with the "<a
2073href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2074continued at the dynamically nearest "exception" label.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2075
2076<h5>Arguments:</h5>
2077
2078<p>This instruction requires several arguments:</p>
2079
2080<ol>
2081 <li>
2082 The optional "cconv" marker indicates which <a href="#callingconv">calling
2083 convention</a> the call should use. If none is specified, the call defaults
2084 to using C calling conventions.
2085 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]2086
2087 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
2088 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>',
2089 and '<tt>inreg</tt>' attributes are valid here.</li>
2090
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2091 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
2092 function value being invoked. In most cases, this is a direct function
2093 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
2094 an arbitrary pointer to function value.
2095 </li>
2096
2097 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
2098 function to be invoked. </li>
2099
2100 <li>'<tt>function args</tt>': argument list whose types match the function
2101 signature argument types. If the function signature indicates the function
2102 accepts a variable number of arguments, the extra arguments can be
2103 specified. </li>
2104
2105 <li>'<tt>normal label</tt>': the label reached when the called function
2106 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
2107
2108 <li>'<tt>exception label</tt>': the label reached when a callee returns with
2109 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
2110
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]2111 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]2112 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
2113 '<tt>readnone</tt>' attributes are valid here.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2114</ol>
2115
2116<h5>Semantics:</h5>
2117
2118<p>This instruction is designed to operate as a standard '<tt><a
2119href="#i_call">call</a></tt>' instruction in most regards. The primary
2120difference is that it establishes an association with a label, which is used by
2121the runtime library to unwind the stack.</p>
2122
2123<p>This instruction is used in languages with destructors to ensure that proper
2124cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
2125exception. Additionally, this is important for implementation of
2126'<tt>catch</tt>' clauses in high-level languages that support them.</p>
2127
2128<h5>Example:</h5>
2129<pre>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2130 %retval = invoke i32 @Test(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2131 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2132 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2133 unwind label %TestCleanup <i>; {i32}:retval set</i>
2134</pre>
2135</div>
2136
2137
2138<!-- _______________________________________________________________________ -->
2139
2140<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
2141Instruction</a> </div>
2142
2143<div class="doc_text">
2144
2145<h5>Syntax:</h5>
2146<pre>
2147 unwind
2148</pre>
2149
2150<h5>Overview:</h5>
2151
2152<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
2153at the first callee in the dynamic call stack which used an <a
2154href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
2155primarily used to implement exception handling.</p>
2156
2157<h5>Semantics:</h5>
2158
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]2159<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2160immediately halt. The dynamic call stack is then searched for the first <a
2161href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
2162execution continues at the "exceptional" destination block specified by the
2163<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
2164dynamic call chain, undefined behavior results.</p>
2165</div>
2166
2167<!-- _______________________________________________________________________ -->
2168
2169<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
2170Instruction</a> </div>
2171
2172<div class="doc_text">
2173
2174<h5>Syntax:</h5>
2175<pre>
2176 unreachable
2177</pre>
2178
2179<h5>Overview:</h5>
2180
2181<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
2182instruction is used to inform the optimizer that a particular portion of the
2183code is not reachable. This can be used to indicate that the code after a
2184no-return function cannot be reached, and other facts.</p>
2185
2186<h5>Semantics:</h5>
2187
2188<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
2189</div>
2190
2191
2192
2193<!-- ======================================================================= -->
2194<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
2195<div class="doc_text">
2196<p>Binary operators are used to do most of the computation in a
Chris Lattnerab596d92008-04-01 18:47:32 +0000[diff] [blame]2197program. They require two operands of the same type, execute an operation on them, and
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2198produce a single value. The operands might represent
2199multiple 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]2200The result value has the same type as its operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2201<p>There are several different binary operators:</p>
2202</div>
2203<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2204<div class="doc_subsubsection">
2205 <a name="i_add">'<tt>add</tt>' Instruction</a>
2206</div>
2207
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2208<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2209
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2210<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2211
2212<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2213 &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]2214</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2215
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2216<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2217
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2218<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2219
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2220<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2221
2222<p>The two arguments to the '<tt>add</tt>' instruction must be <a
2223 href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>, or
2224 <a href="#t_vector">vector</a> values. Both arguments must have identical
2225 types.</p>
2226
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2227<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2228
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2229<p>The value produced is the integer or floating point sum of the two
2230operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2231
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2232<p>If an integer sum has unsigned overflow, the result returned is the
2233mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2234the result.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2235
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2236<p>Because LLVM integers use a two's complement representation, this
2237instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2238
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2239<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2240
2241<pre>
2242 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2243</pre>
2244</div>
2245<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2246<div class="doc_subsubsection">
2247 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
2248</div>
2249
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2250<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2251
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2252<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2253
2254<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2255 &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]2256</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2257
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2258<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2259
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2260<p>The '<tt>sub</tt>' instruction returns the difference of its two
2261operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2262
2263<p>Note that the '<tt>sub</tt>' instruction is used to represent the
2264'<tt>neg</tt>' instruction present in most other intermediate
2265representations.</p>
2266
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2267<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2268
2269<p>The two arguments to the '<tt>sub</tt>' instruction must be <a
2270 href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>,
2271 or <a href="#t_vector">vector</a> values. Both arguments must have identical
2272 types.</p>
2273
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2274<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2275
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2276<p>The value produced is the integer or floating point difference of
2277the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2278
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2279<p>If an integer difference has unsigned overflow, the result returned is the
2280mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2281the result.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2282
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2283<p>Because LLVM integers use a two's complement representation, this
2284instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2285
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2286<h5>Example:</h5>
2287<pre>
2288 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
2289 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
2290</pre>
2291</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2292
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2293<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2294<div class="doc_subsubsection">
2295 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
2296</div>
2297
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2298<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2299
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2300<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2301<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]2302</pre>
2303<h5>Overview:</h5>
2304<p>The '<tt>mul</tt>' instruction returns the product of its two
2305operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2306
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2307<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2308
2309<p>The two arguments to the '<tt>mul</tt>' instruction must be <a
2310href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>,
2311or <a href="#t_vector">vector</a> values. Both arguments must have identical
2312types.</p>
2313
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2314<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2315
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2316<p>The value produced is the integer or floating point product of the
2317two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2318
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2319<p>If the result of an integer multiplication has unsigned overflow,
2320the result returned is the mathematical result modulo
23212<sup>n</sup>, where n is the bit width of the result.</p>
2322<p>Because LLVM integers use a two's complement representation, and the
2323result is the same width as the operands, this instruction returns the
2324correct result for both signed and unsigned integers. If a full product
2325(e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands
2326should be sign-extended or zero-extended as appropriate to the
2327width of the full product.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2328<h5>Example:</h5>
2329<pre> &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
2330</pre>
2331</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2332
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2333<!-- _______________________________________________________________________ -->
2334<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
2335</a></div>
2336<div class="doc_text">
2337<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2338<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]2339</pre>
2340<h5>Overview:</h5>
2341<p>The '<tt>udiv</tt>' instruction returns the quotient of its two
2342operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2343
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2344<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2345
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2346<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2347<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2348values. Both arguments must have identical types.</p>
2349
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2350<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2351
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2352<p>The value produced is the unsigned integer quotient of the two operands.</p>
2353<p>Note that unsigned integer division and signed integer division are distinct
2354operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
2355<p>Division by zero leads to undefined behavior.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2356<h5>Example:</h5>
2357<pre> &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2358</pre>
2359</div>
2360<!-- _______________________________________________________________________ -->
2361<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
2362</a> </div>
2363<div class="doc_text">
2364<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2365<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2366 &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]2367</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2368
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2369<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2370
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2371<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two
2372operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2373
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2374<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2375
2376<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
2377<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2378values. Both arguments must have identical types.</p>
2379
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2380<h5>Semantics:</h5>
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2381<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]2382<p>Note that signed integer division and unsigned integer division are distinct
2383operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
2384<p>Division by zero leads to undefined behavior. Overflow also leads to
2385undefined behavior; this is a rare case, but can occur, for example,
2386by doing a 32-bit division of -2147483648 by -1.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2387<h5>Example:</h5>
2388<pre> &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2389</pre>
2390</div>
2391<!-- _______________________________________________________________________ -->
2392<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
2393Instruction</a> </div>
2394<div class="doc_text">
2395<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2396<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2397 &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]2398</pre>
2399<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2400
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2401<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two
2402operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2403
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2404<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2405
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2406<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2407<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2408of floating point values. Both arguments must have identical types.</p>
2409
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2410<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2411
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2412<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2413
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2414<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2415
2416<pre>
2417 &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]2418</pre>
2419</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2420
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2421<!-- _______________________________________________________________________ -->
2422<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
2423</div>
2424<div class="doc_text">
2425<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2426<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]2427</pre>
2428<h5>Overview:</h5>
2429<p>The '<tt>urem</tt>' instruction returns the remainder from the
2430unsigned division of its two arguments.</p>
2431<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2432<p>The two arguments to the '<tt>urem</tt>' instruction must be
2433<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2434values. Both arguments must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2435<h5>Semantics:</h5>
2436<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2437This instruction always performs an unsigned division to get the remainder.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2438<p>Note that unsigned integer remainder and signed integer remainder are
2439distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
2440<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2441<h5>Example:</h5>
2442<pre> &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2443</pre>
2444
2445</div>
2446<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2447<div class="doc_subsubsection">
2448 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
2449</div>
2450
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2451<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2452
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2453<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2454
2455<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2456 &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]2457</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2458
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2459<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2460
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2461<p>The '<tt>srem</tt>' instruction returns the remainder from the
Dan Gohman3e3fd8c2007-11-05 23:35:22 +0000[diff] [blame]2462signed division of its two operands. This instruction can also take
2463<a href="#t_vector">vector</a> versions of the values in which case
2464the elements must be integers.</p>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]2465
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2466<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2467
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2468<p>The two arguments to the '<tt>srem</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2469<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2470values. Both arguments must have identical types.</p>
2471
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2472<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2473
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2474<p>This instruction returns the <i>remainder</i> of a division (where the result
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2475has the same sign as the dividend, <tt>op1</tt>), not the <i>modulo</i>
2476operator (where the result has the same sign as the divisor, <tt>op2</tt>) of
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2477a value. For more information about the difference, see <a
2478 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
2479Math Forum</a>. For a table of how this is implemented in various languages,
2480please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
2481Wikipedia: modulo operation</a>.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2482<p>Note that signed integer remainder and unsigned integer remainder are
2483distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
2484<p>Taking the remainder of a division by zero leads to undefined behavior.
2485Overflow also leads to undefined behavior; this is a rare case, but can occur,
2486for example, by taking the remainder of a 32-bit division of -2147483648 by -1.
2487(The remainder doesn't actually overflow, but this rule lets srem be
2488implemented using instructions that return both the result of the division
2489and the remainder.)</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2490<h5>Example:</h5>
2491<pre> &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2492</pre>
2493
2494</div>
2495<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2496<div class="doc_subsubsection">
2497 <a name="i_frem">'<tt>frem</tt>' Instruction</a> </div>
2498
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2499<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2500
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2501<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2502<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]2503</pre>
2504<h5>Overview:</h5>
2505<p>The '<tt>frem</tt>' instruction returns the remainder from the
2506division of its two operands.</p>
2507<h5>Arguments:</h5>
2508<p>The two arguments to the '<tt>frem</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2509<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2510of floating point values. Both arguments must have identical types.</p>
2511
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2512<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2513
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2514<p>This instruction returns the <i>remainder</i> of a division.
2515The remainder has the same sign as the dividend.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2516
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2517<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2518
2519<pre>
2520 &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]2521</pre>
2522</div>
2523
2524<!-- ======================================================================= -->
2525<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
2526Operations</a> </div>
2527<div class="doc_text">
2528<p>Bitwise binary operators are used to do various forms of
2529bit-twiddling in a program. They are generally very efficient
2530instructions and can commonly be strength reduced from other
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2531instructions. They require two operands of the same type, execute an operation on them,
2532and 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]2533</div>
2534
2535<!-- _______________________________________________________________________ -->
2536<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
2537Instruction</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; = 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]2541</pre>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2542
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2543<h5>Overview:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2544
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2545<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
2546the left a specified number of bits.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2547
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2548<h5>Arguments:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2549
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2550<p>Both arguments to the '<tt>shl</tt>' instruction must be the same <a
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]2551 href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2552type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2553
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2554<h5>Semantics:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2555
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2556<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod 2<sup>n</sup>,
2557where n is the width of the result. If <tt>op2</tt> is (statically or dynamically) negative or
2558equal to or larger than the number of bits in <tt>op1</tt>, the result is undefined.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2559
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2560<h5>Example:</h5><pre>
2561 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
2562 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
2563 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2564 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2565</pre>
2566</div>
2567<!-- _______________________________________________________________________ -->
2568<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
2569Instruction</a> </div>
2570<div class="doc_text">
2571<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2572<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]2573</pre>
2574
2575<h5>Overview:</h5>
2576<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
2577operand shifted to the right a specified number of bits with zero fill.</p>
2578
2579<h5>Arguments:</h5>
2580<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]2581<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]2582type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2583
2584<h5>Semantics:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2585
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2586<p>This instruction always performs a logical shift right operation. The most
2587significant bits of the result will be filled with zero bits after the
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2588shift. If <tt>op2</tt> is (statically or dynamically) equal to or larger than
2589the number of bits in <tt>op1</tt>, the result is undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2590
2591<h5>Example:</h5>
2592<pre>
2593 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
2594 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
2595 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
2596 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2597 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2598</pre>
2599</div>
2600
2601<!-- _______________________________________________________________________ -->
2602<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
2603Instruction</a> </div>
2604<div class="doc_text">
2605
2606<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2607<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]2608</pre>
2609
2610<h5>Overview:</h5>
2611<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
2612operand shifted to the right a specified number of bits with sign extension.</p>
2613
2614<h5>Arguments:</h5>
2615<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]2616<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]2617type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2618
2619<h5>Semantics:</h5>
2620<p>This instruction always performs an arithmetic shift right operation,
2621The most significant bits of the result will be filled with the sign bit
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2622of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
2623larger than the number of bits in <tt>op1</tt>, the result is undefined.
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2624</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2625
2626<h5>Example:</h5>
2627<pre>
2628 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
2629 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
2630 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
2631 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2632 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2633</pre>
2634</div>
2635
2636<!-- _______________________________________________________________________ -->
2637<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
2638Instruction</a> </div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2639
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2640<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2641
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2642<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2643
2644<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2645 &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]2646</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2647
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2648<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2649
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2650<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
2651its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2652
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2653<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2654
2655<p>The two arguments to the '<tt>and</tt>' instruction must be
2656<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2657values. Both arguments must have identical types.</p>
2658
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2659<h5>Semantics:</h5>
2660<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
2661<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]2662<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2663<table border="1" cellspacing="0" cellpadding="4">
2664 <tbody>
2665 <tr>
2666 <td>In0</td>
2667 <td>In1</td>
2668 <td>Out</td>
2669 </tr>
2670 <tr>
2671 <td>0</td>
2672 <td>0</td>
2673 <td>0</td>
2674 </tr>
2675 <tr>
2676 <td>0</td>
2677 <td>1</td>
2678 <td>0</td>
2679 </tr>
2680 <tr>
2681 <td>1</td>
2682 <td>0</td>
2683 <td>0</td>
2684 </tr>
2685 <tr>
2686 <td>1</td>
2687 <td>1</td>
2688 <td>1</td>
2689 </tr>
2690 </tbody>
2691</table>
2692</div>
2693<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2694<pre>
2695 &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]2696 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
2697 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
2698</pre>
2699</div>
2700<!-- _______________________________________________________________________ -->
2701<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
2702<div class="doc_text">
2703<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2704<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]2705</pre>
2706<h5>Overview:</h5>
2707<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
2708or of its two operands.</p>
2709<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2710
2711<p>The two arguments to the '<tt>or</tt>' instruction must be
2712<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2713values. Both arguments must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2714<h5>Semantics:</h5>
2715<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
2716<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]2717<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2718<table border="1" cellspacing="0" cellpadding="4">
2719 <tbody>
2720 <tr>
2721 <td>In0</td>
2722 <td>In1</td>
2723 <td>Out</td>
2724 </tr>
2725 <tr>
2726 <td>0</td>
2727 <td>0</td>
2728 <td>0</td>
2729 </tr>
2730 <tr>
2731 <td>0</td>
2732 <td>1</td>
2733 <td>1</td>
2734 </tr>
2735 <tr>
2736 <td>1</td>
2737 <td>0</td>
2738 <td>1</td>
2739 </tr>
2740 <tr>
2741 <td>1</td>
2742 <td>1</td>
2743 <td>1</td>
2744 </tr>
2745 </tbody>
2746</table>
2747</div>
2748<h5>Example:</h5>
2749<pre> &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
2750 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
2751 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
2752</pre>
2753</div>
2754<!-- _______________________________________________________________________ -->
2755<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
2756Instruction</a> </div>
2757<div class="doc_text">
2758<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2759<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]2760</pre>
2761<h5>Overview:</h5>
2762<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
2763or of its two operands. The <tt>xor</tt> is used to implement the
2764"one's complement" operation, which is the "~" operator in C.</p>
2765<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2766<p>The two arguments to the '<tt>xor</tt>' instruction must be
2767<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2768values. Both arguments must have identical types.</p>
2769
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2770<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2771
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2772<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
2773<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]2774<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2775<table border="1" cellspacing="0" cellpadding="4">
2776 <tbody>
2777 <tr>
2778 <td>In0</td>
2779 <td>In1</td>
2780 <td>Out</td>
2781 </tr>
2782 <tr>
2783 <td>0</td>
2784 <td>0</td>
2785 <td>0</td>
2786 </tr>
2787 <tr>
2788 <td>0</td>
2789 <td>1</td>
2790 <td>1</td>
2791 </tr>
2792 <tr>
2793 <td>1</td>
2794 <td>0</td>
2795 <td>1</td>
2796 </tr>
2797 <tr>
2798 <td>1</td>
2799 <td>1</td>
2800 <td>0</td>
2801 </tr>
2802 </tbody>
2803</table>
2804</div>
2805<p> </p>
2806<h5>Example:</h5>
2807<pre> &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
2808 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
2809 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
2810 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
2811</pre>
2812</div>
2813
2814<!-- ======================================================================= -->
2815<div class="doc_subsection">
2816 <a name="vectorops">Vector Operations</a>
2817</div>
2818
2819<div class="doc_text">
2820
2821<p>LLVM supports several instructions to represent vector operations in a
2822target-independent manner. These instructions cover the element-access and
2823vector-specific operations needed to process vectors effectively. While LLVM
2824does directly support these vector operations, many sophisticated algorithms
2825will want to use target-specific intrinsics to take full advantage of a specific
2826target.</p>
2827
2828</div>
2829
2830<!-- _______________________________________________________________________ -->
2831<div class="doc_subsubsection">
2832 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
2833</div>
2834
2835<div class="doc_text">
2836
2837<h5>Syntax:</h5>
2838
2839<pre>
2840 &lt;result&gt; = extractelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, i32 &lt;idx&gt; <i>; yields &lt;ty&gt;</i>
2841</pre>
2842
2843<h5>Overview:</h5>
2844
2845<p>
2846The '<tt>extractelement</tt>' instruction extracts a single scalar
2847element from a vector at a specified index.
2848</p>
2849
2850
2851<h5>Arguments:</h5>
2852
2853<p>
2854The first operand of an '<tt>extractelement</tt>' instruction is a
2855value of <a href="#t_vector">vector</a> type. The second operand is
2856an index indicating the position from which to extract the element.
2857The index may be a variable.</p>
2858
2859<h5>Semantics:</h5>
2860
2861<p>
2862The result is a scalar of the same type as the element type of
2863<tt>val</tt>. Its value is the value at position <tt>idx</tt> of
2864<tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
2865results are undefined.
2866</p>
2867
2868<h5>Example:</h5>
2869
2870<pre>
2871 %result = extractelement &lt;4 x i32&gt; %vec, i32 0 <i>; yields i32</i>
2872</pre>
2873</div>
2874
2875
2876<!-- _______________________________________________________________________ -->
2877<div class="doc_subsubsection">
2878 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
2879</div>
2880
2881<div class="doc_text">
2882
2883<h5>Syntax:</h5>
2884
2885<pre>
Dan Gohmanbcc3c502008-05-12 23:38:42 +0000[diff] [blame]2886 &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]2887</pre>
2888
2889<h5>Overview:</h5>
2890
2891<p>
2892The '<tt>insertelement</tt>' instruction inserts a scalar
2893element into a vector at a specified index.
2894</p>
2895
2896
2897<h5>Arguments:</h5>
2898
2899<p>
2900The first operand of an '<tt>insertelement</tt>' instruction is a
2901value of <a href="#t_vector">vector</a> type. The second operand is a
2902scalar value whose type must equal the element type of the first
2903operand. The third operand is an index indicating the position at
2904which to insert the value. The index may be a variable.</p>
2905
2906<h5>Semantics:</h5>
2907
2908<p>
2909The result is a vector of the same type as <tt>val</tt>. Its
2910element values are those of <tt>val</tt> except at position
2911<tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
2912exceeds the length of <tt>val</tt>, the results are undefined.
2913</p>
2914
2915<h5>Example:</h5>
2916
2917<pre>
2918 %result = insertelement &lt;4 x i32&gt; %vec, i32 1, i32 0 <i>; yields &lt;4 x i32&gt;</i>
2919</pre>
2920</div>
2921
2922<!-- _______________________________________________________________________ -->
2923<div class="doc_subsubsection">
2924 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
2925</div>
2926
2927<div class="doc_text">
2928
2929<h5>Syntax:</h5>
2930
2931<pre>
2932 &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;n x i32&gt; &lt;mask&gt; <i>; yields &lt;n x &lt;ty&gt;&gt;</i>
2933</pre>
2934
2935<h5>Overview:</h5>
2936
2937<p>
2938The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
2939from two input vectors, returning a vector of the same type.
2940</p>
2941
2942<h5>Arguments:</h5>
2943
2944<p>
2945The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
2946with types that match each other and types that match the result of the
2947instruction. The third argument is a shuffle mask, which has the same number
2948of elements as the other vector type, but whose element type is always 'i32'.
2949</p>
2950
2951<p>
2952The shuffle mask operand is required to be a constant vector with either
2953constant integer or undef values.
2954</p>
2955
2956<h5>Semantics:</h5>
2957
2958<p>
2959The elements of the two input vectors are numbered from left to right across
2960both of the vectors. The shuffle mask operand specifies, for each element of
2961the result vector, which element of the two input registers the result element
2962gets. The element selector may be undef (meaning "don't care") and the second
2963operand may be undef if performing a shuffle from only one vector.
2964</p>
2965
2966<h5>Example:</h5>
2967
2968<pre>
2969 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
2970 &lt;4 x i32&gt; &lt;i32 0, i32 4, i32 1, i32 5&gt; <i>; yields &lt;4 x i32&gt;</i>
2971 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; undef,
2972 &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.
2973</pre>
2974</div>
2975
2976
2977<!-- ======================================================================= -->
2978<div class="doc_subsection">
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]2979 <a name="aggregateops">Aggregate Operations</a>
2980</div>
2981
2982<div class="doc_text">
2983
2984<p>LLVM supports several instructions for working with aggregate values.
2985</p>
2986
2987</div>
2988
2989<!-- _______________________________________________________________________ -->
2990<div class="doc_subsubsection">
2991 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
2992</div>
2993
2994<div class="doc_text">
2995
2996<h5>Syntax:</h5>
2997
2998<pre>
2999 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
3000</pre>
3001
3002<h5>Overview:</h5>
3003
3004<p>
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3005The '<tt>extractvalue</tt>' instruction extracts the value of a struct field
3006or array element from an aggregate value.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3007</p>
3008
3009
3010<h5>Arguments:</h5>
3011
3012<p>
3013The first operand of an '<tt>extractvalue</tt>' instruction is a
3014value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a>
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3015type. The operands are constant indices to specify which value to extract
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3016in a similar manner as indices in a
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3017'<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
3018</p>
3019
3020<h5>Semantics:</h5>
3021
3022<p>
3023The result is the value at the position in the aggregate specified by
3024the index operands.
3025</p>
3026
3027<h5>Example:</h5>
3028
3029<pre>
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3030 %result = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3031</pre>
3032</div>
3033
3034
3035<!-- _______________________________________________________________________ -->
3036<div class="doc_subsubsection">
3037 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
3038</div>
3039
3040<div class="doc_text">
3041
3042<h5>Syntax:</h5>
3043
3044<pre>
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3045 &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]3046</pre>
3047
3048<h5>Overview:</h5>
3049
3050<p>
3051The '<tt>insertvalue</tt>' instruction inserts a value
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3052into a struct field or array element in an aggregate.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3053</p>
3054
3055
3056<h5>Arguments:</h5>
3057
3058<p>
3059The first operand of an '<tt>insertvalue</tt>' instruction is a
3060value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a> type.
3061The second operand is a first-class value to insert.
Dan Gohman4f29e422008-05-23 21:53:15 +0000[diff] [blame]3062The following operands are constant indices
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3063indicating the position at which to insert the value in a similar manner as
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3064indices in a
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3065'<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
3066The value to insert must have the same type as the value identified
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3067by the indices.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3068
3069<h5>Semantics:</h5>
3070
3071<p>
3072The result is an aggregate of the same type as <tt>val</tt>. Its
3073value is that of <tt>val</tt> except that the value at the position
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3074specified by the indices is that of <tt>elt</tt>.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3075</p>
3076
3077<h5>Example:</h5>
3078
3079<pre>
Dan Gohmanb1aab4e2008-06-23 15:26:37 +0000[diff] [blame]3080 %result = insertvalue {i32, float} %agg, i32 1, 0 <i>; yields {i32, float}</i>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3081</pre>
3082</div>
3083
3084
3085<!-- ======================================================================= -->
3086<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3087 <a name="memoryops">Memory Access and Addressing Operations</a>
3088</div>
3089
3090<div class="doc_text">
3091
3092<p>A key design point of an SSA-based representation is how it
3093represents memory. In LLVM, no memory locations are in SSA form, which
3094makes things very simple. This section describes how to read, write,
3095allocate, and free memory in LLVM.</p>
3096
3097</div>
3098
3099<!-- _______________________________________________________________________ -->
3100<div class="doc_subsubsection">
3101 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
3102</div>
3103
3104<div class="doc_text">
3105
3106<h5>Syntax:</h5>
3107
3108<pre>
3109 &lt;result&gt; = malloc &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
3110</pre>
3111
3112<h5>Overview:</h5>
3113
3114<p>The '<tt>malloc</tt>' instruction allocates memory from the system
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]3115heap and returns a pointer to it. The object is always allocated in the generic
3116address space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3117
3118<h5>Arguments:</h5>
3119
3120<p>The '<tt>malloc</tt>' instruction allocates
3121<tt>sizeof(&lt;type&gt;)*NumElements</tt>
3122bytes of memory from the operating system and returns a pointer of the
3123appropriate type to the program. If "NumElements" is specified, it is the
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3124number of elements allocated, otherwise "NumElements" is defaulted to be one.
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3125If a constant alignment is specified, the value result of the allocation is guaranteed to
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3126be aligned to at least that boundary. If not specified, or if zero, the target can
3127choose to align the allocation on any convenient boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3128
3129<p>'<tt>type</tt>' must be a sized type.</p>
3130
3131<h5>Semantics:</h5>
3132
3133<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]3134a pointer is returned. The result of a zero byte allocattion is undefined. The
3135result is null if there is insufficient memory available.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3136
3137<h5>Example:</h5>
3138
3139<pre>
3140 %array = malloc [4 x i8 ] <i>; yields {[%4 x i8]*}:array</i>
3141
3142 %size = <a href="#i_add">add</a> i32 2, 2 <i>; yields {i32}:size = i32 4</i>
3143 %array1 = malloc i8, i32 4 <i>; yields {i8*}:array1</i>
3144 %array2 = malloc [12 x i8], i32 %size <i>; yields {[12 x i8]*}:array2</i>
3145 %array3 = malloc i32, i32 4, align 1024 <i>; yields {i32*}:array3</i>
3146 %array4 = malloc i32, align 1024 <i>; yields {i32*}:array4</i>
3147</pre>
3148</div>
3149
3150<!-- _______________________________________________________________________ -->
3151<div class="doc_subsubsection">
3152 <a name="i_free">'<tt>free</tt>' Instruction</a>
3153</div>
3154
3155<div class="doc_text">
3156
3157<h5>Syntax:</h5>
3158
3159<pre>
3160 free &lt;type&gt; &lt;value&gt; <i>; yields {void}</i>
3161</pre>
3162
3163<h5>Overview:</h5>
3164
3165<p>The '<tt>free</tt>' instruction returns memory back to the unused
3166memory heap to be reallocated in the future.</p>
3167
3168<h5>Arguments:</h5>
3169
3170<p>'<tt>value</tt>' shall be a pointer value that points to a value
3171that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
3172instruction.</p>
3173
3174<h5>Semantics:</h5>
3175
3176<p>Access to the memory pointed to by the pointer is no longer defined
Chris Lattner329d6532008-04-19 22:41:32 +0000[diff] [blame]3177after this instruction executes. If the pointer is null, the operation
3178is a noop.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3179
3180<h5>Example:</h5>
3181
3182<pre>
3183 %array = <a href="#i_malloc">malloc</a> [4 x i8] <i>; yields {[4 x i8]*}:array</i>
3184 free [4 x i8]* %array
3185</pre>
3186</div>
3187
3188<!-- _______________________________________________________________________ -->
3189<div class="doc_subsubsection">
3190 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
3191</div>
3192
3193<div class="doc_text">
3194
3195<h5>Syntax:</h5>
3196
3197<pre>
3198 &lt;result&gt; = alloca &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
3199</pre>
3200
3201<h5>Overview:</h5>
3202
3203<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
3204currently executing function, to be automatically released when this function
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]3205returns to its caller. The object is always allocated in the generic address
3206space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3207
3208<h5>Arguments:</h5>
3209
3210<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
3211bytes of memory on the runtime stack, returning a pointer of the
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3212appropriate type to the program. If "NumElements" is specified, it is the
3213number of elements allocated, otherwise "NumElements" is defaulted to be one.
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3214If a constant alignment is specified, the value result of the allocation is guaranteed
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3215to be aligned to at least that boundary. If not specified, or if zero, the target
3216can choose to align the allocation on any convenient boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3217
3218<p>'<tt>type</tt>' may be any sized type.</p>
3219
3220<h5>Semantics:</h5>
3221
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]3222<p>Memory is allocated; a pointer is returned. The operation is undefiend if
3223there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3224memory is automatically released when the function returns. The '<tt>alloca</tt>'
3225instruction is commonly used to represent automatic variables that must
3226have an address available. When the function returns (either with the <tt><a
3227 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]3228instructions), the memory is reclaimed. Allocating zero bytes
3229is legal, but the result is undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3230
3231<h5>Example:</h5>
3232
3233<pre>
3234 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
3235 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
3236 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
3237 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
3238</pre>
3239</div>
3240
3241<!-- _______________________________________________________________________ -->
3242<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
3243Instruction</a> </div>
3244<div class="doc_text">
3245<h5>Syntax:</h5>
3246<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>
3247<h5>Overview:</h5>
3248<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
3249<h5>Arguments:</h5>
3250<p>The argument to the '<tt>load</tt>' instruction specifies the memory
3251address from which to load. The pointer must point to a <a
3252 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
3253marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
3254the number or order of execution of this <tt>load</tt> with other
3255volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
3256instructions. </p>
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3257<p>
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3258The optional constant "align" argument specifies the alignment of the operation
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3259(that is, the alignment of the memory address). A value of 0 or an
3260omitted "align" argument means that the operation has the preferential
3261alignment for the target. It is the responsibility of the code emitter
3262to ensure that the alignment information is correct. Overestimating
3263the alignment results in an undefined behavior. Underestimating the
3264alignment may produce less efficient code. An alignment of 1 is always
3265safe.
3266</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3267<h5>Semantics:</h5>
3268<p>The location of memory pointed to is loaded.</p>
3269<h5>Examples:</h5>
3270<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
3271 <a
3272 href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
3273 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
3274</pre>
3275</div>
3276<!-- _______________________________________________________________________ -->
3277<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
3278Instruction</a> </div>
3279<div class="doc_text">
3280<h5>Syntax:</h5>
3281<pre> store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
3282 volatile store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
3283</pre>
3284<h5>Overview:</h5>
3285<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
3286<h5>Arguments:</h5>
3287<p>There are two arguments to the '<tt>store</tt>' instruction: a value
3288to 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]3289operand must be a pointer to the <a href="#t_firstclass">first class</a> type
3290of the '<tt>&lt;value&gt;</tt>'
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3291operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
3292optimizer is not allowed to modify the number or order of execution of
3293this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
3294 href="#i_store">store</a></tt> instructions.</p>
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3295<p>
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3296The optional constant "align" argument specifies the alignment of the operation
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3297(that is, the alignment of the memory address). A value of 0 or an
3298omitted "align" argument means that the operation has the preferential
3299alignment for the target. It is the responsibility of the code emitter
3300to ensure that the alignment information is correct. Overestimating
3301the alignment results in an undefined behavior. Underestimating the
3302alignment may produce less efficient code. An alignment of 1 is always
3303safe.
3304</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3305<h5>Semantics:</h5>
3306<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>'
3307at the location specified by the '<tt>&lt;pointer&gt;</tt>' operand.</p>
3308<h5>Example:</h5>
3309<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling63ffa142007-10-22 05:10:05 +0000[diff] [blame]3310 store i32 3, i32* %ptr <i>; yields {void}</i>
3311 %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]3312</pre>
3313</div>
3314
3315<!-- _______________________________________________________________________ -->
3316<div class="doc_subsubsection">
3317 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
3318</div>
3319
3320<div class="doc_text">
3321<h5>Syntax:</h5>
3322<pre>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3323 &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]3324</pre>
3325
3326<h5>Overview:</h5>
3327
3328<p>
3329The '<tt>getelementptr</tt>' instruction is used to get the address of a
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3330subelement of an aggregate data structure. It performs address calculation only
3331and does not access memory.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3332
3333<h5>Arguments:</h5>
3334
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3335<p>The first argument is always a pointer, and forms the basis of the
3336calculation. The remaining arguments are indices, that indicate which of the
3337elements of the aggregate object are indexed. The interpretation of each index
3338is dependent on the type being indexed into. The first index always indexes the
3339pointer value given as the first argument, the second index indexes a value of
3340the type pointed to (not necessarily the value directly pointed to, since the
3341first index can be non-zero), etc. The first type indexed into must be a pointer
3342value, subsequent types can be arrays, vectors and structs. Note that subsequent
3343types being indexed into can never be pointers, since that would require loading
3344the pointer before continuing calculation.</p>
3345
3346<p>The type of each index argument depends on the type it is indexing into.
3347When indexing into a (packed) structure, only <tt>i32</tt> integer
3348<b>constants</b> are allowed. When indexing into an array, pointer or vector,
3349only integers of 32 or 64 bits are allowed (also non-constants). 32-bit values
3350will be sign extended to 64-bits if required.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3351
3352<p>For example, let's consider a C code fragment and how it gets
3353compiled to LLVM:</p>
3354
3355<div class="doc_code">
3356<pre>
3357struct RT {
3358 char A;
3359 int B[10][20];
3360 char C;
3361};
3362struct ST {
3363 int X;
3364 double Y;
3365 struct RT Z;
3366};
3367
3368int *foo(struct ST *s) {
3369 return &amp;s[1].Z.B[5][13];
3370}
3371</pre>
3372</div>
3373
3374<p>The LLVM code generated by the GCC frontend is:</p>
3375
3376<div class="doc_code">
3377<pre>
3378%RT = type { i8 , [10 x [20 x i32]], i8 }
3379%ST = type { i32, double, %RT }
3380
3381define i32* %foo(%ST* %s) {
3382entry:
3383 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
3384 ret i32* %reg
3385}
3386</pre>
3387</div>
3388
3389<h5>Semantics:</h5>
3390
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3391<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
3392type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
3393}</tt>' type, a structure. The second index indexes into the third element of
3394the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
3395i8 }</tt>' type, another structure. The third index indexes into the second
3396element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
3397array. The two dimensions of the array are subscripted into, yielding an
3398'<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
3399to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
3400
3401<p>Note that it is perfectly legal to index partially through a
3402structure, returning a pointer to an inner element. Because of this,
3403the LLVM code for the given testcase is equivalent to:</p>
3404
3405<pre>
3406 define i32* %foo(%ST* %s) {
3407 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
3408 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
3409 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
3410 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
3411 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
3412 ret i32* %t5
3413 }
3414</pre>
3415
3416<p>Note that it is undefined to access an array out of bounds: array and
3417pointer indexes must always be within the defined bounds of the array type.
Chris Lattnera7d94ba2008-04-24 05:59:56 +0000[diff] [blame]3418The one exception for this rule is zero length arrays. These arrays are
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3419defined to be accessible as variable length arrays, which requires access
3420beyond the zero'th element.</p>
3421
3422<p>The getelementptr instruction is often confusing. For some more insight
3423into how it works, see <a href="GetElementPtr.html">the getelementptr
3424FAQ</a>.</p>
3425
3426<h5>Example:</h5>
3427
3428<pre>
3429 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3430 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
3431 <i>; yields i8*:vptr</i>
3432 %vptr = getelementptr {i32, <2 x i8>}* %svptr, i64 0, i32 1, i32 1
3433 <i>; yields i8*:eptr</i>
3434 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3435</pre>
3436</div>
3437
3438<!-- ======================================================================= -->
3439<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
3440</div>
3441<div class="doc_text">
3442<p>The instructions in this category are the conversion instructions (casting)
3443which all take a single operand and a type. They perform various bit conversions
3444on the operand.</p>
3445</div>
3446
3447<!-- _______________________________________________________________________ -->
3448<div class="doc_subsubsection">
3449 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
3450</div>
3451<div class="doc_text">
3452
3453<h5>Syntax:</h5>
3454<pre>
3455 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3456</pre>
3457
3458<h5>Overview:</h5>
3459<p>
3460The '<tt>trunc</tt>' instruction truncates its operand to the type <tt>ty2</tt>.
3461</p>
3462
3463<h5>Arguments:</h5>
3464<p>
3465The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
3466be an <a href="#t_integer">integer</a> type, and a type that specifies the size
3467and type of the result, which must be an <a href="#t_integer">integer</a>
3468type. The bit size of <tt>value</tt> must be larger than the bit size of
3469<tt>ty2</tt>. Equal sized types are not allowed.</p>
3470
3471<h5>Semantics:</h5>
3472<p>
3473The '<tt>trunc</tt>' instruction truncates the high order bits in <tt>value</tt>
3474and converts the remaining bits to <tt>ty2</tt>. Since the source size must be
3475larger than the destination size, <tt>trunc</tt> cannot be a <i>no-op cast</i>.
3476It will always truncate bits.</p>
3477
3478<h5>Example:</h5>
3479<pre>
3480 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
3481 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
3482 %Y = trunc i32 122 to i1 <i>; yields i1:false</i>
3483</pre>
3484</div>
3485
3486<!-- _______________________________________________________________________ -->
3487<div class="doc_subsubsection">
3488 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
3489</div>
3490<div class="doc_text">
3491
3492<h5>Syntax:</h5>
3493<pre>
3494 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3495</pre>
3496
3497<h5>Overview:</h5>
3498<p>The '<tt>zext</tt>' instruction zero extends its operand to type
3499<tt>ty2</tt>.</p>
3500
3501
3502<h5>Arguments:</h5>
3503<p>The '<tt>zext</tt>' instruction takes a value to cast, which must be of
3504<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3505also be of <a href="#t_integer">integer</a> type. The bit size of the
3506<tt>value</tt> must be smaller than the bit size of the destination type,
3507<tt>ty2</tt>.</p>
3508
3509<h5>Semantics:</h5>
3510<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
3511bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
3512
3513<p>When zero extending from i1, the result will always be either 0 or 1.</p>
3514
3515<h5>Example:</h5>
3516<pre>
3517 %X = zext i32 257 to i64 <i>; yields i64:257</i>
3518 %Y = zext i1 true to i32 <i>; yields i32:1</i>
3519</pre>
3520</div>
3521
3522<!-- _______________________________________________________________________ -->
3523<div class="doc_subsubsection">
3524 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
3525</div>
3526<div class="doc_text">
3527
3528<h5>Syntax:</h5>
3529<pre>
3530 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3531</pre>
3532
3533<h5>Overview:</h5>
3534<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
3535
3536<h5>Arguments:</h5>
3537<p>
3538The '<tt>sext</tt>' instruction takes a value to cast, which must be of
3539<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3540also be of <a href="#t_integer">integer</a> type. The bit size of the
3541<tt>value</tt> must be smaller than the bit size of the destination type,
3542<tt>ty2</tt>.</p>
3543
3544<h5>Semantics:</h5>
3545<p>
3546The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
3547bit (highest order bit) of the <tt>value</tt> until it reaches the bit size of
3548the type <tt>ty2</tt>.</p>
3549
3550<p>When sign extending from i1, the extension always results in -1 or 0.</p>
3551
3552<h5>Example:</h5>
3553<pre>
3554 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
3555 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
3556</pre>
3557</div>
3558
3559<!-- _______________________________________________________________________ -->
3560<div class="doc_subsubsection">
3561 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
3562</div>
3563
3564<div class="doc_text">
3565
3566<h5>Syntax:</h5>
3567
3568<pre>
3569 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3570</pre>
3571
3572<h5>Overview:</h5>
3573<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
3574<tt>ty2</tt>.</p>
3575
3576
3577<h5>Arguments:</h5>
3578<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
3579 point</a> value to cast and a <a href="#t_floating">floating point</a> type to
3580cast it to. The size of <tt>value</tt> must be larger than the size of
3581<tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
3582<i>no-op cast</i>.</p>
3583
3584<h5>Semantics:</h5>
3585<p> The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
3586<a href="#t_floating">floating point</a> type to a smaller
3587<a href="#t_floating">floating point</a> type. If the value cannot fit within
3588the destination type, <tt>ty2</tt>, then the results are undefined.</p>
3589
3590<h5>Example:</h5>
3591<pre>
3592 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
3593 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
3594</pre>
3595</div>
3596
3597<!-- _______________________________________________________________________ -->
3598<div class="doc_subsubsection">
3599 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
3600</div>
3601<div class="doc_text">
3602
3603<h5>Syntax:</h5>
3604<pre>
3605 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3606</pre>
3607
3608<h5>Overview:</h5>
3609<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
3610floating point value.</p>
3611
3612<h5>Arguments:</h5>
3613<p>The '<tt>fpext</tt>' instruction takes a
3614<a href="#t_floating">floating point</a> <tt>value</tt> to cast,
3615and a <a href="#t_floating">floating point</a> type to cast it to. The source
3616type must be smaller than the destination type.</p>
3617
3618<h5>Semantics:</h5>
3619<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
3620<a href="#t_floating">floating point</a> type to a larger
3621<a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
3622used to make a <i>no-op cast</i> because it always changes bits. Use
3623<tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
3624
3625<h5>Example:</h5>
3626<pre>
3627 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
3628 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
3629</pre>
3630</div>
3631
3632<!-- _______________________________________________________________________ -->
3633<div class="doc_subsubsection">
3634 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
3635</div>
3636<div class="doc_text">
3637
3638<h5>Syntax:</h5>
3639<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3640 &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]3641</pre>
3642
3643<h5>Overview:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3644<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3645unsigned integer equivalent of type <tt>ty2</tt>.
3646</p>
3647
3648<h5>Arguments:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3649<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]3650scalar or vector <a href="#t_floating">floating point</a> value, and a type
3651to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
3652type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
3653vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3654
3655<h5>Semantics:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3656<p> The '<tt>fptoui</tt>' instruction converts its
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3657<a href="#t_floating">floating point</a> operand into the nearest (rounding
3658towards zero) unsigned integer value. If the value cannot fit in <tt>ty2</tt>,
3659the results are undefined.</p>
3660
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3661<h5>Example:</h5>
3662<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3663 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]3664 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3665 %X = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3666</pre>
3667</div>
3668
3669<!-- _______________________________________________________________________ -->
3670<div class="doc_subsubsection">
3671 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
3672</div>
3673<div class="doc_text">
3674
3675<h5>Syntax:</h5>
3676<pre>
3677 &lt;result&gt; = fptosi &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3678</pre>
3679
3680<h5>Overview:</h5>
3681<p>The '<tt>fptosi</tt>' instruction converts
3682<a href="#t_floating">floating point</a> <tt>value</tt> to type <tt>ty2</tt>.
3683</p>
3684
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3685<h5>Arguments:</h5>
3686<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]3687scalar or vector <a href="#t_floating">floating point</a> value, and a type
3688to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
3689type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
3690vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3691
3692<h5>Semantics:</h5>
3693<p>The '<tt>fptosi</tt>' instruction converts its
3694<a href="#t_floating">floating point</a> operand into the nearest (rounding
3695towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
3696the results are undefined.</p>
3697
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3698<h5>Example:</h5>
3699<pre>
3700 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]3701 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3702 %X = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
3703</pre>
3704</div>
3705
3706<!-- _______________________________________________________________________ -->
3707<div class="doc_subsubsection">
3708 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
3709</div>
3710<div class="doc_text">
3711
3712<h5>Syntax:</h5>
3713<pre>
3714 &lt;result&gt; = uitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3715</pre>
3716
3717<h5>Overview:</h5>
3718<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
3719integer and converts that value to the <tt>ty2</tt> type.</p>
3720
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3721<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]3722<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be a
3723scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
3724to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
3725type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
3726floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3727
3728<h5>Semantics:</h5>
3729<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
3730integer quantity and converts it to the corresponding floating point value. If
3731the value cannot fit in the floating point value, the results are undefined.</p>
3732
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3733<h5>Example:</h5>
3734<pre>
3735 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
3736 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
3737</pre>
3738</div>
3739
3740<!-- _______________________________________________________________________ -->
3741<div class="doc_subsubsection">
3742 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
3743</div>
3744<div class="doc_text">
3745
3746<h5>Syntax:</h5>
3747<pre>
3748 &lt;result&gt; = sitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3749</pre>
3750
3751<h5>Overview:</h5>
3752<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed
3753integer and converts that value to the <tt>ty2</tt> type.</p>
3754
3755<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]3756<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be a
3757scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
3758to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
3759type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
3760floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3761
3762<h5>Semantics:</h5>
3763<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed
3764integer quantity and converts it to the corresponding floating point value. If
3765the value cannot fit in the floating point value, the results are undefined.</p>
3766
3767<h5>Example:</h5>
3768<pre>
3769 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
3770 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
3771</pre>
3772</div>
3773
3774<!-- _______________________________________________________________________ -->
3775<div class="doc_subsubsection">
3776 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
3777</div>
3778<div class="doc_text">
3779
3780<h5>Syntax:</h5>
3781<pre>
3782 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3783</pre>
3784
3785<h5>Overview:</h5>
3786<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
3787the integer type <tt>ty2</tt>.</p>
3788
3789<h5>Arguments:</h5>
3790<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
3791must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
3792<tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> type.
3793
3794<h5>Semantics:</h5>
3795<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
3796<tt>ty2</tt> by interpreting the pointer value as an integer and either
3797truncating or zero extending that value to the size of the integer type. If
3798<tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
3799<tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
3800are the same size, then nothing is done (<i>no-op cast</i>) other than a type
3801change.</p>
3802
3803<h5>Example:</h5>
3804<pre>
3805 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
3806 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
3807</pre>
3808</div>
3809
3810<!-- _______________________________________________________________________ -->
3811<div class="doc_subsubsection">
3812 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
3813</div>
3814<div class="doc_text">
3815
3816<h5>Syntax:</h5>
3817<pre>
3818 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3819</pre>
3820
3821<h5>Overview:</h5>
3822<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to
3823a pointer type, <tt>ty2</tt>.</p>
3824
3825<h5>Arguments:</h5>
3826<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
3827value to cast, and a type to cast it to, which must be a
3828<a href="#t_pointer">pointer</a> type.
3829
3830<h5>Semantics:</h5>
3831<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
3832<tt>ty2</tt> by applying either a zero extension or a truncation depending on
3833the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
3834size of a pointer then a truncation is done. If <tt>value</tt> is smaller than
3835the size of a pointer then a zero extension is done. If they are the same size,
3836nothing is done (<i>no-op cast</i>).</p>
3837
3838<h5>Example:</h5>
3839<pre>
3840 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
3841 %X = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
3842 %Y = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
3843</pre>
3844</div>
3845
3846<!-- _______________________________________________________________________ -->
3847<div class="doc_subsubsection">
3848 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
3849</div>
3850<div class="doc_text">
3851
3852<h5>Syntax:</h5>
3853<pre>
3854 &lt;result&gt; = bitcast &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3855</pre>
3856
3857<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3858
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3859<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
3860<tt>ty2</tt> without changing any bits.</p>
3861
3862<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3863
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3864<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be
Dan Gohman7305fa02008-09-08 16:45:59 +0000[diff] [blame]3865a non-aggregate first class value, and a type to cast it to, which must also be
3866a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes of
3867<tt>value</tt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3868and the destination type, <tt>ty2</tt>, must be identical. If the source
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3869type is a pointer, the destination type must also be a pointer. This
3870instruction supports bitwise conversion of vectors to integers and to vectors
3871of other types (as long as they have the same size).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3872
3873<h5>Semantics:</h5>
3874<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
3875<tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
3876this conversion. The conversion is done as if the <tt>value</tt> had been
3877stored to memory and read back as type <tt>ty2</tt>. Pointer types may only be
3878converted to other pointer types with this instruction. To convert pointers to
3879other types, use the <a href="#i_inttoptr">inttoptr</a> or
3880<a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
3881
3882<h5>Example:</h5>
3883<pre>
3884 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
3885 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
3886 %Z = bitcast <2xint> %V to i64; <i>; yields i64: %V</i>
3887</pre>
3888</div>
3889
3890<!-- ======================================================================= -->
3891<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
3892<div class="doc_text">
3893<p>The instructions in this category are the "miscellaneous"
3894instructions, which defy better classification.</p>
3895</div>
3896
3897<!-- _______________________________________________________________________ -->
3898<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
3899</div>
3900<div class="doc_text">
3901<h5>Syntax:</h5>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]3902<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]3903</pre>
3904<h5>Overview:</h5>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]3905<p>The '<tt>icmp</tt>' instruction returns a boolean value or
3906a vector of boolean values based on comparison
3907of its two integer, integer vector, or pointer operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3908<h5>Arguments:</h5>
3909<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
3910the condition code indicating the kind of comparison to perform. It is not
3911a value, just a keyword. The possible condition code are:
3912<ol>
3913 <li><tt>eq</tt>: equal</li>
3914 <li><tt>ne</tt>: not equal </li>
3915 <li><tt>ugt</tt>: unsigned greater than</li>
3916 <li><tt>uge</tt>: unsigned greater or equal</li>
3917 <li><tt>ult</tt>: unsigned less than</li>
3918 <li><tt>ule</tt>: unsigned less or equal</li>
3919 <li><tt>sgt</tt>: signed greater than</li>
3920 <li><tt>sge</tt>: signed greater or equal</li>
3921 <li><tt>slt</tt>: signed less than</li>
3922 <li><tt>sle</tt>: signed less or equal</li>
3923</ol>
3924<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]3925<a href="#t_pointer">pointer</a>
3926or integer <a href="#t_vector">vector</a> typed.
3927They must also be identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3928<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3929<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]3930the condition code given as <tt>cond</tt>. The comparison performed always
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]3931yields either an <a href="#t_primitive"><tt>i1</tt></a> or vector of <tt>i1</tt> result, as follows:
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3932<ol>
3933 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
3934 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
3935 </li>
3936 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
3937 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
3938 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3939 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3940 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3941 <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]3942 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3943 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3944 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3945 <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]3946 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3947 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3948 <li><tt>sge</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3949 <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]3950 <li><tt>slt</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3951 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3952 <li><tt>sle</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3953 <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]3954</ol>
3955<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
3956values are compared as if they were integers.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]3957<p>If the operands are integer vectors, then they are compared
3958element by element. The result is an <tt>i1</tt> vector with
3959the same number of elements as the values being compared.
3960Otherwise, the result is an <tt>i1</tt>.
3961</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3962
3963<h5>Example:</h5>
3964<pre> &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
3965 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
3966 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
3967 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
3968 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
3969 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
3970</pre>
3971</div>
3972
3973<!-- _______________________________________________________________________ -->
3974<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
3975</div>
3976<div class="doc_text">
3977<h5>Syntax:</h5>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]3978<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]3979</pre>
3980<h5>Overview:</h5>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]3981<p>The '<tt>fcmp</tt>' instruction returns a boolean value
3982or vector of boolean values based on comparison
3983of its operands.
3984<p>
3985If the operands are floating point scalars, then the result
3986type is a boolean (<a href="#t_primitive"><tt>i1</tt></a>).
3987</p>
3988<p>If the operands are floating point vectors, then the result type
3989is a vector of boolean with the same number of elements as the
3990operands being compared.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3991<h5>Arguments:</h5>
3992<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
3993the condition code indicating the kind of comparison to perform. It is not
3994a value, just a keyword. The possible condition code are:
3995<ol>
3996 <li><tt>false</tt>: no comparison, always returns false</li>
3997 <li><tt>oeq</tt>: ordered and equal</li>
3998 <li><tt>ogt</tt>: ordered and greater than </li>
3999 <li><tt>oge</tt>: ordered and greater than or equal</li>
4000 <li><tt>olt</tt>: ordered and less than </li>
4001 <li><tt>ole</tt>: ordered and less than or equal</li>
4002 <li><tt>one</tt>: ordered and not equal</li>
4003 <li><tt>ord</tt>: ordered (no nans)</li>
4004 <li><tt>ueq</tt>: unordered or equal</li>
4005 <li><tt>ugt</tt>: unordered or greater than </li>
4006 <li><tt>uge</tt>: unordered or greater than or equal</li>
4007 <li><tt>ult</tt>: unordered or less than </li>
4008 <li><tt>ule</tt>: unordered or less than or equal</li>
4009 <li><tt>une</tt>: unordered or not equal</li>
4010 <li><tt>uno</tt>: unordered (either nans)</li>
4011 <li><tt>true</tt>: no comparison, always returns true</li>
4012</ol>
4013<p><i>Ordered</i> means that neither operand is a QNAN while
4014<i>unordered</i> means that either operand may be a QNAN.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4015<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be
4016either a <a href="#t_floating">floating point</a> type
4017or a <a href="#t_vector">vector</a> of floating point type.
4018They must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4019<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4020<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4021according to the condition code given as <tt>cond</tt>.
4022If the operands are vectors, then the vectors are compared
4023element by element.
4024Each comparison performed
4025always yields an <a href="#t_primitive">i1</a> result, as follows:
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4026<ol>
4027 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
4028 <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]4029 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4030 <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]4031 <tt>op1</tt> is greather than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4032 <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]4033 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4034 <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]4035 <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4036 <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]4037 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4038 <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]4039 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4040 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
4041 <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]4042 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4043 <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]4044 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4045 <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]4046 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4047 <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]4048 <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4049 <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]4050 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4051 <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]4052 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4053 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
4054 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
4055</ol>
4056
4057<h5>Example:</h5>
4058<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]4059 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
4060 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
4061 &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]4062</pre>
4063</div>
4064
4065<!-- _______________________________________________________________________ -->
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4066<div class="doc_subsubsection">
4067 <a name="i_vicmp">'<tt>vicmp</tt>' Instruction</a>
4068</div>
4069<div class="doc_text">
4070<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4071<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]4072</pre>
4073<h5>Overview:</h5>
4074<p>The '<tt>vicmp</tt>' instruction returns an integer vector value based on
4075element-wise comparison of its two integer vector operands.</p>
4076<h5>Arguments:</h5>
4077<p>The '<tt>vicmp</tt>' instruction takes three operands. The first operand is
4078the condition code indicating the kind of comparison to perform. It is not
4079a value, just a keyword. The possible condition code are:
4080<ol>
4081 <li><tt>eq</tt>: equal</li>
4082 <li><tt>ne</tt>: not equal </li>
4083 <li><tt>ugt</tt>: unsigned greater than</li>
4084 <li><tt>uge</tt>: unsigned greater or equal</li>
4085 <li><tt>ult</tt>: unsigned less than</li>
4086 <li><tt>ule</tt>: unsigned less or equal</li>
4087 <li><tt>sgt</tt>: signed greater than</li>
4088 <li><tt>sge</tt>: signed greater or equal</li>
4089 <li><tt>slt</tt>: signed less than</li>
4090 <li><tt>sle</tt>: signed less or equal</li>
4091</ol>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4092<p>The remaining two arguments must be <a href="#t_vector">vector</a> or
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4093<a href="#t_integer">integer</a> typed. They must also be identical types.</p>
4094<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4095<p>The '<tt>vicmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4096according to the condition code given as <tt>cond</tt>. The comparison yields a
4097<a href="#t_vector">vector</a> of <a href="#t_integer">integer</a> result, of
4098identical type as the values being compared. The most significant bit in each
4099element is 1 if the element-wise comparison evaluates to true, and is 0
4100otherwise. All other bits of the result are undefined. The condition codes
4101are evaluated identically to the <a href="#i_icmp">'<tt>icmp</tt>'
4102instruction</a>.
4103
4104<h5>Example:</h5>
4105<pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4106 &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>
4107 &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]4108</pre>
4109</div>
4110
4111<!-- _______________________________________________________________________ -->
4112<div class="doc_subsubsection">
4113 <a name="i_vfcmp">'<tt>vfcmp</tt>' Instruction</a>
4114</div>
4115<div class="doc_text">
4116<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4117<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]4118<h5>Overview:</h5>
4119<p>The '<tt>vfcmp</tt>' instruction returns an integer vector value based on
4120element-wise comparison of its two floating point vector operands. The output
4121elements have the same width as the input elements.</p>
4122<h5>Arguments:</h5>
4123<p>The '<tt>vfcmp</tt>' instruction takes three operands. The first operand is
4124the condition code indicating the kind of comparison to perform. It is not
4125a value, just a keyword. The possible condition code are:
4126<ol>
4127 <li><tt>false</tt>: no comparison, always returns false</li>
4128 <li><tt>oeq</tt>: ordered and equal</li>
4129 <li><tt>ogt</tt>: ordered and greater than </li>
4130 <li><tt>oge</tt>: ordered and greater than or equal</li>
4131 <li><tt>olt</tt>: ordered and less than </li>
4132 <li><tt>ole</tt>: ordered and less than or equal</li>
4133 <li><tt>one</tt>: ordered and not equal</li>
4134 <li><tt>ord</tt>: ordered (no nans)</li>
4135 <li><tt>ueq</tt>: unordered or equal</li>
4136 <li><tt>ugt</tt>: unordered or greater than </li>
4137 <li><tt>uge</tt>: unordered or greater than or equal</li>
4138 <li><tt>ult</tt>: unordered or less than </li>
4139 <li><tt>ule</tt>: unordered or less than or equal</li>
4140 <li><tt>une</tt>: unordered or not equal</li>
4141 <li><tt>uno</tt>: unordered (either nans)</li>
4142 <li><tt>true</tt>: no comparison, always returns true</li>
4143</ol>
4144<p>The remaining two arguments must be <a href="#t_vector">vector</a> of
4145<a href="#t_floating">floating point</a> typed. They must also be identical
4146types.</p>
4147<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4148<p>The '<tt>vfcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4149according to the condition code given as <tt>cond</tt>. The comparison yields a
4150<a href="#t_vector">vector</a> of <a href="#t_integer">integer</a> result, with
4151an identical number of elements as the values being compared, and each element
4152having identical with to the width of the floating point elements. The most
4153significant bit in each element is 1 if the element-wise comparison evaluates to
4154true, and is 0 otherwise. All other bits of the result are undefined. The
4155condition codes are evaluated identically to the
4156<a href="#i_fcmp">'<tt>fcmp</tt>' instruction</a>.
4157
4158<h5>Example:</h5>
4159<pre>
Chris Lattner1e5c5cd02008-10-13 16:55:18 +0000[diff] [blame]4160 <i>; yields: result=&lt;2 x i32&gt; &lt; i32 0, i32 -1 &gt;</i>
4161 &lt;result&gt; = vfcmp oeq &lt;2 x float&gt; &lt; float 4, float 0 &gt;, &lt; float 5, float 0 &gt;
4162
4163 <i>; yields: result=&lt;2 x i64&gt; &lt; i64 -1, i64 0 &gt;</i>
4164 &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]4165</pre>
4166</div>
4167
4168<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4169<div class="doc_subsubsection">
4170 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
4171</div>
4172
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4173<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4174
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4175<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4176
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4177<pre> &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...<br></pre>
4178<h5>Overview:</h5>
4179<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in
4180the SSA graph representing the function.</p>
4181<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4182
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4183<p>The type of the incoming values is specified with the first type
4184field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
4185as arguments, with one pair for each predecessor basic block of the
4186current block. Only values of <a href="#t_firstclass">first class</a>
4187type may be used as the value arguments to the PHI node. Only labels
4188may be used as the label arguments.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4189
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4190<p>There must be no non-phi instructions between the start of a basic
4191block and the PHI instructions: i.e. PHI instructions must be first in
4192a basic block.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4193
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4194<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4195
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4196<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
4197specified by the pair corresponding to the predecessor basic block that executed
4198just prior to the current block.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4199
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4200<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4201<pre>
4202Loop: ; Infinite loop that counts from 0 on up...
4203 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
4204 %nextindvar = add i32 %indvar, 1
4205 br label %Loop
4206</pre>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4207</div>
4208
4209<!-- _______________________________________________________________________ -->
4210<div class="doc_subsubsection">
4211 <a name="i_select">'<tt>select</tt>' Instruction</a>
4212</div>
4213
4214<div class="doc_text">
4215
4216<h5>Syntax:</h5>
4217
4218<pre>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4219 &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>
4220
4221 <i>selty</i> is either i1 or {&lt;N x i1&gt}
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4222</pre>
4223
4224<h5>Overview:</h5>
4225
4226<p>
4227The '<tt>select</tt>' instruction is used to choose one value based on a
4228condition, without branching.
4229</p>
4230
4231
4232<h5>Arguments:</h5>
4233
4234<p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4235The '<tt>select</tt>' instruction requires an 'i1' value or
4236a vector of 'i1' values indicating the
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4237condition, and two values of the same <a href="#t_firstclass">first class</a>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4238type. If the val1/val2 are vectors and
4239the condition is a scalar, then entire vectors are selected, not
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4240individual elements.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4241</p>
4242
4243<h5>Semantics:</h5>
4244
4245<p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4246If 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]4247value argument; otherwise, it returns the second value argument.
4248</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4249<p>
4250If the condition is a vector of i1, then the value arguments must
4251be vectors of the same size, and the selection is done element
4252by element.
4253</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4254
4255<h5>Example:</h5>
4256
4257<pre>
4258 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
4259</pre>
4260</div>
4261
4262
4263<!-- _______________________________________________________________________ -->
4264<div class="doc_subsubsection">
4265 <a name="i_call">'<tt>call</tt>' Instruction</a>
4266</div>
4267
4268<div class="doc_text">
4269
4270<h5>Syntax:</h5>
4271<pre>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]4272 &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]4273</pre>
4274
4275<h5>Overview:</h5>
4276
4277<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
4278
4279<h5>Arguments:</h5>
4280
4281<p>This instruction requires several arguments:</p>
4282
4283<ol>
4284 <li>
4285 <p>The optional "tail" marker indicates whether the callee function accesses
4286 any allocas or varargs in the caller. If the "tail" marker is present, the
4287 function call is eligible for tail call optimization. Note that calls may
4288 be marked "tail" even if they do not occur before a <a
4289 href="#i_ret"><tt>ret</tt></a> instruction.
4290 </li>
4291 <li>
4292 <p>The optional "cconv" marker indicates which <a href="#callingconv">calling
4293 convention</a> the call should use. If none is specified, the call defaults
4294 to using C calling conventions.
4295 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4296
4297 <li>
4298 <p>The optional <a href="#paramattrs">Parameter Attributes</a> list for
4299 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>',
4300 and '<tt>inreg</tt>' attributes are valid here.</p>
4301 </li>
4302
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4303 <li>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]4304 <p>'<tt>ty</tt>': the type of the call instruction itself which is also
4305 the type of the return value. Functions that return no value are marked
4306 <tt><a href="#t_void">void</a></tt>.</p>
4307 </li>
4308 <li>
4309 <p>'<tt>fnty</tt>': shall be the signature of the pointer to function
4310 value being invoked. The argument types must match the types implied by
4311 this signature. This type can be omitted if the function is not varargs
4312 and if the function type does not return a pointer to a function.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4313 </li>
4314 <li>
4315 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
4316 be invoked. In most cases, this is a direct function invocation, but
4317 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
4318 to function value.</p>
4319 </li>
4320 <li>
4321 <p>'<tt>function args</tt>': argument list whose types match the
4322 function signature argument types. All arguments must be of
4323 <a href="#t_firstclass">first class</a> type. If the function signature
4324 indicates the function accepts a variable number of arguments, the extra
4325 arguments can be specified.</p>
4326 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4327 <li>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]4328 <p>The optional <a href="#fnattrs">function attributes</a> list. Only
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4329 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
4330 '<tt>readnone</tt>' attributes are valid here.</p>
4331 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4332</ol>
4333
4334<h5>Semantics:</h5>
4335
4336<p>The '<tt>call</tt>' instruction is used to cause control flow to
4337transfer to a specified function, with its incoming arguments bound to
4338the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
4339instruction in the called function, control flow continues with the
4340instruction after the function call, and the return value of the
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]4341function is bound to the result argument.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4342
4343<h5>Example:</h5>
4344
4345<pre>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]4346 %retval = call i32 @test(i32 %argc)
Chris Lattner5e893ef2008-03-21 17:24:17 +0000[diff] [blame]4347 call i32 (i8 *, ...)* @printf(i8 * %msg, i32 12, i8 42) <i>; yields i32</i>
4348 %X = tail call i32 @foo() <i>; yields i32</i>
4349 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
4350 call void %foo(i8 97 signext)
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]4351
4352 %struct.A = type { i32, i8 }
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4353 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]4354 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
4355 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattnerac454b32008-10-08 06:26:11 +0000[diff] [blame]4356 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijman2c4e05a2008-10-07 10:03:45 +0000[diff] [blame]4357 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4358</pre>
4359
4360</div>
4361
4362<!-- _______________________________________________________________________ -->
4363<div class="doc_subsubsection">
4364 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
4365</div>
4366
4367<div class="doc_text">
4368
4369<h5>Syntax:</h5>
4370
4371<pre>
4372 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
4373</pre>
4374
4375<h5>Overview:</h5>
4376
4377<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
4378the "variable argument" area of a function call. It is used to implement the
4379<tt>va_arg</tt> macro in C.</p>
4380
4381<h5>Arguments:</h5>
4382
4383<p>This instruction takes a <tt>va_list*</tt> value and the type of
4384the argument. It returns a value of the specified argument type and
4385increments the <tt>va_list</tt> to point to the next argument. The
4386actual type of <tt>va_list</tt> is target specific.</p>
4387
4388<h5>Semantics:</h5>
4389
4390<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
4391type from the specified <tt>va_list</tt> and causes the
4392<tt>va_list</tt> to point to the next argument. For more information,
4393see the variable argument handling <a href="#int_varargs">Intrinsic
4394Functions</a>.</p>
4395
4396<p>It is legal for this instruction to be called in a function which does not
4397take a variable number of arguments, for example, the <tt>vfprintf</tt>
4398function.</p>
4399
4400<p><tt>va_arg</tt> is an LLVM instruction instead of an <a
4401href="#intrinsics">intrinsic function</a> because it takes a type as an
4402argument.</p>
4403
4404<h5>Example:</h5>
4405
4406<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
4407
4408</div>
4409
4410<!-- *********************************************************************** -->
4411<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
4412<!-- *********************************************************************** -->
4413
4414<div class="doc_text">
4415
4416<p>LLVM supports the notion of an "intrinsic function". These functions have
4417well known names and semantics and are required to follow certain restrictions.
4418Overall, these intrinsics represent an extension mechanism for the LLVM
4419language that does not require changing all of the transformations in LLVM when
4420adding to the language (or the bitcode reader/writer, the parser, etc...).</p>
4421
4422<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
4423prefix is reserved in LLVM for intrinsic names; thus, function names may not
4424begin with this prefix. Intrinsic functions must always be external functions:
4425you cannot define the body of intrinsic functions. Intrinsic functions may
4426only be used in call or invoke instructions: it is illegal to take the address
4427of an intrinsic function. Additionally, because intrinsic functions are part
4428of the LLVM language, it is required if any are added that they be documented
4429here.</p>
4430
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4431<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents
4432a family of functions that perform the same operation but on different data
4433types. Because LLVM can represent over 8 million different integer types,
4434overloading is used commonly to allow an intrinsic function to operate on any
4435integer type. One or more of the argument types or the result type can be
4436overloaded to accept any integer type. Argument types may also be defined as
4437exactly matching a previous argument's type or the result type. This allows an
4438intrinsic function which accepts multiple arguments, but needs all of them to
4439be of the same type, to only be overloaded with respect to a single argument or
4440the result.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4441
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4442<p>Overloaded intrinsics will have the names of its overloaded argument types
4443encoded into its function name, each preceded by a period. Only those types
4444which are overloaded result in a name suffix. Arguments whose type is matched
4445against another type do not. For example, the <tt>llvm.ctpop</tt> function can
4446take an integer of any width and returns an integer of exactly the same integer
4447width. This leads to a family of functions such as
4448<tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29 %val)</tt>.
4449Only one type, the return type, is overloaded, and only one type suffix is
4450required. Because the argument's type is matched against the return type, it
4451does not require its own name suffix.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4452
4453<p>To learn how to add an intrinsic function, please see the
4454<a href="ExtendingLLVM.html">Extending LLVM Guide</a>.
4455</p>
4456
4457</div>
4458
4459<!-- ======================================================================= -->
4460<div class="doc_subsection">
4461 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
4462</div>
4463
4464<div class="doc_text">
4465
4466<p>Variable argument support is defined in LLVM with the <a
4467 href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
4468intrinsic functions. These functions are related to the similarly
4469named macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
4470
4471<p>All of these functions operate on arguments that use a
4472target-specific value type "<tt>va_list</tt>". The LLVM assembly
4473language reference manual does not define what this type is, so all
4474transformations should be prepared to handle these functions regardless of
4475the type used.</p>
4476
4477<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
4478instruction and the variable argument handling intrinsic functions are
4479used.</p>
4480
4481<div class="doc_code">
4482<pre>
4483define i32 @test(i32 %X, ...) {
4484 ; Initialize variable argument processing
4485 %ap = alloca i8*
4486 %ap2 = bitcast i8** %ap to i8*
4487 call void @llvm.va_start(i8* %ap2)
4488
4489 ; Read a single integer argument
4490 %tmp = va_arg i8** %ap, i32
4491
4492 ; Demonstrate usage of llvm.va_copy and llvm.va_end
4493 %aq = alloca i8*
4494 %aq2 = bitcast i8** %aq to i8*
4495 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
4496 call void @llvm.va_end(i8* %aq2)
4497
4498 ; Stop processing of arguments.
4499 call void @llvm.va_end(i8* %ap2)
4500 ret i32 %tmp
4501}
4502
4503declare void @llvm.va_start(i8*)
4504declare void @llvm.va_copy(i8*, i8*)
4505declare void @llvm.va_end(i8*)
4506</pre>
4507</div>
4508
4509</div>
4510
4511<!-- _______________________________________________________________________ -->
4512<div class="doc_subsubsection">
4513 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
4514</div>
4515
4516
4517<div class="doc_text">
4518<h5>Syntax:</h5>
4519<pre> declare void %llvm.va_start(i8* &lt;arglist&gt;)<br></pre>
4520<h5>Overview:</h5>
4521<P>The '<tt>llvm.va_start</tt>' intrinsic initializes
4522<tt>*&lt;arglist&gt;</tt> for subsequent use by <tt><a
4523href="#i_va_arg">va_arg</a></tt>.</p>
4524
4525<h5>Arguments:</h5>
4526
4527<P>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
4528
4529<h5>Semantics:</h5>
4530
4531<P>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
4532macro available in C. In a target-dependent way, it initializes the
4533<tt>va_list</tt> element to which the argument points, so that the next call to
4534<tt>va_arg</tt> will produce the first variable argument passed to the function.
4535Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
4536last argument of the function as the compiler can figure that out.</p>
4537
4538</div>
4539
4540<!-- _______________________________________________________________________ -->
4541<div class="doc_subsubsection">
4542 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
4543</div>
4544
4545<div class="doc_text">
4546<h5>Syntax:</h5>
4547<pre> declare void @llvm.va_end(i8* &lt;arglist&gt;)<br></pre>
4548<h5>Overview:</h5>
4549
4550<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
4551which has been initialized previously with <tt><a href="#int_va_start">llvm.va_start</a></tt>
4552or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
4553
4554<h5>Arguments:</h5>
4555
4556<p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
4557
4558<h5>Semantics:</h5>
4559
4560<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
4561macro available in C. In a target-dependent way, it destroys the
4562<tt>va_list</tt> element to which the argument points. Calls to <a
4563href="#int_va_start"><tt>llvm.va_start</tt></a> and <a href="#int_va_copy">
4564<tt>llvm.va_copy</tt></a> must be matched exactly with calls to
4565<tt>llvm.va_end</tt>.</p>
4566
4567</div>
4568
4569<!-- _______________________________________________________________________ -->
4570<div class="doc_subsubsection">
4571 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
4572</div>
4573
4574<div class="doc_text">
4575
4576<h5>Syntax:</h5>
4577
4578<pre>
4579 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
4580</pre>
4581
4582<h5>Overview:</h5>
4583
4584<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
4585from the source argument list to the destination argument list.</p>
4586
4587<h5>Arguments:</h5>
4588
4589<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
4590The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
4591
4592
4593<h5>Semantics:</h5>
4594
4595<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
4596macro available in C. In a target-dependent way, it copies the source
4597<tt>va_list</tt> element into the destination <tt>va_list</tt> element. This
4598intrinsic is necessary because the <tt><a href="#int_va_start">
4599llvm.va_start</a></tt> intrinsic may be arbitrarily complex and require, for
4600example, memory allocation.</p>
4601
4602</div>
4603
4604<!-- ======================================================================= -->
4605<div class="doc_subsection">
4606 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
4607</div>
4608
4609<div class="doc_text">
4610
4611<p>
4612LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]4613Collection</a> (GC) requires the implementation and generation of these
4614intrinsics.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4615These intrinsics allow identification of <a href="#int_gcroot">GC roots on the
4616stack</a>, as well as garbage collector implementations that require <a
4617href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a> barriers.
4618Front-ends for type-safe garbage collected languages should generate these
4619intrinsics to make use of the LLVM garbage collectors. For more details, see <a
4620href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
4621</p>
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]4622
4623<p>The garbage collection intrinsics only operate on objects in the generic
4624 address space (address space zero).</p>
4625
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4626</div>
4627
4628<!-- _______________________________________________________________________ -->
4629<div class="doc_subsubsection">
4630 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
4631</div>
4632
4633<div class="doc_text">
4634
4635<h5>Syntax:</h5>
4636
4637<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4638 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4639</pre>
4640
4641<h5>Overview:</h5>
4642
4643<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
4644the code generator, and allows some metadata to be associated with it.</p>
4645
4646<h5>Arguments:</h5>
4647
4648<p>The first argument specifies the address of a stack object that contains the
4649root pointer. The second pointer (which must be either a constant or a global
4650value address) contains the meta-data to be associated with the root.</p>
4651
4652<h5>Semantics:</h5>
4653
Chris Lattnera7d94ba2008-04-24 05:59:56 +0000[diff] [blame]4654<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]4655location. At compile-time, the code generator generates information to allow
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4656the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
4657intrinsic may only be used in a function which <a href="#gc">specifies a GC
4658algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4659
4660</div>
4661
4662
4663<!-- _______________________________________________________________________ -->
4664<div class="doc_subsubsection">
4665 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
4666</div>
4667
4668<div class="doc_text">
4669
4670<h5>Syntax:</h5>
4671
4672<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4673 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4674</pre>
4675
4676<h5>Overview:</h5>
4677
4678<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
4679locations, allowing garbage collector implementations that require read
4680barriers.</p>
4681
4682<h5>Arguments:</h5>
4683
4684<p>The second argument is the address to read from, which should be an address
4685allocated from the garbage collector. The first object is a pointer to the
4686start of the referenced object, if needed by the language runtime (otherwise
4687null).</p>
4688
4689<h5>Semantics:</h5>
4690
4691<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
4692instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4693garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
4694may only be used in a function which <a href="#gc">specifies a GC
4695algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4696
4697</div>
4698
4699
4700<!-- _______________________________________________________________________ -->
4701<div class="doc_subsubsection">
4702 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
4703</div>
4704
4705<div class="doc_text">
4706
4707<h5>Syntax:</h5>
4708
4709<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4710 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4711</pre>
4712
4713<h5>Overview:</h5>
4714
4715<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
4716locations, allowing garbage collector implementations that require write
4717barriers (such as generational or reference counting collectors).</p>
4718
4719<h5>Arguments:</h5>
4720
4721<p>The first argument is the reference to store, the second is the start of the
4722object to store it to, and the third is the address of the field of Obj to
4723store to. If the runtime does not require a pointer to the object, Obj may be
4724null.</p>
4725
4726<h5>Semantics:</h5>
4727
4728<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
4729instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4730garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
4731may only be used in a function which <a href="#gc">specifies a GC
4732algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4733
4734</div>
4735
4736
4737
4738<!-- ======================================================================= -->
4739<div class="doc_subsection">
4740 <a name="int_codegen">Code Generator Intrinsics</a>
4741</div>
4742
4743<div class="doc_text">
4744<p>
4745These intrinsics are provided by LLVM to expose special features that may only
4746be implemented with code generator support.
4747</p>
4748
4749</div>
4750
4751<!-- _______________________________________________________________________ -->
4752<div class="doc_subsubsection">
4753 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
4754</div>
4755
4756<div class="doc_text">
4757
4758<h5>Syntax:</h5>
4759<pre>
4760 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
4761</pre>
4762
4763<h5>Overview:</h5>
4764
4765<p>
4766The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
4767target-specific value indicating the return address of the current function
4768or one of its callers.
4769</p>
4770
4771<h5>Arguments:</h5>
4772
4773<p>
4774The argument to this intrinsic indicates which function to return the address
4775for. Zero indicates the calling function, one indicates its caller, etc. The
4776argument is <b>required</b> to be a constant integer value.
4777</p>
4778
4779<h5>Semantics:</h5>
4780
4781<p>
4782The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
4783the return address of the specified call frame, or zero if it cannot be
4784identified. The value returned by this intrinsic is likely to be incorrect or 0
4785for arguments other than zero, so it should only be used for debugging purposes.
4786</p>
4787
4788<p>
4789Note that calling this intrinsic does not prevent function inlining or other
4790aggressive transformations, so the value returned may not be that of the obvious
4791source-language caller.
4792</p>
4793</div>
4794
4795
4796<!-- _______________________________________________________________________ -->
4797<div class="doc_subsubsection">
4798 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
4799</div>
4800
4801<div class="doc_text">
4802
4803<h5>Syntax:</h5>
4804<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4805 declare i8 *@llvm.frameaddress(i32 &lt;level&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4806</pre>
4807
4808<h5>Overview:</h5>
4809
4810<p>
4811The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
4812target-specific frame pointer value for the specified stack frame.
4813</p>
4814
4815<h5>Arguments:</h5>
4816
4817<p>
4818The argument to this intrinsic indicates which function to return the frame
4819pointer for. Zero indicates the calling function, one indicates its caller,
4820etc. The argument is <b>required</b> to be a constant integer value.
4821</p>
4822
4823<h5>Semantics:</h5>
4824
4825<p>
4826The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
4827the frame address of the specified call frame, or zero if it cannot be
4828identified. The value returned by this intrinsic is likely to be incorrect or 0
4829for arguments other than zero, so it should only be used for debugging purposes.
4830</p>
4831
4832<p>
4833Note that calling this intrinsic does not prevent function inlining or other
4834aggressive transformations, so the value returned may not be that of the obvious
4835source-language caller.
4836</p>
4837</div>
4838
4839<!-- _______________________________________________________________________ -->
4840<div class="doc_subsubsection">
4841 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
4842</div>
4843
4844<div class="doc_text">
4845
4846<h5>Syntax:</h5>
4847<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4848 declare i8 *@llvm.stacksave()
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4849</pre>
4850
4851<h5>Overview:</h5>
4852
4853<p>
4854The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
4855the function stack, for use with <a href="#int_stackrestore">
4856<tt>llvm.stackrestore</tt></a>. This is useful for implementing language
4857features like scoped automatic variable sized arrays in C99.
4858</p>
4859
4860<h5>Semantics:</h5>
4861
4862<p>
4863This intrinsic returns a opaque pointer value that can be passed to <a
4864href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
4865<tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
4866<tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
4867state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
4868practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
4869that were allocated after the <tt>llvm.stacksave</tt> was executed.
4870</p>
4871
4872</div>
4873
4874<!-- _______________________________________________________________________ -->
4875<div class="doc_subsubsection">
4876 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
4877</div>
4878
4879<div class="doc_text">
4880
4881<h5>Syntax:</h5>
4882<pre>
4883 declare void @llvm.stackrestore(i8 * %ptr)
4884</pre>
4885
4886<h5>Overview:</h5>
4887
4888<p>
4889The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
4890the function stack to the state it was in when the corresponding <a
4891href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
4892useful for implementing language features like scoped automatic variable sized
4893arrays in C99.
4894</p>
4895
4896<h5>Semantics:</h5>
4897
4898<p>
4899See the description for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.
4900</p>
4901
4902</div>
4903
4904
4905<!-- _______________________________________________________________________ -->
4906<div class="doc_subsubsection">
4907 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
4908</div>
4909
4910<div class="doc_text">
4911
4912<h5>Syntax:</h5>
4913<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4914 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]4915</pre>
4916
4917<h5>Overview:</h5>
4918
4919
4920<p>
4921The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
4922a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
4923no
4924effect on the behavior of the program but can change its performance
4925characteristics.
4926</p>
4927
4928<h5>Arguments:</h5>
4929
4930<p>
4931<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
4932determining if the fetch should be for a read (0) or write (1), and
4933<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
4934locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
4935<tt>locality</tt> arguments must be constant integers.
4936</p>
4937
4938<h5>Semantics:</h5>
4939
4940<p>
4941This intrinsic does not modify the behavior of the program. In particular,
4942prefetches cannot trap and do not produce a value. On targets that support this
4943intrinsic, the prefetch can provide hints to the processor cache for better
4944performance.
4945</p>
4946
4947</div>
4948
4949<!-- _______________________________________________________________________ -->
4950<div class="doc_subsubsection">
4951 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
4952</div>
4953
4954<div class="doc_text">
4955
4956<h5>Syntax:</h5>
4957<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4958 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4959</pre>
4960
4961<h5>Overview:</h5>
4962
4963
4964<p>
4965The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]4966(PC) in a region of
4967code to simulators and other tools. The method is target specific, but it is
4968expected that the marker will use exported symbols to transmit the PC of the
4969marker.
4970The marker makes no guarantees that it will remain with any specific instruction
4971after optimizations. It is possible that the presence of a marker will inhibit
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4972optimizations. The intended use is to be inserted after optimizations to allow
4973correlations of simulation runs.
4974</p>
4975
4976<h5>Arguments:</h5>
4977
4978<p>
4979<tt>id</tt> is a numerical id identifying the marker.
4980</p>
4981
4982<h5>Semantics:</h5>
4983
4984<p>
4985This intrinsic does not modify the behavior of the program. Backends that do not
4986support this intrinisic may ignore it.
4987</p>
4988
4989</div>
4990
4991<!-- _______________________________________________________________________ -->
4992<div class="doc_subsubsection">
4993 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
4994</div>
4995
4996<div class="doc_text">
4997
4998<h5>Syntax:</h5>
4999<pre>
5000 declare i64 @llvm.readcyclecounter( )
5001</pre>
5002
5003<h5>Overview:</h5>
5004
5005
5006<p>
5007The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
5008counter register (or similar low latency, high accuracy clocks) on those targets
5009that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
5010As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
5011should only be used for small timings.
5012</p>
5013
5014<h5>Semantics:</h5>
5015
5016<p>
5017When directly supported, reading the cycle counter should not modify any memory.
5018Implementations are allowed to either return a application specific value or a
5019system wide value. On backends without support, this is lowered to a constant 0.
5020</p>
5021
5022</div>
5023
5024<!-- ======================================================================= -->
5025<div class="doc_subsection">
5026 <a name="int_libc">Standard C Library Intrinsics</a>
5027</div>
5028
5029<div class="doc_text">
5030<p>
5031LLVM provides intrinsics for a few important standard C library functions.
5032These intrinsics allow source-language front-ends to pass information about the
5033alignment of the pointer arguments to the code generator, providing opportunity
5034for more efficient code generation.
5035</p>
5036
5037</div>
5038
5039<!-- _______________________________________________________________________ -->
5040<div class="doc_subsubsection">
5041 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
5042</div>
5043
5044<div class="doc_text">
5045
5046<h5>Syntax:</h5>
5047<pre>
5048 declare void @llvm.memcpy.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5049 i32 &lt;len&gt;, i32 &lt;align&gt;)
5050 declare void @llvm.memcpy.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5051 i64 &lt;len&gt;, i32 &lt;align&gt;)
5052</pre>
5053
5054<h5>Overview:</h5>
5055
5056<p>
5057The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
5058location to the destination location.
5059</p>
5060
5061<p>
5062Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
5063intrinsics do not return a value, and takes an extra alignment argument.
5064</p>
5065
5066<h5>Arguments:</h5>
5067
5068<p>
5069The first argument is a pointer to the destination, the second is a pointer to
5070the source. The third argument is an integer argument
5071specifying the number of bytes to copy, and the fourth argument is the alignment
5072of the source and destination locations.
5073</p>
5074
5075<p>
5076If the call to this intrinisic has an alignment value that is not 0 or 1, then
5077the caller guarantees that both the source and destination pointers are aligned
5078to that boundary.
5079</p>
5080
5081<h5>Semantics:</h5>
5082
5083<p>
5084The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
5085location to the destination location, which are not allowed to overlap. It
5086copies "len" bytes of memory over. If the argument is known to be aligned to
5087some boundary, this can be specified as the fourth argument, otherwise it should
5088be set to 0 or 1.
5089</p>
5090</div>
5091
5092
5093<!-- _______________________________________________________________________ -->
5094<div class="doc_subsubsection">
5095 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
5096</div>
5097
5098<div class="doc_text">
5099
5100<h5>Syntax:</h5>
5101<pre>
5102 declare void @llvm.memmove.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5103 i32 &lt;len&gt;, i32 &lt;align&gt;)
5104 declare void @llvm.memmove.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5105 i64 &lt;len&gt;, i32 &lt;align&gt;)
5106</pre>
5107
5108<h5>Overview:</h5>
5109
5110<p>
5111The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the source
5112location to the destination location. It is similar to the
Chris Lattnerdba16ea2008-01-06 19:51:52 +0000[diff] [blame]5113'<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to overlap.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5114</p>
5115
5116<p>
5117Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
5118intrinsics do not return a value, and takes an extra alignment argument.
5119</p>
5120
5121<h5>Arguments:</h5>
5122
5123<p>
5124The first argument is a pointer to the destination, the second is a pointer to
5125the source. The third argument is an integer argument
5126specifying the number of bytes to copy, and the fourth argument is the alignment
5127of the source and destination locations.
5128</p>
5129
5130<p>
5131If the call to this intrinisic has an alignment value that is not 0 or 1, then
5132the caller guarantees that the source and destination pointers are aligned to
5133that boundary.
5134</p>
5135
5136<h5>Semantics:</h5>
5137
5138<p>
5139The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the source
5140location to the destination location, which may overlap. It
5141copies "len" bytes of memory over. If the argument is known to be aligned to
5142some boundary, this can be specified as the fourth argument, otherwise it should
5143be set to 0 or 1.
5144</p>
5145</div>
5146
5147
5148<!-- _______________________________________________________________________ -->
5149<div class="doc_subsubsection">
5150 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
5151</div>
5152
5153<div class="doc_text">
5154
5155<h5>Syntax:</h5>
5156<pre>
5157 declare void @llvm.memset.i32(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5158 i32 &lt;len&gt;, i32 &lt;align&gt;)
5159 declare void @llvm.memset.i64(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5160 i64 &lt;len&gt;, i32 &lt;align&gt;)
5161</pre>
5162
5163<h5>Overview:</h5>
5164
5165<p>
5166The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a particular
5167byte value.
5168</p>
5169
5170<p>
5171Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
5172does not return a value, and takes an extra alignment argument.
5173</p>
5174
5175<h5>Arguments:</h5>
5176
5177<p>
5178The first argument is a pointer to the destination to fill, the second is the
5179byte value to fill it with, the third argument is an integer
5180argument specifying the number of bytes to fill, and the fourth argument is the
5181known alignment of destination location.
5182</p>
5183
5184<p>
5185If the call to this intrinisic has an alignment value that is not 0 or 1, then
5186the caller guarantees that the destination pointer is aligned to that boundary.
5187</p>
5188
5189<h5>Semantics:</h5>
5190
5191<p>
5192The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting at
5193the
5194destination location. If the argument is known to be aligned to some boundary,
5195this can be specified as the fourth argument, otherwise it should be set to 0 or
51961.
5197</p>
5198</div>
5199
5200
5201<!-- _______________________________________________________________________ -->
5202<div class="doc_subsubsection">
5203 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
5204</div>
5205
5206<div class="doc_text">
5207
5208<h5>Syntax:</h5>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5209<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]5210floating point or vector of floating point type. Not all targets support all
5211types however.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5212<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5213 declare float @llvm.sqrt.f32(float %Val)
5214 declare double @llvm.sqrt.f64(double %Val)
5215 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
5216 declare fp128 @llvm.sqrt.f128(fp128 %Val)
5217 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5218</pre>
5219
5220<h5>Overview:</h5>
5221
5222<p>
5223The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5224returning the same value as the libm '<tt>sqrt</tt>' functions would. Unlike
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5225<tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
Chris Lattnerf914a002008-01-29 07:00:44 +0000[diff] [blame]5226negative numbers other than -0.0 (which allows for better optimization, because
5227there is no need to worry about errno being set). <tt>llvm.sqrt(-0.0)</tt> is
5228defined to return -0.0 like IEEE sqrt.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5229</p>
5230
5231<h5>Arguments:</h5>
5232
5233<p>
5234The argument and return value are floating point numbers of the same type.
5235</p>
5236
5237<h5>Semantics:</h5>
5238
5239<p>
5240This function returns the sqrt of the specified operand if it is a nonnegative
5241floating point number.
5242</p>
5243</div>
5244
5245<!-- _______________________________________________________________________ -->
5246<div class="doc_subsubsection">
5247 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
5248</div>
5249
5250<div class="doc_text">
5251
5252<h5>Syntax:</h5>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5253<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]5254floating point or vector of floating point type. Not all targets support all
5255types however.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5256<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5257 declare float @llvm.powi.f32(float %Val, i32 %power)
5258 declare double @llvm.powi.f64(double %Val, i32 %power)
5259 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
5260 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
5261 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5262</pre>
5263
5264<h5>Overview:</h5>
5265
5266<p>
5267The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
5268specified (positive or negative) power. The order of evaluation of
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5269multiplications is not defined. When a vector of floating point type is
5270used, the second argument remains a scalar integer value.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5271</p>
5272
5273<h5>Arguments:</h5>
5274
5275<p>
5276The second argument is an integer power, and the first is a value to raise to
5277that power.
5278</p>
5279
5280<h5>Semantics:</h5>
5281
5282<p>
5283This function returns the first value raised to the second power with an
5284unspecified sequence of rounding operations.</p>
5285</div>
5286
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5287<!-- _______________________________________________________________________ -->
5288<div class="doc_subsubsection">
5289 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
5290</div>
5291
5292<div class="doc_text">
5293
5294<h5>Syntax:</h5>
5295<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
5296floating point or vector of floating point type. Not all targets support all
5297types however.
5298<pre>
5299 declare float @llvm.sin.f32(float %Val)
5300 declare double @llvm.sin.f64(double %Val)
5301 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
5302 declare fp128 @llvm.sin.f128(fp128 %Val)
5303 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
5304</pre>
5305
5306<h5>Overview:</h5>
5307
5308<p>
5309The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.
5310</p>
5311
5312<h5>Arguments:</h5>
5313
5314<p>
5315The argument and return value are floating point numbers of the same type.
5316</p>
5317
5318<h5>Semantics:</h5>
5319
5320<p>
5321This function returns the sine of the specified operand, returning the
5322same values as the libm <tt>sin</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5323conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5324</div>
5325
5326<!-- _______________________________________________________________________ -->
5327<div class="doc_subsubsection">
5328 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
5329</div>
5330
5331<div class="doc_text">
5332
5333<h5>Syntax:</h5>
5334<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
5335floating point or vector of floating point type. Not all targets support all
5336types however.
5337<pre>
5338 declare float @llvm.cos.f32(float %Val)
5339 declare double @llvm.cos.f64(double %Val)
5340 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
5341 declare fp128 @llvm.cos.f128(fp128 %Val)
5342 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
5343</pre>
5344
5345<h5>Overview:</h5>
5346
5347<p>
5348The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.
5349</p>
5350
5351<h5>Arguments:</h5>
5352
5353<p>
5354The argument and return value are floating point numbers of the same type.
5355</p>
5356
5357<h5>Semantics:</h5>
5358
5359<p>
5360This function returns the cosine of the specified operand, returning the
5361same values as the libm <tt>cos</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5362conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5363</div>
5364
5365<!-- _______________________________________________________________________ -->
5366<div class="doc_subsubsection">
5367 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
5368</div>
5369
5370<div class="doc_text">
5371
5372<h5>Syntax:</h5>
5373<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
5374floating point or vector of floating point type. Not all targets support all
5375types however.
5376<pre>
5377 declare float @llvm.pow.f32(float %Val, float %Power)
5378 declare double @llvm.pow.f64(double %Val, double %Power)
5379 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
5380 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
5381 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
5382</pre>
5383
5384<h5>Overview:</h5>
5385
5386<p>
5387The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
5388specified (positive or negative) power.
5389</p>
5390
5391<h5>Arguments:</h5>
5392
5393<p>
5394The second argument is a floating point power, and the first is a value to
5395raise to that power.
5396</p>
5397
5398<h5>Semantics:</h5>
5399
5400<p>
5401This function returns the first value raised to the second power,
5402returning the
5403same values as the libm <tt>pow</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5404conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5405</div>
5406
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5407
5408<!-- ======================================================================= -->
5409<div class="doc_subsection">
5410 <a name="int_manip">Bit Manipulation Intrinsics</a>
5411</div>
5412
5413<div class="doc_text">
5414<p>
5415LLVM provides intrinsics for a few important bit manipulation operations.
5416These allow efficient code generation for some algorithms.
5417</p>
5418
5419</div>
5420
5421<!-- _______________________________________________________________________ -->
5422<div class="doc_subsubsection">
5423 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
5424</div>
5425
5426<div class="doc_text">
5427
5428<h5>Syntax:</h5>
5429<p>This is an overloaded intrinsic function. You can use bswap on any integer
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5430type that is an even number of bytes (i.e. BitWidth % 16 == 0).
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5431<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5432 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
5433 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
5434 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5435</pre>
5436
5437<h5>Overview:</h5>
5438
5439<p>
5440The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
5441values with an even number of bytes (positive multiple of 16 bits). These are
5442useful for performing operations on data that is not in the target's native
5443byte order.
5444</p>
5445
5446<h5>Semantics:</h5>
5447
5448<p>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5449The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5450and low byte of the input i16 swapped. Similarly, the <tt>llvm.bswap.i32</tt>
5451intrinsic returns an i32 value that has the four bytes of the input i32
5452swapped, 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]5453i32 will have its bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i48</tt>,
5454<tt>llvm.bswap.i64</tt> and other intrinsics extend this concept to
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5455additional even-byte lengths (6 bytes, 8 bytes and more, respectively).
5456</p>
5457
5458</div>
5459
5460<!-- _______________________________________________________________________ -->
5461<div class="doc_subsubsection">
5462 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
5463</div>
5464
5465<div class="doc_text">
5466
5467<h5>Syntax:</h5>
5468<p>This is an overloaded intrinsic. You can use llvm.ctpop on any integer bit
5469width. Not all targets support all bit widths however.
5470<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5471 declare i8 @llvm.ctpop.i8 (i8 &lt;src&gt;)
5472 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5473 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5474 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
5475 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5476</pre>
5477
5478<h5>Overview:</h5>
5479
5480<p>
5481The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
5482value.
5483</p>
5484
5485<h5>Arguments:</h5>
5486
5487<p>
5488The only argument is the value to be counted. The argument may be of any
5489integer type. The return type must match the argument type.
5490</p>
5491
5492<h5>Semantics:</h5>
5493
5494<p>
5495The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
5496</p>
5497</div>
5498
5499<!-- _______________________________________________________________________ -->
5500<div class="doc_subsubsection">
5501 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
5502</div>
5503
5504<div class="doc_text">
5505
5506<h5>Syntax:</h5>
5507<p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
5508integer bit width. Not all targets support all bit widths however.
5509<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5510 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
5511 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5512 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5513 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
5514 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5515</pre>
5516
5517<h5>Overview:</h5>
5518
5519<p>
5520The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
5521leading zeros in a variable.
5522</p>
5523
5524<h5>Arguments:</h5>
5525
5526<p>
5527The only argument is the value to be counted. The argument may be of any
5528integer type. The return type must match the argument type.
5529</p>
5530
5531<h5>Semantics:</h5>
5532
5533<p>
5534The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
5535in a variable. If the src == 0 then the result is the size in bits of the type
5536of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.
5537</p>
5538</div>
5539
5540
5541
5542<!-- _______________________________________________________________________ -->
5543<div class="doc_subsubsection">
5544 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
5545</div>
5546
5547<div class="doc_text">
5548
5549<h5>Syntax:</h5>
5550<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
5551integer bit width. Not all targets support all bit widths however.
5552<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5553 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
5554 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5555 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5556 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
5557 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5558</pre>
5559
5560<h5>Overview:</h5>
5561
5562<p>
5563The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
5564trailing zeros.
5565</p>
5566
5567<h5>Arguments:</h5>
5568
5569<p>
5570The only argument is the value to be counted. The argument may be of any
5571integer type. The return type must match the argument type.
5572</p>
5573
5574<h5>Semantics:</h5>
5575
5576<p>
5577The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
5578in a variable. If the src == 0 then the result is the size in bits of the type
5579of src. For example, <tt>llvm.cttz(2) = 1</tt>.
5580</p>
5581</div>
5582
5583<!-- _______________________________________________________________________ -->
5584<div class="doc_subsubsection">
5585 <a name="int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic</a>
5586</div>
5587
5588<div class="doc_text">
5589
5590<h5>Syntax:</h5>
5591<p>This is an overloaded intrinsic. You can use <tt>llvm.part.select</tt>
5592on any integer bit width.
5593<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5594 declare i17 @llvm.part.select.i17 (i17 %val, i32 %loBit, i32 %hiBit)
5595 declare i29 @llvm.part.select.i29 (i29 %val, i32 %loBit, i32 %hiBit)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5596</pre>
5597
5598<h5>Overview:</h5>
5599<p>The '<tt>llvm.part.select</tt>' family of intrinsic functions selects a
5600range of bits from an integer value and returns them in the same bit width as
5601the original value.</p>
5602
5603<h5>Arguments:</h5>
5604<p>The first argument, <tt>%val</tt> and the result may be integer types of
5605any bit width but they must have the same bit width. The second and third
5606arguments must be <tt>i32</tt> type since they specify only a bit index.</p>
5607
5608<h5>Semantics:</h5>
5609<p>The operation of the '<tt>llvm.part.select</tt>' intrinsic has two modes
5610of operation: forwards and reverse. If <tt>%loBit</tt> is greater than
5611<tt>%hiBits</tt> then the intrinsic operates in reverse mode. Otherwise it
5612operates in forward mode.</p>
5613<p>In forward mode, this intrinsic is the equivalent of shifting <tt>%val</tt>
5614right by <tt>%loBit</tt> bits and then ANDing it with a mask with
5615only the <tt>%hiBit - %loBit</tt> bits set, as follows:</p>
5616<ol>
5617 <li>The <tt>%val</tt> is shifted right (LSHR) by the number of bits specified
5618 by <tt>%loBits</tt>. This normalizes the value to the low order bits.</li>
5619 <li>The <tt>%loBits</tt> value is subtracted from the <tt>%hiBits</tt> value
5620 to determine the number of bits to retain.</li>
5621 <li>A mask of the retained bits is created by shifting a -1 value.</li>
5622 <li>The mask is ANDed with <tt>%val</tt> to produce the result.
5623</ol>
5624<p>In reverse mode, a similar computation is made except that the bits are
5625returned in the reverse order. So, for example, if <tt>X</tt> has the value
5626<tt>i16 0x0ACF (101011001111)</tt> and we apply
5627<tt>part.select(i16 X, 8, 3)</tt> to it, we get back the value
5628<tt>i16 0x0026 (000000100110)</tt>.</p>
5629</div>
5630
5631<div class="doc_subsubsection">
5632 <a name="int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic</a>
5633</div>
5634
5635<div class="doc_text">
5636
5637<h5>Syntax:</h5>
5638<p>This is an overloaded intrinsic. You can use <tt>llvm.part.set</tt>
5639on any integer bit width.
5640<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5641 declare i17 @llvm.part.set.i17.i9 (i17 %val, i9 %repl, i32 %lo, i32 %hi)
5642 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]5643</pre>
5644
5645<h5>Overview:</h5>
5646<p>The '<tt>llvm.part.set</tt>' family of intrinsic functions replaces a range
5647of bits in an integer value with another integer value. It returns the integer
5648with the replaced bits.</p>
5649
5650<h5>Arguments:</h5>
5651<p>The first argument, <tt>%val</tt> and the result may be integer types of
5652any bit width but they must have the same bit width. <tt>%val</tt> is the value
5653whose bits will be replaced. The second argument, <tt>%repl</tt> may be an
5654integer of any bit width. The third and fourth arguments must be <tt>i32</tt>
5655type since they specify only a bit index.</p>
5656
5657<h5>Semantics:</h5>
5658<p>The operation of the '<tt>llvm.part.set</tt>' intrinsic has two modes
5659of operation: forwards and reverse. If <tt>%lo</tt> is greater than
5660<tt>%hi</tt> then the intrinsic operates in reverse mode. Otherwise it
5661operates in forward mode.</p>
5662<p>For both modes, the <tt>%repl</tt> value is prepared for use by either
5663truncating it down to the size of the replacement area or zero extending it
5664up to that size.</p>
5665<p>In forward mode, the bits between <tt>%lo</tt> and <tt>%hi</tt> (inclusive)
5666are replaced with corresponding bits from <tt>%repl</tt>. That is the 0th bit
5667in <tt>%repl</tt> replaces the <tt>%lo</tt>th bit in <tt>%val</tt> and etc. up
5668to the <tt>%hi</tt>th bit.
5669<p>In reverse mode, a similar computation is made except that the bits are
5670reversed. That is, the <tt>0</tt>th bit in <tt>%repl</tt> replaces the
5671<tt>%hi</tt> bit in <tt>%val</tt> and etc. down to the <tt>%lo</tt>th bit.
5672<h5>Examples:</h5>
5673<pre>
5674 llvm.part.set(0xFFFF, 0, 4, 7) -&gt; 0xFF0F
5675 llvm.part.set(0xFFFF, 0, 7, 4) -&gt; 0xFF0F
5676 llvm.part.set(0xFFFF, 1, 7, 4) -&gt; 0xFF8F
5677 llvm.part.set(0xFFFF, F, 8, 3) -&gt; 0xFFE7
5678 llvm.part.set(0xFFFF, 0, 3, 8) -&gt; 0xFE07
5679</pre>
5680</div>
5681
5682<!-- ======================================================================= -->
5683<div class="doc_subsection">
5684 <a name="int_debugger">Debugger Intrinsics</a>
5685</div>
5686
5687<div class="doc_text">
5688<p>
5689The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
5690are described in the <a
5691href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
5692Debugging</a> document.
5693</p>
5694</div>
5695
5696
5697<!-- ======================================================================= -->
5698<div class="doc_subsection">
5699 <a name="int_eh">Exception Handling Intrinsics</a>
5700</div>
5701
5702<div class="doc_text">
5703<p> The LLVM exception handling intrinsics (which all start with
5704<tt>llvm.eh.</tt> prefix), are described in the <a
5705href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
5706Handling</a> document. </p>
5707</div>
5708
5709<!-- ======================================================================= -->
5710<div class="doc_subsection">
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5711 <a name="int_trampoline">Trampoline Intrinsic</a>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5712</div>
5713
5714<div class="doc_text">
5715<p>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5716 This intrinsic makes it possible to excise one parameter, marked with
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5717 the <tt>nest</tt> attribute, from a function. The result is a callable
5718 function pointer lacking the nest parameter - the caller does not need
5719 to provide a value for it. Instead, the value to use is stored in
5720 advance in a "trampoline", a block of memory usually allocated
5721 on the stack, which also contains code to splice the nest value into the
5722 argument list. This is used to implement the GCC nested function address
5723 extension.
5724</p>
5725<p>
5726 For example, if the function is
5727 <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]5728 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]5729<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5730 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
5731 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
5732 %p = call i8* @llvm.init.trampoline( i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8*), i8* %nval )
5733 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5734</pre>
Bill Wendlinge262b4c2007-09-22 09:23:55 +0000[diff] [blame]5735 <p>The call <tt>%val = call i32 %fp( i32 %x, i32 %y )</tt> is then equivalent
5736 to <tt>%val = call i32 %f( i8* %nval, i32 %x, i32 %y )</tt>.</p>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5737</div>
5738
5739<!-- _______________________________________________________________________ -->
5740<div class="doc_subsubsection">
5741 <a name="int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a>
5742</div>
5743<div class="doc_text">
5744<h5>Syntax:</h5>
5745<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5746declare 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]5747</pre>
5748<h5>Overview:</h5>
5749<p>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5750 This fills the memory pointed to by <tt>tramp</tt> with code
5751 and returns a function pointer suitable for executing it.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5752</p>
5753<h5>Arguments:</h5>
5754<p>
5755 The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
5756 pointers. The <tt>tramp</tt> argument must point to a sufficiently large
5757 and sufficiently aligned block of memory; this memory is written to by the
Duncan Sands35012212007-08-22 23:39:54 +0000[diff] [blame]5758 intrinsic. Note that the size and the alignment are target-specific - LLVM
5759 currently provides no portable way of determining them, so a front-end that
5760 generates this intrinsic needs to have some target-specific knowledge.
5761 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]5762</p>
5763<h5>Semantics:</h5>
5764<p>
5765 The block of memory pointed to by <tt>tramp</tt> is filled with target
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5766 dependent code, turning it into a function. A pointer to this function is
5767 returned, but needs to be bitcast to an
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5768 <a href="#int_trampoline">appropriate function pointer type</a>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5769 before being called. The new function's signature is the same as that of
5770 <tt>func</tt> with any arguments marked with the <tt>nest</tt> attribute
5771 removed. At most one such <tt>nest</tt> argument is allowed, and it must be
5772 of pointer type. Calling the new function is equivalent to calling
5773 <tt>func</tt> with the same argument list, but with <tt>nval</tt> used for the
5774 missing <tt>nest</tt> argument. If, after calling
5775 <tt>llvm.init.trampoline</tt>, the memory pointed to by <tt>tramp</tt> is
5776 modified, then the effect of any later call to the returned function pointer is
5777 undefined.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5778</p>
5779</div>
5780
5781<!-- ======================================================================= -->
5782<div class="doc_subsection">
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]5783 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
5784</div>
5785
5786<div class="doc_text">
5787<p>
5788 These intrinsic functions expand the "universal IR" of LLVM to represent
5789 hardware constructs for atomic operations and memory synchronization. This
5790 provides an interface to the hardware, not an interface to the programmer. It
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]5791 is aimed at a low enough level to allow any programming models or APIs
5792 (Application Programming Interfaces) which
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]5793 need atomic behaviors to map cleanly onto it. It is also modeled primarily on
5794 hardware behavior. Just as hardware provides a "universal IR" for source
5795 languages, it also provides a starting point for developing a "universal"
5796 atomic operation and synchronization IR.
5797</p>
5798<p>
5799 These do <em>not</em> form an API such as high-level threading libraries,
5800 software transaction memory systems, atomic primitives, and intrinsic
5801 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
5802 application libraries. The hardware interface provided by LLVM should allow
5803 a clean implementation of all of these APIs and parallel programming models.
5804 No one model or paradigm should be selected above others unless the hardware
5805 itself ubiquitously does so.
5806
5807</p>
5808</div>
5809
5810<!-- _______________________________________________________________________ -->
5811<div class="doc_subsubsection">
5812 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
5813</div>
5814<div class="doc_text">
5815<h5>Syntax:</h5>
5816<pre>
5817declare void @llvm.memory.barrier( i1 &lt;ll&gt;, i1 &lt;ls&gt;, i1 &lt;sl&gt;, i1 &lt;ss&gt;,
5818i1 &lt;device&gt; )
5819
5820</pre>
5821<h5>Overview:</h5>
5822<p>
5823 The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
5824 specific pairs of memory access types.
5825</p>
5826<h5>Arguments:</h5>
5827<p>
5828 The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
5829 The first four arguments enables a specific barrier as listed below. The fith
5830 argument specifies that the barrier applies to io or device or uncached memory.
5831
5832</p>
5833 <ul>
5834 <li><tt>ll</tt>: load-load barrier</li>
5835 <li><tt>ls</tt>: load-store barrier</li>
5836 <li><tt>sl</tt>: store-load barrier</li>
5837 <li><tt>ss</tt>: store-store barrier</li>
5838 <li><tt>device</tt>: barrier applies to device and uncached memory also.
5839 </ul>
5840<h5>Semantics:</h5>
5841<p>
5842 This intrinsic causes the system to enforce some ordering constraints upon
5843 the loads and stores of the program. This barrier does not indicate
5844 <em>when</em> any events will occur, it only enforces an <em>order</em> in
5845 which they occur. For any of the specified pairs of load and store operations
5846 (f.ex. load-load, or store-load), all of the first operations preceding the
5847 barrier will complete before any of the second operations succeeding the
5848 barrier begin. Specifically the semantics for each pairing is as follows:
5849</p>
5850 <ul>
5851 <li><tt>ll</tt>: All loads before the barrier must complete before any load
5852 after the barrier begins.</li>
5853
5854 <li><tt>ls</tt>: All loads before the barrier must complete before any
5855 store after the barrier begins.</li>
5856 <li><tt>ss</tt>: All stores before the barrier must complete before any
5857 store after the barrier begins.</li>
5858 <li><tt>sl</tt>: All stores before the barrier must complete before any
5859 load after the barrier begins.</li>
5860 </ul>
5861<p>
5862 These semantics are applied with a logical "and" behavior when more than one
5863 is enabled in a single memory barrier intrinsic.
5864</p>
5865<p>
5866 Backends may implement stronger barriers than those requested when they do not
5867 support as fine grained a barrier as requested. Some architectures do not
5868 need all types of barriers and on such architectures, these become noops.
5869</p>
5870<h5>Example:</h5>
5871<pre>
5872%ptr = malloc i32
5873 store i32 4, %ptr
5874
5875%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
5876 call void @llvm.memory.barrier( i1 false, i1 true, i1 false, i1 false )
5877 <i>; guarantee the above finishes</i>
5878 store i32 8, %ptr <i>; before this begins</i>
5879</pre>
5880</div>
5881
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5882<!-- _______________________________________________________________________ -->
5883<div class="doc_subsubsection">
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]5884 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5885</div>
5886<div class="doc_text">
5887<h5>Syntax:</h5>
5888<p>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]5889 This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
5890 any integer bit width and for different address spaces. Not all targets
5891 support all bit widths however.</p>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5892
5893<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]5894declare i8 @llvm.atomic.cmp.swap.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt; )
5895declare i16 @llvm.atomic.cmp.swap.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt; )
5896declare i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt; )
5897declare 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]5898
5899</pre>
5900<h5>Overview:</h5>
5901<p>
5902 This loads a value in memory and compares it to a given value. If they are
5903 equal, it stores a new value into the memory.
5904</p>
5905<h5>Arguments:</h5>
5906<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]5907 The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result as
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5908 well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
5909 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
5910 this integer type. While any bit width integer may be used, targets may only
5911 lower representations they support in hardware.
5912
5913</p>
5914<h5>Semantics:</h5>
5915<p>
5916 This entire intrinsic must be executed atomically. It first loads the value
5917 in memory pointed to by <tt>ptr</tt> and compares it with the value
5918 <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the memory. The
5919 loaded value is yielded in all cases. This provides the equivalent of an
5920 atomic compare-and-swap operation within the SSA framework.
5921</p>
5922<h5>Examples:</h5>
5923
5924<pre>
5925%ptr = malloc i32
5926 store i32 4, %ptr
5927
5928%val1 = add i32 4, 4
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]5929%result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 4, %val1 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5930 <i>; yields {i32}:result1 = 4</i>
5931%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
5932%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
5933
5934%val2 = add i32 1, 1
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]5935%result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 5, %val2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5936 <i>; yields {i32}:result2 = 8</i>
5937%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
5938
5939%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
5940</pre>
5941</div>
5942
5943<!-- _______________________________________________________________________ -->
5944<div class="doc_subsubsection">
5945 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
5946</div>
5947<div class="doc_text">
5948<h5>Syntax:</h5>
5949
5950<p>
5951 This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
5952 integer bit width. Not all targets support all bit widths however.</p>
5953<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]5954declare i8 @llvm.atomic.swap.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;val&gt; )
5955declare i16 @llvm.atomic.swap.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;val&gt; )
5956declare i32 @llvm.atomic.swap.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;val&gt; )
5957declare i64 @llvm.atomic.swap.i64.p0i64( i64* &lt;ptr&gt;, i64 &lt;val&gt; )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5958
5959</pre>
5960<h5>Overview:</h5>
5961<p>
5962 This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
5963 the value from memory. It then stores the value in <tt>val</tt> in the memory
5964 at <tt>ptr</tt>.
5965</p>
5966<h5>Arguments:</h5>
5967
5968<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]5969 The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both the
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5970 <tt>val</tt> argument and the result must be integers of the same bit width.
5971 The first argument, <tt>ptr</tt>, must be a pointer to a value of this
5972 integer type. The targets may only lower integer representations they
5973 support.
5974</p>
5975<h5>Semantics:</h5>
5976<p>
5977 This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
5978 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
5979 equivalent of an atomic swap operation within the SSA framework.
5980
5981</p>
5982<h5>Examples:</h5>
5983<pre>
5984%ptr = malloc i32
5985 store i32 4, %ptr
5986
5987%val1 = add i32 4, 4
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]5988%result1 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val1 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5989 <i>; yields {i32}:result1 = 4</i>
5990%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
5991%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
5992
5993%val2 = add i32 1, 1
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]5994%result2 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5995 <i>; yields {i32}:result2 = 8</i>
5996
5997%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
5998%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
5999</pre>
6000</div>
6001
6002<!-- _______________________________________________________________________ -->
6003<div class="doc_subsubsection">
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6004 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6005
6006</div>
6007<div class="doc_text">
6008<h5>Syntax:</h5>
6009<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6010 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]6011 integer bit width. Not all targets support all bit widths however.</p>
6012<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6013declare i8 @llvm.atomic.load.add.i8..p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6014declare i16 @llvm.atomic.load.add.i16..p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6015declare i32 @llvm.atomic.load.add.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6016declare 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]6017
6018</pre>
6019<h5>Overview:</h5>
6020<p>
6021 This intrinsic adds <tt>delta</tt> to the value stored in memory at
6022 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
6023</p>
6024<h5>Arguments:</h5>
6025<p>
6026
6027 The intrinsic takes two arguments, the first a pointer to an integer value
6028 and the second an integer value. The result is also an integer value. These
6029 integer types can have any bit width, but they must all have the same bit
6030 width. The targets may only lower integer representations they support.
6031</p>
6032<h5>Semantics:</h5>
6033<p>
6034 This intrinsic does a series of operations atomically. It first loads the
6035 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
6036 to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
6037</p>
6038
6039<h5>Examples:</h5>
6040<pre>
6041%ptr = malloc i32
6042 store i32 4, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6043%result1 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 4 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6044 <i>; yields {i32}:result1 = 4</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6045%result2 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6046 <i>; yields {i32}:result2 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6047%result3 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 5 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6048 <i>; yields {i32}:result3 = 10</i>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6049%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6050</pre>
6051</div>
6052
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6053<!-- _______________________________________________________________________ -->
6054<div class="doc_subsubsection">
6055 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
6056
6057</div>
6058<div class="doc_text">
6059<h5>Syntax:</h5>
6060<p>
6061 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]6062 any integer bit width and for different address spaces. Not all targets
6063 support all bit widths however.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6064<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6065declare i8 @llvm.atomic.load.sub.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6066declare i16 @llvm.atomic.load.sub.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6067declare i32 @llvm.atomic.load.sub.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6068declare 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]6069
6070</pre>
6071<h5>Overview:</h5>
6072<p>
6073 This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
6074 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
6075</p>
6076<h5>Arguments:</h5>
6077<p>
6078
6079 The intrinsic takes two arguments, the first a pointer to an integer value
6080 and the second an integer value. The result is also an integer value. These
6081 integer types can have any bit width, but they must all have the same bit
6082 width. The targets may only lower integer representations they support.
6083</p>
6084<h5>Semantics:</h5>
6085<p>
6086 This intrinsic does a series of operations atomically. It first loads the
6087 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
6088 result to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
6089</p>
6090
6091<h5>Examples:</h5>
6092<pre>
6093%ptr = malloc i32
6094 store i32 8, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6095%result1 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 4 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6096 <i>; yields {i32}:result1 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6097%result2 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 2 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6098 <i>; yields {i32}:result2 = 4</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6099%result3 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 5 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6100 <i>; yields {i32}:result3 = 2</i>
6101%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
6102</pre>
6103</div>
6104
6105<!-- _______________________________________________________________________ -->
6106<div class="doc_subsubsection">
6107 <a name="int_atomic_load_and">'<tt>llvm.atomic.load.and.*</tt>' Intrinsic</a><br>
6108 <a name="int_atomic_load_nand">'<tt>llvm.atomic.load.nand.*</tt>' Intrinsic</a><br>
6109 <a name="int_atomic_load_or">'<tt>llvm.atomic.load.or.*</tt>' Intrinsic</a><br>
6110 <a name="int_atomic_load_xor">'<tt>llvm.atomic.load.xor.*</tt>' Intrinsic</a><br>
6111
6112</div>
6113<div class="doc_text">
6114<h5>Syntax:</h5>
6115<p>
6116 These are overloaded intrinsics. You can use <tt>llvm.atomic.load_and</tt>,
6117 <tt>llvm.atomic.load_nand</tt>, <tt>llvm.atomic.load_or</tt>, and
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6118 <tt>llvm.atomic.load_xor</tt> on any integer bit width and for different
6119 address spaces. Not all targets support all bit widths however.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6120<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6121declare i8 @llvm.atomic.load.and.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6122declare i16 @llvm.atomic.load.and.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6123declare i32 @llvm.atomic.load.and.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6124declare 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]6125
6126</pre>
6127
6128<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6129declare i8 @llvm.atomic.load.or.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6130declare i16 @llvm.atomic.load.or.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6131declare i32 @llvm.atomic.load.or.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6132declare 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]6133
6134</pre>
6135
6136<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6137declare i8 @llvm.atomic.load.nand.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6138declare i16 @llvm.atomic.load.nand.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6139declare i32 @llvm.atomic.load.nand.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6140declare 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]6141
6142</pre>
6143
6144<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6145declare i8 @llvm.atomic.load.xor.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6146declare i16 @llvm.atomic.load.xor.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6147declare i32 @llvm.atomic.load.xor.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6148declare 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]6149
6150</pre>
6151<h5>Overview:</h5>
6152<p>
6153 These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
6154 the value stored in memory at <tt>ptr</tt>. It yields the original value
6155 at <tt>ptr</tt>.
6156</p>
6157<h5>Arguments:</h5>
6158<p>
6159
6160 These intrinsics take two arguments, the first a pointer to an integer value
6161 and the second an integer value. The result is also an integer value. These
6162 integer types can have any bit width, but they must all have the same bit
6163 width. The targets may only lower integer representations they support.
6164</p>
6165<h5>Semantics:</h5>
6166<p>
6167 These intrinsics does a series of operations atomically. They first load the
6168 value stored at <tt>ptr</tt>. They then do the bitwise operation
6169 <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the original
6170 value stored at <tt>ptr</tt>.
6171</p>
6172
6173<h5>Examples:</h5>
6174<pre>
6175%ptr = malloc i32
6176 store i32 0x0F0F, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6177%result0 = call i32 @llvm.atomic.load.nand.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6178 <i>; yields {i32}:result0 = 0x0F0F</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6179%result1 = call i32 @llvm.atomic.load.and.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6180 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6181%result2 = call i32 @llvm.atomic.load.or.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6182 <i>; yields {i32}:result2 = 0xF0</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6183%result3 = call i32 @llvm.atomic.load.xor.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6184 <i>; yields {i32}:result3 = FF</i>
6185%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
6186</pre>
6187</div>
6188
6189
6190<!-- _______________________________________________________________________ -->
6191<div class="doc_subsubsection">
6192 <a name="int_atomic_load_max">'<tt>llvm.atomic.load.max.*</tt>' Intrinsic</a><br>
6193 <a name="int_atomic_load_min">'<tt>llvm.atomic.load.min.*</tt>' Intrinsic</a><br>
6194 <a name="int_atomic_load_umax">'<tt>llvm.atomic.load.umax.*</tt>' Intrinsic</a><br>
6195 <a name="int_atomic_load_umin">'<tt>llvm.atomic.load.umin.*</tt>' Intrinsic</a><br>
6196
6197</div>
6198<div class="doc_text">
6199<h5>Syntax:</h5>
6200<p>
6201 These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
6202 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6203 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
6204 address spaces. Not all targets
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6205 support all bit widths however.</p>
6206<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6207declare i8 @llvm.atomic.load.max.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6208declare i16 @llvm.atomic.load.max.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6209declare i32 @llvm.atomic.load.max.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6210declare 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]6211
6212</pre>
6213
6214<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6215declare i8 @llvm.atomic.load.min.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6216declare i16 @llvm.atomic.load.min.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6217declare i32 @llvm.atomic.load.min.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6218declare 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]6219
6220</pre>
6221
6222<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6223declare i8 @llvm.atomic.load.umax.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6224declare i16 @llvm.atomic.load.umax.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6225declare i32 @llvm.atomic.load.umax.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6226declare 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]6227
6228</pre>
6229
6230<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6231declare i8 @llvm.atomic.load.umin.i8..p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6232declare i16 @llvm.atomic.load.umin.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6233declare i32 @llvm.atomic.load.umin.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6234declare 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]6235
6236</pre>
6237<h5>Overview:</h5>
6238<p>
6239 These intrinsics takes the signed or unsigned minimum or maximum of
6240 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
6241 original value at <tt>ptr</tt>.
6242</p>
6243<h5>Arguments:</h5>
6244<p>
6245
6246 These intrinsics take two arguments, the first a pointer to an integer value
6247 and the second an integer value. The result is also an integer value. These
6248 integer types can have any bit width, but they must all have the same bit
6249 width. The targets may only lower integer representations they support.
6250</p>
6251<h5>Semantics:</h5>
6252<p>
6253 These intrinsics does a series of operations atomically. They first load the
6254 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or max
6255 <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They yield
6256 the original value stored at <tt>ptr</tt>.
6257</p>
6258
6259<h5>Examples:</h5>
6260<pre>
6261%ptr = malloc i32
6262 store i32 7, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6263%result0 = call i32 @llvm.atomic.load.min.i32.p0i32( i32* %ptr, i32 -2 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6264 <i>; yields {i32}:result0 = 7</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6265%result1 = call i32 @llvm.atomic.load.max.i32.p0i32( i32* %ptr, i32 8 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6266 <i>; yields {i32}:result1 = -2</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6267%result2 = call i32 @llvm.atomic.load.umin.i32.p0i32( i32* %ptr, i32 10 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6268 <i>; yields {i32}:result2 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6269%result3 = call i32 @llvm.atomic.load.umax.i32.p0i32( i32* %ptr, i32 30 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6270 <i>; yields {i32}:result3 = 8</i>
6271%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
6272</pre>
6273</div>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6274
6275<!-- ======================================================================= -->
6276<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6277 <a name="int_general">General Intrinsics</a>
6278</div>
6279
6280<div class="doc_text">
6281<p> This class of intrinsics is designed to be generic and has
6282no specific purpose. </p>
6283</div>
6284
6285<!-- _______________________________________________________________________ -->
6286<div class="doc_subsubsection">
6287 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
6288</div>
6289
6290<div class="doc_text">
6291
6292<h5>Syntax:</h5>
6293<pre>
6294 declare void @llvm.var.annotation(i8* &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6295</pre>
6296
6297<h5>Overview:</h5>
6298
6299<p>
6300The '<tt>llvm.var.annotation</tt>' intrinsic
6301</p>
6302
6303<h5>Arguments:</h5>
6304
6305<p>
6306The first argument is a pointer to a value, the second is a pointer to a
6307global string, the third is a pointer to a global string which is the source
6308file name, and the last argument is the line number.
6309</p>
6310
6311<h5>Semantics:</h5>
6312
6313<p>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]6314This intrinsic allows annotation of local variables with arbitrary strings.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6315This can be useful for special purpose optimizations that want to look for these
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]6316annotations. These have no other defined use, they are ignored by code
6317generation and optimization.
6318</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6319</div>
6320
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6321<!-- _______________________________________________________________________ -->
6322<div class="doc_subsubsection">
Tanya Lattnerc9869b12007-09-21 23:57:59 +0000[diff] [blame]6323 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6324</div>
6325
6326<div class="doc_text">
6327
6328<h5>Syntax:</h5>
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]6329<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
6330any integer bit width.
6331</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6332<pre>
Tanya Lattner09161fe2007-09-22 00:03:01 +0000[diff] [blame]6333 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6334 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6335 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6336 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6337 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]6338</pre>
6339
6340<h5>Overview:</h5>
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]6341
6342<p>
6343The '<tt>llvm.annotation</tt>' intrinsic.
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6344</p>
6345
6346<h5>Arguments:</h5>
6347
6348<p>
6349The first argument is an integer value (result of some expression),
6350the second is a pointer to a global string, the third is a pointer to a global
6351string which is the source file name, and the last argument is the line number.
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]6352It returns the value of the first argument.
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6353</p>
6354
6355<h5>Semantics:</h5>
6356
6357<p>
6358This intrinsic allows annotations to be put on arbitrary expressions
6359with arbitrary strings. This can be useful for special purpose optimizations
6360that want to look for these annotations. These have no other defined use, they
6361are ignored by code generation and optimization.
6362</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6363
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]6364<!-- _______________________________________________________________________ -->
6365<div class="doc_subsubsection">
6366 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
6367</div>
6368
6369<div class="doc_text">
6370
6371<h5>Syntax:</h5>
6372<pre>
6373 declare void @llvm.trap()
6374</pre>
6375
6376<h5>Overview:</h5>
6377
6378<p>
6379The '<tt>llvm.trap</tt>' intrinsic
6380</p>
6381
6382<h5>Arguments:</h5>
6383
6384<p>
6385None
6386</p>
6387
6388<h5>Semantics:</h5>
6389
6390<p>
6391This intrinsics is lowered to the target dependent trap instruction. If the
6392target does not have a trap instruction, this intrinsic will be lowered to the
6393call of the abort() function.
6394</p>
6395</div>
6396
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6397<!-- *********************************************************************** -->
6398<hr>
6399<address>
6400 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
6401 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
6402 <a href="http://validator.w3.org/check/referer"><img
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]6403 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!"></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6404
6405 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
6406 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
6407 Last modified: $Date$
6408</address>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]6409
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6410</body>
6411</html>