Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / c4a04309a954095b6b49330e46434464a1e7b51c / . / docs / LangRef.html
blob: f8ca2879c44743480160fe5e97cf9469706d9096 [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>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]27 <li><a href="#aliasstructure">Aliases</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]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>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]61 <li><a href="#simpleconstants">Simple Constants</a></li>
62 <li><a href="#aggregateconstants">Aggregate Constants</a></li>
63 <li><a href="#globalconstants">Global Variable and Function Addresses</a></li>
64 <li><a href="#undefvalues">Undefined Values</a></li>
65 <li><a href="#constantexprs">Constant Expressions</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]66 </ol>
67 </li>
68 <li><a href="#othervalues">Other Values</a>
69 <ol>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]70 <li><a href="#inlineasm">Inline Assembler Expressions</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]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>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]146 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]147 <li><a href="#otherops">Other Operations</a>
148 <ol>
149 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
150 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]151 <li><a href="#i_vicmp">'<tt>vicmp</tt>' Instruction</a></li>
152 <li><a href="#i_vfcmp">'<tt>vfcmp</tt>' Instruction</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]153 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
154 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
155 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
156 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
157 </ol>
158 </li>
159 </ol>
160 </li>
161 <li><a href="#intrinsics">Intrinsic Functions</a>
162 <ol>
163 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
164 <ol>
165 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
166 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
167 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
168 </ol>
169 </li>
170 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
171 <ol>
172 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
173 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
174 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
175 </ol>
176 </li>
177 <li><a href="#int_codegen">Code Generator Intrinsics</a>
178 <ol>
179 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
180 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
181 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
182 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
183 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
184 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
185 <li><a href="#int_readcyclecounter"><tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
186 </ol>
187 </li>
188 <li><a href="#int_libc">Standard C Library Intrinsics</a>
189 <ol>
190 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
191 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
192 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
193 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
194 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]195 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
196 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
197 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]198 </ol>
199 </li>
200 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
201 <ol>
202 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
203 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
204 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
205 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
206 <li><a href="#int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic </a></li>
207 <li><a href="#int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic </a></li>
208 </ol>
209 </li>
210 <li><a href="#int_debugger">Debugger intrinsics</a></li>
211 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]212 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]213 <ol>
214 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]215 </ol>
216 </li>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]217 <li><a href="#int_atomics">Atomic intrinsics</a>
218 <ol>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]219 <li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]220 <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]221 <li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]222 <li><a href="#int_atomic_load_add"><tt>llvm.atomic.load.add</tt></a></li>
223 <li><a href="#int_atomic_load_sub"><tt>llvm.atomic.load.sub</tt></a></li>
224 <li><a href="#int_atomic_load_and"><tt>llvm.atomic.load.and</tt></a></li>
225 <li><a href="#int_atomic_load_nand"><tt>llvm.atomic.load.nand</tt></a></li>
226 <li><a href="#int_atomic_load_or"><tt>llvm.atomic.load.or</tt></a></li>
227 <li><a href="#int_atomic_load_xor"><tt>llvm.atomic.load.xor</tt></a></li>
228 <li><a href="#int_atomic_load_max"><tt>llvm.atomic.load.max</tt></a></li>
229 <li><a href="#int_atomic_load_min"><tt>llvm.atomic.load.min</tt></a></li>
230 <li><a href="#int_atomic_load_umax"><tt>llvm.atomic.load.umax</tt></a></li>
231 <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]232 </ol>
233 </li>
Reid Spencerb043f672007-07-20 19:59:11 +0000[diff] [blame]234 <li><a href="#int_general">General intrinsics</a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]235 <ol>
Reid Spencerb043f672007-07-20 19:59:11 +0000[diff] [blame]236 <li><a href="#int_var_annotation">
Tanya Lattner51369f32007-09-22 00:01:26 +0000[diff] [blame]237 <tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]238 <li><a href="#int_annotation">
Tanya Lattner51369f32007-09-22 00:01:26 +0000[diff] [blame]239 <tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]240 <li><a href="#int_trap">
241 <tt>llvm.trap</tt>' Intrinsic</a></li>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]242 </ol>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]243 </li>
244 </ol>
245 </li>
246</ol>
247
248<div class="doc_author">
249 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
250 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
251</div>
252
253<!-- *********************************************************************** -->
254<div class="doc_section"> <a name="abstract">Abstract </a></div>
255<!-- *********************************************************************** -->
256
257<div class="doc_text">
258<p>This document is a reference manual for the LLVM assembly language.
Bill Wendlinge7846a52008-08-05 22:29:16 +0000[diff] [blame]259LLVM is a Static Single Assignment (SSA) based representation that provides
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]260type safety, low-level operations, flexibility, and the capability of
261representing 'all' high-level languages cleanly. It is the common code
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]262representation used throughout all phases of the LLVM compilation
263strategy.</p>
264</div>
265
266<!-- *********************************************************************** -->
267<div class="doc_section"> <a name="introduction">Introduction</a> </div>
268<!-- *********************************************************************** -->
269
270<div class="doc_text">
271
272<p>The LLVM code representation is designed to be used in three
273different forms: as an in-memory compiler IR, as an on-disk bitcode
274representation (suitable for fast loading by a Just-In-Time compiler),
275and as a human readable assembly language representation. This allows
276LLVM to provide a powerful intermediate representation for efficient
277compiler transformations and analysis, while providing a natural means
278to debug and visualize the transformations. The three different forms
279of LLVM are all equivalent. This document describes the human readable
280representation and notation.</p>
281
282<p>The LLVM representation aims to be light-weight and low-level
283while being expressive, typed, and extensible at the same time. It
284aims to be a "universal IR" of sorts, by being at a low enough level
285that high-level ideas may be cleanly mapped to it (similar to how
286microprocessors are "universal IR's", allowing many source languages to
287be mapped to them). By providing type information, LLVM can be used as
288the target of optimizations: for example, through pointer analysis, it
289can be proven that a C automatic variable is never accessed outside of
290the current function... allowing it to be promoted to a simple SSA
291value instead of a memory location.</p>
292
293</div>
294
295<!-- _______________________________________________________________________ -->
296<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
297
298<div class="doc_text">
299
300<p>It is important to note that this document describes 'well formed'
301LLVM assembly language. There is a difference between what the parser
302accepts and what is considered 'well formed'. For example, the
303following instruction is syntactically okay, but not well formed:</p>
304
305<div class="doc_code">
306<pre>
307%x = <a href="#i_add">add</a> i32 1, %x
308</pre>
309</div>
310
311<p>...because the definition of <tt>%x</tt> does not dominate all of
312its uses. The LLVM infrastructure provides a verification pass that may
313be used to verify that an LLVM module is well formed. This pass is
314automatically run by the parser after parsing input assembly and by
315the optimizer before it outputs bitcode. The violations pointed out
316by the verifier pass indicate bugs in transformation passes or input to
317the parser.</p>
318</div>
319
Chris Lattnera83fdc02007-10-03 17:34:29 +0000[diff] [blame]320<!-- Describe the typesetting conventions here. -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]321
322<!-- *********************************************************************** -->
323<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
324<!-- *********************************************************************** -->
325
326<div class="doc_text">
327
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]328 <p>LLVM identifiers come in two basic types: global and local. Global
329 identifiers (functions, global variables) begin with the @ character. Local
330 identifiers (register names, types) begin with the % character. Additionally,
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]331 there are three different formats for identifiers, for different purposes:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]332
333<ol>
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]334 <li>Named values are represented as a string of characters with their prefix.
335 For example, %foo, @DivisionByZero, %a.really.long.identifier. The actual
336 regular expression used is '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]337 Identifiers which require other characters in their names can be surrounded
Daniel Dunbar73f9a772008-10-14 23:51:43 +0000[diff] [blame]338 with quotes. Special characters may be escaped using "\xx" where xx is the
339 ASCII code for the character in hexadecimal. In this way, any character can
340 be used in a name value, even quotes themselves.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]341
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]342 <li>Unnamed values are represented as an unsigned numeric value with their
343 prefix. For example, %12, @2, %44.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]344
345 <li>Constants, which are described in a <a href="#constants">section about
346 constants</a>, below.</li>
347</ol>
348
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]349<p>LLVM requires that values start with a prefix for two reasons: Compilers
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]350don't need to worry about name clashes with reserved words, and the set of
351reserved words may be expanded in the future without penalty. Additionally,
352unnamed identifiers allow a compiler to quickly come up with a temporary
353variable without having to avoid symbol table conflicts.</p>
354
355<p>Reserved words in LLVM are very similar to reserved words in other
356languages. There are keywords for different opcodes
357('<tt><a href="#i_add">add</a></tt>',
358 '<tt><a href="#i_bitcast">bitcast</a></tt>',
359 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
360href="#t_void">void</a></tt>', '<tt><a href="#t_primitive">i32</a></tt>', etc...),
361and others. These reserved words cannot conflict with variable names, because
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]362none of them start with a prefix character ('%' or '@').</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]363
364<p>Here is an example of LLVM code to multiply the integer variable
365'<tt>%X</tt>' by 8:</p>
366
367<p>The easy way:</p>
368
369<div class="doc_code">
370<pre>
371%result = <a href="#i_mul">mul</a> i32 %X, 8
372</pre>
373</div>
374
375<p>After strength reduction:</p>
376
377<div class="doc_code">
378<pre>
379%result = <a href="#i_shl">shl</a> i32 %X, i8 3
380</pre>
381</div>
382
383<p>And the hard way:</p>
384
385<div class="doc_code">
386<pre>
387<a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
388<a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
389%result = <a href="#i_add">add</a> i32 %1, %1
390</pre>
391</div>
392
393<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
394important lexical features of LLVM:</p>
395
396<ol>
397
398 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
399 line.</li>
400
401 <li>Unnamed temporaries are created when the result of a computation is not
402 assigned to a named value.</li>
403
404 <li>Unnamed temporaries are numbered sequentially</li>
405
406</ol>
407
408<p>...and it also shows a convention that we follow in this document. When
409demonstrating instructions, we will follow an instruction with a comment that
410defines the type and name of value produced. Comments are shown in italic
411text.</p>
412
413</div>
414
415<!-- *********************************************************************** -->
416<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
417<!-- *********************************************************************** -->
418
419<!-- ======================================================================= -->
420<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
421</div>
422
423<div class="doc_text">
424
425<p>LLVM programs are composed of "Module"s, each of which is a
426translation unit of the input programs. Each module consists of
427functions, global variables, and symbol table entries. Modules may be
428combined together with the LLVM linker, which merges function (and
429global variable) definitions, resolves forward declarations, and merges
430symbol table entries. Here is an example of the "hello world" module:</p>
431
432<div class="doc_code">
433<pre><i>; Declare the string constant as a global constant...</i>
434<a href="#identifiers">@.LC0</a> = <a href="#linkage_internal">internal</a> <a
435 href="#globalvars">constant</a> <a href="#t_array">[13 x i8]</a> c"hello world\0A\00" <i>; [13 x i8]*</i>
436
437<i>; External declaration of the puts function</i>
438<a href="#functionstructure">declare</a> i32 @puts(i8 *) <i>; i32(i8 *)* </i>
439
440<i>; Definition of main function</i>
441define i32 @main() { <i>; i32()* </i>
442 <i>; Convert [13x i8 ]* to i8 *...</i>
443 %cast210 = <a
444 href="#i_getelementptr">getelementptr</a> [13 x i8 ]* @.LC0, i64 0, i64 0 <i>; i8 *</i>
445
446 <i>; Call puts function to write out the string to stdout...</i>
447 <a
448 href="#i_call">call</a> i32 @puts(i8 * %cast210) <i>; i32</i>
449 <a
450 href="#i_ret">ret</a> i32 0<br>}<br>
451</pre>
452</div>
453
454<p>This example is made up of a <a href="#globalvars">global variable</a>
455named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
456function, and a <a href="#functionstructure">function definition</a>
457for "<tt>main</tt>".</p>
458
459<p>In general, a module is made up of a list of global values,
460where both functions and global variables are global values. Global values are
461represented by a pointer to a memory location (in this case, a pointer to an
462array of char, and a pointer to a function), and have one of the following <a
463href="#linkage">linkage types</a>.</p>
464
465</div>
466
467<!-- ======================================================================= -->
468<div class="doc_subsection">
469 <a name="linkage">Linkage Types</a>
470</div>
471
472<div class="doc_text">
473
474<p>
475All Global Variables and Functions have one of the following types of linkage:
476</p>
477
478<dl>
479
Dale Johannesen96e7e092008-05-23 23:13:41 +0000[diff] [blame]480 <dt><tt><b><a name="linkage_internal">internal</a></b></tt>: </dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]481
482 <dd>Global values with internal linkage are only directly accessible by
483 objects in the current module. In particular, linking code into a module with
484 an internal global value may cause the internal to be renamed as necessary to
485 avoid collisions. Because the symbol is internal to the module, all
486 references can be updated. This corresponds to the notion of the
487 '<tt>static</tt>' keyword in C.
488 </dd>
489
490 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
491
492 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
493 the same name when linkage occurs. This is typically used to implement
494 inline functions, templates, or other code which must be generated in each
495 translation unit that uses it. Unreferenced <tt>linkonce</tt> globals are
496 allowed to be discarded.
497 </dd>
498
Dale Johannesen96e7e092008-05-23 23:13:41 +0000[diff] [blame]499 <dt><tt><b><a name="linkage_common">common</a></b></tt>: </dt>
500
501 <dd>"<tt>common</tt>" linkage is exactly the same as <tt>linkonce</tt>
502 linkage, except that unreferenced <tt>common</tt> globals may not be
503 discarded. This is used for globals that may be emitted in multiple
504 translation units, but that are not guaranteed to be emitted into every
505 translation unit that uses them. One example of this is tentative
506 definitions in C, such as "<tt>int X;</tt>" at global scope.
507 </dd>
508
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]509 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
510
Dale Johannesen96e7e092008-05-23 23:13:41 +0000[diff] [blame]511 <dd>"<tt>weak</tt>" linkage is the same as <tt>common</tt> linkage, except
512 that some targets may choose to emit different assembly sequences for them
513 for target-dependent reasons. This is used for globals that are declared
514 "weak" in C source code.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]515 </dd>
516
517 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
518
519 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
520 pointer to array type. When two global variables with appending linkage are
521 linked together, the two global arrays are appended together. This is the
522 LLVM, typesafe, equivalent of having the system linker append together
523 "sections" with identical names when .o files are linked.
524 </dd>
525
526 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt>: </dt>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]527 <dd>The semantics of this linkage follow the ELF object file model: the
528 symbol is weak until linked, if not linked, the symbol becomes null instead
529 of being an undefined reference.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]530 </dd>
531
532 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
533
534 <dd>If none of the above identifiers are used, the global is externally
535 visible, meaning that it participates in linkage and can be used to resolve
536 external symbol references.
537 </dd>
538</dl>
539
540 <p>
541 The next two types of linkage are targeted for Microsoft Windows platform
542 only. They are designed to support importing (exporting) symbols from (to)
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]543 DLLs (Dynamic Link Libraries).
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]544 </p>
545
546 <dl>
547 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt>: </dt>
548
549 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
550 or variable via a global pointer to a pointer that is set up by the DLL
551 exporting the symbol. On Microsoft Windows targets, the pointer name is
552 formed by combining <code>_imp__</code> and the function or variable name.
553 </dd>
554
555 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt>: </dt>
556
557 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
558 pointer to a pointer in a DLL, so that it can be referenced with the
559 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
560 name is formed by combining <code>_imp__</code> and the function or variable
561 name.
562 </dd>
563
564</dl>
565
566<p><a name="linkage_external"></a>For example, since the "<tt>.LC0</tt>"
567variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
568variable and was linked with this one, one of the two would be renamed,
569preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
570external (i.e., lacking any linkage declarations), they are accessible
571outside of the current module.</p>
572<p>It is illegal for a function <i>declaration</i>
573to have any linkage type other than "externally visible", <tt>dllimport</tt>,
574or <tt>extern_weak</tt>.</p>
575<p>Aliases can have only <tt>external</tt>, <tt>internal</tt> and <tt>weak</tt>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]576linkages.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]577</div>
578
579<!-- ======================================================================= -->
580<div class="doc_subsection">
581 <a name="callingconv">Calling Conventions</a>
582</div>
583
584<div class="doc_text">
585
586<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
587and <a href="#i_invoke">invokes</a> can all have an optional calling convention
588specified for the call. The calling convention of any pair of dynamic
589caller/callee must match, or the behavior of the program is undefined. The
590following calling conventions are supported by LLVM, and more may be added in
591the future:</p>
592
593<dl>
594 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
595
596 <dd>This calling convention (the default if no other calling convention is
597 specified) matches the target C calling conventions. This calling convention
598 supports varargs function calls and tolerates some mismatch in the declared
599 prototype and implemented declaration of the function (as does normal C).
600 </dd>
601
602 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
603
604 <dd>This calling convention attempts to make calls as fast as possible
605 (e.g. by passing things in registers). This calling convention allows the
606 target to use whatever tricks it wants to produce fast code for the target,
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]607 without having to conform to an externally specified ABI (Application Binary
608 Interface). Implementations of this convention should allow arbitrary
Arnold Schwaighofer07444922008-05-14 09:17:12 +0000[diff] [blame]609 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> to be
610 supported. This calling convention does not support varargs and requires the
611 prototype of all callees to exactly match the prototype of the function
612 definition.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]613 </dd>
614
615 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
616
617 <dd>This calling convention attempts to make code in the caller as efficient
618 as possible under the assumption that the call is not commonly executed. As
619 such, these calls often preserve all registers so that the call does not break
620 any live ranges in the caller side. This calling convention does not support
621 varargs and requires the prototype of all callees to exactly match the
622 prototype of the function definition.
623 </dd>
624
625 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
626
627 <dd>Any calling convention may be specified by number, allowing
628 target-specific calling conventions to be used. Target specific calling
629 conventions start at 64.
630 </dd>
631</dl>
632
633<p>More calling conventions can be added/defined on an as-needed basis, to
634support pascal conventions or any other well-known target-independent
635convention.</p>
636
637</div>
638
639<!-- ======================================================================= -->
640<div class="doc_subsection">
641 <a name="visibility">Visibility Styles</a>
642</div>
643
644<div class="doc_text">
645
646<p>
647All Global Variables and Functions have one of the following visibility styles:
648</p>
649
650<dl>
651 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
652
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]653 <dd>On targets that use the ELF object file format, default visibility means
654 that the declaration is visible to other
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]655 modules and, in shared libraries, means that the declared entity may be
656 overridden. On Darwin, default visibility means that the declaration is
657 visible to other modules. Default visibility corresponds to "external
658 linkage" in the language.
659 </dd>
660
661 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
662
663 <dd>Two declarations of an object with hidden visibility refer to the same
664 object if they are in the same shared object. Usually, hidden visibility
665 indicates that the symbol will not be placed into the dynamic symbol table,
666 so no other module (executable or shared library) can reference it
667 directly.
668 </dd>
669
670 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
671
672 <dd>On ELF, protected visibility indicates that the symbol will be placed in
673 the dynamic symbol table, but that references within the defining module will
674 bind to the local symbol. That is, the symbol cannot be overridden by another
675 module.
676 </dd>
677</dl>
678
679</div>
680
681<!-- ======================================================================= -->
682<div class="doc_subsection">
683 <a name="globalvars">Global Variables</a>
684</div>
685
686<div class="doc_text">
687
688<p>Global variables define regions of memory allocated at compilation time
689instead of run-time. Global variables may optionally be initialized, may have
690an explicit section to be placed in, and may have an optional explicit alignment
691specified. A variable may be defined as "thread_local", which means that it
692will not be shared by threads (each thread will have a separated copy of the
693variable). A variable may be defined as a global "constant," which indicates
694that the contents of the variable will <b>never</b> be modified (enabling better
695optimization, allowing the global data to be placed in the read-only section of
696an executable, etc). Note that variables that need runtime initialization
697cannot be marked "constant" as there is a store to the variable.</p>
698
699<p>
700LLVM explicitly allows <em>declarations</em> of global variables to be marked
701constant, even if the final definition of the global is not. This capability
702can be used to enable slightly better optimization of the program, but requires
703the language definition to guarantee that optimizations based on the
704'constantness' are valid for the translation units that do not include the
705definition.
706</p>
707
708<p>As SSA values, global variables define pointer values that are in
709scope (i.e. they dominate) all basic blocks in the program. Global
710variables always define a pointer to their "content" type because they
711describe a region of memory, and all memory objects in LLVM are
712accessed through pointers.</p>
713
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]714<p>A global variable may be declared to reside in a target-specifc numbered
715address space. For targets that support them, address spaces may affect how
716optimizations are performed and/or what target instructions are used to access
Christopher Lamb20a39e92007-12-12 08:44:39 +0000[diff] [blame]717the variable. The default address space is zero. The address space qualifier
718must precede any other attributes.</p>
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]719
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]720<p>LLVM allows an explicit section to be specified for globals. If the target
721supports it, it will emit globals to the section specified.</p>
722
723<p>An explicit alignment may be specified for a global. If not present, or if
724the alignment is set to zero, the alignment of the global is set by the target
725to whatever it feels convenient. If an explicit alignment is specified, the
726global is forced to have at least that much alignment. All alignments must be
727a power of 2.</p>
728
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]729<p>For example, the following defines a global in a numbered address space with
730an initializer, section, and alignment:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]731
732<div class="doc_code">
733<pre>
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]734@G = constant float 1.0 addrspace(5), section "foo", align 4
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]735</pre>
736</div>
737
738</div>
739
740
741<!-- ======================================================================= -->
742<div class="doc_subsection">
743 <a name="functionstructure">Functions</a>
744</div>
745
746<div class="doc_text">
747
748<p>LLVM function definitions consist of the "<tt>define</tt>" keyord,
749an optional <a href="#linkage">linkage type</a>, an optional
750<a href="#visibility">visibility style</a>, an optional
751<a href="#callingconv">calling convention</a>, a return type, an optional
752<a href="#paramattrs">parameter attribute</a> for the return type, a function
753name, a (possibly empty) argument list (each with optional
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]754<a href="#paramattrs">parameter attributes</a>), optional
755<a href="#fnattrs">function attributes</a>, an optional section,
756an optional alignment, an optional <a href="#gc">garbage collector name</a>,
Chris Lattneree202542008-10-04 18:10:21 +0000[diff] [blame]757an opening curly brace, a list of basic blocks, and a closing curly brace.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]758
759LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
760optional <a href="#linkage">linkage type</a>, an optional
761<a href="#visibility">visibility style</a>, an optional
762<a href="#callingconv">calling convention</a>, a return type, an optional
763<a href="#paramattrs">parameter attribute</a> for the return type, a function
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]764name, a possibly empty list of arguments, an optional alignment, and an optional
Gordon Henriksend606f5b2007-12-10 03:30:21 +0000[diff] [blame]765<a href="#gc">garbage collector name</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]766
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]767<p>A function definition contains a list of basic blocks, forming the CFG
768(Control Flow Graph) for
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]769the function. Each basic block may optionally start with a label (giving the
770basic block a symbol table entry), contains a list of instructions, and ends
771with a <a href="#terminators">terminator</a> instruction (such as a branch or
772function return).</p>
773
774<p>The first basic block in a function is special in two ways: it is immediately
775executed on entrance to the function, and it is not allowed to have predecessor
776basic blocks (i.e. there can not be any branches to the entry block of a
777function). Because the block can have no predecessors, it also cannot have any
778<a href="#i_phi">PHI nodes</a>.</p>
779
780<p>LLVM allows an explicit section to be specified for functions. If the target
781supports it, it will emit functions to the section specified.</p>
782
783<p>An explicit alignment may be specified for a function. If not present, or if
784the alignment is set to zero, the alignment of the function is set by the target
785to whatever it feels convenient. If an explicit alignment is specified, the
786function is forced to have at least that much alignment. All alignments must be
787a power of 2.</p>
788
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]789 <h5>Syntax:</h5>
790
791<div class="doc_code">
Chris Lattner1e5c5cd02008-10-13 16:55:18 +0000[diff] [blame]792<tt>
793define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
794 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
795 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
796 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
797 [<a href="#gc">gc</a>] { ... }
798</tt>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]799</div>
800
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]801</div>
802
803
804<!-- ======================================================================= -->
805<div class="doc_subsection">
806 <a name="aliasstructure">Aliases</a>
807</div>
808<div class="doc_text">
809 <p>Aliases act as "second name" for the aliasee value (which can be either
Anton Korobeynikov96822812008-03-22 08:36:14 +0000[diff] [blame]810 function, global variable, another alias or bitcast of global value). Aliases
811 may have an optional <a href="#linkage">linkage type</a>, and an
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]812 optional <a href="#visibility">visibility style</a>.</p>
813
814 <h5>Syntax:</h5>
815
816<div class="doc_code">
817<pre>
Duncan Sandsd7bfabf2008-09-12 20:48:21 +0000[diff] [blame]818@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]819</pre>
820</div>
821
822</div>
823
824
825
826<!-- ======================================================================= -->
827<div class="doc_subsection"><a name="paramattrs">Parameter Attributes</a></div>
828<div class="doc_text">
829 <p>The return type and each parameter of a function type may have a set of
830 <i>parameter attributes</i> associated with them. Parameter attributes are
831 used to communicate additional information about the result or parameters of
Duncan Sandsf5588dc2007-11-27 13:23:08 +0000[diff] [blame]832 a function. Parameter attributes are considered to be part of the function,
833 not of the function type, so functions with different parameter attributes
834 can have the same function type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]835
836 <p>Parameter attributes are simple keywords that follow the type specified. If
837 multiple parameter attributes are needed, they are space separated. For
838 example:</p>
839
840<div class="doc_code">
841<pre>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]842declare i32 @printf(i8* noalias , ...)
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]843declare i32 @atoi(i8 zeroext)
844declare signext i8 @returns_signed_char()
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]845</pre>
846</div>
847
Duncan Sandsf5588dc2007-11-27 13:23:08 +0000[diff] [blame]848 <p>Note that any attributes for the function result (<tt>nounwind</tt>,
849 <tt>readonly</tt>) come immediately after the argument list.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]850
851 <p>Currently, only the following parameter attributes are defined:</p>
852 <dl>
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]853 <dt><tt>zeroext</tt></dt>
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]854 <dd>This indicates to the code generator that the parameter or return value
855 should be zero-extended to a 32-bit value by the caller (for a parameter)
856 or the callee (for a return value).</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]857
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]858 <dt><tt>signext</tt></dt>
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]859 <dd>This indicates to the code generator that the parameter or return value
860 should be sign-extended to a 32-bit value by the caller (for a parameter)
861 or the callee (for a return value).</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]862
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]863 <dt><tt>inreg</tt></dt>
Dale Johannesenc08a0e22008-09-25 20:47:45 +0000[diff] [blame]864 <dd>This indicates that this parameter or return value should be treated
865 in a special target-dependent fashion during while emitting code for a
866 function call or return (usually, by putting it in a register as opposed
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]867 to memory, though some targets use it to distinguish between two different
868 kinds of registers). Use of this attribute is target-specific.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]869
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]870 <dt><tt><a name="byval">byval</a></tt></dt>
Chris Lattner04c86182008-01-15 04:34:22 +0000[diff] [blame]871 <dd>This indicates that the pointer parameter should really be passed by
872 value to the function. The attribute implies that a hidden copy of the
873 pointee is made between the caller and the callee, so the callee is unable
Chris Lattner6a9f3c42008-08-05 18:21:08 +0000[diff] [blame]874 to modify the value in the callee. This attribute is only valid on LLVM
Chris Lattner04c86182008-01-15 04:34:22 +0000[diff] [blame]875 pointer arguments. It is generally used to pass structs and arrays by
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]876 value, but is also valid on pointers to scalars. The copy is considered to
877 belong to the caller not the callee (for example,
878 <tt><a href="#readonly">readonly</a></tt> functions should not write to
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]879 <tt>byval</tt> parameters). This is not a valid attribute for return
880 values. </dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]881
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]882 <dt><tt>sret</tt></dt>
Duncan Sands616cc032008-02-18 04:19:38 +0000[diff] [blame]883 <dd>This indicates that the pointer parameter specifies the address of a
884 structure that is the return value of the function in the source program.
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]885 This pointer must be guaranteed by the caller to be valid: loads and stores
886 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]887 be applied to the first parameter. This is not a valid attribute for
888 return values. </dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]889
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]890 <dt><tt>noalias</tt></dt>
Owen Andersonc4fc4cd2008-02-18 04:09:01 +0000[diff] [blame]891 <dd>This indicates that the parameter does not alias any global or any other
892 parameter. The caller is responsible for ensuring that this is the case,
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]893 usually by placing the value in a stack allocation. This is not a valid
894 attribute for return values.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]895
Duncan Sands4ee46812007-07-27 19:57:41 +0000[diff] [blame]896 <dt><tt>nest</tt></dt>
Duncan Sandsf1a7d4c2008-07-08 09:27:25 +0000[diff] [blame]897 <dd>This indicates that the pointer parameter can be excised using the
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]898 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
899 attribute for return values.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]900 </dl>
901
902</div>
903
904<!-- ======================================================================= -->
905<div class="doc_subsection">
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]906 <a name="gc">Garbage Collector Names</a>
907</div>
908
909<div class="doc_text">
910<p>Each function may specify a garbage collector name, which is simply a
911string.</p>
912
913<div class="doc_code"><pre
914>define void @f() gc "name" { ...</pre></div>
915
916<p>The compiler declares the supported values of <i>name</i>. Specifying a
917collector which will cause the compiler to alter its output in order to support
918the named garbage collection algorithm.</p>
919</div>
920
921<!-- ======================================================================= -->
922<div class="doc_subsection">
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]923 <a name="fnattrs">Function Attributes</a>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]924</div>
925
926<div class="doc_text">
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]927
928<p>Function attributes are set to communicate additional information about
929 a function. Function attributes are considered to be part of the function,
930 not of the function type, so functions with different parameter attributes
931 can have the same function type.</p>
932
933 <p>Function attributes are simple keywords that follow the type specified. If
934 multiple attributes are needed, they are space separated. For
935 example:</p>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]936
937<div class="doc_code">
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]938<pre>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]939define void @f() noinline { ... }
940define void @f() alwaysinline { ... }
941define void @f() alwaysinline optsize { ... }
942define void @f() optsize
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]943</pre>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]944</div>
945
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]946<dl>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]947<dt><tt>alwaysinline</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]948<dd>This attribute indicates that the inliner should attempt to inline this
949function into callers whenever possible, ignoring any active inlining size
950threshold for this caller.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]951
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]952<dt><tt>noinline</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]953<dd>This attribute indicates that the inliner should never inline this function
Chris Lattner377f3ae2008-10-05 17:14:59 +0000[diff] [blame]954in any situation. This attribute may not be used together with the
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]955<tt>alwaysinline</tt> attribute.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]956
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]957<dt><tt>optsize</tt></dt>
Devang Patelc29099b2008-09-29 18:34:44 +0000[diff] [blame]958<dd>This attribute suggests that optimization passes and code generator passes
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]959make choices that keep the code size of this function low, and otherwise do
960optimizations specifically to reduce code size.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]961
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]962<dt><tt>noreturn</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]963<dd>This function attribute indicates that the function never returns normally.
964This produces undefined behavior at runtime if the function ever does
965dynamically return.</dd>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]966
967<dt><tt>nounwind</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]968<dd>This function attribute indicates that the function never returns with an
969unwind or exceptional control flow. If the function does unwind, its runtime
970behavior is undefined.</dd>
971
972<dt><tt>readnone</tt></dt>
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]973<dd>This attribute indicates that the function computes its result (or the
974exception it throws) based strictly on its arguments, without dereferencing any
975pointer arguments or otherwise accessing any mutable state (e.g. memory, control
976registers, etc) visible to caller functions. It does not write through any
977pointer arguments (including <tt><a href="#byval">byval</a></tt> arguments) and
978never changes any state visible to callers.</dd>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]979
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]980<dt><tt><a name="readonly">readonly</a></tt></dt>
981<dd>This attribute indicates that the function does not write through any
982pointer arguments (including <tt><a href="#byval">byval</a></tt> arguments)
983or otherwise modify any state (e.g. memory, control registers, etc) visible to
984caller functions. It may dereference pointer arguments and read state that may
985be set in the caller. A readonly function always returns the same value (or
986throws the same exception) when called with the same set of arguments and global
987state.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]988</dl>
989
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]990</div>
991
992<!-- ======================================================================= -->
993<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]994 <a name="moduleasm">Module-Level Inline Assembly</a>
995</div>
996
997<div class="doc_text">
998<p>
999Modules may contain "module-level inline asm" blocks, which corresponds to the
1000GCC "file scope inline asm" blocks. These blocks are internally concatenated by
1001LLVM and treated as a single unit, but may be separated in the .ll file if
1002desired. The syntax is very simple:
1003</p>
1004
1005<div class="doc_code">
1006<pre>
1007module asm "inline asm code goes here"
1008module asm "more can go here"
1009</pre>
1010</div>
1011
1012<p>The strings can contain any character by escaping non-printable characters.
1013 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
1014 for the number.
1015</p>
1016
1017<p>
1018 The inline asm code is simply printed to the machine code .s file when
1019 assembly code is generated.
1020</p>
1021</div>
1022
1023<!-- ======================================================================= -->
1024<div class="doc_subsection">
1025 <a name="datalayout">Data Layout</a>
1026</div>
1027
1028<div class="doc_text">
1029<p>A module may specify a target specific data layout string that specifies how
1030data is to be laid out in memory. The syntax for the data layout is simply:</p>
1031<pre> target datalayout = "<i>layout specification</i>"</pre>
1032<p>The <i>layout specification</i> consists of a list of specifications
1033separated by the minus sign character ('-'). Each specification starts with a
1034letter and may include other information after the letter to define some
1035aspect of the data layout. The specifications accepted are as follows: </p>
1036<dl>
1037 <dt><tt>E</tt></dt>
1038 <dd>Specifies that the target lays out data in big-endian form. That is, the
1039 bits with the most significance have the lowest address location.</dd>
1040 <dt><tt>e</tt></dt>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1041 <dd>Specifies that the target lays out data in little-endian form. That is,
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1042 the bits with the least significance have the lowest address location.</dd>
1043 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1044 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
1045 <i>preferred</i> alignments. All sizes are in bits. Specifying the <i>pref</i>
1046 alignment is optional. If omitted, the preceding <tt>:</tt> should be omitted
1047 too.</dd>
1048 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1049 <dd>This specifies the alignment for an integer type of a given bit
1050 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1051 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1052 <dd>This specifies the alignment for a vector type of a given bit
1053 <i>size</i>.</dd>
1054 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1055 <dd>This specifies the alignment for a floating point type of a given bit
1056 <i>size</i>. The value of <i>size</i> must be either 32 (float) or 64
1057 (double).</dd>
1058 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1059 <dd>This specifies the alignment for an aggregate type of a given bit
1060 <i>size</i>.</dd>
1061</dl>
1062<p>When constructing the data layout for a given target, LLVM starts with a
1063default set of specifications which are then (possibly) overriden by the
1064specifications in the <tt>datalayout</tt> keyword. The default specifications
1065are given in this list:</p>
1066<ul>
1067 <li><tt>E</tt> - big endian</li>
1068 <li><tt>p:32:64:64</tt> - 32-bit pointers with 64-bit alignment</li>
1069 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1070 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1071 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1072 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1073 <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]1074 alignment of 64-bits</li>
1075 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1076 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1077 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1078 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1079 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
1080</ul>
Chris Lattner6a9f3c42008-08-05 18:21:08 +0000[diff] [blame]1081<p>When LLVM is determining the alignment for a given type, it uses the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]1082following rules:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1083<ol>
1084 <li>If the type sought is an exact match for one of the specifications, that
1085 specification is used.</li>
1086 <li>If no match is found, and the type sought is an integer type, then the
1087 smallest integer type that is larger than the bitwidth of the sought type is
1088 used. If none of the specifications are larger than the bitwidth then the the
1089 largest integer type is used. For example, given the default specifications
1090 above, the i7 type will use the alignment of i8 (next largest) while both
1091 i65 and i256 will use the alignment of i64 (largest specified).</li>
1092 <li>If no match is found, and the type sought is a vector type, then the
1093 largest vector type that is smaller than the sought vector type will be used
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]1094 as a fall back. This happens because &lt;128 x double&gt; can be implemented
1095 in terms of 64 &lt;2 x double&gt;, for example.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1096</ol>
1097</div>
1098
1099<!-- *********************************************************************** -->
1100<div class="doc_section"> <a name="typesystem">Type System</a> </div>
1101<!-- *********************************************************************** -->
1102
1103<div class="doc_text">
1104
1105<p>The LLVM type system is one of the most important features of the
1106intermediate representation. Being typed enables a number of
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1107optimizations to be performed on the intermediate representation directly,
1108without having to do
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1109extra analyses on the side before the transformation. A strong type
1110system makes it easier to read the generated code and enables novel
1111analyses and transformations that are not feasible to perform on normal
1112three address code representations.</p>
1113
1114</div>
1115
1116<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1117<div class="doc_subsection"> <a name="t_classifications">Type
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1118Classifications</a> </div>
1119<div class="doc_text">
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1120<p>The types fall into a few useful
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1121classifications:</p>
1122
1123<table border="1" cellspacing="0" cellpadding="4">
1124 <tbody>
1125 <tr><th>Classification</th><th>Types</th></tr>
1126 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1127 <td><a href="#t_integer">integer</a></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1128 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
1129 </tr>
1130 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1131 <td><a href="#t_floating">floating point</a></td>
1132 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1133 </tr>
1134 <tr>
1135 <td><a name="t_firstclass">first class</a></td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1136 <td><a href="#t_integer">integer</a>,
1137 <a href="#t_floating">floating point</a>,
1138 <a href="#t_pointer">pointer</a>,
Dan Gohmanf6237db2008-06-18 18:42:13 +0000[diff] [blame]1139 <a href="#t_vector">vector</a>,
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]1140 <a href="#t_struct">structure</a>,
1141 <a href="#t_array">array</a>,
Dan Gohmand13951e2008-05-23 22:50:26 +0000[diff] [blame]1142 <a href="#t_label">label</a>.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1143 </td>
1144 </tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1145 <tr>
1146 <td><a href="#t_primitive">primitive</a></td>
1147 <td><a href="#t_label">label</a>,
1148 <a href="#t_void">void</a>,
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1149 <a href="#t_floating">floating point</a>.</td>
1150 </tr>
1151 <tr>
1152 <td><a href="#t_derived">derived</a></td>
1153 <td><a href="#t_integer">integer</a>,
1154 <a href="#t_array">array</a>,
1155 <a href="#t_function">function</a>,
1156 <a href="#t_pointer">pointer</a>,
1157 <a href="#t_struct">structure</a>,
1158 <a href="#t_pstruct">packed structure</a>,
1159 <a href="#t_vector">vector</a>,
1160 <a href="#t_opaque">opaque</a>.
Dan Gohman032ba852008-10-14 16:32:04 +0000[diff] [blame]1161 </td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1162 </tr>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1163 </tbody>
1164</table>
1165
1166<p>The <a href="#t_firstclass">first class</a> types are perhaps the
1167most important. Values of these types are the only ones which can be
1168produced by instructions, passed as arguments, or used as operands to
Dan Gohman4f29e422008-05-23 21:53:15 +0000[diff] [blame]1169instructions.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1170</div>
1171
1172<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1173<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1174
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1175<div class="doc_text">
1176<p>The primitive types are the fundamental building blocks of the LLVM
1177system.</p>
1178
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1179</div>
1180
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1181<!-- _______________________________________________________________________ -->
1182<div class="doc_subsubsection"> <a name="t_floating">Floating Point Types</a> </div>
1183
1184<div class="doc_text">
1185 <table>
1186 <tbody>
1187 <tr><th>Type</th><th>Description</th></tr>
1188 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1189 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1190 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1191 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1192 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1193 </tbody>
1194 </table>
1195</div>
1196
1197<!-- _______________________________________________________________________ -->
1198<div class="doc_subsubsection"> <a name="t_void">Void Type</a> </div>
1199
1200<div class="doc_text">
1201<h5>Overview:</h5>
1202<p>The void type does not represent any value and has no size.</p>
1203
1204<h5>Syntax:</h5>
1205
1206<pre>
1207 void
1208</pre>
1209</div>
1210
1211<!-- _______________________________________________________________________ -->
1212<div class="doc_subsubsection"> <a name="t_label">Label Type</a> </div>
1213
1214<div class="doc_text">
1215<h5>Overview:</h5>
1216<p>The label type represents code labels.</p>
1217
1218<h5>Syntax:</h5>
1219
1220<pre>
1221 label
1222</pre>
1223</div>
1224
1225
1226<!-- ======================================================================= -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1227<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
1228
1229<div class="doc_text">
1230
1231<p>The real power in LLVM comes from the derived types in the system.
1232This is what allows a programmer to represent arrays, functions,
1233pointers, and other useful types. Note that these derived types may be
1234recursive: For example, it is possible to have a two dimensional array.</p>
1235
1236</div>
1237
1238<!-- _______________________________________________________________________ -->
1239<div class="doc_subsubsection"> <a name="t_integer">Integer Type</a> </div>
1240
1241<div class="doc_text">
1242
1243<h5>Overview:</h5>
1244<p>The integer type is a very simple derived type that simply specifies an
1245arbitrary bit width for the integer type desired. Any bit width from 1 bit to
12462^23-1 (about 8 million) can be specified.</p>
1247
1248<h5>Syntax:</h5>
1249
1250<pre>
1251 iN
1252</pre>
1253
1254<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1255value.</p>
1256
1257<h5>Examples:</h5>
1258<table class="layout">
Chris Lattner251ab812007-12-18 06:18:21 +0000[diff] [blame]1259 <tbody>
1260 <tr>
1261 <td><tt>i1</tt></td>
1262 <td>a single-bit integer.</td>
1263 </tr><tr>
1264 <td><tt>i32</tt></td>
1265 <td>a 32-bit integer.</td>
1266 </tr><tr>
1267 <td><tt>i1942652</tt></td>
1268 <td>a really big integer of over 1 million bits.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1269 </tr>
Chris Lattner251ab812007-12-18 06:18:21 +0000[diff] [blame]1270 </tbody>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1271</table>
1272</div>
1273
1274<!-- _______________________________________________________________________ -->
1275<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
1276
1277<div class="doc_text">
1278
1279<h5>Overview:</h5>
1280
1281<p>The array type is a very simple derived type that arranges elements
1282sequentially in memory. The array type requires a size (number of
1283elements) and an underlying data type.</p>
1284
1285<h5>Syntax:</h5>
1286
1287<pre>
1288 [&lt;# elements&gt; x &lt;elementtype&gt;]
1289</pre>
1290
1291<p>The number of elements is a constant integer value; elementtype may
1292be any type with a size.</p>
1293
1294<h5>Examples:</h5>
1295<table class="layout">
1296 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1297 <td class="left"><tt>[40 x i32]</tt></td>
1298 <td class="left">Array of 40 32-bit integer values.</td>
1299 </tr>
1300 <tr class="layout">
1301 <td class="left"><tt>[41 x i32]</tt></td>
1302 <td class="left">Array of 41 32-bit integer values.</td>
1303 </tr>
1304 <tr class="layout">
1305 <td class="left"><tt>[4 x i8]</tt></td>
1306 <td class="left">Array of 4 8-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1307 </tr>
1308</table>
1309<p>Here are some examples of multidimensional arrays:</p>
1310<table class="layout">
1311 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1312 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1313 <td class="left">3x4 array of 32-bit integer values.</td>
1314 </tr>
1315 <tr class="layout">
1316 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1317 <td class="left">12x10 array of single precision floating point values.</td>
1318 </tr>
1319 <tr class="layout">
1320 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1321 <td class="left">2x3x4 array of 16-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1322 </tr>
1323</table>
1324
1325<p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
1326length array. Normally, accesses past the end of an array are undefined in
1327LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
1328As a special case, however, zero length arrays are recognized to be variable
1329length. This allows implementation of 'pascal style arrays' with the LLVM
1330type "{ i32, [0 x float]}", for example.</p>
1331
1332</div>
1333
1334<!-- _______________________________________________________________________ -->
1335<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
1336<div class="doc_text">
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1337
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1338<h5>Overview:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1339
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1340<p>The function type can be thought of as a function signature. It
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1341consists of a return type and a list of formal parameter types. The
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1342return 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]1343If 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]1344class types, and the struct must have at least one element.</p>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1345
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1346<h5>Syntax:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1347
1348<pre>
1349 &lt;returntype list&gt; (&lt;parameter list&gt;)
1350</pre>
1351
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1352<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
1353specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1354which indicates that the function takes a variable number of arguments.
1355Variable argument functions can access their arguments with the <a
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1356 href="#int_varargs">variable argument handling intrinsic</a> functions.
1357'<tt>&lt;returntype list&gt;</tt>' is a comma-separated list of
1358<a href="#t_firstclass">first class</a> type specifiers.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1359
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1360<h5>Examples:</h5>
1361<table class="layout">
1362 <tr class="layout">
1363 <td class="left"><tt>i32 (i32)</tt></td>
1364 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
1365 </td>
1366 </tr><tr class="layout">
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]1367 <td class="left"><tt>float&nbsp;(i16&nbsp;signext,&nbsp;i32&nbsp;*)&nbsp;*
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1368 </tt></td>
1369 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
1370 an <tt>i16</tt> that should be sign extended and a
1371 <a href="#t_pointer">pointer</a> to <tt>i32</tt>, returning
1372 <tt>float</tt>.
1373 </td>
1374 </tr><tr class="layout">
1375 <td class="left"><tt>i32 (i8*, ...)</tt></td>
1376 <td class="left">A vararg function that takes at least one
1377 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1378 which returns an integer. This is the signature for <tt>printf</tt> in
1379 LLVM.
1380 </td>
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1381 </tr><tr class="layout">
1382 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Devang Patel6dddba22008-03-24 18:10:52 +0000[diff] [blame]1383 <td class="left">A function taking an <tt>i32></tt>, returning two
1384 <tt> i32 </tt> values as an aggregate of type <tt>{ i32, i32 }</tt>
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1385 </td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1386 </tr>
1387</table>
1388
1389</div>
1390<!-- _______________________________________________________________________ -->
1391<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
1392<div class="doc_text">
1393<h5>Overview:</h5>
1394<p>The structure type is used to represent a collection of data members
1395together in memory. The packing of the field types is defined to match
1396the ABI of the underlying processor. The elements of a structure may
1397be any type that has a size.</p>
1398<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1399and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1400field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1401instruction.</p>
1402<h5>Syntax:</h5>
1403<pre> { &lt;type list&gt; }<br></pre>
1404<h5>Examples:</h5>
1405<table class="layout">
1406 <tr class="layout">
1407 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1408 <td class="left">A triple of three <tt>i32</tt> values</td>
1409 </tr><tr class="layout">
1410 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1411 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1412 second element is a <a href="#t_pointer">pointer</a> to a
1413 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1414 an <tt>i32</tt>.</td>
1415 </tr>
1416</table>
1417</div>
1418
1419<!-- _______________________________________________________________________ -->
1420<div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
1421</div>
1422<div class="doc_text">
1423<h5>Overview:</h5>
1424<p>The packed structure type is used to represent a collection of data members
1425together in memory. There is no padding between fields. Further, the alignment
1426of a packed structure is 1 byte. The elements of a packed structure may
1427be any type that has a size.</p>
1428<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1429and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1430field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1431instruction.</p>
1432<h5>Syntax:</h5>
1433<pre> &lt; { &lt;type list&gt; } &gt; <br></pre>
1434<h5>Examples:</h5>
1435<table class="layout">
1436 <tr class="layout">
1437 <td class="left"><tt>&lt; { i32, i32, i32 } &gt;</tt></td>
1438 <td class="left">A triple of three <tt>i32</tt> values</td>
1439 </tr><tr class="layout">
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1440 <td class="left">
1441<tt>&lt;&nbsp;{&nbsp;float,&nbsp;i32&nbsp;(i32)*&nbsp;}&nbsp;&gt;</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1442 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1443 second element is a <a href="#t_pointer">pointer</a> to a
1444 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1445 an <tt>i32</tt>.</td>
1446 </tr>
1447</table>
1448</div>
1449
1450<!-- _______________________________________________________________________ -->
1451<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
1452<div class="doc_text">
1453<h5>Overview:</h5>
1454<p>As in many languages, the pointer type represents a pointer or
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]1455reference to another object, which must live in memory. Pointer types may have
1456an optional address space attribute defining the target-specific numbered
1457address space where the pointed-to object resides. The default address space is
1458zero.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1459<h5>Syntax:</h5>
1460<pre> &lt;type&gt; *<br></pre>
1461<h5>Examples:</h5>
1462<table class="layout">
1463 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1464 <td class="left"><tt>[4x i32]*</tt></td>
1465 <td class="left">A <a href="#t_pointer">pointer</a> to <a
1466 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
1467 </tr>
1468 <tr class="layout">
1469 <td class="left"><tt>i32 (i32 *) *</tt></td>
1470 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1471 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1472 <tt>i32</tt>.</td>
1473 </tr>
1474 <tr class="layout">
1475 <td class="left"><tt>i32 addrspace(5)*</tt></td>
1476 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
1477 that resides in address space #5.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1478 </tr>
1479</table>
1480</div>
1481
1482<!-- _______________________________________________________________________ -->
1483<div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
1484<div class="doc_text">
1485
1486<h5>Overview:</h5>
1487
1488<p>A vector type is a simple derived type that represents a vector
1489of elements. Vector types are used when multiple primitive data
1490are operated in parallel using a single instruction (SIMD).
1491A vector type requires a size (number of
1492elements) and an underlying primitive data type. Vectors must have a power
1493of two length (1, 2, 4, 8, 16 ...). Vector types are
1494considered <a href="#t_firstclass">first class</a>.</p>
1495
1496<h5>Syntax:</h5>
1497
1498<pre>
1499 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1500</pre>
1501
1502<p>The number of elements is a constant integer value; elementtype may
1503be any integer or floating point type.</p>
1504
1505<h5>Examples:</h5>
1506
1507<table class="layout">
1508 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1509 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
1510 <td class="left">Vector of 4 32-bit integer values.</td>
1511 </tr>
1512 <tr class="layout">
1513 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
1514 <td class="left">Vector of 8 32-bit floating-point values.</td>
1515 </tr>
1516 <tr class="layout">
1517 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
1518 <td class="left">Vector of 2 64-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1519 </tr>
1520</table>
1521</div>
1522
1523<!-- _______________________________________________________________________ -->
1524<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1525<div class="doc_text">
1526
1527<h5>Overview:</h5>
1528
1529<p>Opaque types are used to represent unknown types in the system. This
Gordon Henriksenda0706e2007-10-14 00:34:53 +0000[diff] [blame]1530corresponds (for example) to the C notion of a forward declared structure type.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1531In LLVM, opaque types can eventually be resolved to any type (not just a
1532structure type).</p>
1533
1534<h5>Syntax:</h5>
1535
1536<pre>
1537 opaque
1538</pre>
1539
1540<h5>Examples:</h5>
1541
1542<table class="layout">
1543 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1544 <td class="left"><tt>opaque</tt></td>
1545 <td class="left">An opaque type.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1546 </tr>
1547</table>
1548</div>
1549
1550
1551<!-- *********************************************************************** -->
1552<div class="doc_section"> <a name="constants">Constants</a> </div>
1553<!-- *********************************************************************** -->
1554
1555<div class="doc_text">
1556
1557<p>LLVM has several different basic types of constants. This section describes
1558them all and their syntax.</p>
1559
1560</div>
1561
1562<!-- ======================================================================= -->
1563<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
1564
1565<div class="doc_text">
1566
1567<dl>
1568 <dt><b>Boolean constants</b></dt>
1569
1570 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
1571 constants of the <tt><a href="#t_primitive">i1</a></tt> type.
1572 </dd>
1573
1574 <dt><b>Integer constants</b></dt>
1575
1576 <dd>Standard integers (such as '4') are constants of the <a
1577 href="#t_integer">integer</a> type. Negative numbers may be used with
1578 integer types.
1579 </dd>
1580
1581 <dt><b>Floating point constants</b></dt>
1582
1583 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
1584 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]1585 notation (see below). The assembler requires the exact decimal value of
1586 a floating-point constant. For example, the assembler accepts 1.25 but
1587 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
1588 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1589
1590 <dt><b>Null pointer constants</b></dt>
1591
1592 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
1593 and must be of <a href="#t_pointer">pointer type</a>.</dd>
1594
1595</dl>
1596
1597<p>The one non-intuitive notation for constants is the optional hexadecimal form
1598of floating point constants. For example, the form '<tt>double
15990x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
16004.5e+15</tt>'. The only time hexadecimal floating point constants are required
1601(and the only time that they are generated by the disassembler) is when a
1602floating point constant must be emitted but it cannot be represented as a
1603decimal floating point number. For example, NaN's, infinities, and other
1604special values are represented in their IEEE hexadecimal format so that
1605assembly and disassembly do not cause any bits to change in the constants.</p>
1606
1607</div>
1608
1609<!-- ======================================================================= -->
1610<div class="doc_subsection"><a name="aggregateconstants">Aggregate Constants</a>
1611</div>
1612
1613<div class="doc_text">
1614<p>Aggregate constants arise from aggregation of simple constants
1615and smaller aggregate constants.</p>
1616
1617<dl>
1618 <dt><b>Structure constants</b></dt>
1619
1620 <dd>Structure constants are represented with notation similar to structure
1621 type definitions (a comma separated list of elements, surrounded by braces
Chris Lattner6c8de962007-12-25 20:34:52 +0000[diff] [blame]1622 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
1623 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]1624 must have <a href="#t_struct">structure type</a>, and the number and
1625 types of elements must match those specified by the type.
1626 </dd>
1627
1628 <dt><b>Array constants</b></dt>
1629
1630 <dd>Array constants are represented with notation similar to array type
1631 definitions (a comma separated list of elements, surrounded by square brackets
1632 (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74 ]</tt>". Array
1633 constants must have <a href="#t_array">array type</a>, and the number and
1634 types of elements must match those specified by the type.
1635 </dd>
1636
1637 <dt><b>Vector constants</b></dt>
1638
1639 <dd>Vector constants are represented with notation similar to vector type
1640 definitions (a comma separated list of elements, surrounded by
1641 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32 42,
1642 i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must have <a
1643 href="#t_vector">vector type</a>, and the number and types of elements must
1644 match those specified by the type.
1645 </dd>
1646
1647 <dt><b>Zero initialization</b></dt>
1648
1649 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1650 value to zero of <em>any</em> type, including scalar and aggregate types.
1651 This is often used to avoid having to print large zero initializers (e.g. for
1652 large arrays) and is always exactly equivalent to using explicit zero
1653 initializers.
1654 </dd>
1655</dl>
1656
1657</div>
1658
1659<!-- ======================================================================= -->
1660<div class="doc_subsection">
1661 <a name="globalconstants">Global Variable and Function Addresses</a>
1662</div>
1663
1664<div class="doc_text">
1665
1666<p>The addresses of <a href="#globalvars">global variables</a> and <a
1667href="#functionstructure">functions</a> are always implicitly valid (link-time)
1668constants. These constants are explicitly referenced when the <a
1669href="#identifiers">identifier for the global</a> is used and always have <a
1670href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1671file:</p>
1672
1673<div class="doc_code">
1674<pre>
1675@X = global i32 17
1676@Y = global i32 42
1677@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
1678</pre>
1679</div>
1680
1681</div>
1682
1683<!-- ======================================================================= -->
1684<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
1685<div class="doc_text">
1686 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
1687 no specific value. Undefined values may be of any type and be used anywhere
1688 a constant is permitted.</p>
1689
1690 <p>Undefined values indicate to the compiler that the program is well defined
1691 no matter what value is used, giving the compiler more freedom to optimize.
1692 </p>
1693</div>
1694
1695<!-- ======================================================================= -->
1696<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1697</div>
1698
1699<div class="doc_text">
1700
1701<p>Constant expressions are used to allow expressions involving other constants
1702to be used as constants. Constant expressions may be of any <a
1703href="#t_firstclass">first class</a> type and may involve any LLVM operation
1704that does not have side effects (e.g. load and call are not supported). The
1705following is the syntax for constant expressions:</p>
1706
1707<dl>
1708 <dt><b><tt>trunc ( CST to TYPE )</tt></b></dt>
1709 <dd>Truncate a constant to another type. The bit size of CST must be larger
1710 than the bit size of TYPE. Both types must be integers.</dd>
1711
1712 <dt><b><tt>zext ( CST to TYPE )</tt></b></dt>
1713 <dd>Zero extend a constant to another type. The bit size of CST must be
1714 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1715
1716 <dt><b><tt>sext ( CST to TYPE )</tt></b></dt>
1717 <dd>Sign extend a constant to another type. The bit size of CST must be
1718 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1719
1720 <dt><b><tt>fptrunc ( CST to TYPE )</tt></b></dt>
1721 <dd>Truncate a floating point constant to another floating point type. The
1722 size of CST must be larger than the size of TYPE. Both types must be
1723 floating point.</dd>
1724
1725 <dt><b><tt>fpext ( CST to TYPE )</tt></b></dt>
1726 <dd>Floating point extend a constant to another type. The size of CST must be
1727 smaller or equal to the size of TYPE. Both types must be floating point.</dd>
1728
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]1729 <dt><b><tt>fptoui ( CST to TYPE )</tt></b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1730 <dd>Convert a floating point constant to the corresponding unsigned integer
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1731 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1732 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1733 of the same number of elements. If the value won't fit in the integer type,
1734 the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1735
1736 <dt><b><tt>fptosi ( CST to TYPE )</tt></b></dt>
1737 <dd>Convert a floating point constant to the corresponding signed integer
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1738 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1739 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1740 of the same number of elements. If the value won't fit in the integer type,
1741 the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1742
1743 <dt><b><tt>uitofp ( CST to TYPE )</tt></b></dt>
1744 <dd>Convert an unsigned integer constant to the corresponding floating point
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1745 constant. TYPE must be a scalar or vector floating point type. CST must be of
1746 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1747 of the same number of elements. If the value won't fit in the floating point
1748 type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1749
1750 <dt><b><tt>sitofp ( CST to TYPE )</tt></b></dt>
1751 <dd>Convert a signed integer constant to the corresponding floating point
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1752 constant. TYPE must be a scalar or vector floating point type. CST must be of
1753 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1754 of the same number of elements. If the value won't fit in the floating point
1755 type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1756
1757 <dt><b><tt>ptrtoint ( CST to TYPE )</tt></b></dt>
1758 <dd>Convert a pointer typed constant to the corresponding integer constant
1759 TYPE must be an integer type. CST must be of pointer type. The CST value is
1760 zero extended, truncated, or unchanged to make it fit in TYPE.</dd>
1761
1762 <dt><b><tt>inttoptr ( CST to TYPE )</tt></b></dt>
1763 <dd>Convert a integer constant to a pointer constant. TYPE must be a
1764 pointer type. CST must be of integer type. The CST value is zero extended,
1765 truncated, or unchanged to make it fit in a pointer size. This one is
1766 <i>really</i> dangerous!</dd>
1767
1768 <dt><b><tt>bitcast ( CST to TYPE )</tt></b></dt>
1769 <dd>Convert a constant, CST, to another TYPE. The size of CST and TYPE must be
1770 identical (same number of bits). The conversion is done as if the CST value
1771 was stored to memory and read back as TYPE. In other words, no bits change
1772 with this operator, just the type. This can be used for conversion of
1773 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]1774 pointers it is only valid to cast to another pointer type. It is not valid
1775 to bitcast to or from an aggregate type.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1776 </dd>
1777
1778 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
1779
1780 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
1781 constants. As with the <a href="#i_getelementptr">getelementptr</a>
1782 instruction, the index list may have zero or more indexes, which are required
1783 to make sense for the type of "CSTPTR".</dd>
1784
1785 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
1786
1787 <dd>Perform the <a href="#i_select">select operation</a> on
1788 constants.</dd>
1789
1790 <dt><b><tt>icmp COND ( VAL1, VAL2 )</tt></b></dt>
1791 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
1792
1793 <dt><b><tt>fcmp COND ( VAL1, VAL2 )</tt></b></dt>
1794 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
1795
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]1796 <dt><b><tt>vicmp COND ( VAL1, VAL2 )</tt></b></dt>
1797 <dd>Performs the <a href="#i_vicmp">vicmp operation</a> on constants.</dd>
1798
1799 <dt><b><tt>vfcmp COND ( VAL1, VAL2 )</tt></b></dt>
1800 <dd>Performs the <a href="#i_vfcmp">vfcmp operation</a> on constants.</dd>
1801
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1802 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
1803
1804 <dd>Perform the <a href="#i_extractelement">extractelement
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]1805 operation</a> on constants.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1806
1807 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
1808
1809 <dd>Perform the <a href="#i_insertelement">insertelement
1810 operation</a> on constants.</dd>
1811
1812
1813 <dt><b><tt>shufflevector ( VEC1, VEC2, IDXMASK )</tt></b></dt>
1814
1815 <dd>Perform the <a href="#i_shufflevector">shufflevector
1816 operation</a> on constants.</dd>
1817
1818 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
1819
1820 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
1821 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
1822 binary</a> operations. The constraints on operands are the same as those for
1823 the corresponding instruction (e.g. no bitwise operations on floating point
1824 values are allowed).</dd>
1825</dl>
1826</div>
1827
1828<!-- *********************************************************************** -->
1829<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
1830<!-- *********************************************************************** -->
1831
1832<!-- ======================================================================= -->
1833<div class="doc_subsection">
1834<a name="inlineasm">Inline Assembler Expressions</a>
1835</div>
1836
1837<div class="doc_text">
1838
1839<p>
1840LLVM supports inline assembler expressions (as opposed to <a href="#moduleasm">
1841Module-Level Inline Assembly</a>) through the use of a special value. This
1842value represents the inline assembler as a string (containing the instructions
1843to emit), a list of operand constraints (stored as a string), and a flag that
1844indicates whether or not the inline asm expression has side effects. An example
1845inline assembler expression is:
1846</p>
1847
1848<div class="doc_code">
1849<pre>
1850i32 (i32) asm "bswap $0", "=r,r"
1851</pre>
1852</div>
1853
1854<p>
1855Inline assembler expressions may <b>only</b> be used as the callee operand of
1856a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we have:
1857</p>
1858
1859<div class="doc_code">
1860<pre>
1861%X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
1862</pre>
1863</div>
1864
1865<p>
1866Inline asms with side effects not visible in the constraint list must be marked
1867as having side effects. This is done through the use of the
1868'<tt>sideeffect</tt>' keyword, like so:
1869</p>
1870
1871<div class="doc_code">
1872<pre>
1873call void asm sideeffect "eieio", ""()
1874</pre>
1875</div>
1876
1877<p>TODO: The format of the asm and constraints string still need to be
1878documented here. Constraints on what can be done (e.g. duplication, moving, etc
Chris Lattner1cdd64e2008-10-04 18:36:02 +0000[diff] [blame]1879need to be documented). This is probably best done by reference to another
1880document that covers inline asm from a holistic perspective.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1881</p>
1882
1883</div>
1884
1885<!-- *********************************************************************** -->
1886<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
1887<!-- *********************************************************************** -->
1888
1889<div class="doc_text">
1890
1891<p>The LLVM instruction set consists of several different
1892classifications of instructions: <a href="#terminators">terminator
1893instructions</a>, <a href="#binaryops">binary instructions</a>,
1894<a href="#bitwiseops">bitwise binary instructions</a>, <a
1895 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
1896instructions</a>.</p>
1897
1898</div>
1899
1900<!-- ======================================================================= -->
1901<div class="doc_subsection"> <a name="terminators">Terminator
1902Instructions</a> </div>
1903
1904<div class="doc_text">
1905
1906<p>As mentioned <a href="#functionstructure">previously</a>, every
1907basic block in a program ends with a "Terminator" instruction, which
1908indicates which block should be executed after the current block is
1909finished. These terminator instructions typically yield a '<tt>void</tt>'
1910value: they produce control flow, not values (the one exception being
1911the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
1912<p>There are six different terminator instructions: the '<a
1913 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
1914instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
1915the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
1916 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
1917 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
1918
1919</div>
1920
1921<!-- _______________________________________________________________________ -->
1922<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
1923Instruction</a> </div>
1924<div class="doc_text">
1925<h5>Syntax:</h5>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]1926<pre>
1927 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]1928 ret void <i>; Return from void function</i>
1929</pre>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1930
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1931<h5>Overview:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1932
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]1933<p>The '<tt>ret</tt>' instruction is used to return control flow (and
1934optionally a value) from a function back to the caller.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1935<p>There are two forms of the '<tt>ret</tt>' instruction: one that
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]1936returns a value and then causes control flow, and one that just causes
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1937control flow to occur.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1938
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1939<h5>Arguments:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1940
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]1941<p>The '<tt>ret</tt>' instruction optionally accepts a single argument,
1942the return value. The type of the return value must be a
1943'<a href="#t_firstclass">first class</a>' type.</p>
1944
1945<p>A function is not <a href="#wellformed">well formed</a> if
1946it it has a non-void return type and contains a '<tt>ret</tt>'
1947instruction with no return value or a return value with a type that
1948does not match its type, or if it has a void return type and contains
1949a '<tt>ret</tt>' instruction with a return value.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1950
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1951<h5>Semantics:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1952
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1953<p>When the '<tt>ret</tt>' instruction is executed, control flow
1954returns back to the calling function's context. If the caller is a "<a
1955 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
1956the instruction after the call. If the caller was an "<a
1957 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
1958at the beginning of the "normal" destination block. If the instruction
1959returns a value, that value shall set the call or invoke instruction's
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]1960return value.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1961
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1962<h5>Example:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1963
1964<pre>
1965 ret i32 5 <i>; Return an integer value of 5</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1966 ret void <i>; Return from a void function</i>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]1967 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]1968</pre>
1969</div>
1970<!-- _______________________________________________________________________ -->
1971<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
1972<div class="doc_text">
1973<h5>Syntax:</h5>
1974<pre> br i1 &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;<br> br label &lt;dest&gt; <i>; Unconditional branch</i>
1975</pre>
1976<h5>Overview:</h5>
1977<p>The '<tt>br</tt>' instruction is used to cause control flow to
1978transfer to a different basic block in the current function. There are
1979two forms of this instruction, corresponding to a conditional branch
1980and an unconditional branch.</p>
1981<h5>Arguments:</h5>
1982<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
1983single '<tt>i1</tt>' value and two '<tt>label</tt>' values. The
1984unconditional form of the '<tt>br</tt>' instruction takes a single
1985'<tt>label</tt>' value as a target.</p>
1986<h5>Semantics:</h5>
1987<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
1988argument is evaluated. If the value is <tt>true</tt>, control flows
1989to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
1990control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
1991<h5>Example:</h5>
1992<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
1993 href="#i_ret">ret</a> i32 1<br>IfUnequal:<br> <a href="#i_ret">ret</a> i32 0<br></pre>
1994</div>
1995<!-- _______________________________________________________________________ -->
1996<div class="doc_subsubsection">
1997 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
1998</div>
1999
2000<div class="doc_text">
2001<h5>Syntax:</h5>
2002
2003<pre>
2004 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
2005</pre>
2006
2007<h5>Overview:</h5>
2008
2009<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
2010several different places. It is a generalization of the '<tt>br</tt>'
2011instruction, allowing a branch to occur to one of many possible
2012destinations.</p>
2013
2014
2015<h5>Arguments:</h5>
2016
2017<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
2018comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
2019an array of pairs of comparison value constants and '<tt>label</tt>'s. The
2020table is not allowed to contain duplicate constant entries.</p>
2021
2022<h5>Semantics:</h5>
2023
2024<p>The <tt>switch</tt> instruction specifies a table of values and
2025destinations. When the '<tt>switch</tt>' instruction is executed, this
2026table is searched for the given value. If the value is found, control flow is
2027transfered to the corresponding destination; otherwise, control flow is
2028transfered to the default destination.</p>
2029
2030<h5>Implementation:</h5>
2031
2032<p>Depending on properties of the target machine and the particular
2033<tt>switch</tt> instruction, this instruction may be code generated in different
2034ways. For example, it could be generated as a series of chained conditional
2035branches or with a lookup table.</p>
2036
2037<h5>Example:</h5>
2038
2039<pre>
2040 <i>; Emulate a conditional br instruction</i>
2041 %Val = <a href="#i_zext">zext</a> i1 %value to i32
2042 switch i32 %Val, label %truedest [i32 0, label %falsedest ]
2043
2044 <i>; Emulate an unconditional br instruction</i>
2045 switch i32 0, label %dest [ ]
2046
2047 <i>; Implement a jump table:</i>
2048 switch i32 %val, label %otherwise [ i32 0, label %onzero
2049 i32 1, label %onone
2050 i32 2, label %ontwo ]
2051</pre>
2052</div>
2053
2054<!-- _______________________________________________________________________ -->
2055<div class="doc_subsubsection">
2056 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
2057</div>
2058
2059<div class="doc_text">
2060
2061<h5>Syntax:</h5>
2062
2063<pre>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]2064 &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]2065 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
2066</pre>
2067
2068<h5>Overview:</h5>
2069
2070<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
2071function, with the possibility of control flow transfer to either the
2072'<tt>normal</tt>' label or the
2073'<tt>exception</tt>' label. If the callee function returns with the
2074"<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
2075"normal" label. If the callee (or any indirect callees) returns with the "<a
2076href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]2077continued at the dynamically nearest "exception" label.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2078
2079<h5>Arguments:</h5>
2080
2081<p>This instruction requires several arguments:</p>
2082
2083<ol>
2084 <li>
2085 The optional "cconv" marker indicates which <a href="#callingconv">calling
2086 convention</a> the call should use. If none is specified, the call defaults
2087 to using C calling conventions.
2088 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]2089
2090 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
2091 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>',
2092 and '<tt>inreg</tt>' attributes are valid here.</li>
2093
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2094 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
2095 function value being invoked. In most cases, this is a direct function
2096 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
2097 an arbitrary pointer to function value.
2098 </li>
2099
2100 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
2101 function to be invoked. </li>
2102
2103 <li>'<tt>function args</tt>': argument list whose types match the function
2104 signature argument types. If the function signature indicates the function
2105 accepts a variable number of arguments, the extra arguments can be
2106 specified. </li>
2107
2108 <li>'<tt>normal label</tt>': the label reached when the called function
2109 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
2110
2111 <li>'<tt>exception label</tt>': the label reached when a callee returns with
2112 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
2113
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]2114 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]2115 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
2116 '<tt>readnone</tt>' attributes are valid here.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2117</ol>
2118
2119<h5>Semantics:</h5>
2120
2121<p>This instruction is designed to operate as a standard '<tt><a
2122href="#i_call">call</a></tt>' instruction in most regards. The primary
2123difference is that it establishes an association with a label, which is used by
2124the runtime library to unwind the stack.</p>
2125
2126<p>This instruction is used in languages with destructors to ensure that proper
2127cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
2128exception. Additionally, this is important for implementation of
2129'<tt>catch</tt>' clauses in high-level languages that support them.</p>
2130
2131<h5>Example:</h5>
2132<pre>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2133 %retval = invoke i32 @Test(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2134 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2135 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2136 unwind label %TestCleanup <i>; {i32}:retval set</i>
2137</pre>
2138</div>
2139
2140
2141<!-- _______________________________________________________________________ -->
2142
2143<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
2144Instruction</a> </div>
2145
2146<div class="doc_text">
2147
2148<h5>Syntax:</h5>
2149<pre>
2150 unwind
2151</pre>
2152
2153<h5>Overview:</h5>
2154
2155<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
2156at the first callee in the dynamic call stack which used an <a
2157href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
2158primarily used to implement exception handling.</p>
2159
2160<h5>Semantics:</h5>
2161
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]2162<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2163immediately halt. The dynamic call stack is then searched for the first <a
2164href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
2165execution continues at the "exceptional" destination block specified by the
2166<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
2167dynamic call chain, undefined behavior results.</p>
2168</div>
2169
2170<!-- _______________________________________________________________________ -->
2171
2172<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
2173Instruction</a> </div>
2174
2175<div class="doc_text">
2176
2177<h5>Syntax:</h5>
2178<pre>
2179 unreachable
2180</pre>
2181
2182<h5>Overview:</h5>
2183
2184<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
2185instruction is used to inform the optimizer that a particular portion of the
2186code is not reachable. This can be used to indicate that the code after a
2187no-return function cannot be reached, and other facts.</p>
2188
2189<h5>Semantics:</h5>
2190
2191<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
2192</div>
2193
2194
2195
2196<!-- ======================================================================= -->
2197<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
2198<div class="doc_text">
2199<p>Binary operators are used to do most of the computation in a
Chris Lattnerab596d92008-04-01 18:47:32 +0000[diff] [blame]2200program. They require two operands of the same type, execute an operation on them, and
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2201produce a single value. The operands might represent
2202multiple 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]2203The result value has the same type as its operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2204<p>There are several different binary operators:</p>
2205</div>
2206<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2207<div class="doc_subsubsection">
2208 <a name="i_add">'<tt>add</tt>' Instruction</a>
2209</div>
2210
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2211<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2212
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2213<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2214
2215<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2216 &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]2217</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2218
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2219<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2220
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2221<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2222
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2223<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2224
2225<p>The two arguments to the '<tt>add</tt>' instruction must be <a
2226 href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>, or
2227 <a href="#t_vector">vector</a> values. Both arguments must have identical
2228 types.</p>
2229
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2230<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2231
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2232<p>The value produced is the integer or floating point sum of the two
2233operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2234
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2235<p>If an integer sum has unsigned overflow, the result returned is the
2236mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2237the result.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2238
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2239<p>Because LLVM integers use a two's complement representation, this
2240instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2241
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2242<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2243
2244<pre>
2245 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2246</pre>
2247</div>
2248<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2249<div class="doc_subsubsection">
2250 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
2251</div>
2252
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2253<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2254
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2255<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2256
2257<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2258 &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]2259</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2260
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2261<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2262
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2263<p>The '<tt>sub</tt>' instruction returns the difference of its two
2264operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2265
2266<p>Note that the '<tt>sub</tt>' instruction is used to represent the
2267'<tt>neg</tt>' instruction present in most other intermediate
2268representations.</p>
2269
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2270<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2271
2272<p>The two arguments to the '<tt>sub</tt>' instruction must be <a
2273 href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>,
2274 or <a href="#t_vector">vector</a> values. Both arguments must have identical
2275 types.</p>
2276
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2277<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2278
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2279<p>The value produced is the integer or floating point difference of
2280the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2281
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2282<p>If an integer difference has unsigned overflow, the result returned is the
2283mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2284the result.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2285
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2286<p>Because LLVM integers use a two's complement representation, this
2287instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2288
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2289<h5>Example:</h5>
2290<pre>
2291 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
2292 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
2293</pre>
2294</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2295
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2296<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2297<div class="doc_subsubsection">
2298 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
2299</div>
2300
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2301<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2302
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2303<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2304<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]2305</pre>
2306<h5>Overview:</h5>
2307<p>The '<tt>mul</tt>' instruction returns the product of its two
2308operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2309
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2310<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2311
2312<p>The two arguments to the '<tt>mul</tt>' instruction must be <a
2313href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>,
2314or <a href="#t_vector">vector</a> values. Both arguments must have identical
2315types.</p>
2316
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2317<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2318
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2319<p>The value produced is the integer or floating point product of the
2320two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2321
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2322<p>If the result of an integer multiplication has unsigned overflow,
2323the result returned is the mathematical result modulo
23242<sup>n</sup>, where n is the bit width of the result.</p>
2325<p>Because LLVM integers use a two's complement representation, and the
2326result is the same width as the operands, this instruction returns the
2327correct result for both signed and unsigned integers. If a full product
2328(e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands
2329should be sign-extended or zero-extended as appropriate to the
2330width of the full product.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2331<h5>Example:</h5>
2332<pre> &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
2333</pre>
2334</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2335
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2336<!-- _______________________________________________________________________ -->
2337<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
2338</a></div>
2339<div class="doc_text">
2340<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2341<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]2342</pre>
2343<h5>Overview:</h5>
2344<p>The '<tt>udiv</tt>' instruction returns the quotient of its two
2345operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2346
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2347<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2348
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2349<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2350<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2351values. Both arguments must have identical types.</p>
2352
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2353<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2354
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2355<p>The value produced is the unsigned integer quotient of the two operands.</p>
2356<p>Note that unsigned integer division and signed integer division are distinct
2357operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
2358<p>Division by zero leads to undefined behavior.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2359<h5>Example:</h5>
2360<pre> &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2361</pre>
2362</div>
2363<!-- _______________________________________________________________________ -->
2364<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
2365</a> </div>
2366<div class="doc_text">
2367<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2368<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2369 &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]2370</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2371
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2372<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2373
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2374<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two
2375operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2376
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2377<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2378
2379<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
2380<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2381values. Both arguments must have identical types.</p>
2382
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2383<h5>Semantics:</h5>
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2384<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]2385<p>Note that signed integer division and unsigned integer division are distinct
2386operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
2387<p>Division by zero leads to undefined behavior. Overflow also leads to
2388undefined behavior; this is a rare case, but can occur, for example,
2389by doing a 32-bit division of -2147483648 by -1.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2390<h5>Example:</h5>
2391<pre> &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2392</pre>
2393</div>
2394<!-- _______________________________________________________________________ -->
2395<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
2396Instruction</a> </div>
2397<div class="doc_text">
2398<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2399<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2400 &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]2401</pre>
2402<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2403
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2404<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two
2405operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2406
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2407<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2408
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2409<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2410<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2411of floating point values. Both arguments must have identical types.</p>
2412
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2413<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2414
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2415<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2416
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2417<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2418
2419<pre>
2420 &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]2421</pre>
2422</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2423
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2424<!-- _______________________________________________________________________ -->
2425<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
2426</div>
2427<div class="doc_text">
2428<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2429<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]2430</pre>
2431<h5>Overview:</h5>
2432<p>The '<tt>urem</tt>' instruction returns the remainder from the
2433unsigned division of its two arguments.</p>
2434<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2435<p>The two arguments to the '<tt>urem</tt>' instruction must be
2436<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2437values. Both arguments must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2438<h5>Semantics:</h5>
2439<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2440This instruction always performs an unsigned division to get the remainder.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2441<p>Note that unsigned integer remainder and signed integer remainder are
2442distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
2443<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2444<h5>Example:</h5>
2445<pre> &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2446</pre>
2447
2448</div>
2449<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2450<div class="doc_subsubsection">
2451 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
2452</div>
2453
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2454<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2455
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2456<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2457
2458<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2459 &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]2460</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2461
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2462<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2463
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2464<p>The '<tt>srem</tt>' instruction returns the remainder from the
Dan Gohman3e3fd8c2007-11-05 23:35:22 +0000[diff] [blame]2465signed division of its two operands. This instruction can also take
2466<a href="#t_vector">vector</a> versions of the values in which case
2467the elements must be integers.</p>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]2468
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2469<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2470
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2471<p>The two arguments to the '<tt>srem</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2472<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2473values. Both arguments must have identical types.</p>
2474
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2475<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2476
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2477<p>This instruction returns the <i>remainder</i> of a division (where the result
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2478has the same sign as the dividend, <tt>op1</tt>), not the <i>modulo</i>
2479operator (where the result has the same sign as the divisor, <tt>op2</tt>) of
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2480a value. For more information about the difference, see <a
2481 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
2482Math Forum</a>. For a table of how this is implemented in various languages,
2483please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
2484Wikipedia: modulo operation</a>.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2485<p>Note that signed integer remainder and unsigned integer remainder are
2486distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
2487<p>Taking the remainder of a division by zero leads to undefined behavior.
2488Overflow also leads to undefined behavior; this is a rare case, but can occur,
2489for example, by taking the remainder of a 32-bit division of -2147483648 by -1.
2490(The remainder doesn't actually overflow, but this rule lets srem be
2491implemented using instructions that return both the result of the division
2492and the remainder.)</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2493<h5>Example:</h5>
2494<pre> &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2495</pre>
2496
2497</div>
2498<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2499<div class="doc_subsubsection">
2500 <a name="i_frem">'<tt>frem</tt>' Instruction</a> </div>
2501
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2502<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2503
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2504<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2505<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]2506</pre>
2507<h5>Overview:</h5>
2508<p>The '<tt>frem</tt>' instruction returns the remainder from the
2509division of its two operands.</p>
2510<h5>Arguments:</h5>
2511<p>The two arguments to the '<tt>frem</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2512<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2513of floating point values. Both arguments must have identical types.</p>
2514
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2515<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2516
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2517<p>This instruction returns the <i>remainder</i> of a division.
2518The remainder has the same sign as the dividend.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2519
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2520<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2521
2522<pre>
2523 &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]2524</pre>
2525</div>
2526
2527<!-- ======================================================================= -->
2528<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
2529Operations</a> </div>
2530<div class="doc_text">
2531<p>Bitwise binary operators are used to do various forms of
2532bit-twiddling in a program. They are generally very efficient
2533instructions and can commonly be strength reduced from other
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2534instructions. They require two operands of the same type, execute an operation on them,
2535and 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]2536</div>
2537
2538<!-- _______________________________________________________________________ -->
2539<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
2540Instruction</a> </div>
2541<div class="doc_text">
2542<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2543<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]2544</pre>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2545
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2546<h5>Overview:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2547
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2548<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
2549the left a specified number of bits.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2550
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2551<h5>Arguments:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2552
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2553<p>Both arguments to the '<tt>shl</tt>' instruction must be the same <a
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]2554 href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2555type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2556
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2557<h5>Semantics:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2558
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2559<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod 2<sup>n</sup>,
2560where n is the width of the result. If <tt>op2</tt> is (statically or dynamically) negative or
2561equal 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]2562
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2563<h5>Example:</h5><pre>
2564 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
2565 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
2566 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2567 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2568</pre>
2569</div>
2570<!-- _______________________________________________________________________ -->
2571<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
2572Instruction</a> </div>
2573<div class="doc_text">
2574<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2575<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]2576</pre>
2577
2578<h5>Overview:</h5>
2579<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
2580operand shifted to the right a specified number of bits with zero fill.</p>
2581
2582<h5>Arguments:</h5>
2583<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]2584<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]2585type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2586
2587<h5>Semantics:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2588
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2589<p>This instruction always performs a logical shift right operation. The most
2590significant bits of the result will be filled with zero bits after the
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2591shift. If <tt>op2</tt> is (statically or dynamically) equal to or larger than
2592the number of bits in <tt>op1</tt>, the result is undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2593
2594<h5>Example:</h5>
2595<pre>
2596 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
2597 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
2598 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
2599 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2600 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2601</pre>
2602</div>
2603
2604<!-- _______________________________________________________________________ -->
2605<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
2606Instruction</a> </div>
2607<div class="doc_text">
2608
2609<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2610<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]2611</pre>
2612
2613<h5>Overview:</h5>
2614<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
2615operand shifted to the right a specified number of bits with sign extension.</p>
2616
2617<h5>Arguments:</h5>
2618<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]2619<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]2620type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2621
2622<h5>Semantics:</h5>
2623<p>This instruction always performs an arithmetic shift right operation,
2624The most significant bits of the result will be filled with the sign bit
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2625of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
2626larger than the number of bits in <tt>op1</tt>, the result is undefined.
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2627</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2628
2629<h5>Example:</h5>
2630<pre>
2631 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
2632 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
2633 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
2634 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2635 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2636</pre>
2637</div>
2638
2639<!-- _______________________________________________________________________ -->
2640<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
2641Instruction</a> </div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2642
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2643<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2644
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2645<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2646
2647<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2648 &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]2649</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2650
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2651<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2652
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2653<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
2654its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2655
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2656<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2657
2658<p>The two arguments to the '<tt>and</tt>' instruction must be
2659<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2660values. Both arguments must have identical types.</p>
2661
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2662<h5>Semantics:</h5>
2663<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
2664<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]2665<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2666<table border="1" cellspacing="0" cellpadding="4">
2667 <tbody>
2668 <tr>
2669 <td>In0</td>
2670 <td>In1</td>
2671 <td>Out</td>
2672 </tr>
2673 <tr>
2674 <td>0</td>
2675 <td>0</td>
2676 <td>0</td>
2677 </tr>
2678 <tr>
2679 <td>0</td>
2680 <td>1</td>
2681 <td>0</td>
2682 </tr>
2683 <tr>
2684 <td>1</td>
2685 <td>0</td>
2686 <td>0</td>
2687 </tr>
2688 <tr>
2689 <td>1</td>
2690 <td>1</td>
2691 <td>1</td>
2692 </tr>
2693 </tbody>
2694</table>
2695</div>
2696<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2697<pre>
2698 &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]2699 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
2700 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
2701</pre>
2702</div>
2703<!-- _______________________________________________________________________ -->
2704<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
2705<div class="doc_text">
2706<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2707<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]2708</pre>
2709<h5>Overview:</h5>
2710<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
2711or of its two operands.</p>
2712<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2713
2714<p>The two arguments to the '<tt>or</tt>' instruction must be
2715<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2716values. Both arguments must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2717<h5>Semantics:</h5>
2718<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
2719<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]2720<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2721<table border="1" cellspacing="0" cellpadding="4">
2722 <tbody>
2723 <tr>
2724 <td>In0</td>
2725 <td>In1</td>
2726 <td>Out</td>
2727 </tr>
2728 <tr>
2729 <td>0</td>
2730 <td>0</td>
2731 <td>0</td>
2732 </tr>
2733 <tr>
2734 <td>0</td>
2735 <td>1</td>
2736 <td>1</td>
2737 </tr>
2738 <tr>
2739 <td>1</td>
2740 <td>0</td>
2741 <td>1</td>
2742 </tr>
2743 <tr>
2744 <td>1</td>
2745 <td>1</td>
2746 <td>1</td>
2747 </tr>
2748 </tbody>
2749</table>
2750</div>
2751<h5>Example:</h5>
2752<pre> &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
2753 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
2754 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
2755</pre>
2756</div>
2757<!-- _______________________________________________________________________ -->
2758<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
2759Instruction</a> </div>
2760<div class="doc_text">
2761<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2762<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]2763</pre>
2764<h5>Overview:</h5>
2765<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
2766or of its two operands. The <tt>xor</tt> is used to implement the
2767"one's complement" operation, which is the "~" operator in C.</p>
2768<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2769<p>The two arguments to the '<tt>xor</tt>' instruction must be
2770<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2771values. Both arguments must have identical types.</p>
2772
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2773<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2774
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2775<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
2776<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]2777<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2778<table border="1" cellspacing="0" cellpadding="4">
2779 <tbody>
2780 <tr>
2781 <td>In0</td>
2782 <td>In1</td>
2783 <td>Out</td>
2784 </tr>
2785 <tr>
2786 <td>0</td>
2787 <td>0</td>
2788 <td>0</td>
2789 </tr>
2790 <tr>
2791 <td>0</td>
2792 <td>1</td>
2793 <td>1</td>
2794 </tr>
2795 <tr>
2796 <td>1</td>
2797 <td>0</td>
2798 <td>1</td>
2799 </tr>
2800 <tr>
2801 <td>1</td>
2802 <td>1</td>
2803 <td>0</td>
2804 </tr>
2805 </tbody>
2806</table>
2807</div>
2808<p> </p>
2809<h5>Example:</h5>
2810<pre> &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
2811 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
2812 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
2813 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
2814</pre>
2815</div>
2816
2817<!-- ======================================================================= -->
2818<div class="doc_subsection">
2819 <a name="vectorops">Vector Operations</a>
2820</div>
2821
2822<div class="doc_text">
2823
2824<p>LLVM supports several instructions to represent vector operations in a
2825target-independent manner. These instructions cover the element-access and
2826vector-specific operations needed to process vectors effectively. While LLVM
2827does directly support these vector operations, many sophisticated algorithms
2828will want to use target-specific intrinsics to take full advantage of a specific
2829target.</p>
2830
2831</div>
2832
2833<!-- _______________________________________________________________________ -->
2834<div class="doc_subsubsection">
2835 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
2836</div>
2837
2838<div class="doc_text">
2839
2840<h5>Syntax:</h5>
2841
2842<pre>
2843 &lt;result&gt; = extractelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, i32 &lt;idx&gt; <i>; yields &lt;ty&gt;</i>
2844</pre>
2845
2846<h5>Overview:</h5>
2847
2848<p>
2849The '<tt>extractelement</tt>' instruction extracts a single scalar
2850element from a vector at a specified index.
2851</p>
2852
2853
2854<h5>Arguments:</h5>
2855
2856<p>
2857The first operand of an '<tt>extractelement</tt>' instruction is a
2858value of <a href="#t_vector">vector</a> type. The second operand is
2859an index indicating the position from which to extract the element.
2860The index may be a variable.</p>
2861
2862<h5>Semantics:</h5>
2863
2864<p>
2865The result is a scalar of the same type as the element type of
2866<tt>val</tt>. Its value is the value at position <tt>idx</tt> of
2867<tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
2868results are undefined.
2869</p>
2870
2871<h5>Example:</h5>
2872
2873<pre>
2874 %result = extractelement &lt;4 x i32&gt; %vec, i32 0 <i>; yields i32</i>
2875</pre>
2876</div>
2877
2878
2879<!-- _______________________________________________________________________ -->
2880<div class="doc_subsubsection">
2881 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
2882</div>
2883
2884<div class="doc_text">
2885
2886<h5>Syntax:</h5>
2887
2888<pre>
Dan Gohmanbcc3c502008-05-12 23:38:42 +0000[diff] [blame]2889 &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]2890</pre>
2891
2892<h5>Overview:</h5>
2893
2894<p>
2895The '<tt>insertelement</tt>' instruction inserts a scalar
2896element into a vector at a specified index.
2897</p>
2898
2899
2900<h5>Arguments:</h5>
2901
2902<p>
2903The first operand of an '<tt>insertelement</tt>' instruction is a
2904value of <a href="#t_vector">vector</a> type. The second operand is a
2905scalar value whose type must equal the element type of the first
2906operand. The third operand is an index indicating the position at
2907which to insert the value. The index may be a variable.</p>
2908
2909<h5>Semantics:</h5>
2910
2911<p>
2912The result is a vector of the same type as <tt>val</tt>. Its
2913element values are those of <tt>val</tt> except at position
2914<tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
2915exceeds the length of <tt>val</tt>, the results are undefined.
2916</p>
2917
2918<h5>Example:</h5>
2919
2920<pre>
2921 %result = insertelement &lt;4 x i32&gt; %vec, i32 1, i32 0 <i>; yields &lt;4 x i32&gt;</i>
2922</pre>
2923</div>
2924
2925<!-- _______________________________________________________________________ -->
2926<div class="doc_subsubsection">
2927 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
2928</div>
2929
2930<div class="doc_text">
2931
2932<h5>Syntax:</h5>
2933
2934<pre>
2935 &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>
2936</pre>
2937
2938<h5>Overview:</h5>
2939
2940<p>
2941The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
2942from two input vectors, returning a vector of the same type.
2943</p>
2944
2945<h5>Arguments:</h5>
2946
2947<p>
2948The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
2949with types that match each other and types that match the result of the
2950instruction. The third argument is a shuffle mask, which has the same number
2951of elements as the other vector type, but whose element type is always 'i32'.
2952</p>
2953
2954<p>
2955The shuffle mask operand is required to be a constant vector with either
2956constant integer or undef values.
2957</p>
2958
2959<h5>Semantics:</h5>
2960
2961<p>
2962The elements of the two input vectors are numbered from left to right across
2963both of the vectors. The shuffle mask operand specifies, for each element of
2964the result vector, which element of the two input registers the result element
2965gets. The element selector may be undef (meaning "don't care") and the second
2966operand may be undef if performing a shuffle from only one vector.
2967</p>
2968
2969<h5>Example:</h5>
2970
2971<pre>
2972 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
2973 &lt;4 x i32&gt; &lt;i32 0, i32 4, i32 1, i32 5&gt; <i>; yields &lt;4 x i32&gt;</i>
2974 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; undef,
2975 &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.
2976</pre>
2977</div>
2978
2979
2980<!-- ======================================================================= -->
2981<div class="doc_subsection">
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]2982 <a name="aggregateops">Aggregate Operations</a>
2983</div>
2984
2985<div class="doc_text">
2986
2987<p>LLVM supports several instructions for working with aggregate values.
2988</p>
2989
2990</div>
2991
2992<!-- _______________________________________________________________________ -->
2993<div class="doc_subsubsection">
2994 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
2995</div>
2996
2997<div class="doc_text">
2998
2999<h5>Syntax:</h5>
3000
3001<pre>
3002 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
3003</pre>
3004
3005<h5>Overview:</h5>
3006
3007<p>
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3008The '<tt>extractvalue</tt>' instruction extracts the value of a struct field
3009or array element from an aggregate value.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3010</p>
3011
3012
3013<h5>Arguments:</h5>
3014
3015<p>
3016The first operand of an '<tt>extractvalue</tt>' instruction is a
3017value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a>
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3018type. The operands are constant indices to specify which value to extract
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3019in a similar manner as indices in a
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3020'<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
3021</p>
3022
3023<h5>Semantics:</h5>
3024
3025<p>
3026The result is the value at the position in the aggregate specified by
3027the index operands.
3028</p>
3029
3030<h5>Example:</h5>
3031
3032<pre>
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3033 %result = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3034</pre>
3035</div>
3036
3037
3038<!-- _______________________________________________________________________ -->
3039<div class="doc_subsubsection">
3040 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
3041</div>
3042
3043<div class="doc_text">
3044
3045<h5>Syntax:</h5>
3046
3047<pre>
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3048 &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]3049</pre>
3050
3051<h5>Overview:</h5>
3052
3053<p>
3054The '<tt>insertvalue</tt>' instruction inserts a value
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3055into a struct field or array element in an aggregate.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3056</p>
3057
3058
3059<h5>Arguments:</h5>
3060
3061<p>
3062The first operand of an '<tt>insertvalue</tt>' instruction is a
3063value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a> type.
3064The second operand is a first-class value to insert.
Dan Gohman4f29e422008-05-23 21:53:15 +0000[diff] [blame]3065The following operands are constant indices
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3066indicating the position at which to insert the value in a similar manner as
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3067indices in a
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3068'<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
3069The value to insert must have the same type as the value identified
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3070by the indices.
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3071</p>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3072
3073<h5>Semantics:</h5>
3074
3075<p>
3076The result is an aggregate of the same type as <tt>val</tt>. Its
3077value is that of <tt>val</tt> except that the value at the position
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3078specified by the indices is that of <tt>elt</tt>.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3079</p>
3080
3081<h5>Example:</h5>
3082
3083<pre>
Dan Gohmanb1aab4e2008-06-23 15:26:37 +0000[diff] [blame]3084 %result = insertvalue {i32, float} %agg, i32 1, 0 <i>; yields {i32, float}</i>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3085</pre>
3086</div>
3087
3088
3089<!-- ======================================================================= -->
3090<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3091 <a name="memoryops">Memory Access and Addressing Operations</a>
3092</div>
3093
3094<div class="doc_text">
3095
3096<p>A key design point of an SSA-based representation is how it
3097represents memory. In LLVM, no memory locations are in SSA form, which
3098makes things very simple. This section describes how to read, write,
3099allocate, and free memory in LLVM.</p>
3100
3101</div>
3102
3103<!-- _______________________________________________________________________ -->
3104<div class="doc_subsubsection">
3105 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
3106</div>
3107
3108<div class="doc_text">
3109
3110<h5>Syntax:</h5>
3111
3112<pre>
3113 &lt;result&gt; = malloc &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
3114</pre>
3115
3116<h5>Overview:</h5>
3117
3118<p>The '<tt>malloc</tt>' instruction allocates memory from the system
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]3119heap and returns a pointer to it. The object is always allocated in the generic
3120address space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3121
3122<h5>Arguments:</h5>
3123
3124<p>The '<tt>malloc</tt>' instruction allocates
3125<tt>sizeof(&lt;type&gt;)*NumElements</tt>
3126bytes of memory from the operating system and returns a pointer of the
3127appropriate type to the program. If "NumElements" is specified, it is the
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3128number of elements allocated, otherwise "NumElements" is defaulted to be one.
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3129If a constant alignment is specified, the value result of the allocation is guaranteed to
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3130be aligned to at least that boundary. If not specified, or if zero, the target can
3131choose to align the allocation on any convenient boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3132
3133<p>'<tt>type</tt>' must be a sized type.</p>
3134
3135<h5>Semantics:</h5>
3136
3137<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]3138a pointer is returned. The result of a zero byte allocattion is undefined. The
3139result is null if there is insufficient memory available.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3140
3141<h5>Example:</h5>
3142
3143<pre>
3144 %array = malloc [4 x i8 ] <i>; yields {[%4 x i8]*}:array</i>
3145
3146 %size = <a href="#i_add">add</a> i32 2, 2 <i>; yields {i32}:size = i32 4</i>
3147 %array1 = malloc i8, i32 4 <i>; yields {i8*}:array1</i>
3148 %array2 = malloc [12 x i8], i32 %size <i>; yields {[12 x i8]*}:array2</i>
3149 %array3 = malloc i32, i32 4, align 1024 <i>; yields {i32*}:array3</i>
3150 %array4 = malloc i32, align 1024 <i>; yields {i32*}:array4</i>
3151</pre>
3152</div>
3153
3154<!-- _______________________________________________________________________ -->
3155<div class="doc_subsubsection">
3156 <a name="i_free">'<tt>free</tt>' Instruction</a>
3157</div>
3158
3159<div class="doc_text">
3160
3161<h5>Syntax:</h5>
3162
3163<pre>
3164 free &lt;type&gt; &lt;value&gt; <i>; yields {void}</i>
3165</pre>
3166
3167<h5>Overview:</h5>
3168
3169<p>The '<tt>free</tt>' instruction returns memory back to the unused
3170memory heap to be reallocated in the future.</p>
3171
3172<h5>Arguments:</h5>
3173
3174<p>'<tt>value</tt>' shall be a pointer value that points to a value
3175that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
3176instruction.</p>
3177
3178<h5>Semantics:</h5>
3179
3180<p>Access to the memory pointed to by the pointer is no longer defined
Chris Lattner329d6532008-04-19 22:41:32 +0000[diff] [blame]3181after this instruction executes. If the pointer is null, the operation
3182is a noop.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3183
3184<h5>Example:</h5>
3185
3186<pre>
3187 %array = <a href="#i_malloc">malloc</a> [4 x i8] <i>; yields {[4 x i8]*}:array</i>
3188 free [4 x i8]* %array
3189</pre>
3190</div>
3191
3192<!-- _______________________________________________________________________ -->
3193<div class="doc_subsubsection">
3194 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
3195</div>
3196
3197<div class="doc_text">
3198
3199<h5>Syntax:</h5>
3200
3201<pre>
3202 &lt;result&gt; = alloca &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
3203</pre>
3204
3205<h5>Overview:</h5>
3206
3207<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
3208currently executing function, to be automatically released when this function
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]3209returns to its caller. The object is always allocated in the generic address
3210space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3211
3212<h5>Arguments:</h5>
3213
3214<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
3215bytes of memory on the runtime stack, returning a pointer of the
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3216appropriate type to the program. If "NumElements" is specified, it is the
3217number of elements allocated, otherwise "NumElements" is defaulted to be one.
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3218If a constant alignment is specified, the value result of the allocation is guaranteed
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3219to be aligned to at least that boundary. If not specified, or if zero, the target
3220can choose to align the allocation on any convenient boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3221
3222<p>'<tt>type</tt>' may be any sized type.</p>
3223
3224<h5>Semantics:</h5>
3225
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]3226<p>Memory is allocated; a pointer is returned. The operation is undefiend if
3227there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3228memory is automatically released when the function returns. The '<tt>alloca</tt>'
3229instruction is commonly used to represent automatic variables that must
3230have an address available. When the function returns (either with the <tt><a
3231 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]3232instructions), the memory is reclaimed. Allocating zero bytes
3233is legal, but the result is undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3234
3235<h5>Example:</h5>
3236
3237<pre>
3238 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
3239 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
3240 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
3241 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
3242</pre>
3243</div>
3244
3245<!-- _______________________________________________________________________ -->
3246<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
3247Instruction</a> </div>
3248<div class="doc_text">
3249<h5>Syntax:</h5>
3250<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>
3251<h5>Overview:</h5>
3252<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
3253<h5>Arguments:</h5>
3254<p>The argument to the '<tt>load</tt>' instruction specifies the memory
3255address from which to load. The pointer must point to a <a
3256 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
3257marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
3258the number or order of execution of this <tt>load</tt> with other
3259volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
3260instructions. </p>
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3261<p>
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3262The optional constant "align" argument specifies the alignment of the operation
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3263(that is, the alignment of the memory address). A value of 0 or an
3264omitted "align" argument means that the operation has the preferential
3265alignment for the target. It is the responsibility of the code emitter
3266to ensure that the alignment information is correct. Overestimating
3267the alignment results in an undefined behavior. Underestimating the
3268alignment may produce less efficient code. An alignment of 1 is always
3269safe.
3270</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3271<h5>Semantics:</h5>
3272<p>The location of memory pointed to is loaded.</p>
3273<h5>Examples:</h5>
3274<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
3275 <a
3276 href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
3277 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
3278</pre>
3279</div>
3280<!-- _______________________________________________________________________ -->
3281<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
3282Instruction</a> </div>
3283<div class="doc_text">
3284<h5>Syntax:</h5>
3285<pre> store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
3286 volatile store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
3287</pre>
3288<h5>Overview:</h5>
3289<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
3290<h5>Arguments:</h5>
3291<p>There are two arguments to the '<tt>store</tt>' instruction: a value
3292to 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]3293operand must be a pointer to the <a href="#t_firstclass">first class</a> type
3294of the '<tt>&lt;value&gt;</tt>'
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3295operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
3296optimizer is not allowed to modify the number or order of execution of
3297this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
3298 href="#i_store">store</a></tt> instructions.</p>
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3299<p>
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3300The optional constant "align" argument specifies the alignment of the operation
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3301(that is, the alignment of the memory address). A value of 0 or an
3302omitted "align" argument means that the operation has the preferential
3303alignment for the target. It is the responsibility of the code emitter
3304to ensure that the alignment information is correct. Overestimating
3305the alignment results in an undefined behavior. Underestimating the
3306alignment may produce less efficient code. An alignment of 1 is always
3307safe.
3308</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3309<h5>Semantics:</h5>
3310<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>'
3311at the location specified by the '<tt>&lt;pointer&gt;</tt>' operand.</p>
3312<h5>Example:</h5>
3313<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling63ffa142007-10-22 05:10:05 +0000[diff] [blame]3314 store i32 3, i32* %ptr <i>; yields {void}</i>
3315 %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]3316</pre>
3317</div>
3318
3319<!-- _______________________________________________________________________ -->
3320<div class="doc_subsubsection">
3321 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
3322</div>
3323
3324<div class="doc_text">
3325<h5>Syntax:</h5>
3326<pre>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3327 &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]3328</pre>
3329
3330<h5>Overview:</h5>
3331
3332<p>
3333The '<tt>getelementptr</tt>' instruction is used to get the address of a
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3334subelement of an aggregate data structure. It performs address calculation only
3335and does not access memory.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3336
3337<h5>Arguments:</h5>
3338
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3339<p>The first argument is always a pointer, and forms the basis of the
3340calculation. The remaining arguments are indices, that indicate which of the
3341elements of the aggregate object are indexed. The interpretation of each index
3342is dependent on the type being indexed into. The first index always indexes the
3343pointer value given as the first argument, the second index indexes a value of
3344the type pointed to (not necessarily the value directly pointed to, since the
3345first index can be non-zero), etc. The first type indexed into must be a pointer
3346value, subsequent types can be arrays, vectors and structs. Note that subsequent
3347types being indexed into can never be pointers, since that would require loading
3348the pointer before continuing calculation.</p>
3349
3350<p>The type of each index argument depends on the type it is indexing into.
3351When indexing into a (packed) structure, only <tt>i32</tt> integer
3352<b>constants</b> are allowed. When indexing into an array, pointer or vector,
3353only integers of 32 or 64 bits are allowed (also non-constants). 32-bit values
3354will be sign extended to 64-bits if required.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3355
3356<p>For example, let's consider a C code fragment and how it gets
3357compiled to LLVM:</p>
3358
3359<div class="doc_code">
3360<pre>
3361struct RT {
3362 char A;
3363 int B[10][20];
3364 char C;
3365};
3366struct ST {
3367 int X;
3368 double Y;
3369 struct RT Z;
3370};
3371
3372int *foo(struct ST *s) {
3373 return &amp;s[1].Z.B[5][13];
3374}
3375</pre>
3376</div>
3377
3378<p>The LLVM code generated by the GCC frontend is:</p>
3379
3380<div class="doc_code">
3381<pre>
3382%RT = type { i8 , [10 x [20 x i32]], i8 }
3383%ST = type { i32, double, %RT }
3384
3385define i32* %foo(%ST* %s) {
3386entry:
3387 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
3388 ret i32* %reg
3389}
3390</pre>
3391</div>
3392
3393<h5>Semantics:</h5>
3394
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3395<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
3396type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
3397}</tt>' type, a structure. The second index indexes into the third element of
3398the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
3399i8 }</tt>' type, another structure. The third index indexes into the second
3400element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
3401array. The two dimensions of the array are subscripted into, yielding an
3402'<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
3403to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
3404
3405<p>Note that it is perfectly legal to index partially through a
3406structure, returning a pointer to an inner element. Because of this,
3407the LLVM code for the given testcase is equivalent to:</p>
3408
3409<pre>
3410 define i32* %foo(%ST* %s) {
3411 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
3412 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
3413 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
3414 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
3415 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
3416 ret i32* %t5
3417 }
3418</pre>
3419
3420<p>Note that it is undefined to access an array out of bounds: array and
3421pointer indexes must always be within the defined bounds of the array type.
Chris Lattnera7d94ba2008-04-24 05:59:56 +0000[diff] [blame]3422The one exception for this rule is zero length arrays. These arrays are
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3423defined to be accessible as variable length arrays, which requires access
3424beyond the zero'th element.</p>
3425
3426<p>The getelementptr instruction is often confusing. For some more insight
3427into how it works, see <a href="GetElementPtr.html">the getelementptr
3428FAQ</a>.</p>
3429
3430<h5>Example:</h5>
3431
3432<pre>
3433 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3434 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
3435 <i>; yields i8*:vptr</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3436 %vptr = getelementptr {i32, &lt;2 x i8&gt;}* %svptr, i64 0, i32 1, i32 1
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3437 <i>; yields i8*:eptr</i>
3438 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3439</pre>
3440</div>
3441
3442<!-- ======================================================================= -->
3443<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
3444</div>
3445<div class="doc_text">
3446<p>The instructions in this category are the conversion instructions (casting)
3447which all take a single operand and a type. They perform various bit conversions
3448on the operand.</p>
3449</div>
3450
3451<!-- _______________________________________________________________________ -->
3452<div class="doc_subsubsection">
3453 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
3454</div>
3455<div class="doc_text">
3456
3457<h5>Syntax:</h5>
3458<pre>
3459 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3460</pre>
3461
3462<h5>Overview:</h5>
3463<p>
3464The '<tt>trunc</tt>' instruction truncates its operand to the type <tt>ty2</tt>.
3465</p>
3466
3467<h5>Arguments:</h5>
3468<p>
3469The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
3470be an <a href="#t_integer">integer</a> type, and a type that specifies the size
3471and type of the result, which must be an <a href="#t_integer">integer</a>
3472type. The bit size of <tt>value</tt> must be larger than the bit size of
3473<tt>ty2</tt>. Equal sized types are not allowed.</p>
3474
3475<h5>Semantics:</h5>
3476<p>
3477The '<tt>trunc</tt>' instruction truncates the high order bits in <tt>value</tt>
3478and converts the remaining bits to <tt>ty2</tt>. Since the source size must be
3479larger than the destination size, <tt>trunc</tt> cannot be a <i>no-op cast</i>.
3480It will always truncate bits.</p>
3481
3482<h5>Example:</h5>
3483<pre>
3484 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
3485 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
3486 %Y = trunc i32 122 to i1 <i>; yields i1:false</i>
3487</pre>
3488</div>
3489
3490<!-- _______________________________________________________________________ -->
3491<div class="doc_subsubsection">
3492 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
3493</div>
3494<div class="doc_text">
3495
3496<h5>Syntax:</h5>
3497<pre>
3498 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3499</pre>
3500
3501<h5>Overview:</h5>
3502<p>The '<tt>zext</tt>' instruction zero extends its operand to type
3503<tt>ty2</tt>.</p>
3504
3505
3506<h5>Arguments:</h5>
3507<p>The '<tt>zext</tt>' instruction takes a value to cast, which must be of
3508<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3509also be of <a href="#t_integer">integer</a> type. The bit size of the
3510<tt>value</tt> must be smaller than the bit size of the destination type,
3511<tt>ty2</tt>.</p>
3512
3513<h5>Semantics:</h5>
3514<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
3515bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
3516
3517<p>When zero extending from i1, the result will always be either 0 or 1.</p>
3518
3519<h5>Example:</h5>
3520<pre>
3521 %X = zext i32 257 to i64 <i>; yields i64:257</i>
3522 %Y = zext i1 true to i32 <i>; yields i32:1</i>
3523</pre>
3524</div>
3525
3526<!-- _______________________________________________________________________ -->
3527<div class="doc_subsubsection">
3528 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
3529</div>
3530<div class="doc_text">
3531
3532<h5>Syntax:</h5>
3533<pre>
3534 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3535</pre>
3536
3537<h5>Overview:</h5>
3538<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
3539
3540<h5>Arguments:</h5>
3541<p>
3542The '<tt>sext</tt>' instruction takes a value to cast, which must be of
3543<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3544also be of <a href="#t_integer">integer</a> type. The bit size of the
3545<tt>value</tt> must be smaller than the bit size of the destination type,
3546<tt>ty2</tt>.</p>
3547
3548<h5>Semantics:</h5>
3549<p>
3550The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
3551bit (highest order bit) of the <tt>value</tt> until it reaches the bit size of
3552the type <tt>ty2</tt>.</p>
3553
3554<p>When sign extending from i1, the extension always results in -1 or 0.</p>
3555
3556<h5>Example:</h5>
3557<pre>
3558 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
3559 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
3560</pre>
3561</div>
3562
3563<!-- _______________________________________________________________________ -->
3564<div class="doc_subsubsection">
3565 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
3566</div>
3567
3568<div class="doc_text">
3569
3570<h5>Syntax:</h5>
3571
3572<pre>
3573 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3574</pre>
3575
3576<h5>Overview:</h5>
3577<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
3578<tt>ty2</tt>.</p>
3579
3580
3581<h5>Arguments:</h5>
3582<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
3583 point</a> value to cast and a <a href="#t_floating">floating point</a> type to
3584cast it to. The size of <tt>value</tt> must be larger than the size of
3585<tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
3586<i>no-op cast</i>.</p>
3587
3588<h5>Semantics:</h5>
3589<p> The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
3590<a href="#t_floating">floating point</a> type to a smaller
3591<a href="#t_floating">floating point</a> type. If the value cannot fit within
3592the destination type, <tt>ty2</tt>, then the results are undefined.</p>
3593
3594<h5>Example:</h5>
3595<pre>
3596 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
3597 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
3598</pre>
3599</div>
3600
3601<!-- _______________________________________________________________________ -->
3602<div class="doc_subsubsection">
3603 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
3604</div>
3605<div class="doc_text">
3606
3607<h5>Syntax:</h5>
3608<pre>
3609 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3610</pre>
3611
3612<h5>Overview:</h5>
3613<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
3614floating point value.</p>
3615
3616<h5>Arguments:</h5>
3617<p>The '<tt>fpext</tt>' instruction takes a
3618<a href="#t_floating">floating point</a> <tt>value</tt> to cast,
3619and a <a href="#t_floating">floating point</a> type to cast it to. The source
3620type must be smaller than the destination type.</p>
3621
3622<h5>Semantics:</h5>
3623<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
3624<a href="#t_floating">floating point</a> type to a larger
3625<a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
3626used to make a <i>no-op cast</i> because it always changes bits. Use
3627<tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
3628
3629<h5>Example:</h5>
3630<pre>
3631 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
3632 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
3633</pre>
3634</div>
3635
3636<!-- _______________________________________________________________________ -->
3637<div class="doc_subsubsection">
3638 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
3639</div>
3640<div class="doc_text">
3641
3642<h5>Syntax:</h5>
3643<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3644 &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]3645</pre>
3646
3647<h5>Overview:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3648<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3649unsigned integer equivalent of type <tt>ty2</tt>.
3650</p>
3651
3652<h5>Arguments:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3653<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]3654scalar or vector <a href="#t_floating">floating point</a> value, and a type
3655to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
3656type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
3657vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3658
3659<h5>Semantics:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3660<p> The '<tt>fptoui</tt>' instruction converts its
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3661<a href="#t_floating">floating point</a> operand into the nearest (rounding
3662towards zero) unsigned integer value. If the value cannot fit in <tt>ty2</tt>,
3663the results are undefined.</p>
3664
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3665<h5>Example:</h5>
3666<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3667 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]3668 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3669 %X = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3670</pre>
3671</div>
3672
3673<!-- _______________________________________________________________________ -->
3674<div class="doc_subsubsection">
3675 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
3676</div>
3677<div class="doc_text">
3678
3679<h5>Syntax:</h5>
3680<pre>
3681 &lt;result&gt; = fptosi &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3682</pre>
3683
3684<h5>Overview:</h5>
3685<p>The '<tt>fptosi</tt>' instruction converts
3686<a href="#t_floating">floating point</a> <tt>value</tt> to type <tt>ty2</tt>.
3687</p>
3688
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3689<h5>Arguments:</h5>
3690<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]3691scalar or vector <a href="#t_floating">floating point</a> value, and a type
3692to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
3693type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
3694vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3695
3696<h5>Semantics:</h5>
3697<p>The '<tt>fptosi</tt>' instruction converts its
3698<a href="#t_floating">floating point</a> operand into the nearest (rounding
3699towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
3700the results are undefined.</p>
3701
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3702<h5>Example:</h5>
3703<pre>
3704 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]3705 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3706 %X = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
3707</pre>
3708</div>
3709
3710<!-- _______________________________________________________________________ -->
3711<div class="doc_subsubsection">
3712 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
3713</div>
3714<div class="doc_text">
3715
3716<h5>Syntax:</h5>
3717<pre>
3718 &lt;result&gt; = uitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3719</pre>
3720
3721<h5>Overview:</h5>
3722<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
3723integer and converts that value to the <tt>ty2</tt> type.</p>
3724
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3725<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]3726<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be a
3727scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
3728to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
3729type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
3730floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3731
3732<h5>Semantics:</h5>
3733<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
3734integer quantity and converts it to the corresponding floating point value. If
3735the value cannot fit in the floating point value, the results are undefined.</p>
3736
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3737<h5>Example:</h5>
3738<pre>
3739 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3740 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3741</pre>
3742</div>
3743
3744<!-- _______________________________________________________________________ -->
3745<div class="doc_subsubsection">
3746 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
3747</div>
3748<div class="doc_text">
3749
3750<h5>Syntax:</h5>
3751<pre>
3752 &lt;result&gt; = sitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3753</pre>
3754
3755<h5>Overview:</h5>
3756<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed
3757integer and converts that value to the <tt>ty2</tt> type.</p>
3758
3759<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]3760<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be a
3761scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
3762to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
3763type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
3764floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3765
3766<h5>Semantics:</h5>
3767<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed
3768integer quantity and converts it to the corresponding floating point value. If
3769the value cannot fit in the floating point value, the results are undefined.</p>
3770
3771<h5>Example:</h5>
3772<pre>
3773 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3774 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3775</pre>
3776</div>
3777
3778<!-- _______________________________________________________________________ -->
3779<div class="doc_subsubsection">
3780 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
3781</div>
3782<div class="doc_text">
3783
3784<h5>Syntax:</h5>
3785<pre>
3786 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3787</pre>
3788
3789<h5>Overview:</h5>
3790<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
3791the integer type <tt>ty2</tt>.</p>
3792
3793<h5>Arguments:</h5>
3794<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
3795must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3796<tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3797
3798<h5>Semantics:</h5>
3799<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
3800<tt>ty2</tt> by interpreting the pointer value as an integer and either
3801truncating or zero extending that value to the size of the integer type. If
3802<tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
3803<tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
3804are the same size, then nothing is done (<i>no-op cast</i>) other than a type
3805change.</p>
3806
3807<h5>Example:</h5>
3808<pre>
3809 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
3810 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
3811</pre>
3812</div>
3813
3814<!-- _______________________________________________________________________ -->
3815<div class="doc_subsubsection">
3816 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
3817</div>
3818<div class="doc_text">
3819
3820<h5>Syntax:</h5>
3821<pre>
3822 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3823</pre>
3824
3825<h5>Overview:</h5>
3826<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to
3827a pointer type, <tt>ty2</tt>.</p>
3828
3829<h5>Arguments:</h5>
3830<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
3831value to cast, and a type to cast it to, which must be a
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3832<a href="#t_pointer">pointer</a> type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3833
3834<h5>Semantics:</h5>
3835<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
3836<tt>ty2</tt> by applying either a zero extension or a truncation depending on
3837the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
3838size of a pointer then a truncation is done. If <tt>value</tt> is smaller than
3839the size of a pointer then a zero extension is done. If they are the same size,
3840nothing is done (<i>no-op cast</i>).</p>
3841
3842<h5>Example:</h5>
3843<pre>
3844 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
3845 %X = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
3846 %Y = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
3847</pre>
3848</div>
3849
3850<!-- _______________________________________________________________________ -->
3851<div class="doc_subsubsection">
3852 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
3853</div>
3854<div class="doc_text">
3855
3856<h5>Syntax:</h5>
3857<pre>
3858 &lt;result&gt; = bitcast &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3859</pre>
3860
3861<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3862
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3863<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
3864<tt>ty2</tt> without changing any bits.</p>
3865
3866<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3867
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3868<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be
Dan Gohman7305fa02008-09-08 16:45:59 +0000[diff] [blame]3869a non-aggregate first class value, and a type to cast it to, which must also be
3870a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes of
3871<tt>value</tt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3872and the destination type, <tt>ty2</tt>, must be identical. If the source
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3873type is a pointer, the destination type must also be a pointer. This
3874instruction supports bitwise conversion of vectors to integers and to vectors
3875of other types (as long as they have the same size).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3876
3877<h5>Semantics:</h5>
3878<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
3879<tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
3880this conversion. The conversion is done as if the <tt>value</tt> had been
3881stored to memory and read back as type <tt>ty2</tt>. Pointer types may only be
3882converted to other pointer types with this instruction. To convert pointers to
3883other types, use the <a href="#i_inttoptr">inttoptr</a> or
3884<a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
3885
3886<h5>Example:</h5>
3887<pre>
3888 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
3889 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3890 %Z = bitcast &lt;2 x int&gt; %V to i64; <i>; yields i64: %V</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3891</pre>
3892</div>
3893
3894<!-- ======================================================================= -->
3895<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
3896<div class="doc_text">
3897<p>The instructions in this category are the "miscellaneous"
3898instructions, which defy better classification.</p>
3899</div>
3900
3901<!-- _______________________________________________________________________ -->
3902<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
3903</div>
3904<div class="doc_text">
3905<h5>Syntax:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3906<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]3907</pre>
3908<h5>Overview:</h5>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]3909<p>The '<tt>icmp</tt>' instruction returns a boolean value or
3910a vector of boolean values based on comparison
3911of its two integer, integer vector, or pointer operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3912<h5>Arguments:</h5>
3913<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
3914the condition code indicating the kind of comparison to perform. It is not
3915a value, just a keyword. The possible condition code are:
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3916</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3917<ol>
3918 <li><tt>eq</tt>: equal</li>
3919 <li><tt>ne</tt>: not equal </li>
3920 <li><tt>ugt</tt>: unsigned greater than</li>
3921 <li><tt>uge</tt>: unsigned greater or equal</li>
3922 <li><tt>ult</tt>: unsigned less than</li>
3923 <li><tt>ule</tt>: unsigned less or equal</li>
3924 <li><tt>sgt</tt>: signed greater than</li>
3925 <li><tt>sge</tt>: signed greater or equal</li>
3926 <li><tt>slt</tt>: signed less than</li>
3927 <li><tt>sle</tt>: signed less or equal</li>
3928</ol>
3929<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]3930<a href="#t_pointer">pointer</a>
3931or integer <a href="#t_vector">vector</a> typed.
3932They must also be identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3933<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3934<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]3935the condition code given as <tt>cond</tt>. The comparison performed always
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]3936yields either an <a href="#t_primitive"><tt>i1</tt></a> or vector of <tt>i1</tt> result, as follows:
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3937</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3938<ol>
3939 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
3940 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
3941 </li>
3942 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3943 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3944 <li><tt>ugt</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 greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3946 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3947 <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]3948 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3949 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3950 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3951 <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]3952 <li><tt>sgt</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 greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3954 <li><tt>sge</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3955 <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]3956 <li><tt>slt</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3957 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3958 <li><tt>sle</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3959 <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]3960</ol>
3961<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
3962values are compared as if they were integers.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]3963<p>If the operands are integer vectors, then they are compared
3964element by element. The result is an <tt>i1</tt> vector with
3965the same number of elements as the values being compared.
3966Otherwise, the result is an <tt>i1</tt>.
3967</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3968
3969<h5>Example:</h5>
3970<pre> &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
3971 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
3972 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
3973 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
3974 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
3975 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
3976</pre>
3977</div>
3978
3979<!-- _______________________________________________________________________ -->
3980<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
3981</div>
3982<div class="doc_text">
3983<h5>Syntax:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3984<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]3985</pre>
3986<h5>Overview:</h5>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]3987<p>The '<tt>fcmp</tt>' instruction returns a boolean value
3988or vector of boolean values based on comparison
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3989of its operands.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]3990<p>
3991If the operands are floating point scalars, then the result
3992type is a boolean (<a href="#t_primitive"><tt>i1</tt></a>).
3993</p>
3994<p>If the operands are floating point vectors, then the result type
3995is a vector of boolean with the same number of elements as the
3996operands being compared.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3997<h5>Arguments:</h5>
3998<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
3999the condition code indicating the kind of comparison to perform. It is not
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4000a value, just a keyword. The possible condition code are:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4001<ol>
4002 <li><tt>false</tt>: no comparison, always returns false</li>
4003 <li><tt>oeq</tt>: ordered and equal</li>
4004 <li><tt>ogt</tt>: ordered and greater than </li>
4005 <li><tt>oge</tt>: ordered and greater than or equal</li>
4006 <li><tt>olt</tt>: ordered and less than </li>
4007 <li><tt>ole</tt>: ordered and less than or equal</li>
4008 <li><tt>one</tt>: ordered and not equal</li>
4009 <li><tt>ord</tt>: ordered (no nans)</li>
4010 <li><tt>ueq</tt>: unordered or equal</li>
4011 <li><tt>ugt</tt>: unordered or greater than </li>
4012 <li><tt>uge</tt>: unordered or greater than or equal</li>
4013 <li><tt>ult</tt>: unordered or less than </li>
4014 <li><tt>ule</tt>: unordered or less than or equal</li>
4015 <li><tt>une</tt>: unordered or not equal</li>
4016 <li><tt>uno</tt>: unordered (either nans)</li>
4017 <li><tt>true</tt>: no comparison, always returns true</li>
4018</ol>
4019<p><i>Ordered</i> means that neither operand is a QNAN while
4020<i>unordered</i> means that either operand may be a QNAN.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4021<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be
4022either a <a href="#t_floating">floating point</a> type
4023or a <a href="#t_vector">vector</a> of floating point type.
4024They must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4025<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4026<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4027according to the condition code given as <tt>cond</tt>.
4028If the operands are vectors, then the vectors are compared
4029element by element.
4030Each comparison performed
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4031always yields an <a href="#t_primitive">i1</a> result, as follows:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4032<ol>
4033 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
4034 <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]4035 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4036 <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]4037 <tt>op1</tt> is greather than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4038 <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]4039 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4040 <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]4041 <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4042 <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]4043 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4044 <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]4045 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4046 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
4047 <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]4048 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4049 <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]4050 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4051 <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]4052 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4053 <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]4054 <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4055 <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]4056 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4057 <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]4058 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4059 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
4060 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
4061</ol>
4062
4063<h5>Example:</h5>
4064<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]4065 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
4066 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
4067 &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]4068</pre>
4069</div>
4070
4071<!-- _______________________________________________________________________ -->
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4072<div class="doc_subsubsection">
4073 <a name="i_vicmp">'<tt>vicmp</tt>' Instruction</a>
4074</div>
4075<div class="doc_text">
4076<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4077<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]4078</pre>
4079<h5>Overview:</h5>
4080<p>The '<tt>vicmp</tt>' instruction returns an integer vector value based on
4081element-wise comparison of its two integer vector operands.</p>
4082<h5>Arguments:</h5>
4083<p>The '<tt>vicmp</tt>' instruction takes three operands. The first operand is
4084the condition code indicating the kind of comparison to perform. It is not
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4085a value, just a keyword. The possible condition code are:</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4086<ol>
4087 <li><tt>eq</tt>: equal</li>
4088 <li><tt>ne</tt>: not equal </li>
4089 <li><tt>ugt</tt>: unsigned greater than</li>
4090 <li><tt>uge</tt>: unsigned greater or equal</li>
4091 <li><tt>ult</tt>: unsigned less than</li>
4092 <li><tt>ule</tt>: unsigned less or equal</li>
4093 <li><tt>sgt</tt>: signed greater than</li>
4094 <li><tt>sge</tt>: signed greater or equal</li>
4095 <li><tt>slt</tt>: signed less than</li>
4096 <li><tt>sle</tt>: signed less or equal</li>
4097</ol>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4098<p>The remaining two arguments must be <a href="#t_vector">vector</a> or
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4099<a href="#t_integer">integer</a> typed. They must also be identical types.</p>
4100<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4101<p>The '<tt>vicmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4102according to the condition code given as <tt>cond</tt>. The comparison yields a
4103<a href="#t_vector">vector</a> of <a href="#t_integer">integer</a> result, of
4104identical type as the values being compared. The most significant bit in each
4105element is 1 if the element-wise comparison evaluates to true, and is 0
4106otherwise. All other bits of the result are undefined. The condition codes
4107are evaluated identically to the <a href="#i_icmp">'<tt>icmp</tt>'
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4108instruction</a>.</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4109
4110<h5>Example:</h5>
4111<pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4112 &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>
4113 &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]4114</pre>
4115</div>
4116
4117<!-- _______________________________________________________________________ -->
4118<div class="doc_subsubsection">
4119 <a name="i_vfcmp">'<tt>vfcmp</tt>' Instruction</a>
4120</div>
4121<div class="doc_text">
4122<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4123<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]4124<h5>Overview:</h5>
4125<p>The '<tt>vfcmp</tt>' instruction returns an integer vector value based on
4126element-wise comparison of its two floating point vector operands. The output
4127elements have the same width as the input elements.</p>
4128<h5>Arguments:</h5>
4129<p>The '<tt>vfcmp</tt>' instruction takes three operands. The first operand is
4130the condition code indicating the kind of comparison to perform. It is not
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4131a value, just a keyword. The possible condition code are:</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4132<ol>
4133 <li><tt>false</tt>: no comparison, always returns false</li>
4134 <li><tt>oeq</tt>: ordered and equal</li>
4135 <li><tt>ogt</tt>: ordered and greater than </li>
4136 <li><tt>oge</tt>: ordered and greater than or equal</li>
4137 <li><tt>olt</tt>: ordered and less than </li>
4138 <li><tt>ole</tt>: ordered and less than or equal</li>
4139 <li><tt>one</tt>: ordered and not equal</li>
4140 <li><tt>ord</tt>: ordered (no nans)</li>
4141 <li><tt>ueq</tt>: unordered or equal</li>
4142 <li><tt>ugt</tt>: unordered or greater than </li>
4143 <li><tt>uge</tt>: unordered or greater than or equal</li>
4144 <li><tt>ult</tt>: unordered or less than </li>
4145 <li><tt>ule</tt>: unordered or less than or equal</li>
4146 <li><tt>une</tt>: unordered or not equal</li>
4147 <li><tt>uno</tt>: unordered (either nans)</li>
4148 <li><tt>true</tt>: no comparison, always returns true</li>
4149</ol>
4150<p>The remaining two arguments must be <a href="#t_vector">vector</a> of
4151<a href="#t_floating">floating point</a> typed. They must also be identical
4152types.</p>
4153<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4154<p>The '<tt>vfcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4155according to the condition code given as <tt>cond</tt>. The comparison yields a
4156<a href="#t_vector">vector</a> of <a href="#t_integer">integer</a> result, with
4157an identical number of elements as the values being compared, and each element
4158having identical with to the width of the floating point elements. The most
4159significant bit in each element is 1 if the element-wise comparison evaluates to
4160true, and is 0 otherwise. All other bits of the result are undefined. The
4161condition codes are evaluated identically to the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4162<a href="#i_fcmp">'<tt>fcmp</tt>' instruction</a>.</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4163
4164<h5>Example:</h5>
4165<pre>
Chris Lattner1e5c5cd02008-10-13 16:55:18 +0000[diff] [blame]4166 <i>; yields: result=&lt;2 x i32&gt; &lt; i32 0, i32 -1 &gt;</i>
4167 &lt;result&gt; = vfcmp oeq &lt;2 x float&gt; &lt; float 4, float 0 &gt;, &lt; float 5, float 0 &gt;
4168
4169 <i>; yields: result=&lt;2 x i64&gt; &lt; i64 -1, i64 0 &gt;</i>
4170 &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]4171</pre>
4172</div>
4173
4174<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4175<div class="doc_subsubsection">
4176 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
4177</div>
4178
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4179<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4180
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4181<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4182
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4183<pre> &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...<br></pre>
4184<h5>Overview:</h5>
4185<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in
4186the SSA graph representing the function.</p>
4187<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4188
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4189<p>The type of the incoming values is specified with the first type
4190field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
4191as arguments, with one pair for each predecessor basic block of the
4192current block. Only values of <a href="#t_firstclass">first class</a>
4193type may be used as the value arguments to the PHI node. Only labels
4194may be used as the label arguments.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4195
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4196<p>There must be no non-phi instructions between the start of a basic
4197block and the PHI instructions: i.e. PHI instructions must be first in
4198a basic 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>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4201
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4202<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
4203specified by the pair corresponding to the predecessor basic block that executed
4204just prior to the current block.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4205
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4206<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4207<pre>
4208Loop: ; Infinite loop that counts from 0 on up...
4209 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
4210 %nextindvar = add i32 %indvar, 1
4211 br label %Loop
4212</pre>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4213</div>
4214
4215<!-- _______________________________________________________________________ -->
4216<div class="doc_subsubsection">
4217 <a name="i_select">'<tt>select</tt>' Instruction</a>
4218</div>
4219
4220<div class="doc_text">
4221
4222<h5>Syntax:</h5>
4223
4224<pre>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4225 &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>
4226
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4227 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4228</pre>
4229
4230<h5>Overview:</h5>
4231
4232<p>
4233The '<tt>select</tt>' instruction is used to choose one value based on a
4234condition, without branching.
4235</p>
4236
4237
4238<h5>Arguments:</h5>
4239
4240<p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4241The '<tt>select</tt>' instruction requires an 'i1' value or
4242a vector of 'i1' values indicating the
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4243condition, and two values of the same <a href="#t_firstclass">first class</a>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4244type. If the val1/val2 are vectors and
4245the condition is a scalar, then entire vectors are selected, not
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4246individual elements.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4247</p>
4248
4249<h5>Semantics:</h5>
4250
4251<p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4252If 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]4253value argument; otherwise, it returns the second value argument.
4254</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4255<p>
4256If the condition is a vector of i1, then the value arguments must
4257be vectors of the same size, and the selection is done element
4258by element.
4259</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4260
4261<h5>Example:</h5>
4262
4263<pre>
4264 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
4265</pre>
4266</div>
4267
4268
4269<!-- _______________________________________________________________________ -->
4270<div class="doc_subsubsection">
4271 <a name="i_call">'<tt>call</tt>' Instruction</a>
4272</div>
4273
4274<div class="doc_text">
4275
4276<h5>Syntax:</h5>
4277<pre>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]4278 &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]4279</pre>
4280
4281<h5>Overview:</h5>
4282
4283<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
4284
4285<h5>Arguments:</h5>
4286
4287<p>This instruction requires several arguments:</p>
4288
4289<ol>
4290 <li>
4291 <p>The optional "tail" marker indicates whether the callee function accesses
4292 any allocas or varargs in the caller. If the "tail" marker is present, the
4293 function call is eligible for tail call optimization. Note that calls may
4294 be marked "tail" even if they do not occur before a <a
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4295 href="#i_ret"><tt>ret</tt></a> instruction.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4296 </li>
4297 <li>
4298 <p>The optional "cconv" marker indicates which <a href="#callingconv">calling
4299 convention</a> the call should use. If none is specified, the call defaults
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4300 to using C calling conventions.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4301 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4302
4303 <li>
4304 <p>The optional <a href="#paramattrs">Parameter Attributes</a> list for
4305 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>',
4306 and '<tt>inreg</tt>' attributes are valid here.</p>
4307 </li>
4308
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4309 <li>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]4310 <p>'<tt>ty</tt>': the type of the call instruction itself which is also
4311 the type of the return value. Functions that return no value are marked
4312 <tt><a href="#t_void">void</a></tt>.</p>
4313 </li>
4314 <li>
4315 <p>'<tt>fnty</tt>': shall be the signature of the pointer to function
4316 value being invoked. The argument types must match the types implied by
4317 this signature. This type can be omitted if the function is not varargs
4318 and if the function type does not return a pointer to a function.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4319 </li>
4320 <li>
4321 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
4322 be invoked. In most cases, this is a direct function invocation, but
4323 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
4324 to function value.</p>
4325 </li>
4326 <li>
4327 <p>'<tt>function args</tt>': argument list whose types match the
4328 function signature argument types. All arguments must be of
4329 <a href="#t_firstclass">first class</a> type. If the function signature
4330 indicates the function accepts a variable number of arguments, the extra
4331 arguments can be specified.</p>
4332 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4333 <li>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]4334 <p>The optional <a href="#fnattrs">function attributes</a> list. Only
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4335 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
4336 '<tt>readnone</tt>' attributes are valid here.</p>
4337 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4338</ol>
4339
4340<h5>Semantics:</h5>
4341
4342<p>The '<tt>call</tt>' instruction is used to cause control flow to
4343transfer to a specified function, with its incoming arguments bound to
4344the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
4345instruction in the called function, control flow continues with the
4346instruction after the function call, and the return value of the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4347function is bound to the result argument.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4348
4349<h5>Example:</h5>
4350
4351<pre>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]4352 %retval = call i32 @test(i32 %argc)
Chris Lattner5e893ef2008-03-21 17:24:17 +0000[diff] [blame]4353 call i32 (i8 *, ...)* @printf(i8 * %msg, i32 12, i8 42) <i>; yields i32</i>
4354 %X = tail call i32 @foo() <i>; yields i32</i>
4355 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
4356 call void %foo(i8 97 signext)
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]4357
4358 %struct.A = type { i32, i8 }
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4359 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]4360 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
4361 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattnerac454b32008-10-08 06:26:11 +0000[diff] [blame]4362 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijman2c4e05a2008-10-07 10:03:45 +0000[diff] [blame]4363 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4364</pre>
4365
4366</div>
4367
4368<!-- _______________________________________________________________________ -->
4369<div class="doc_subsubsection">
4370 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
4371</div>
4372
4373<div class="doc_text">
4374
4375<h5>Syntax:</h5>
4376
4377<pre>
4378 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
4379</pre>
4380
4381<h5>Overview:</h5>
4382
4383<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
4384the "variable argument" area of a function call. It is used to implement the
4385<tt>va_arg</tt> macro in C.</p>
4386
4387<h5>Arguments:</h5>
4388
4389<p>This instruction takes a <tt>va_list*</tt> value and the type of
4390the argument. It returns a value of the specified argument type and
4391increments the <tt>va_list</tt> to point to the next argument. The
4392actual type of <tt>va_list</tt> is target specific.</p>
4393
4394<h5>Semantics:</h5>
4395
4396<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
4397type from the specified <tt>va_list</tt> and causes the
4398<tt>va_list</tt> to point to the next argument. For more information,
4399see the variable argument handling <a href="#int_varargs">Intrinsic
4400Functions</a>.</p>
4401
4402<p>It is legal for this instruction to be called in a function which does not
4403take a variable number of arguments, for example, the <tt>vfprintf</tt>
4404function.</p>
4405
4406<p><tt>va_arg</tt> is an LLVM instruction instead of an <a
4407href="#intrinsics">intrinsic function</a> because it takes a type as an
4408argument.</p>
4409
4410<h5>Example:</h5>
4411
4412<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
4413
4414</div>
4415
4416<!-- *********************************************************************** -->
4417<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
4418<!-- *********************************************************************** -->
4419
4420<div class="doc_text">
4421
4422<p>LLVM supports the notion of an "intrinsic function". These functions have
4423well known names and semantics and are required to follow certain restrictions.
4424Overall, these intrinsics represent an extension mechanism for the LLVM
4425language that does not require changing all of the transformations in LLVM when
4426adding to the language (or the bitcode reader/writer, the parser, etc...).</p>
4427
4428<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
4429prefix is reserved in LLVM for intrinsic names; thus, function names may not
4430begin with this prefix. Intrinsic functions must always be external functions:
4431you cannot define the body of intrinsic functions. Intrinsic functions may
4432only be used in call or invoke instructions: it is illegal to take the address
4433of an intrinsic function. Additionally, because intrinsic functions are part
4434of the LLVM language, it is required if any are added that they be documented
4435here.</p>
4436
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4437<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents
4438a family of functions that perform the same operation but on different data
4439types. Because LLVM can represent over 8 million different integer types,
4440overloading is used commonly to allow an intrinsic function to operate on any
4441integer type. One or more of the argument types or the result type can be
4442overloaded to accept any integer type. Argument types may also be defined as
4443exactly matching a previous argument's type or the result type. This allows an
4444intrinsic function which accepts multiple arguments, but needs all of them to
4445be of the same type, to only be overloaded with respect to a single argument or
4446the result.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4447
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4448<p>Overloaded intrinsics will have the names of its overloaded argument types
4449encoded into its function name, each preceded by a period. Only those types
4450which are overloaded result in a name suffix. Arguments whose type is matched
4451against another type do not. For example, the <tt>llvm.ctpop</tt> function can
4452take an integer of any width and returns an integer of exactly the same integer
4453width. This leads to a family of functions such as
4454<tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29 %val)</tt>.
4455Only one type, the return type, is overloaded, and only one type suffix is
4456required. Because the argument's type is matched against the return type, it
4457does not require its own name suffix.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4458
4459<p>To learn how to add an intrinsic function, please see the
4460<a href="ExtendingLLVM.html">Extending LLVM Guide</a>.
4461</p>
4462
4463</div>
4464
4465<!-- ======================================================================= -->
4466<div class="doc_subsection">
4467 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
4468</div>
4469
4470<div class="doc_text">
4471
4472<p>Variable argument support is defined in LLVM with the <a
4473 href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
4474intrinsic functions. These functions are related to the similarly
4475named macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
4476
4477<p>All of these functions operate on arguments that use a
4478target-specific value type "<tt>va_list</tt>". The LLVM assembly
4479language reference manual does not define what this type is, so all
4480transformations should be prepared to handle these functions regardless of
4481the type used.</p>
4482
4483<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
4484instruction and the variable argument handling intrinsic functions are
4485used.</p>
4486
4487<div class="doc_code">
4488<pre>
4489define i32 @test(i32 %X, ...) {
4490 ; Initialize variable argument processing
4491 %ap = alloca i8*
4492 %ap2 = bitcast i8** %ap to i8*
4493 call void @llvm.va_start(i8* %ap2)
4494
4495 ; Read a single integer argument
4496 %tmp = va_arg i8** %ap, i32
4497
4498 ; Demonstrate usage of llvm.va_copy and llvm.va_end
4499 %aq = alloca i8*
4500 %aq2 = bitcast i8** %aq to i8*
4501 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
4502 call void @llvm.va_end(i8* %aq2)
4503
4504 ; Stop processing of arguments.
4505 call void @llvm.va_end(i8* %ap2)
4506 ret i32 %tmp
4507}
4508
4509declare void @llvm.va_start(i8*)
4510declare void @llvm.va_copy(i8*, i8*)
4511declare void @llvm.va_end(i8*)
4512</pre>
4513</div>
4514
4515</div>
4516
4517<!-- _______________________________________________________________________ -->
4518<div class="doc_subsubsection">
4519 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
4520</div>
4521
4522
4523<div class="doc_text">
4524<h5>Syntax:</h5>
4525<pre> declare void %llvm.va_start(i8* &lt;arglist&gt;)<br></pre>
4526<h5>Overview:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4527<p>The '<tt>llvm.va_start</tt>' intrinsic initializes
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4528<tt>*&lt;arglist&gt;</tt> for subsequent use by <tt><a
4529href="#i_va_arg">va_arg</a></tt>.</p>
4530
4531<h5>Arguments:</h5>
4532
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4533<p>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4534
4535<h5>Semantics:</h5>
4536
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4537<p>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4538macro available in C. In a target-dependent way, it initializes the
4539<tt>va_list</tt> element to which the argument points, so that the next call to
4540<tt>va_arg</tt> will produce the first variable argument passed to the function.
4541Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
4542last argument of the function as the compiler can figure that out.</p>
4543
4544</div>
4545
4546<!-- _______________________________________________________________________ -->
4547<div class="doc_subsubsection">
4548 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
4549</div>
4550
4551<div class="doc_text">
4552<h5>Syntax:</h5>
4553<pre> declare void @llvm.va_end(i8* &lt;arglist&gt;)<br></pre>
4554<h5>Overview:</h5>
4555
4556<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
4557which has been initialized previously with <tt><a href="#int_va_start">llvm.va_start</a></tt>
4558or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
4559
4560<h5>Arguments:</h5>
4561
4562<p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
4563
4564<h5>Semantics:</h5>
4565
4566<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
4567macro available in C. In a target-dependent way, it destroys the
4568<tt>va_list</tt> element to which the argument points. Calls to <a
4569href="#int_va_start"><tt>llvm.va_start</tt></a> and <a href="#int_va_copy">
4570<tt>llvm.va_copy</tt></a> must be matched exactly with calls to
4571<tt>llvm.va_end</tt>.</p>
4572
4573</div>
4574
4575<!-- _______________________________________________________________________ -->
4576<div class="doc_subsubsection">
4577 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
4578</div>
4579
4580<div class="doc_text">
4581
4582<h5>Syntax:</h5>
4583
4584<pre>
4585 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
4586</pre>
4587
4588<h5>Overview:</h5>
4589
4590<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
4591from the source argument list to the destination argument list.</p>
4592
4593<h5>Arguments:</h5>
4594
4595<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
4596The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
4597
4598
4599<h5>Semantics:</h5>
4600
4601<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
4602macro available in C. In a target-dependent way, it copies the source
4603<tt>va_list</tt> element into the destination <tt>va_list</tt> element. This
4604intrinsic is necessary because the <tt><a href="#int_va_start">
4605llvm.va_start</a></tt> intrinsic may be arbitrarily complex and require, for
4606example, memory allocation.</p>
4607
4608</div>
4609
4610<!-- ======================================================================= -->
4611<div class="doc_subsection">
4612 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
4613</div>
4614
4615<div class="doc_text">
4616
4617<p>
4618LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]4619Collection</a> (GC) requires the implementation and generation of these
4620intrinsics.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4621These intrinsics allow identification of <a href="#int_gcroot">GC roots on the
4622stack</a>, as well as garbage collector implementations that require <a
4623href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a> barriers.
4624Front-ends for type-safe garbage collected languages should generate these
4625intrinsics to make use of the LLVM garbage collectors. For more details, see <a
4626href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
4627</p>
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]4628
4629<p>The garbage collection intrinsics only operate on objects in the generic
4630 address space (address space zero).</p>
4631
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4632</div>
4633
4634<!-- _______________________________________________________________________ -->
4635<div class="doc_subsubsection">
4636 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
4637</div>
4638
4639<div class="doc_text">
4640
4641<h5>Syntax:</h5>
4642
4643<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4644 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4645</pre>
4646
4647<h5>Overview:</h5>
4648
4649<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
4650the code generator, and allows some metadata to be associated with it.</p>
4651
4652<h5>Arguments:</h5>
4653
4654<p>The first argument specifies the address of a stack object that contains the
4655root pointer. The second pointer (which must be either a constant or a global
4656value address) contains the meta-data to be associated with the root.</p>
4657
4658<h5>Semantics:</h5>
4659
Chris Lattnera7d94ba2008-04-24 05:59:56 +0000[diff] [blame]4660<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]4661location. At compile-time, the code generator generates information to allow
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4662the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
4663intrinsic may only be used in a function which <a href="#gc">specifies a GC
4664algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4665
4666</div>
4667
4668
4669<!-- _______________________________________________________________________ -->
4670<div class="doc_subsubsection">
4671 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
4672</div>
4673
4674<div class="doc_text">
4675
4676<h5>Syntax:</h5>
4677
4678<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4679 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4680</pre>
4681
4682<h5>Overview:</h5>
4683
4684<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
4685locations, allowing garbage collector implementations that require read
4686barriers.</p>
4687
4688<h5>Arguments:</h5>
4689
4690<p>The second argument is the address to read from, which should be an address
4691allocated from the garbage collector. The first object is a pointer to the
4692start of the referenced object, if needed by the language runtime (otherwise
4693null).</p>
4694
4695<h5>Semantics:</h5>
4696
4697<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
4698instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4699garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
4700may only be used in a function which <a href="#gc">specifies a GC
4701algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4702
4703</div>
4704
4705
4706<!-- _______________________________________________________________________ -->
4707<div class="doc_subsubsection">
4708 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
4709</div>
4710
4711<div class="doc_text">
4712
4713<h5>Syntax:</h5>
4714
4715<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4716 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4717</pre>
4718
4719<h5>Overview:</h5>
4720
4721<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
4722locations, allowing garbage collector implementations that require write
4723barriers (such as generational or reference counting collectors).</p>
4724
4725<h5>Arguments:</h5>
4726
4727<p>The first argument is the reference to store, the second is the start of the
4728object to store it to, and the third is the address of the field of Obj to
4729store to. If the runtime does not require a pointer to the object, Obj may be
4730null.</p>
4731
4732<h5>Semantics:</h5>
4733
4734<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
4735instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4736garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
4737may only be used in a function which <a href="#gc">specifies a GC
4738algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4739
4740</div>
4741
4742
4743
4744<!-- ======================================================================= -->
4745<div class="doc_subsection">
4746 <a name="int_codegen">Code Generator Intrinsics</a>
4747</div>
4748
4749<div class="doc_text">
4750<p>
4751These intrinsics are provided by LLVM to expose special features that may only
4752be implemented with code generator support.
4753</p>
4754
4755</div>
4756
4757<!-- _______________________________________________________________________ -->
4758<div class="doc_subsubsection">
4759 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
4760</div>
4761
4762<div class="doc_text">
4763
4764<h5>Syntax:</h5>
4765<pre>
4766 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
4767</pre>
4768
4769<h5>Overview:</h5>
4770
4771<p>
4772The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
4773target-specific value indicating the return address of the current function
4774or one of its callers.
4775</p>
4776
4777<h5>Arguments:</h5>
4778
4779<p>
4780The argument to this intrinsic indicates which function to return the address
4781for. Zero indicates the calling function, one indicates its caller, etc. The
4782argument is <b>required</b> to be a constant integer value.
4783</p>
4784
4785<h5>Semantics:</h5>
4786
4787<p>
4788The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
4789the return address of the specified call frame, or zero if it cannot be
4790identified. The value returned by this intrinsic is likely to be incorrect or 0
4791for arguments other than zero, so it should only be used for debugging purposes.
4792</p>
4793
4794<p>
4795Note that calling this intrinsic does not prevent function inlining or other
4796aggressive transformations, so the value returned may not be that of the obvious
4797source-language caller.
4798</p>
4799</div>
4800
4801
4802<!-- _______________________________________________________________________ -->
4803<div class="doc_subsubsection">
4804 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
4805</div>
4806
4807<div class="doc_text">
4808
4809<h5>Syntax:</h5>
4810<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4811 declare i8 *@llvm.frameaddress(i32 &lt;level&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4812</pre>
4813
4814<h5>Overview:</h5>
4815
4816<p>
4817The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
4818target-specific frame pointer value for the specified stack frame.
4819</p>
4820
4821<h5>Arguments:</h5>
4822
4823<p>
4824The argument to this intrinsic indicates which function to return the frame
4825pointer for. Zero indicates the calling function, one indicates its caller,
4826etc. The argument is <b>required</b> to be a constant integer value.
4827</p>
4828
4829<h5>Semantics:</h5>
4830
4831<p>
4832The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
4833the frame address of the specified call frame, or zero if it cannot be
4834identified. The value returned by this intrinsic is likely to be incorrect or 0
4835for arguments other than zero, so it should only be used for debugging purposes.
4836</p>
4837
4838<p>
4839Note that calling this intrinsic does not prevent function inlining or other
4840aggressive transformations, so the value returned may not be that of the obvious
4841source-language caller.
4842</p>
4843</div>
4844
4845<!-- _______________________________________________________________________ -->
4846<div class="doc_subsubsection">
4847 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
4848</div>
4849
4850<div class="doc_text">
4851
4852<h5>Syntax:</h5>
4853<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4854 declare i8 *@llvm.stacksave()
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4855</pre>
4856
4857<h5>Overview:</h5>
4858
4859<p>
4860The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
4861the function stack, for use with <a href="#int_stackrestore">
4862<tt>llvm.stackrestore</tt></a>. This is useful for implementing language
4863features like scoped automatic variable sized arrays in C99.
4864</p>
4865
4866<h5>Semantics:</h5>
4867
4868<p>
4869This intrinsic returns a opaque pointer value that can be passed to <a
4870href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
4871<tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
4872<tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
4873state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
4874practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
4875that were allocated after the <tt>llvm.stacksave</tt> was executed.
4876</p>
4877
4878</div>
4879
4880<!-- _______________________________________________________________________ -->
4881<div class="doc_subsubsection">
4882 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
4883</div>
4884
4885<div class="doc_text">
4886
4887<h5>Syntax:</h5>
4888<pre>
4889 declare void @llvm.stackrestore(i8 * %ptr)
4890</pre>
4891
4892<h5>Overview:</h5>
4893
4894<p>
4895The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
4896the function stack to the state it was in when the corresponding <a
4897href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
4898useful for implementing language features like scoped automatic variable sized
4899arrays in C99.
4900</p>
4901
4902<h5>Semantics:</h5>
4903
4904<p>
4905See the description for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.
4906</p>
4907
4908</div>
4909
4910
4911<!-- _______________________________________________________________________ -->
4912<div class="doc_subsubsection">
4913 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
4914</div>
4915
4916<div class="doc_text">
4917
4918<h5>Syntax:</h5>
4919<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4920 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]4921</pre>
4922
4923<h5>Overview:</h5>
4924
4925
4926<p>
4927The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
4928a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
4929no
4930effect on the behavior of the program but can change its performance
4931characteristics.
4932</p>
4933
4934<h5>Arguments:</h5>
4935
4936<p>
4937<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
4938determining if the fetch should be for a read (0) or write (1), and
4939<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
4940locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
4941<tt>locality</tt> arguments must be constant integers.
4942</p>
4943
4944<h5>Semantics:</h5>
4945
4946<p>
4947This intrinsic does not modify the behavior of the program. In particular,
4948prefetches cannot trap and do not produce a value. On targets that support this
4949intrinsic, the prefetch can provide hints to the processor cache for better
4950performance.
4951</p>
4952
4953</div>
4954
4955<!-- _______________________________________________________________________ -->
4956<div class="doc_subsubsection">
4957 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
4958</div>
4959
4960<div class="doc_text">
4961
4962<h5>Syntax:</h5>
4963<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4964 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4965</pre>
4966
4967<h5>Overview:</h5>
4968
4969
4970<p>
4971The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]4972(PC) in a region of
4973code to simulators and other tools. The method is target specific, but it is
4974expected that the marker will use exported symbols to transmit the PC of the
4975marker.
4976The marker makes no guarantees that it will remain with any specific instruction
4977after optimizations. It is possible that the presence of a marker will inhibit
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4978optimizations. The intended use is to be inserted after optimizations to allow
4979correlations of simulation runs.
4980</p>
4981
4982<h5>Arguments:</h5>
4983
4984<p>
4985<tt>id</tt> is a numerical id identifying the marker.
4986</p>
4987
4988<h5>Semantics:</h5>
4989
4990<p>
4991This intrinsic does not modify the behavior of the program. Backends that do not
4992support this intrinisic may ignore it.
4993</p>
4994
4995</div>
4996
4997<!-- _______________________________________________________________________ -->
4998<div class="doc_subsubsection">
4999 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
5000</div>
5001
5002<div class="doc_text">
5003
5004<h5>Syntax:</h5>
5005<pre>
5006 declare i64 @llvm.readcyclecounter( )
5007</pre>
5008
5009<h5>Overview:</h5>
5010
5011
5012<p>
5013The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
5014counter register (or similar low latency, high accuracy clocks) on those targets
5015that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
5016As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
5017should only be used for small timings.
5018</p>
5019
5020<h5>Semantics:</h5>
5021
5022<p>
5023When directly supported, reading the cycle counter should not modify any memory.
5024Implementations are allowed to either return a application specific value or a
5025system wide value. On backends without support, this is lowered to a constant 0.
5026</p>
5027
5028</div>
5029
5030<!-- ======================================================================= -->
5031<div class="doc_subsection">
5032 <a name="int_libc">Standard C Library Intrinsics</a>
5033</div>
5034
5035<div class="doc_text">
5036<p>
5037LLVM provides intrinsics for a few important standard C library functions.
5038These intrinsics allow source-language front-ends to pass information about the
5039alignment of the pointer arguments to the code generator, providing opportunity
5040for more efficient code generation.
5041</p>
5042
5043</div>
5044
5045<!-- _______________________________________________________________________ -->
5046<div class="doc_subsubsection">
5047 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
5048</div>
5049
5050<div class="doc_text">
5051
5052<h5>Syntax:</h5>
5053<pre>
5054 declare void @llvm.memcpy.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5055 i32 &lt;len&gt;, i32 &lt;align&gt;)
5056 declare void @llvm.memcpy.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5057 i64 &lt;len&gt;, i32 &lt;align&gt;)
5058</pre>
5059
5060<h5>Overview:</h5>
5061
5062<p>
5063The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
5064location to the destination location.
5065</p>
5066
5067<p>
5068Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
5069intrinsics do not return a value, and takes an extra alignment argument.
5070</p>
5071
5072<h5>Arguments:</h5>
5073
5074<p>
5075The first argument is a pointer to the destination, the second is a pointer to
5076the source. The third argument is an integer argument
5077specifying the number of bytes to copy, and the fourth argument is the alignment
5078of the source and destination locations.
5079</p>
5080
5081<p>
5082If the call to this intrinisic has an alignment value that is not 0 or 1, then
5083the caller guarantees that both the source and destination pointers are aligned
5084to that boundary.
5085</p>
5086
5087<h5>Semantics:</h5>
5088
5089<p>
5090The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
5091location to the destination location, which are not allowed to overlap. It
5092copies "len" bytes of memory over. If the argument is known to be aligned to
5093some boundary, this can be specified as the fourth argument, otherwise it should
5094be set to 0 or 1.
5095</p>
5096</div>
5097
5098
5099<!-- _______________________________________________________________________ -->
5100<div class="doc_subsubsection">
5101 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
5102</div>
5103
5104<div class="doc_text">
5105
5106<h5>Syntax:</h5>
5107<pre>
5108 declare void @llvm.memmove.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5109 i32 &lt;len&gt;, i32 &lt;align&gt;)
5110 declare void @llvm.memmove.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5111 i64 &lt;len&gt;, i32 &lt;align&gt;)
5112</pre>
5113
5114<h5>Overview:</h5>
5115
5116<p>
5117The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the source
5118location to the destination location. It is similar to the
Chris Lattnerdba16ea2008-01-06 19:51:52 +0000[diff] [blame]5119'<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to overlap.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5120</p>
5121
5122<p>
5123Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
5124intrinsics do not return a value, and takes an extra alignment argument.
5125</p>
5126
5127<h5>Arguments:</h5>
5128
5129<p>
5130The first argument is a pointer to the destination, the second is a pointer to
5131the source. The third argument is an integer argument
5132specifying the number of bytes to copy, and the fourth argument is the alignment
5133of the source and destination locations.
5134</p>
5135
5136<p>
5137If the call to this intrinisic has an alignment value that is not 0 or 1, then
5138the caller guarantees that the source and destination pointers are aligned to
5139that boundary.
5140</p>
5141
5142<h5>Semantics:</h5>
5143
5144<p>
5145The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the source
5146location to the destination location, which may overlap. It
5147copies "len" bytes of memory over. If the argument is known to be aligned to
5148some boundary, this can be specified as the fourth argument, otherwise it should
5149be set to 0 or 1.
5150</p>
5151</div>
5152
5153
5154<!-- _______________________________________________________________________ -->
5155<div class="doc_subsubsection">
5156 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
5157</div>
5158
5159<div class="doc_text">
5160
5161<h5>Syntax:</h5>
5162<pre>
5163 declare void @llvm.memset.i32(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5164 i32 &lt;len&gt;, i32 &lt;align&gt;)
5165 declare void @llvm.memset.i64(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5166 i64 &lt;len&gt;, i32 &lt;align&gt;)
5167</pre>
5168
5169<h5>Overview:</h5>
5170
5171<p>
5172The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a particular
5173byte value.
5174</p>
5175
5176<p>
5177Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
5178does not return a value, and takes an extra alignment argument.
5179</p>
5180
5181<h5>Arguments:</h5>
5182
5183<p>
5184The first argument is a pointer to the destination to fill, the second is the
5185byte value to fill it with, the third argument is an integer
5186argument specifying the number of bytes to fill, and the fourth argument is the
5187known alignment of destination location.
5188</p>
5189
5190<p>
5191If the call to this intrinisic has an alignment value that is not 0 or 1, then
5192the caller guarantees that the destination pointer is aligned to that boundary.
5193</p>
5194
5195<h5>Semantics:</h5>
5196
5197<p>
5198The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting at
5199the
5200destination location. If the argument is known to be aligned to some boundary,
5201this can be specified as the fourth argument, otherwise it should be set to 0 or
52021.
5203</p>
5204</div>
5205
5206
5207<!-- _______________________________________________________________________ -->
5208<div class="doc_subsubsection">
5209 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
5210</div>
5211
5212<div class="doc_text">
5213
5214<h5>Syntax:</h5>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5215<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]5216floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5217types however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5218<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5219 declare float @llvm.sqrt.f32(float %Val)
5220 declare double @llvm.sqrt.f64(double %Val)
5221 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
5222 declare fp128 @llvm.sqrt.f128(fp128 %Val)
5223 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5224</pre>
5225
5226<h5>Overview:</h5>
5227
5228<p>
5229The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5230returning the same value as the libm '<tt>sqrt</tt>' functions would. Unlike
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5231<tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
Chris Lattnerf914a002008-01-29 07:00:44 +0000[diff] [blame]5232negative numbers other than -0.0 (which allows for better optimization, because
5233there is no need to worry about errno being set). <tt>llvm.sqrt(-0.0)</tt> is
5234defined to return -0.0 like IEEE sqrt.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5235</p>
5236
5237<h5>Arguments:</h5>
5238
5239<p>
5240The argument and return value are floating point numbers of the same type.
5241</p>
5242
5243<h5>Semantics:</h5>
5244
5245<p>
5246This function returns the sqrt of the specified operand if it is a nonnegative
5247floating point number.
5248</p>
5249</div>
5250
5251<!-- _______________________________________________________________________ -->
5252<div class="doc_subsubsection">
5253 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
5254</div>
5255
5256<div class="doc_text">
5257
5258<h5>Syntax:</h5>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5259<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]5260floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5261types however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5262<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5263 declare float @llvm.powi.f32(float %Val, i32 %power)
5264 declare double @llvm.powi.f64(double %Val, i32 %power)
5265 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
5266 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
5267 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5268</pre>
5269
5270<h5>Overview:</h5>
5271
5272<p>
5273The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
5274specified (positive or negative) power. The order of evaluation of
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5275multiplications is not defined. When a vector of floating point type is
5276used, the second argument remains a scalar integer value.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5277</p>
5278
5279<h5>Arguments:</h5>
5280
5281<p>
5282The second argument is an integer power, and the first is a value to raise to
5283that power.
5284</p>
5285
5286<h5>Semantics:</h5>
5287
5288<p>
5289This function returns the first value raised to the second power with an
5290unspecified sequence of rounding operations.</p>
5291</div>
5292
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5293<!-- _______________________________________________________________________ -->
5294<div class="doc_subsubsection">
5295 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
5296</div>
5297
5298<div class="doc_text">
5299
5300<h5>Syntax:</h5>
5301<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
5302floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5303types however.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5304<pre>
5305 declare float @llvm.sin.f32(float %Val)
5306 declare double @llvm.sin.f64(double %Val)
5307 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
5308 declare fp128 @llvm.sin.f128(fp128 %Val)
5309 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
5310</pre>
5311
5312<h5>Overview:</h5>
5313
5314<p>
5315The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.
5316</p>
5317
5318<h5>Arguments:</h5>
5319
5320<p>
5321The argument and return value are floating point numbers of the same type.
5322</p>
5323
5324<h5>Semantics:</h5>
5325
5326<p>
5327This function returns the sine of the specified operand, returning the
5328same values as the libm <tt>sin</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5329conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5330</div>
5331
5332<!-- _______________________________________________________________________ -->
5333<div class="doc_subsubsection">
5334 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
5335</div>
5336
5337<div class="doc_text">
5338
5339<h5>Syntax:</h5>
5340<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
5341floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5342types however.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5343<pre>
5344 declare float @llvm.cos.f32(float %Val)
5345 declare double @llvm.cos.f64(double %Val)
5346 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
5347 declare fp128 @llvm.cos.f128(fp128 %Val)
5348 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
5349</pre>
5350
5351<h5>Overview:</h5>
5352
5353<p>
5354The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.
5355</p>
5356
5357<h5>Arguments:</h5>
5358
5359<p>
5360The argument and return value are floating point numbers of the same type.
5361</p>
5362
5363<h5>Semantics:</h5>
5364
5365<p>
5366This function returns the cosine of the specified operand, returning the
5367same values as the libm <tt>cos</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5368conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5369</div>
5370
5371<!-- _______________________________________________________________________ -->
5372<div class="doc_subsubsection">
5373 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
5374</div>
5375
5376<div class="doc_text">
5377
5378<h5>Syntax:</h5>
5379<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
5380floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5381types however.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5382<pre>
5383 declare float @llvm.pow.f32(float %Val, float %Power)
5384 declare double @llvm.pow.f64(double %Val, double %Power)
5385 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
5386 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
5387 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
5388</pre>
5389
5390<h5>Overview:</h5>
5391
5392<p>
5393The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
5394specified (positive or negative) power.
5395</p>
5396
5397<h5>Arguments:</h5>
5398
5399<p>
5400The second argument is a floating point power, and the first is a value to
5401raise to that power.
5402</p>
5403
5404<h5>Semantics:</h5>
5405
5406<p>
5407This function returns the first value raised to the second power,
5408returning the
5409same values as the libm <tt>pow</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5410conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5411</div>
5412
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5413
5414<!-- ======================================================================= -->
5415<div class="doc_subsection">
5416 <a name="int_manip">Bit Manipulation Intrinsics</a>
5417</div>
5418
5419<div class="doc_text">
5420<p>
5421LLVM provides intrinsics for a few important bit manipulation operations.
5422These allow efficient code generation for some algorithms.
5423</p>
5424
5425</div>
5426
5427<!-- _______________________________________________________________________ -->
5428<div class="doc_subsubsection">
5429 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
5430</div>
5431
5432<div class="doc_text">
5433
5434<h5>Syntax:</h5>
5435<p>This is an overloaded intrinsic function. You can use bswap on any integer
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5436type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5437<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5438 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
5439 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
5440 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5441</pre>
5442
5443<h5>Overview:</h5>
5444
5445<p>
5446The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
5447values with an even number of bytes (positive multiple of 16 bits). These are
5448useful for performing operations on data that is not in the target's native
5449byte order.
5450</p>
5451
5452<h5>Semantics:</h5>
5453
5454<p>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5455The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5456and low byte of the input i16 swapped. Similarly, the <tt>llvm.bswap.i32</tt>
5457intrinsic returns an i32 value that has the four bytes of the input i32
5458swapped, 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]5459i32 will have its bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i48</tt>,
5460<tt>llvm.bswap.i64</tt> and other intrinsics extend this concept to
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5461additional even-byte lengths (6 bytes, 8 bytes and more, respectively).
5462</p>
5463
5464</div>
5465
5466<!-- _______________________________________________________________________ -->
5467<div class="doc_subsubsection">
5468 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
5469</div>
5470
5471<div class="doc_text">
5472
5473<h5>Syntax:</h5>
5474<p>This is an overloaded intrinsic. You can use llvm.ctpop on any integer bit
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5475width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5476<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5477 declare i8 @llvm.ctpop.i8 (i8 &lt;src&gt;)
5478 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5479 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5480 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
5481 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5482</pre>
5483
5484<h5>Overview:</h5>
5485
5486<p>
5487The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
5488value.
5489</p>
5490
5491<h5>Arguments:</h5>
5492
5493<p>
5494The only argument is the value to be counted. The argument may be of any
5495integer type. The return type must match the argument type.
5496</p>
5497
5498<h5>Semantics:</h5>
5499
5500<p>
5501The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
5502</p>
5503</div>
5504
5505<!-- _______________________________________________________________________ -->
5506<div class="doc_subsubsection">
5507 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
5508</div>
5509
5510<div class="doc_text">
5511
5512<h5>Syntax:</h5>
5513<p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5514integer bit width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5515<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5516 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
5517 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5518 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5519 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
5520 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5521</pre>
5522
5523<h5>Overview:</h5>
5524
5525<p>
5526The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
5527leading zeros in a variable.
5528</p>
5529
5530<h5>Arguments:</h5>
5531
5532<p>
5533The only argument is the value to be counted. The argument may be of any
5534integer type. The return type must match the argument type.
5535</p>
5536
5537<h5>Semantics:</h5>
5538
5539<p>
5540The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
5541in a variable. If the src == 0 then the result is the size in bits of the type
5542of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.
5543</p>
5544</div>
5545
5546
5547
5548<!-- _______________________________________________________________________ -->
5549<div class="doc_subsubsection">
5550 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
5551</div>
5552
5553<div class="doc_text">
5554
5555<h5>Syntax:</h5>
5556<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5557integer bit width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5558<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5559 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
5560 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5561 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5562 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
5563 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5564</pre>
5565
5566<h5>Overview:</h5>
5567
5568<p>
5569The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
5570trailing zeros.
5571</p>
5572
5573<h5>Arguments:</h5>
5574
5575<p>
5576The only argument is the value to be counted. The argument may be of any
5577integer type. The return type must match the argument type.
5578</p>
5579
5580<h5>Semantics:</h5>
5581
5582<p>
5583The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
5584in a variable. If the src == 0 then the result is the size in bits of the type
5585of src. For example, <tt>llvm.cttz(2) = 1</tt>.
5586</p>
5587</div>
5588
5589<!-- _______________________________________________________________________ -->
5590<div class="doc_subsubsection">
5591 <a name="int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic</a>
5592</div>
5593
5594<div class="doc_text">
5595
5596<h5>Syntax:</h5>
5597<p>This is an overloaded intrinsic. You can use <tt>llvm.part.select</tt>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5598on any integer bit width.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5599<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5600 declare i17 @llvm.part.select.i17 (i17 %val, i32 %loBit, i32 %hiBit)
5601 declare i29 @llvm.part.select.i29 (i29 %val, i32 %loBit, i32 %hiBit)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5602</pre>
5603
5604<h5>Overview:</h5>
5605<p>The '<tt>llvm.part.select</tt>' family of intrinsic functions selects a
5606range of bits from an integer value and returns them in the same bit width as
5607the original value.</p>
5608
5609<h5>Arguments:</h5>
5610<p>The first argument, <tt>%val</tt> and the result may be integer types of
5611any bit width but they must have the same bit width. The second and third
5612arguments must be <tt>i32</tt> type since they specify only a bit index.</p>
5613
5614<h5>Semantics:</h5>
5615<p>The operation of the '<tt>llvm.part.select</tt>' intrinsic has two modes
5616of operation: forwards and reverse. If <tt>%loBit</tt> is greater than
5617<tt>%hiBits</tt> then the intrinsic operates in reverse mode. Otherwise it
5618operates in forward mode.</p>
5619<p>In forward mode, this intrinsic is the equivalent of shifting <tt>%val</tt>
5620right by <tt>%loBit</tt> bits and then ANDing it with a mask with
5621only the <tt>%hiBit - %loBit</tt> bits set, as follows:</p>
5622<ol>
5623 <li>The <tt>%val</tt> is shifted right (LSHR) by the number of bits specified
5624 by <tt>%loBits</tt>. This normalizes the value to the low order bits.</li>
5625 <li>The <tt>%loBits</tt> value is subtracted from the <tt>%hiBits</tt> value
5626 to determine the number of bits to retain.</li>
5627 <li>A mask of the retained bits is created by shifting a -1 value.</li>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5628 <li>The mask is ANDed with <tt>%val</tt> to produce the result.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5629</ol>
5630<p>In reverse mode, a similar computation is made except that the bits are
5631returned in the reverse order. So, for example, if <tt>X</tt> has the value
5632<tt>i16 0x0ACF (101011001111)</tt> and we apply
5633<tt>part.select(i16 X, 8, 3)</tt> to it, we get back the value
5634<tt>i16 0x0026 (000000100110)</tt>.</p>
5635</div>
5636
5637<div class="doc_subsubsection">
5638 <a name="int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic</a>
5639</div>
5640
5641<div class="doc_text">
5642
5643<h5>Syntax:</h5>
5644<p>This is an overloaded intrinsic. You can use <tt>llvm.part.set</tt>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5645on any integer bit width.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5646<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5647 declare i17 @llvm.part.set.i17.i9 (i17 %val, i9 %repl, i32 %lo, i32 %hi)
5648 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]5649</pre>
5650
5651<h5>Overview:</h5>
5652<p>The '<tt>llvm.part.set</tt>' family of intrinsic functions replaces a range
5653of bits in an integer value with another integer value. It returns the integer
5654with the replaced bits.</p>
5655
5656<h5>Arguments:</h5>
5657<p>The first argument, <tt>%val</tt> and the result may be integer types of
5658any bit width but they must have the same bit width. <tt>%val</tt> is the value
5659whose bits will be replaced. The second argument, <tt>%repl</tt> may be an
5660integer of any bit width. The third and fourth arguments must be <tt>i32</tt>
5661type since they specify only a bit index.</p>
5662
5663<h5>Semantics:</h5>
5664<p>The operation of the '<tt>llvm.part.set</tt>' intrinsic has two modes
5665of operation: forwards and reverse. If <tt>%lo</tt> is greater than
5666<tt>%hi</tt> then the intrinsic operates in reverse mode. Otherwise it
5667operates in forward mode.</p>
5668<p>For both modes, the <tt>%repl</tt> value is prepared for use by either
5669truncating it down to the size of the replacement area or zero extending it
5670up to that size.</p>
5671<p>In forward mode, the bits between <tt>%lo</tt> and <tt>%hi</tt> (inclusive)
5672are replaced with corresponding bits from <tt>%repl</tt>. That is the 0th bit
5673in <tt>%repl</tt> replaces the <tt>%lo</tt>th bit in <tt>%val</tt> and etc. up
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5674to the <tt>%hi</tt>th bit.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5675<p>In reverse mode, a similar computation is made except that the bits are
5676reversed. That is, the <tt>0</tt>th bit in <tt>%repl</tt> replaces the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5677<tt>%hi</tt> bit in <tt>%val</tt> and etc. down to the <tt>%lo</tt>th bit.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5678<h5>Examples:</h5>
5679<pre>
5680 llvm.part.set(0xFFFF, 0, 4, 7) -&gt; 0xFF0F
5681 llvm.part.set(0xFFFF, 0, 7, 4) -&gt; 0xFF0F
5682 llvm.part.set(0xFFFF, 1, 7, 4) -&gt; 0xFF8F
5683 llvm.part.set(0xFFFF, F, 8, 3) -&gt; 0xFFE7
5684 llvm.part.set(0xFFFF, 0, 3, 8) -&gt; 0xFE07
5685</pre>
5686</div>
5687
5688<!-- ======================================================================= -->
5689<div class="doc_subsection">
5690 <a name="int_debugger">Debugger Intrinsics</a>
5691</div>
5692
5693<div class="doc_text">
5694<p>
5695The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
5696are described in the <a
5697href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
5698Debugging</a> document.
5699</p>
5700</div>
5701
5702
5703<!-- ======================================================================= -->
5704<div class="doc_subsection">
5705 <a name="int_eh">Exception Handling Intrinsics</a>
5706</div>
5707
5708<div class="doc_text">
5709<p> The LLVM exception handling intrinsics (which all start with
5710<tt>llvm.eh.</tt> prefix), are described in the <a
5711href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
5712Handling</a> document. </p>
5713</div>
5714
5715<!-- ======================================================================= -->
5716<div class="doc_subsection">
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5717 <a name="int_trampoline">Trampoline Intrinsic</a>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5718</div>
5719
5720<div class="doc_text">
5721<p>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5722 This intrinsic makes it possible to excise one parameter, marked with
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5723 the <tt>nest</tt> attribute, from a function. The result is a callable
5724 function pointer lacking the nest parameter - the caller does not need
5725 to provide a value for it. Instead, the value to use is stored in
5726 advance in a "trampoline", a block of memory usually allocated
5727 on the stack, which also contains code to splice the nest value into the
5728 argument list. This is used to implement the GCC nested function address
5729 extension.
5730</p>
5731<p>
5732 For example, if the function is
5733 <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]5734 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]5735<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5736 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
5737 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
5738 %p = call i8* @llvm.init.trampoline( i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8*), i8* %nval )
5739 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5740</pre>
Bill Wendlinge262b4c2007-09-22 09:23:55 +0000[diff] [blame]5741 <p>The call <tt>%val = call i32 %fp( i32 %x, i32 %y )</tt> is then equivalent
5742 to <tt>%val = call i32 %f( i8* %nval, i32 %x, i32 %y )</tt>.</p>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5743</div>
5744
5745<!-- _______________________________________________________________________ -->
5746<div class="doc_subsubsection">
5747 <a name="int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a>
5748</div>
5749<div class="doc_text">
5750<h5>Syntax:</h5>
5751<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5752declare 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]5753</pre>
5754<h5>Overview:</h5>
5755<p>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5756 This fills the memory pointed to by <tt>tramp</tt> with code
5757 and returns a function pointer suitable for executing it.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5758</p>
5759<h5>Arguments:</h5>
5760<p>
5761 The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
5762 pointers. The <tt>tramp</tt> argument must point to a sufficiently large
5763 and sufficiently aligned block of memory; this memory is written to by the
Duncan Sands35012212007-08-22 23:39:54 +0000[diff] [blame]5764 intrinsic. Note that the size and the alignment are target-specific - LLVM
5765 currently provides no portable way of determining them, so a front-end that
5766 generates this intrinsic needs to have some target-specific knowledge.
5767 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]5768</p>
5769<h5>Semantics:</h5>
5770<p>
5771 The block of memory pointed to by <tt>tramp</tt> is filled with target
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5772 dependent code, turning it into a function. A pointer to this function is
5773 returned, but needs to be bitcast to an
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5774 <a href="#int_trampoline">appropriate function pointer type</a>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5775 before being called. The new function's signature is the same as that of
5776 <tt>func</tt> with any arguments marked with the <tt>nest</tt> attribute
5777 removed. At most one such <tt>nest</tt> argument is allowed, and it must be
5778 of pointer type. Calling the new function is equivalent to calling
5779 <tt>func</tt> with the same argument list, but with <tt>nval</tt> used for the
5780 missing <tt>nest</tt> argument. If, after calling
5781 <tt>llvm.init.trampoline</tt>, the memory pointed to by <tt>tramp</tt> is
5782 modified, then the effect of any later call to the returned function pointer is
5783 undefined.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5784</p>
5785</div>
5786
5787<!-- ======================================================================= -->
5788<div class="doc_subsection">
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]5789 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
5790</div>
5791
5792<div class="doc_text">
5793<p>
5794 These intrinsic functions expand the "universal IR" of LLVM to represent
5795 hardware constructs for atomic operations and memory synchronization. This
5796 provides an interface to the hardware, not an interface to the programmer. It
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]5797 is aimed at a low enough level to allow any programming models or APIs
5798 (Application Programming Interfaces) which
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]5799 need atomic behaviors to map cleanly onto it. It is also modeled primarily on
5800 hardware behavior. Just as hardware provides a "universal IR" for source
5801 languages, it also provides a starting point for developing a "universal"
5802 atomic operation and synchronization IR.
5803</p>
5804<p>
5805 These do <em>not</em> form an API such as high-level threading libraries,
5806 software transaction memory systems, atomic primitives, and intrinsic
5807 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
5808 application libraries. The hardware interface provided by LLVM should allow
5809 a clean implementation of all of these APIs and parallel programming models.
5810 No one model or paradigm should be selected above others unless the hardware
5811 itself ubiquitously does so.
5812
5813</p>
5814</div>
5815
5816<!-- _______________________________________________________________________ -->
5817<div class="doc_subsubsection">
5818 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
5819</div>
5820<div class="doc_text">
5821<h5>Syntax:</h5>
5822<pre>
5823declare void @llvm.memory.barrier( i1 &lt;ll&gt;, i1 &lt;ls&gt;, i1 &lt;sl&gt;, i1 &lt;ss&gt;,
5824i1 &lt;device&gt; )
5825
5826</pre>
5827<h5>Overview:</h5>
5828<p>
5829 The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
5830 specific pairs of memory access types.
5831</p>
5832<h5>Arguments:</h5>
5833<p>
5834 The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
5835 The first four arguments enables a specific barrier as listed below. The fith
5836 argument specifies that the barrier applies to io or device or uncached memory.
5837
5838</p>
5839 <ul>
5840 <li><tt>ll</tt>: load-load barrier</li>
5841 <li><tt>ls</tt>: load-store barrier</li>
5842 <li><tt>sl</tt>: store-load barrier</li>
5843 <li><tt>ss</tt>: store-store barrier</li>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5844 <li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]5845 </ul>
5846<h5>Semantics:</h5>
5847<p>
5848 This intrinsic causes the system to enforce some ordering constraints upon
5849 the loads and stores of the program. This barrier does not indicate
5850 <em>when</em> any events will occur, it only enforces an <em>order</em> in
5851 which they occur. For any of the specified pairs of load and store operations
5852 (f.ex. load-load, or store-load), all of the first operations preceding the
5853 barrier will complete before any of the second operations succeeding the
5854 barrier begin. Specifically the semantics for each pairing is as follows:
5855</p>
5856 <ul>
5857 <li><tt>ll</tt>: All loads before the barrier must complete before any load
5858 after the barrier begins.</li>
5859
5860 <li><tt>ls</tt>: All loads before the barrier must complete before any
5861 store after the barrier begins.</li>
5862 <li><tt>ss</tt>: All stores before the barrier must complete before any
5863 store after the barrier begins.</li>
5864 <li><tt>sl</tt>: All stores before the barrier must complete before any
5865 load after the barrier begins.</li>
5866 </ul>
5867<p>
5868 These semantics are applied with a logical "and" behavior when more than one
5869 is enabled in a single memory barrier intrinsic.
5870</p>
5871<p>
5872 Backends may implement stronger barriers than those requested when they do not
5873 support as fine grained a barrier as requested. Some architectures do not
5874 need all types of barriers and on such architectures, these become noops.
5875</p>
5876<h5>Example:</h5>
5877<pre>
5878%ptr = malloc i32
5879 store i32 4, %ptr
5880
5881%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
5882 call void @llvm.memory.barrier( i1 false, i1 true, i1 false, i1 false )
5883 <i>; guarantee the above finishes</i>
5884 store i32 8, %ptr <i>; before this begins</i>
5885</pre>
5886</div>
5887
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5888<!-- _______________________________________________________________________ -->
5889<div class="doc_subsubsection">
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]5890 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5891</div>
5892<div class="doc_text">
5893<h5>Syntax:</h5>
5894<p>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]5895 This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
5896 any integer bit width and for different address spaces. Not all targets
5897 support all bit widths however.</p>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5898
5899<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]5900declare i8 @llvm.atomic.cmp.swap.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt; )
5901declare i16 @llvm.atomic.cmp.swap.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt; )
5902declare i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt; )
5903declare 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]5904
5905</pre>
5906<h5>Overview:</h5>
5907<p>
5908 This loads a value in memory and compares it to a given value. If they are
5909 equal, it stores a new value into the memory.
5910</p>
5911<h5>Arguments:</h5>
5912<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]5913 The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result as
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5914 well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
5915 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
5916 this integer type. While any bit width integer may be used, targets may only
5917 lower representations they support in hardware.
5918
5919</p>
5920<h5>Semantics:</h5>
5921<p>
5922 This entire intrinsic must be executed atomically. It first loads the value
5923 in memory pointed to by <tt>ptr</tt> and compares it with the value
5924 <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the memory. The
5925 loaded value is yielded in all cases. This provides the equivalent of an
5926 atomic compare-and-swap operation within the SSA framework.
5927</p>
5928<h5>Examples:</h5>
5929
5930<pre>
5931%ptr = malloc i32
5932 store i32 4, %ptr
5933
5934%val1 = add i32 4, 4
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]5935%result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 4, %val1 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5936 <i>; yields {i32}:result1 = 4</i>
5937%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
5938%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
5939
5940%val2 = add i32 1, 1
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]5941%result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 5, %val2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5942 <i>; yields {i32}:result2 = 8</i>
5943%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
5944
5945%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
5946</pre>
5947</div>
5948
5949<!-- _______________________________________________________________________ -->
5950<div class="doc_subsubsection">
5951 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
5952</div>
5953<div class="doc_text">
5954<h5>Syntax:</h5>
5955
5956<p>
5957 This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
5958 integer bit width. Not all targets support all bit widths however.</p>
5959<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]5960declare i8 @llvm.atomic.swap.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;val&gt; )
5961declare i16 @llvm.atomic.swap.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;val&gt; )
5962declare i32 @llvm.atomic.swap.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;val&gt; )
5963declare i64 @llvm.atomic.swap.i64.p0i64( i64* &lt;ptr&gt;, i64 &lt;val&gt; )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5964
5965</pre>
5966<h5>Overview:</h5>
5967<p>
5968 This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
5969 the value from memory. It then stores the value in <tt>val</tt> in the memory
5970 at <tt>ptr</tt>.
5971</p>
5972<h5>Arguments:</h5>
5973
5974<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]5975 The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both the
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5976 <tt>val</tt> argument and the result must be integers of the same bit width.
5977 The first argument, <tt>ptr</tt>, must be a pointer to a value of this
5978 integer type. The targets may only lower integer representations they
5979 support.
5980</p>
5981<h5>Semantics:</h5>
5982<p>
5983 This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
5984 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
5985 equivalent of an atomic swap operation within the SSA framework.
5986
5987</p>
5988<h5>Examples:</h5>
5989<pre>
5990%ptr = malloc i32
5991 store i32 4, %ptr
5992
5993%val1 = add i32 4, 4
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]5994%result1 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val1 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5995 <i>; yields {i32}:result1 = 4</i>
5996%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
5997%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
5998
5999%val2 = add i32 1, 1
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6000%result2 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6001 <i>; yields {i32}:result2 = 8</i>
6002
6003%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
6004%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
6005</pre>
6006</div>
6007
6008<!-- _______________________________________________________________________ -->
6009<div class="doc_subsubsection">
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6010 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6011
6012</div>
6013<div class="doc_text">
6014<h5>Syntax:</h5>
6015<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6016 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]6017 integer bit width. Not all targets support all bit widths however.</p>
6018<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6019declare i8 @llvm.atomic.load.add.i8..p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6020declare i16 @llvm.atomic.load.add.i16..p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6021declare i32 @llvm.atomic.load.add.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6022declare 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]6023
6024</pre>
6025<h5>Overview:</h5>
6026<p>
6027 This intrinsic adds <tt>delta</tt> to the value stored in memory at
6028 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
6029</p>
6030<h5>Arguments:</h5>
6031<p>
6032
6033 The intrinsic takes two arguments, the first a pointer to an integer value
6034 and the second an integer value. The result is also an integer value. These
6035 integer types can have any bit width, but they must all have the same bit
6036 width. The targets may only lower integer representations they support.
6037</p>
6038<h5>Semantics:</h5>
6039<p>
6040 This intrinsic does a series of operations atomically. It first loads the
6041 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
6042 to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
6043</p>
6044
6045<h5>Examples:</h5>
6046<pre>
6047%ptr = malloc i32
6048 store i32 4, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6049%result1 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 4 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6050 <i>; yields {i32}:result1 = 4</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6051%result2 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6052 <i>; yields {i32}:result2 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6053%result3 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 5 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6054 <i>; yields {i32}:result3 = 10</i>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6055%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6056</pre>
6057</div>
6058
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6059<!-- _______________________________________________________________________ -->
6060<div class="doc_subsubsection">
6061 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
6062
6063</div>
6064<div class="doc_text">
6065<h5>Syntax:</h5>
6066<p>
6067 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]6068 any integer bit width and for different address spaces. Not all targets
6069 support all bit widths however.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6070<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6071declare i8 @llvm.atomic.load.sub.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6072declare i16 @llvm.atomic.load.sub.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6073declare i32 @llvm.atomic.load.sub.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6074declare 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]6075
6076</pre>
6077<h5>Overview:</h5>
6078<p>
6079 This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
6080 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
6081</p>
6082<h5>Arguments:</h5>
6083<p>
6084
6085 The intrinsic takes two arguments, the first a pointer to an integer value
6086 and the second an integer value. The result is also an integer value. These
6087 integer types can have any bit width, but they must all have the same bit
6088 width. The targets may only lower integer representations they support.
6089</p>
6090<h5>Semantics:</h5>
6091<p>
6092 This intrinsic does a series of operations atomically. It first loads the
6093 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
6094 result to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
6095</p>
6096
6097<h5>Examples:</h5>
6098<pre>
6099%ptr = malloc i32
6100 store i32 8, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6101%result1 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 4 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6102 <i>; yields {i32}:result1 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6103%result2 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 2 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6104 <i>; yields {i32}:result2 = 4</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6105%result3 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 5 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6106 <i>; yields {i32}:result3 = 2</i>
6107%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
6108</pre>
6109</div>
6110
6111<!-- _______________________________________________________________________ -->
6112<div class="doc_subsubsection">
6113 <a name="int_atomic_load_and">'<tt>llvm.atomic.load.and.*</tt>' Intrinsic</a><br>
6114 <a name="int_atomic_load_nand">'<tt>llvm.atomic.load.nand.*</tt>' Intrinsic</a><br>
6115 <a name="int_atomic_load_or">'<tt>llvm.atomic.load.or.*</tt>' Intrinsic</a><br>
6116 <a name="int_atomic_load_xor">'<tt>llvm.atomic.load.xor.*</tt>' Intrinsic</a><br>
6117
6118</div>
6119<div class="doc_text">
6120<h5>Syntax:</h5>
6121<p>
6122 These are overloaded intrinsics. You can use <tt>llvm.atomic.load_and</tt>,
6123 <tt>llvm.atomic.load_nand</tt>, <tt>llvm.atomic.load_or</tt>, and
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6124 <tt>llvm.atomic.load_xor</tt> on any integer bit width and for different
6125 address spaces. Not all targets support all bit widths however.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6126<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6127declare i8 @llvm.atomic.load.and.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6128declare i16 @llvm.atomic.load.and.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6129declare i32 @llvm.atomic.load.and.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6130declare 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]6131
6132</pre>
6133
6134<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6135declare i8 @llvm.atomic.load.or.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6136declare i16 @llvm.atomic.load.or.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6137declare i32 @llvm.atomic.load.or.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6138declare 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]6139
6140</pre>
6141
6142<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6143declare i8 @llvm.atomic.load.nand.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6144declare i16 @llvm.atomic.load.nand.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6145declare i32 @llvm.atomic.load.nand.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6146declare 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]6147
6148</pre>
6149
6150<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6151declare i8 @llvm.atomic.load.xor.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6152declare i16 @llvm.atomic.load.xor.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6153declare i32 @llvm.atomic.load.xor.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6154declare 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]6155
6156</pre>
6157<h5>Overview:</h5>
6158<p>
6159 These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
6160 the value stored in memory at <tt>ptr</tt>. It yields the original value
6161 at <tt>ptr</tt>.
6162</p>
6163<h5>Arguments:</h5>
6164<p>
6165
6166 These intrinsics take two arguments, the first a pointer to an integer value
6167 and the second an integer value. The result is also an integer value. These
6168 integer types can have any bit width, but they must all have the same bit
6169 width. The targets may only lower integer representations they support.
6170</p>
6171<h5>Semantics:</h5>
6172<p>
6173 These intrinsics does a series of operations atomically. They first load the
6174 value stored at <tt>ptr</tt>. They then do the bitwise operation
6175 <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the original
6176 value stored at <tt>ptr</tt>.
6177</p>
6178
6179<h5>Examples:</h5>
6180<pre>
6181%ptr = malloc i32
6182 store i32 0x0F0F, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6183%result0 = call i32 @llvm.atomic.load.nand.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6184 <i>; yields {i32}:result0 = 0x0F0F</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6185%result1 = call i32 @llvm.atomic.load.and.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6186 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6187%result2 = call i32 @llvm.atomic.load.or.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6188 <i>; yields {i32}:result2 = 0xF0</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6189%result3 = call i32 @llvm.atomic.load.xor.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6190 <i>; yields {i32}:result3 = FF</i>
6191%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
6192</pre>
6193</div>
6194
6195
6196<!-- _______________________________________________________________________ -->
6197<div class="doc_subsubsection">
6198 <a name="int_atomic_load_max">'<tt>llvm.atomic.load.max.*</tt>' Intrinsic</a><br>
6199 <a name="int_atomic_load_min">'<tt>llvm.atomic.load.min.*</tt>' Intrinsic</a><br>
6200 <a name="int_atomic_load_umax">'<tt>llvm.atomic.load.umax.*</tt>' Intrinsic</a><br>
6201 <a name="int_atomic_load_umin">'<tt>llvm.atomic.load.umin.*</tt>' Intrinsic</a><br>
6202
6203</div>
6204<div class="doc_text">
6205<h5>Syntax:</h5>
6206<p>
6207 These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
6208 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6209 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
6210 address spaces. Not all targets
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6211 support all bit widths however.</p>
6212<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6213declare i8 @llvm.atomic.load.max.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6214declare i16 @llvm.atomic.load.max.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6215declare i32 @llvm.atomic.load.max.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6216declare 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]6217
6218</pre>
6219
6220<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6221declare i8 @llvm.atomic.load.min.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6222declare i16 @llvm.atomic.load.min.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6223declare i32 @llvm.atomic.load.min.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6224declare 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]6225
6226</pre>
6227
6228<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6229declare i8 @llvm.atomic.load.umax.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6230declare i16 @llvm.atomic.load.umax.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6231declare i32 @llvm.atomic.load.umax.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6232declare 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]6233
6234</pre>
6235
6236<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6237declare i8 @llvm.atomic.load.umin.i8..p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6238declare i16 @llvm.atomic.load.umin.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6239declare i32 @llvm.atomic.load.umin.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6240declare 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]6241
6242</pre>
6243<h5>Overview:</h5>
6244<p>
6245 These intrinsics takes the signed or unsigned minimum or maximum of
6246 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
6247 original value at <tt>ptr</tt>.
6248</p>
6249<h5>Arguments:</h5>
6250<p>
6251
6252 These intrinsics take two arguments, the first a pointer to an integer value
6253 and the second an integer value. The result is also an integer value. These
6254 integer types can have any bit width, but they must all have the same bit
6255 width. The targets may only lower integer representations they support.
6256</p>
6257<h5>Semantics:</h5>
6258<p>
6259 These intrinsics does a series of operations atomically. They first load the
6260 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or max
6261 <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They yield
6262 the original value stored at <tt>ptr</tt>.
6263</p>
6264
6265<h5>Examples:</h5>
6266<pre>
6267%ptr = malloc i32
6268 store i32 7, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6269%result0 = call i32 @llvm.atomic.load.min.i32.p0i32( i32* %ptr, i32 -2 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6270 <i>; yields {i32}:result0 = 7</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6271%result1 = call i32 @llvm.atomic.load.max.i32.p0i32( i32* %ptr, i32 8 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6272 <i>; yields {i32}:result1 = -2</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6273%result2 = call i32 @llvm.atomic.load.umin.i32.p0i32( i32* %ptr, i32 10 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6274 <i>; yields {i32}:result2 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6275%result3 = call i32 @llvm.atomic.load.umax.i32.p0i32( i32* %ptr, i32 30 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6276 <i>; yields {i32}:result3 = 8</i>
6277%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
6278</pre>
6279</div>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6280
6281<!-- ======================================================================= -->
6282<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6283 <a name="int_general">General Intrinsics</a>
6284</div>
6285
6286<div class="doc_text">
6287<p> This class of intrinsics is designed to be generic and has
6288no specific purpose. </p>
6289</div>
6290
6291<!-- _______________________________________________________________________ -->
6292<div class="doc_subsubsection">
6293 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
6294</div>
6295
6296<div class="doc_text">
6297
6298<h5>Syntax:</h5>
6299<pre>
6300 declare void @llvm.var.annotation(i8* &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6301</pre>
6302
6303<h5>Overview:</h5>
6304
6305<p>
6306The '<tt>llvm.var.annotation</tt>' intrinsic
6307</p>
6308
6309<h5>Arguments:</h5>
6310
6311<p>
6312The first argument is a pointer to a value, the second is a pointer to a
6313global string, the third is a pointer to a global string which is the source
6314file name, and the last argument is the line number.
6315</p>
6316
6317<h5>Semantics:</h5>
6318
6319<p>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]6320This intrinsic allows annotation of local variables with arbitrary strings.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6321This can be useful for special purpose optimizations that want to look for these
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]6322annotations. These have no other defined use, they are ignored by code
6323generation and optimization.
6324</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6325</div>
6326
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6327<!-- _______________________________________________________________________ -->
6328<div class="doc_subsubsection">
Tanya Lattnerc9869b12007-09-21 23:57:59 +0000[diff] [blame]6329 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6330</div>
6331
6332<div class="doc_text">
6333
6334<h5>Syntax:</h5>
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]6335<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
6336any integer bit width.
6337</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6338<pre>
Tanya Lattner09161fe2007-09-22 00:03:01 +0000[diff] [blame]6339 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6340 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6341 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6342 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6343 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]6344</pre>
6345
6346<h5>Overview:</h5>
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]6347
6348<p>
6349The '<tt>llvm.annotation</tt>' intrinsic.
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6350</p>
6351
6352<h5>Arguments:</h5>
6353
6354<p>
6355The first argument is an integer value (result of some expression),
6356the second is a pointer to a global string, the third is a pointer to a global
6357string which is the source file name, and the last argument is the line number.
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]6358It returns the value of the first argument.
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6359</p>
6360
6361<h5>Semantics:</h5>
6362
6363<p>
6364This intrinsic allows annotations to be put on arbitrary expressions
6365with arbitrary strings. This can be useful for special purpose optimizations
6366that want to look for these annotations. These have no other defined use, they
6367are ignored by code generation and optimization.
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]6368</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6369</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6370
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]6371<!-- _______________________________________________________________________ -->
6372<div class="doc_subsubsection">
6373 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
6374</div>
6375
6376<div class="doc_text">
6377
6378<h5>Syntax:</h5>
6379<pre>
6380 declare void @llvm.trap()
6381</pre>
6382
6383<h5>Overview:</h5>
6384
6385<p>
6386The '<tt>llvm.trap</tt>' intrinsic
6387</p>
6388
6389<h5>Arguments:</h5>
6390
6391<p>
6392None
6393</p>
6394
6395<h5>Semantics:</h5>
6396
6397<p>
6398This intrinsics is lowered to the target dependent trap instruction. If the
6399target does not have a trap instruction, this intrinsic will be lowered to the
6400call of the abort() function.
6401</p>
6402</div>
6403
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6404<!-- *********************************************************************** -->
6405<hr>
6406<address>
6407 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
6408 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
6409 <a href="http://validator.w3.org/check/referer"><img
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]6410 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!"></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6411
6412 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
6413 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
6414 Last modified: $Date$
6415</address>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]6416
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6417</body>
6418</html>