Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / llvm / 6bae0d0d235b13c2503eb4f7503bda066c7a008e / . / docs / LangRef.html
blob: 27bcb920025cf9a029262bdcc200d9a9f92157e3 [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
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]338 with quotes. In this way, anything except a <tt>&quot;</tt> character can
339 be used in a named value.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]340
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]341 <li>Unnamed values are represented as an unsigned numeric value with their
342 prefix. For example, %12, @2, %44.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]343
344 <li>Constants, which are described in a <a href="#constants">section about
345 constants</a>, below.</li>
346</ol>
347
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]348<p>LLVM requires that values start with a prefix for two reasons: Compilers
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]349don't need to worry about name clashes with reserved words, and the set of
350reserved words may be expanded in the future without penalty. Additionally,
351unnamed identifiers allow a compiler to quickly come up with a temporary
352variable without having to avoid symbol table conflicts.</p>
353
354<p>Reserved words in LLVM are very similar to reserved words in other
355languages. There are keywords for different opcodes
356('<tt><a href="#i_add">add</a></tt>',
357 '<tt><a href="#i_bitcast">bitcast</a></tt>',
358 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
359href="#t_void">void</a></tt>', '<tt><a href="#t_primitive">i32</a></tt>', etc...),
360and others. These reserved words cannot conflict with variable names, because
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]361none of them start with a prefix character ('%' or '@').</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]362
363<p>Here is an example of LLVM code to multiply the integer variable
364'<tt>%X</tt>' by 8:</p>
365
366<p>The easy way:</p>
367
368<div class="doc_code">
369<pre>
370%result = <a href="#i_mul">mul</a> i32 %X, 8
371</pre>
372</div>
373
374<p>After strength reduction:</p>
375
376<div class="doc_code">
377<pre>
378%result = <a href="#i_shl">shl</a> i32 %X, i8 3
379</pre>
380</div>
381
382<p>And the hard way:</p>
383
384<div class="doc_code">
385<pre>
386<a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
387<a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
388%result = <a href="#i_add">add</a> i32 %1, %1
389</pre>
390</div>
391
392<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
393important lexical features of LLVM:</p>
394
395<ol>
396
397 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
398 line.</li>
399
400 <li>Unnamed temporaries are created when the result of a computation is not
401 assigned to a named value.</li>
402
403 <li>Unnamed temporaries are numbered sequentially</li>
404
405</ol>
406
407<p>...and it also shows a convention that we follow in this document. When
408demonstrating instructions, we will follow an instruction with a comment that
409defines the type and name of value produced. Comments are shown in italic
410text.</p>
411
412</div>
413
414<!-- *********************************************************************** -->
415<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
416<!-- *********************************************************************** -->
417
418<!-- ======================================================================= -->
419<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
420</div>
421
422<div class="doc_text">
423
424<p>LLVM programs are composed of "Module"s, each of which is a
425translation unit of the input programs. Each module consists of
426functions, global variables, and symbol table entries. Modules may be
427combined together with the LLVM linker, which merges function (and
428global variable) definitions, resolves forward declarations, and merges
429symbol table entries. Here is an example of the "hello world" module:</p>
430
431<div class="doc_code">
432<pre><i>; Declare the string constant as a global constant...</i>
433<a href="#identifiers">@.LC0</a> = <a href="#linkage_internal">internal</a> <a
434 href="#globalvars">constant</a> <a href="#t_array">[13 x i8]</a> c"hello world\0A\00" <i>; [13 x i8]*</i>
435
436<i>; External declaration of the puts function</i>
437<a href="#functionstructure">declare</a> i32 @puts(i8 *) <i>; i32(i8 *)* </i>
438
439<i>; Definition of main function</i>
440define i32 @main() { <i>; i32()* </i>
441 <i>; Convert [13x i8 ]* to i8 *...</i>
442 %cast210 = <a
443 href="#i_getelementptr">getelementptr</a> [13 x i8 ]* @.LC0, i64 0, i64 0 <i>; i8 *</i>
444
445 <i>; Call puts function to write out the string to stdout...</i>
446 <a
447 href="#i_call">call</a> i32 @puts(i8 * %cast210) <i>; i32</i>
448 <a
449 href="#i_ret">ret</a> i32 0<br>}<br>
450</pre>
451</div>
452
453<p>This example is made up of a <a href="#globalvars">global variable</a>
454named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
455function, and a <a href="#functionstructure">function definition</a>
456for "<tt>main</tt>".</p>
457
458<p>In general, a module is made up of a list of global values,
459where both functions and global variables are global values. Global values are
460represented by a pointer to a memory location (in this case, a pointer to an
461array of char, and a pointer to a function), and have one of the following <a
462href="#linkage">linkage types</a>.</p>
463
464</div>
465
466<!-- ======================================================================= -->
467<div class="doc_subsection">
468 <a name="linkage">Linkage Types</a>
469</div>
470
471<div class="doc_text">
472
473<p>
474All Global Variables and Functions have one of the following types of linkage:
475</p>
476
477<dl>
478
Dale Johannesen96e7e092008-05-23 23:13:41 +0000[diff] [blame]479 <dt><tt><b><a name="linkage_internal">internal</a></b></tt>: </dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]480
481 <dd>Global values with internal linkage are only directly accessible by
482 objects in the current module. In particular, linking code into a module with
483 an internal global value may cause the internal to be renamed as necessary to
484 avoid collisions. Because the symbol is internal to the module, all
485 references can be updated. This corresponds to the notion of the
486 '<tt>static</tt>' keyword in C.
487 </dd>
488
489 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
490
491 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
492 the same name when linkage occurs. This is typically used to implement
493 inline functions, templates, or other code which must be generated in each
494 translation unit that uses it. Unreferenced <tt>linkonce</tt> globals are
495 allowed to be discarded.
496 </dd>
497
Dale Johannesen96e7e092008-05-23 23:13:41 +0000[diff] [blame]498 <dt><tt><b><a name="linkage_common">common</a></b></tt>: </dt>
499
500 <dd>"<tt>common</tt>" linkage is exactly the same as <tt>linkonce</tt>
501 linkage, except that unreferenced <tt>common</tt> globals may not be
502 discarded. This is used for globals that may be emitted in multiple
503 translation units, but that are not guaranteed to be emitted into every
504 translation unit that uses them. One example of this is tentative
505 definitions in C, such as "<tt>int X;</tt>" at global scope.
506 </dd>
507
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]508 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
509
Dale Johannesen96e7e092008-05-23 23:13:41 +0000[diff] [blame]510 <dd>"<tt>weak</tt>" linkage is the same as <tt>common</tt> linkage, except
511 that some targets may choose to emit different assembly sequences for them
512 for target-dependent reasons. This is used for globals that are declared
513 "weak" in C source code.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]514 </dd>
515
516 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
517
518 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
519 pointer to array type. When two global variables with appending linkage are
520 linked together, the two global arrays are appended together. This is the
521 LLVM, typesafe, equivalent of having the system linker append together
522 "sections" with identical names when .o files are linked.
523 </dd>
524
525 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt>: </dt>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]526 <dd>The semantics of this linkage follow the ELF object file model: the
527 symbol is weak until linked, if not linked, the symbol becomes null instead
528 of being an undefined reference.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]529 </dd>
530
531 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
532
533 <dd>If none of the above identifiers are used, the global is externally
534 visible, meaning that it participates in linkage and can be used to resolve
535 external symbol references.
536 </dd>
537</dl>
538
539 <p>
540 The next two types of linkage are targeted for Microsoft Windows platform
541 only. They are designed to support importing (exporting) symbols from (to)
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]542 DLLs (Dynamic Link Libraries).
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]543 </p>
544
545 <dl>
546 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt>: </dt>
547
548 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
549 or variable via a global pointer to a pointer that is set up by the DLL
550 exporting the symbol. On Microsoft Windows targets, the pointer name is
551 formed by combining <code>_imp__</code> and the function or variable name.
552 </dd>
553
554 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt>: </dt>
555
556 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
557 pointer to a pointer in a DLL, so that it can be referenced with the
558 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
559 name is formed by combining <code>_imp__</code> and the function or variable
560 name.
561 </dd>
562
563</dl>
564
565<p><a name="linkage_external"></a>For example, since the "<tt>.LC0</tt>"
566variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
567variable and was linked with this one, one of the two would be renamed,
568preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
569external (i.e., lacking any linkage declarations), they are accessible
570outside of the current module.</p>
571<p>It is illegal for a function <i>declaration</i>
572to have any linkage type other than "externally visible", <tt>dllimport</tt>,
573or <tt>extern_weak</tt>.</p>
574<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]575linkages.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]576</div>
577
578<!-- ======================================================================= -->
579<div class="doc_subsection">
580 <a name="callingconv">Calling Conventions</a>
581</div>
582
583<div class="doc_text">
584
585<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
586and <a href="#i_invoke">invokes</a> can all have an optional calling convention
587specified for the call. The calling convention of any pair of dynamic
588caller/callee must match, or the behavior of the program is undefined. The
589following calling conventions are supported by LLVM, and more may be added in
590the future:</p>
591
592<dl>
593 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
594
595 <dd>This calling convention (the default if no other calling convention is
596 specified) matches the target C calling conventions. This calling convention
597 supports varargs function calls and tolerates some mismatch in the declared
598 prototype and implemented declaration of the function (as does normal C).
599 </dd>
600
601 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
602
603 <dd>This calling convention attempts to make calls as fast as possible
604 (e.g. by passing things in registers). This calling convention allows the
605 target to use whatever tricks it wants to produce fast code for the target,
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]606 without having to conform to an externally specified ABI (Application Binary
607 Interface). Implementations of this convention should allow arbitrary
Arnold Schwaighofer07444922008-05-14 09:17:12 +0000[diff] [blame]608 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> to be
609 supported. This calling convention does not support varargs and requires the
610 prototype of all callees to exactly match the prototype of the function
611 definition.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]612 </dd>
613
614 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
615
616 <dd>This calling convention attempts to make code in the caller as efficient
617 as possible under the assumption that the call is not commonly executed. As
618 such, these calls often preserve all registers so that the call does not break
619 any live ranges in the caller side. This calling convention does not support
620 varargs and requires the prototype of all callees to exactly match the
621 prototype of the function definition.
622 </dd>
623
624 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
625
626 <dd>Any calling convention may be specified by number, allowing
627 target-specific calling conventions to be used. Target specific calling
628 conventions start at 64.
629 </dd>
630</dl>
631
632<p>More calling conventions can be added/defined on an as-needed basis, to
633support pascal conventions or any other well-known target-independent
634convention.</p>
635
636</div>
637
638<!-- ======================================================================= -->
639<div class="doc_subsection">
640 <a name="visibility">Visibility Styles</a>
641</div>
642
643<div class="doc_text">
644
645<p>
646All Global Variables and Functions have one of the following visibility styles:
647</p>
648
649<dl>
650 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
651
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]652 <dd>On targets that use the ELF object file format, default visibility means
653 that the declaration is visible to other
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]654 modules and, in shared libraries, means that the declared entity may be
655 overridden. On Darwin, default visibility means that the declaration is
656 visible to other modules. Default visibility corresponds to "external
657 linkage" in the language.
658 </dd>
659
660 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
661
662 <dd>Two declarations of an object with hidden visibility refer to the same
663 object if they are in the same shared object. Usually, hidden visibility
664 indicates that the symbol will not be placed into the dynamic symbol table,
665 so no other module (executable or shared library) can reference it
666 directly.
667 </dd>
668
669 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
670
671 <dd>On ELF, protected visibility indicates that the symbol will be placed in
672 the dynamic symbol table, but that references within the defining module will
673 bind to the local symbol. That is, the symbol cannot be overridden by another
674 module.
675 </dd>
676</dl>
677
678</div>
679
680<!-- ======================================================================= -->
681<div class="doc_subsection">
682 <a name="globalvars">Global Variables</a>
683</div>
684
685<div class="doc_text">
686
687<p>Global variables define regions of memory allocated at compilation time
688instead of run-time. Global variables may optionally be initialized, may have
689an explicit section to be placed in, and may have an optional explicit alignment
690specified. A variable may be defined as "thread_local", which means that it
691will not be shared by threads (each thread will have a separated copy of the
692variable). A variable may be defined as a global "constant," which indicates
693that the contents of the variable will <b>never</b> be modified (enabling better
694optimization, allowing the global data to be placed in the read-only section of
695an executable, etc). Note that variables that need runtime initialization
696cannot be marked "constant" as there is a store to the variable.</p>
697
698<p>
699LLVM explicitly allows <em>declarations</em> of global variables to be marked
700constant, even if the final definition of the global is not. This capability
701can be used to enable slightly better optimization of the program, but requires
702the language definition to guarantee that optimizations based on the
703'constantness' are valid for the translation units that do not include the
704definition.
705</p>
706
707<p>As SSA values, global variables define pointer values that are in
708scope (i.e. they dominate) all basic blocks in the program. Global
709variables always define a pointer to their "content" type because they
710describe a region of memory, and all memory objects in LLVM are
711accessed through pointers.</p>
712
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]713<p>A global variable may be declared to reside in a target-specifc numbered
714address space. For targets that support them, address spaces may affect how
715optimizations are performed and/or what target instructions are used to access
Christopher Lamb20a39e92007-12-12 08:44:39 +0000[diff] [blame]716the variable. The default address space is zero. The address space qualifier
717must precede any other attributes.</p>
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]718
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]719<p>LLVM allows an explicit section to be specified for globals. If the target
720supports it, it will emit globals to the section specified.</p>
721
722<p>An explicit alignment may be specified for a global. If not present, or if
723the alignment is set to zero, the alignment of the global is set by the target
724to whatever it feels convenient. If an explicit alignment is specified, the
725global is forced to have at least that much alignment. All alignments must be
726a power of 2.</p>
727
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]728<p>For example, the following defines a global in a numbered address space with
729an initializer, section, and alignment:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]730
731<div class="doc_code">
732<pre>
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]733@G = constant float 1.0 addrspace(5), section "foo", align 4
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]734</pre>
735</div>
736
737</div>
738
739
740<!-- ======================================================================= -->
741<div class="doc_subsection">
742 <a name="functionstructure">Functions</a>
743</div>
744
745<div class="doc_text">
746
747<p>LLVM function definitions consist of the "<tt>define</tt>" keyord,
748an optional <a href="#linkage">linkage type</a>, an optional
749<a href="#visibility">visibility style</a>, an optional
750<a href="#callingconv">calling convention</a>, a return type, an optional
751<a href="#paramattrs">parameter attribute</a> for the return type, a function
752name, a (possibly empty) argument list (each with optional
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]753<a href="#paramattrs">parameter attributes</a>), optional
754<a href="#fnattrs">function attributes</a>, an optional section,
755an optional alignment, an optional <a href="#gc">garbage collector name</a>,
Chris Lattneree202542008-10-04 18:10:21 +0000[diff] [blame]756an opening curly brace, a list of basic blocks, and a closing curly brace.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]757
758LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
759optional <a href="#linkage">linkage type</a>, an optional
760<a href="#visibility">visibility style</a>, an optional
761<a href="#callingconv">calling convention</a>, a return type, an optional
762<a href="#paramattrs">parameter attribute</a> for the return type, a function
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]763name, a possibly empty list of arguments, an optional alignment, and an optional
Gordon Henriksend606f5b2007-12-10 03:30:21 +0000[diff] [blame]764<a href="#gc">garbage collector name</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]765
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]766<p>A function definition contains a list of basic blocks, forming the CFG
767(Control Flow Graph) for
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]768the function. Each basic block may optionally start with a label (giving the
769basic block a symbol table entry), contains a list of instructions, and ends
770with a <a href="#terminators">terminator</a> instruction (such as a branch or
771function return).</p>
772
773<p>The first basic block in a function is special in two ways: it is immediately
774executed on entrance to the function, and it is not allowed to have predecessor
775basic blocks (i.e. there can not be any branches to the entry block of a
776function). Because the block can have no predecessors, it also cannot have any
777<a href="#i_phi">PHI nodes</a>.</p>
778
779<p>LLVM allows an explicit section to be specified for functions. If the target
780supports it, it will emit functions to the section specified.</p>
781
782<p>An explicit alignment may be specified for a function. If not present, or if
783the alignment is set to zero, the alignment of the function is set by the target
784to whatever it feels convenient. If an explicit alignment is specified, the
785function is forced to have at least that much alignment. All alignments must be
786a power of 2.</p>
787
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]788 <h5>Syntax:</h5>
789
790<div class="doc_code">
Chris Lattner1e5c5cd02008-10-13 16:55:18 +0000[diff] [blame]791<tt>
792define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
793 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
794 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
795 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
796 [<a href="#gc">gc</a>] { ... }
797</tt>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]798</div>
799
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]800</div>
801
802
803<!-- ======================================================================= -->
804<div class="doc_subsection">
805 <a name="aliasstructure">Aliases</a>
806</div>
807<div class="doc_text">
808 <p>Aliases act as "second name" for the aliasee value (which can be either
Anton Korobeynikov96822812008-03-22 08:36:14 +0000[diff] [blame]809 function, global variable, another alias or bitcast of global value). Aliases
810 may have an optional <a href="#linkage">linkage type</a>, and an
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]811 optional <a href="#visibility">visibility style</a>.</p>
812
813 <h5>Syntax:</h5>
814
815<div class="doc_code">
816<pre>
Duncan Sandsd7bfabf2008-09-12 20:48:21 +0000[diff] [blame]817@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]818</pre>
819</div>
820
821</div>
822
823
824
825<!-- ======================================================================= -->
826<div class="doc_subsection"><a name="paramattrs">Parameter Attributes</a></div>
827<div class="doc_text">
828 <p>The return type and each parameter of a function type may have a set of
829 <i>parameter attributes</i> associated with them. Parameter attributes are
830 used to communicate additional information about the result or parameters of
Duncan Sandsf5588dc2007-11-27 13:23:08 +0000[diff] [blame]831 a function. Parameter attributes are considered to be part of the function,
832 not of the function type, so functions with different parameter attributes
833 can have the same function type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]834
835 <p>Parameter attributes are simple keywords that follow the type specified. If
836 multiple parameter attributes are needed, they are space separated. For
837 example:</p>
838
839<div class="doc_code">
840<pre>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]841declare i32 @printf(i8* noalias , ...)
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]842declare i32 @atoi(i8 zeroext)
843declare signext i8 @returns_signed_char()
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]844</pre>
845</div>
846
Duncan Sandsf5588dc2007-11-27 13:23:08 +0000[diff] [blame]847 <p>Note that any attributes for the function result (<tt>nounwind</tt>,
848 <tt>readonly</tt>) come immediately after the argument list.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]849
850 <p>Currently, only the following parameter attributes are defined:</p>
851 <dl>
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]852 <dt><tt>zeroext</tt></dt>
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]853 <dd>This indicates to the code generator that the parameter or return value
854 should be zero-extended to a 32-bit value by the caller (for a parameter)
855 or the callee (for a return value).</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]856
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]857 <dt><tt>signext</tt></dt>
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]858 <dd>This indicates to the code generator that the parameter or return value
859 should be sign-extended to a 32-bit value by the caller (for a parameter)
860 or the callee (for a return value).</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]861
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]862 <dt><tt>inreg</tt></dt>
Dale Johannesenc08a0e22008-09-25 20:47:45 +0000[diff] [blame]863 <dd>This indicates that this parameter or return value should be treated
864 in a special target-dependent fashion during while emitting code for a
865 function call or return (usually, by putting it in a register as opposed
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]866 to memory, though some targets use it to distinguish between two different
867 kinds of registers). Use of this attribute is target-specific.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]868
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]869 <dt><tt><a name="byval">byval</a></tt></dt>
Chris Lattner04c86182008-01-15 04:34:22 +0000[diff] [blame]870 <dd>This indicates that the pointer parameter should really be passed by
871 value to the function. The attribute implies that a hidden copy of the
872 pointee is made between the caller and the callee, so the callee is unable
Chris Lattner6a9f3c42008-08-05 18:21:08 +0000[diff] [blame]873 to modify the value in the callee. This attribute is only valid on LLVM
Chris Lattner04c86182008-01-15 04:34:22 +0000[diff] [blame]874 pointer arguments. It is generally used to pass structs and arrays by
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]875 value, but is also valid on pointers to scalars. The copy is considered to
876 belong to the caller not the callee (for example,
877 <tt><a href="#readonly">readonly</a></tt> functions should not write to
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]878 <tt>byval</tt> parameters). This is not a valid attribute for return
879 values. </dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]880
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]881 <dt><tt>sret</tt></dt>
Duncan Sands616cc032008-02-18 04:19:38 +0000[diff] [blame]882 <dd>This indicates that the pointer parameter specifies the address of a
883 structure that is the return value of the function in the source program.
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]884 This pointer must be guaranteed by the caller to be valid: loads and stores
885 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]886 be applied to the first parameter. This is not a valid attribute for
887 return values. </dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]888
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]889 <dt><tt>noalias</tt></dt>
Owen Andersonc4fc4cd2008-02-18 04:09:01 +0000[diff] [blame]890 <dd>This indicates that the parameter does not alias any global or any other
891 parameter. The caller is responsible for ensuring that this is the case,
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]892 usually by placing the value in a stack allocation. This is not a valid
893 attribute for return values.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]894
Duncan Sands4ee46812007-07-27 19:57:41 +0000[diff] [blame]895 <dt><tt>nest</tt></dt>
Duncan Sandsf1a7d4c2008-07-08 09:27:25 +0000[diff] [blame]896 <dd>This indicates that the pointer parameter can be excised using the
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]897 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
898 attribute for return values.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]899 </dl>
900
901</div>
902
903<!-- ======================================================================= -->
904<div class="doc_subsection">
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]905 <a name="gc">Garbage Collector Names</a>
906</div>
907
908<div class="doc_text">
909<p>Each function may specify a garbage collector name, which is simply a
910string.</p>
911
912<div class="doc_code"><pre
913>define void @f() gc "name" { ...</pre></div>
914
915<p>The compiler declares the supported values of <i>name</i>. Specifying a
916collector which will cause the compiler to alter its output in order to support
917the named garbage collection algorithm.</p>
918</div>
919
920<!-- ======================================================================= -->
921<div class="doc_subsection">
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]922 <a name="fnattrs">Function Attributes</a>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]923</div>
924
925<div class="doc_text">
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]926
927<p>Function attributes are set to communicate additional information about
928 a function. Function attributes are considered to be part of the function,
929 not of the function type, so functions with different parameter attributes
930 can have the same function type.</p>
931
932 <p>Function attributes are simple keywords that follow the type specified. If
933 multiple attributes are needed, they are space separated. For
934 example:</p>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]935
936<div class="doc_code">
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]937<pre>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]938define void @f() noinline { ... }
939define void @f() alwaysinline { ... }
940define void @f() alwaysinline optsize { ... }
941define void @f() optsize
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]942</pre>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]943</div>
944
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]945<dl>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]946<dt><tt>alwaysinline</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]947<dd>This attribute indicates that the inliner should attempt to inline this
948function into callers whenever possible, ignoring any active inlining size
949threshold for this caller.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]950
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]951<dt><tt>noinline</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]952<dd>This attribute indicates that the inliner should never inline this function
Chris Lattner377f3ae2008-10-05 17:14:59 +0000[diff] [blame]953in any situation. This attribute may not be used together with the
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]954<tt>alwaysinline</tt> attribute.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]955
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]956<dt><tt>optsize</tt></dt>
Devang Patelc29099b2008-09-29 18:34:44 +0000[diff] [blame]957<dd>This attribute suggests that optimization passes and code generator passes
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]958make choices that keep the code size of this function low, and otherwise do
959optimizations specifically to reduce code size.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]960
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]961<dt><tt>noreturn</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]962<dd>This function attribute indicates that the function never returns normally.
963This produces undefined behavior at runtime if the function ever does
964dynamically return.</dd>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]965
966<dt><tt>nounwind</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]967<dd>This function attribute indicates that the function never returns with an
968unwind or exceptional control flow. If the function does unwind, its runtime
969behavior is undefined.</dd>
970
971<dt><tt>readnone</tt></dt>
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]972<dd>This attribute indicates that the function computes its result (or the
973exception it throws) based strictly on its arguments, without dereferencing any
974pointer arguments or otherwise accessing any mutable state (e.g. memory, control
975registers, etc) visible to caller functions. It does not write through any
976pointer arguments (including <tt><a href="#byval">byval</a></tt> arguments) and
977never changes any state visible to callers.</dd>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]978
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]979<dt><tt><a name="readonly">readonly</a></tt></dt>
980<dd>This attribute indicates that the function does not write through any
981pointer arguments (including <tt><a href="#byval">byval</a></tt> arguments)
982or otherwise modify any state (e.g. memory, control registers, etc) visible to
983caller functions. It may dereference pointer arguments and read state that may
984be set in the caller. A readonly function always returns the same value (or
985throws the same exception) when called with the same set of arguments and global
986state.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]987</dl>
988
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]989</div>
990
991<!-- ======================================================================= -->
992<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]993 <a name="moduleasm">Module-Level Inline Assembly</a>
994</div>
995
996<div class="doc_text">
997<p>
998Modules may contain "module-level inline asm" blocks, which corresponds to the
999GCC "file scope inline asm" blocks. These blocks are internally concatenated by
1000LLVM and treated as a single unit, but may be separated in the .ll file if
1001desired. The syntax is very simple:
1002</p>
1003
1004<div class="doc_code">
1005<pre>
1006module asm "inline asm code goes here"
1007module asm "more can go here"
1008</pre>
1009</div>
1010
1011<p>The strings can contain any character by escaping non-printable characters.
1012 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
1013 for the number.
1014</p>
1015
1016<p>
1017 The inline asm code is simply printed to the machine code .s file when
1018 assembly code is generated.
1019</p>
1020</div>
1021
1022<!-- ======================================================================= -->
1023<div class="doc_subsection">
1024 <a name="datalayout">Data Layout</a>
1025</div>
1026
1027<div class="doc_text">
1028<p>A module may specify a target specific data layout string that specifies how
1029data is to be laid out in memory. The syntax for the data layout is simply:</p>
1030<pre> target datalayout = "<i>layout specification</i>"</pre>
1031<p>The <i>layout specification</i> consists of a list of specifications
1032separated by the minus sign character ('-'). Each specification starts with a
1033letter and may include other information after the letter to define some
1034aspect of the data layout. The specifications accepted are as follows: </p>
1035<dl>
1036 <dt><tt>E</tt></dt>
1037 <dd>Specifies that the target lays out data in big-endian form. That is, the
1038 bits with the most significance have the lowest address location.</dd>
1039 <dt><tt>e</tt></dt>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1040 <dd>Specifies that the target lays out data in little-endian form. That is,
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1041 the bits with the least significance have the lowest address location.</dd>
1042 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1043 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
1044 <i>preferred</i> alignments. All sizes are in bits. Specifying the <i>pref</i>
1045 alignment is optional. If omitted, the preceding <tt>:</tt> should be omitted
1046 too.</dd>
1047 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1048 <dd>This specifies the alignment for an integer type of a given bit
1049 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1050 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1051 <dd>This specifies the alignment for a vector type of a given bit
1052 <i>size</i>.</dd>
1053 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1054 <dd>This specifies the alignment for a floating point type of a given bit
1055 <i>size</i>. The value of <i>size</i> must be either 32 (float) or 64
1056 (double).</dd>
1057 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1058 <dd>This specifies the alignment for an aggregate type of a given bit
1059 <i>size</i>.</dd>
1060</dl>
1061<p>When constructing the data layout for a given target, LLVM starts with a
1062default set of specifications which are then (possibly) overriden by the
1063specifications in the <tt>datalayout</tt> keyword. The default specifications
1064are given in this list:</p>
1065<ul>
1066 <li><tt>E</tt> - big endian</li>
1067 <li><tt>p:32:64:64</tt> - 32-bit pointers with 64-bit alignment</li>
1068 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1069 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1070 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1071 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1072 <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]1073 alignment of 64-bits</li>
1074 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1075 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1076 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1077 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1078 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
1079</ul>
Chris Lattner6a9f3c42008-08-05 18:21:08 +0000[diff] [blame]1080<p>When LLVM is determining the alignment for a given type, it uses the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]1081following rules:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1082<ol>
1083 <li>If the type sought is an exact match for one of the specifications, that
1084 specification is used.</li>
1085 <li>If no match is found, and the type sought is an integer type, then the
1086 smallest integer type that is larger than the bitwidth of the sought type is
1087 used. If none of the specifications are larger than the bitwidth then the the
1088 largest integer type is used. For example, given the default specifications
1089 above, the i7 type will use the alignment of i8 (next largest) while both
1090 i65 and i256 will use the alignment of i64 (largest specified).</li>
1091 <li>If no match is found, and the type sought is a vector type, then the
1092 largest vector type that is smaller than the sought vector type will be used
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]1093 as a fall back. This happens because &lt;128 x double&gt; can be implemented
1094 in terms of 64 &lt;2 x double&gt;, for example.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1095</ol>
1096</div>
1097
1098<!-- *********************************************************************** -->
1099<div class="doc_section"> <a name="typesystem">Type System</a> </div>
1100<!-- *********************************************************************** -->
1101
1102<div class="doc_text">
1103
1104<p>The LLVM type system is one of the most important features of the
1105intermediate representation. Being typed enables a number of
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1106optimizations to be performed on the intermediate representation directly,
1107without having to do
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1108extra analyses on the side before the transformation. A strong type
1109system makes it easier to read the generated code and enables novel
1110analyses and transformations that are not feasible to perform on normal
1111three address code representations.</p>
1112
1113</div>
1114
1115<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1116<div class="doc_subsection"> <a name="t_classifications">Type
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1117Classifications</a> </div>
1118<div class="doc_text">
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1119<p>The types fall into a few useful
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1120classifications:</p>
1121
1122<table border="1" cellspacing="0" cellpadding="4">
1123 <tbody>
1124 <tr><th>Classification</th><th>Types</th></tr>
1125 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1126 <td><a href="#t_integer">integer</a></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1127 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
1128 </tr>
1129 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1130 <td><a href="#t_floating">floating point</a></td>
1131 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1132 </tr>
1133 <tr>
1134 <td><a name="t_firstclass">first class</a></td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1135 <td><a href="#t_integer">integer</a>,
1136 <a href="#t_floating">floating point</a>,
1137 <a href="#t_pointer">pointer</a>,
Dan Gohmanf6237db2008-06-18 18:42:13 +0000[diff] [blame]1138 <a href="#t_vector">vector</a>,
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]1139 <a href="#t_struct">structure</a>,
1140 <a href="#t_array">array</a>,
Dan Gohmand13951e2008-05-23 22:50:26 +0000[diff] [blame]1141 <a href="#t_label">label</a>.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1142 </td>
1143 </tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1144 <tr>
1145 <td><a href="#t_primitive">primitive</a></td>
1146 <td><a href="#t_label">label</a>,
1147 <a href="#t_void">void</a>,
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1148 <a href="#t_floating">floating point</a>.</td>
1149 </tr>
1150 <tr>
1151 <td><a href="#t_derived">derived</a></td>
1152 <td><a href="#t_integer">integer</a>,
1153 <a href="#t_array">array</a>,
1154 <a href="#t_function">function</a>,
1155 <a href="#t_pointer">pointer</a>,
1156 <a href="#t_struct">structure</a>,
1157 <a href="#t_pstruct">packed structure</a>,
1158 <a href="#t_vector">vector</a>,
1159 <a href="#t_opaque">opaque</a>.
Dan Gohman032ba852008-10-14 16:32:04 +0000[diff] [blame]1160 </td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1161 </tr>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1162 </tbody>
1163</table>
1164
1165<p>The <a href="#t_firstclass">first class</a> types are perhaps the
1166most important. Values of these types are the only ones which can be
1167produced by instructions, passed as arguments, or used as operands to
Dan Gohman4f29e422008-05-23 21:53:15 +0000[diff] [blame]1168instructions.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1169</div>
1170
1171<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1172<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1173
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1174<div class="doc_text">
1175<p>The primitive types are the fundamental building blocks of the LLVM
1176system.</p>
1177
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1178</div>
1179
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1180<!-- _______________________________________________________________________ -->
1181<div class="doc_subsubsection"> <a name="t_floating">Floating Point Types</a> </div>
1182
1183<div class="doc_text">
1184 <table>
1185 <tbody>
1186 <tr><th>Type</th><th>Description</th></tr>
1187 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1188 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1189 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1190 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1191 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1192 </tbody>
1193 </table>
1194</div>
1195
1196<!-- _______________________________________________________________________ -->
1197<div class="doc_subsubsection"> <a name="t_void">Void Type</a> </div>
1198
1199<div class="doc_text">
1200<h5>Overview:</h5>
1201<p>The void type does not represent any value and has no size.</p>
1202
1203<h5>Syntax:</h5>
1204
1205<pre>
1206 void
1207</pre>
1208</div>
1209
1210<!-- _______________________________________________________________________ -->
1211<div class="doc_subsubsection"> <a name="t_label">Label Type</a> </div>
1212
1213<div class="doc_text">
1214<h5>Overview:</h5>
1215<p>The label type represents code labels.</p>
1216
1217<h5>Syntax:</h5>
1218
1219<pre>
1220 label
1221</pre>
1222</div>
1223
1224
1225<!-- ======================================================================= -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1226<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
1227
1228<div class="doc_text">
1229
1230<p>The real power in LLVM comes from the derived types in the system.
1231This is what allows a programmer to represent arrays, functions,
1232pointers, and other useful types. Note that these derived types may be
1233recursive: For example, it is possible to have a two dimensional array.</p>
1234
1235</div>
1236
1237<!-- _______________________________________________________________________ -->
1238<div class="doc_subsubsection"> <a name="t_integer">Integer Type</a> </div>
1239
1240<div class="doc_text">
1241
1242<h5>Overview:</h5>
1243<p>The integer type is a very simple derived type that simply specifies an
1244arbitrary bit width for the integer type desired. Any bit width from 1 bit to
12452^23-1 (about 8 million) can be specified.</p>
1246
1247<h5>Syntax:</h5>
1248
1249<pre>
1250 iN
1251</pre>
1252
1253<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1254value.</p>
1255
1256<h5>Examples:</h5>
1257<table class="layout">
Chris Lattner251ab812007-12-18 06:18:21 +0000[diff] [blame]1258 <tbody>
1259 <tr>
1260 <td><tt>i1</tt></td>
1261 <td>a single-bit integer.</td>
1262 </tr><tr>
1263 <td><tt>i32</tt></td>
1264 <td>a 32-bit integer.</td>
1265 </tr><tr>
1266 <td><tt>i1942652</tt></td>
1267 <td>a really big integer of over 1 million bits.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1268 </tr>
Chris Lattner251ab812007-12-18 06:18:21 +0000[diff] [blame]1269 </tbody>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1270</table>
1271</div>
1272
1273<!-- _______________________________________________________________________ -->
1274<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
1275
1276<div class="doc_text">
1277
1278<h5>Overview:</h5>
1279
1280<p>The array type is a very simple derived type that arranges elements
1281sequentially in memory. The array type requires a size (number of
1282elements) and an underlying data type.</p>
1283
1284<h5>Syntax:</h5>
1285
1286<pre>
1287 [&lt;# elements&gt; x &lt;elementtype&gt;]
1288</pre>
1289
1290<p>The number of elements is a constant integer value; elementtype may
1291be any type with a size.</p>
1292
1293<h5>Examples:</h5>
1294<table class="layout">
1295 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1296 <td class="left"><tt>[40 x i32]</tt></td>
1297 <td class="left">Array of 40 32-bit integer values.</td>
1298 </tr>
1299 <tr class="layout">
1300 <td class="left"><tt>[41 x i32]</tt></td>
1301 <td class="left">Array of 41 32-bit integer values.</td>
1302 </tr>
1303 <tr class="layout">
1304 <td class="left"><tt>[4 x i8]</tt></td>
1305 <td class="left">Array of 4 8-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1306 </tr>
1307</table>
1308<p>Here are some examples of multidimensional arrays:</p>
1309<table class="layout">
1310 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1311 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1312 <td class="left">3x4 array of 32-bit integer values.</td>
1313 </tr>
1314 <tr class="layout">
1315 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1316 <td class="left">12x10 array of single precision floating point values.</td>
1317 </tr>
1318 <tr class="layout">
1319 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1320 <td class="left">2x3x4 array of 16-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1321 </tr>
1322</table>
1323
1324<p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
1325length array. Normally, accesses past the end of an array are undefined in
1326LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
1327As a special case, however, zero length arrays are recognized to be variable
1328length. This allows implementation of 'pascal style arrays' with the LLVM
1329type "{ i32, [0 x float]}", for example.</p>
1330
1331</div>
1332
1333<!-- _______________________________________________________________________ -->
1334<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
1335<div class="doc_text">
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1336
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1337<h5>Overview:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1338
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1339<p>The function type can be thought of as a function signature. It
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1340consists of a return type and a list of formal parameter types. The
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1341return 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]1342If 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]1343class types, and the struct must have at least one element.</p>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1344
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1345<h5>Syntax:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1346
1347<pre>
1348 &lt;returntype list&gt; (&lt;parameter list&gt;)
1349</pre>
1350
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1351<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
1352specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1353which indicates that the function takes a variable number of arguments.
1354Variable argument functions can access their arguments with the <a
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1355 href="#int_varargs">variable argument handling intrinsic</a> functions.
1356'<tt>&lt;returntype list&gt;</tt>' is a comma-separated list of
1357<a href="#t_firstclass">first class</a> type specifiers.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1358
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1359<h5>Examples:</h5>
1360<table class="layout">
1361 <tr class="layout">
1362 <td class="left"><tt>i32 (i32)</tt></td>
1363 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
1364 </td>
1365 </tr><tr class="layout">
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]1366 <td class="left"><tt>float&nbsp;(i16&nbsp;signext,&nbsp;i32&nbsp;*)&nbsp;*
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1367 </tt></td>
1368 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
1369 an <tt>i16</tt> that should be sign extended and a
1370 <a href="#t_pointer">pointer</a> to <tt>i32</tt>, returning
1371 <tt>float</tt>.
1372 </td>
1373 </tr><tr class="layout">
1374 <td class="left"><tt>i32 (i8*, ...)</tt></td>
1375 <td class="left">A vararg function that takes at least one
1376 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1377 which returns an integer. This is the signature for <tt>printf</tt> in
1378 LLVM.
1379 </td>
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1380 </tr><tr class="layout">
1381 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Devang Patel6dddba22008-03-24 18:10:52 +0000[diff] [blame]1382 <td class="left">A function taking an <tt>i32></tt>, returning two
1383 <tt> i32 </tt> values as an aggregate of type <tt>{ i32, i32 }</tt>
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1384 </td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1385 </tr>
1386</table>
1387
1388</div>
1389<!-- _______________________________________________________________________ -->
1390<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
1391<div class="doc_text">
1392<h5>Overview:</h5>
1393<p>The structure type is used to represent a collection of data members
1394together in memory. The packing of the field types is defined to match
1395the ABI of the underlying processor. The elements of a structure may
1396be any type that has a size.</p>
1397<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1398and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1399field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1400instruction.</p>
1401<h5>Syntax:</h5>
1402<pre> { &lt;type list&gt; }<br></pre>
1403<h5>Examples:</h5>
1404<table class="layout">
1405 <tr class="layout">
1406 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1407 <td class="left">A triple of three <tt>i32</tt> values</td>
1408 </tr><tr class="layout">
1409 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1410 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1411 second element is a <a href="#t_pointer">pointer</a> to a
1412 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1413 an <tt>i32</tt>.</td>
1414 </tr>
1415</table>
1416</div>
1417
1418<!-- _______________________________________________________________________ -->
1419<div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
1420</div>
1421<div class="doc_text">
1422<h5>Overview:</h5>
1423<p>The packed structure type is used to represent a collection of data members
1424together in memory. There is no padding between fields. Further, the alignment
1425of a packed structure is 1 byte. The elements of a packed structure may
1426be any type that has a size.</p>
1427<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1428and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1429field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1430instruction.</p>
1431<h5>Syntax:</h5>
1432<pre> &lt; { &lt;type list&gt; } &gt; <br></pre>
1433<h5>Examples:</h5>
1434<table class="layout">
1435 <tr class="layout">
1436 <td class="left"><tt>&lt; { i32, i32, i32 } &gt;</tt></td>
1437 <td class="left">A triple of three <tt>i32</tt> values</td>
1438 </tr><tr class="layout">
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1439 <td class="left">
1440<tt>&lt;&nbsp;{&nbsp;float,&nbsp;i32&nbsp;(i32)*&nbsp;}&nbsp;&gt;</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1441 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1442 second element is a <a href="#t_pointer">pointer</a> to a
1443 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1444 an <tt>i32</tt>.</td>
1445 </tr>
1446</table>
1447</div>
1448
1449<!-- _______________________________________________________________________ -->
1450<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
1451<div class="doc_text">
1452<h5>Overview:</h5>
1453<p>As in many languages, the pointer type represents a pointer or
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]1454reference to another object, which must live in memory. Pointer types may have
1455an optional address space attribute defining the target-specific numbered
1456address space where the pointed-to object resides. The default address space is
1457zero.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1458<h5>Syntax:</h5>
1459<pre> &lt;type&gt; *<br></pre>
1460<h5>Examples:</h5>
1461<table class="layout">
1462 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1463 <td class="left"><tt>[4x i32]*</tt></td>
1464 <td class="left">A <a href="#t_pointer">pointer</a> to <a
1465 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
1466 </tr>
1467 <tr class="layout">
1468 <td class="left"><tt>i32 (i32 *) *</tt></td>
1469 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1470 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1471 <tt>i32</tt>.</td>
1472 </tr>
1473 <tr class="layout">
1474 <td class="left"><tt>i32 addrspace(5)*</tt></td>
1475 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
1476 that resides in address space #5.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1477 </tr>
1478</table>
1479</div>
1480
1481<!-- _______________________________________________________________________ -->
1482<div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
1483<div class="doc_text">
1484
1485<h5>Overview:</h5>
1486
1487<p>A vector type is a simple derived type that represents a vector
1488of elements. Vector types are used when multiple primitive data
1489are operated in parallel using a single instruction (SIMD).
1490A vector type requires a size (number of
1491elements) and an underlying primitive data type. Vectors must have a power
1492of two length (1, 2, 4, 8, 16 ...). Vector types are
1493considered <a href="#t_firstclass">first class</a>.</p>
1494
1495<h5>Syntax:</h5>
1496
1497<pre>
1498 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1499</pre>
1500
1501<p>The number of elements is a constant integer value; elementtype may
1502be any integer or floating point type.</p>
1503
1504<h5>Examples:</h5>
1505
1506<table class="layout">
1507 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1508 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
1509 <td class="left">Vector of 4 32-bit integer values.</td>
1510 </tr>
1511 <tr class="layout">
1512 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
1513 <td class="left">Vector of 8 32-bit floating-point values.</td>
1514 </tr>
1515 <tr class="layout">
1516 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
1517 <td class="left">Vector of 2 64-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1518 </tr>
1519</table>
1520</div>
1521
1522<!-- _______________________________________________________________________ -->
1523<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1524<div class="doc_text">
1525
1526<h5>Overview:</h5>
1527
1528<p>Opaque types are used to represent unknown types in the system. This
Gordon Henriksenda0706e2007-10-14 00:34:53 +0000[diff] [blame]1529corresponds (for example) to the C notion of a forward declared structure type.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1530In LLVM, opaque types can eventually be resolved to any type (not just a
1531structure type).</p>
1532
1533<h5>Syntax:</h5>
1534
1535<pre>
1536 opaque
1537</pre>
1538
1539<h5>Examples:</h5>
1540
1541<table class="layout">
1542 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1543 <td class="left"><tt>opaque</tt></td>
1544 <td class="left">An opaque type.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1545 </tr>
1546</table>
1547</div>
1548
1549
1550<!-- *********************************************************************** -->
1551<div class="doc_section"> <a name="constants">Constants</a> </div>
1552<!-- *********************************************************************** -->
1553
1554<div class="doc_text">
1555
1556<p>LLVM has several different basic types of constants. This section describes
1557them all and their syntax.</p>
1558
1559</div>
1560
1561<!-- ======================================================================= -->
1562<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
1563
1564<div class="doc_text">
1565
1566<dl>
1567 <dt><b>Boolean constants</b></dt>
1568
1569 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
1570 constants of the <tt><a href="#t_primitive">i1</a></tt> type.
1571 </dd>
1572
1573 <dt><b>Integer constants</b></dt>
1574
1575 <dd>Standard integers (such as '4') are constants of the <a
1576 href="#t_integer">integer</a> type. Negative numbers may be used with
1577 integer types.
1578 </dd>
1579
1580 <dt><b>Floating point constants</b></dt>
1581
1582 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
1583 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]1584 notation (see below). The assembler requires the exact decimal value of
1585 a floating-point constant. For example, the assembler accepts 1.25 but
1586 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
1587 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1588
1589 <dt><b>Null pointer constants</b></dt>
1590
1591 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
1592 and must be of <a href="#t_pointer">pointer type</a>.</dd>
1593
1594</dl>
1595
1596<p>The one non-intuitive notation for constants is the optional hexadecimal form
1597of floating point constants. For example, the form '<tt>double
15980x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
15994.5e+15</tt>'. The only time hexadecimal floating point constants are required
1600(and the only time that they are generated by the disassembler) is when a
1601floating point constant must be emitted but it cannot be represented as a
1602decimal floating point number. For example, NaN's, infinities, and other
1603special values are represented in their IEEE hexadecimal format so that
1604assembly and disassembly do not cause any bits to change in the constants.</p>
1605
1606</div>
1607
1608<!-- ======================================================================= -->
1609<div class="doc_subsection"><a name="aggregateconstants">Aggregate Constants</a>
1610</div>
1611
1612<div class="doc_text">
1613<p>Aggregate constants arise from aggregation of simple constants
1614and smaller aggregate constants.</p>
1615
1616<dl>
1617 <dt><b>Structure constants</b></dt>
1618
1619 <dd>Structure constants are represented with notation similar to structure
1620 type definitions (a comma separated list of elements, surrounded by braces
Chris Lattner6c8de962007-12-25 20:34:52 +0000[diff] [blame]1621 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
1622 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]1623 must have <a href="#t_struct">structure type</a>, and the number and
1624 types of elements must match those specified by the type.
1625 </dd>
1626
1627 <dt><b>Array constants</b></dt>
1628
1629 <dd>Array constants are represented with notation similar to array type
1630 definitions (a comma separated list of elements, surrounded by square brackets
1631 (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74 ]</tt>". Array
1632 constants must have <a href="#t_array">array type</a>, and the number and
1633 types of elements must match those specified by the type.
1634 </dd>
1635
1636 <dt><b>Vector constants</b></dt>
1637
1638 <dd>Vector constants are represented with notation similar to vector type
1639 definitions (a comma separated list of elements, surrounded by
1640 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32 42,
1641 i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must have <a
1642 href="#t_vector">vector type</a>, and the number and types of elements must
1643 match those specified by the type.
1644 </dd>
1645
1646 <dt><b>Zero initialization</b></dt>
1647
1648 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1649 value to zero of <em>any</em> type, including scalar and aggregate types.
1650 This is often used to avoid having to print large zero initializers (e.g. for
1651 large arrays) and is always exactly equivalent to using explicit zero
1652 initializers.
1653 </dd>
1654</dl>
1655
1656</div>
1657
1658<!-- ======================================================================= -->
1659<div class="doc_subsection">
1660 <a name="globalconstants">Global Variable and Function Addresses</a>
1661</div>
1662
1663<div class="doc_text">
1664
1665<p>The addresses of <a href="#globalvars">global variables</a> and <a
1666href="#functionstructure">functions</a> are always implicitly valid (link-time)
1667constants. These constants are explicitly referenced when the <a
1668href="#identifiers">identifier for the global</a> is used and always have <a
1669href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1670file:</p>
1671
1672<div class="doc_code">
1673<pre>
1674@X = global i32 17
1675@Y = global i32 42
1676@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
1677</pre>
1678</div>
1679
1680</div>
1681
1682<!-- ======================================================================= -->
1683<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
1684<div class="doc_text">
1685 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
1686 no specific value. Undefined values may be of any type and be used anywhere
1687 a constant is permitted.</p>
1688
1689 <p>Undefined values indicate to the compiler that the program is well defined
1690 no matter what value is used, giving the compiler more freedom to optimize.
1691 </p>
1692</div>
1693
1694<!-- ======================================================================= -->
1695<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1696</div>
1697
1698<div class="doc_text">
1699
1700<p>Constant expressions are used to allow expressions involving other constants
1701to be used as constants. Constant expressions may be of any <a
1702href="#t_firstclass">first class</a> type and may involve any LLVM operation
1703that does not have side effects (e.g. load and call are not supported). The
1704following is the syntax for constant expressions:</p>
1705
1706<dl>
1707 <dt><b><tt>trunc ( CST to TYPE )</tt></b></dt>
1708 <dd>Truncate a constant to another type. The bit size of CST must be larger
1709 than the bit size of TYPE. Both types must be integers.</dd>
1710
1711 <dt><b><tt>zext ( CST to TYPE )</tt></b></dt>
1712 <dd>Zero extend a constant to another type. The bit size of CST must be
1713 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1714
1715 <dt><b><tt>sext ( CST to TYPE )</tt></b></dt>
1716 <dd>Sign extend a constant to another type. The bit size of CST must be
1717 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1718
1719 <dt><b><tt>fptrunc ( CST to TYPE )</tt></b></dt>
1720 <dd>Truncate a floating point constant to another floating point type. The
1721 size of CST must be larger than the size of TYPE. Both types must be
1722 floating point.</dd>
1723
1724 <dt><b><tt>fpext ( CST to TYPE )</tt></b></dt>
1725 <dd>Floating point extend a constant to another type. The size of CST must be
1726 smaller or equal to the size of TYPE. Both types must be floating point.</dd>
1727
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]1728 <dt><b><tt>fptoui ( CST to TYPE )</tt></b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1729 <dd>Convert a floating point constant to the corresponding unsigned integer
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1730 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1731 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1732 of the same number of elements. If the value won't fit in the integer type,
1733 the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1734
1735 <dt><b><tt>fptosi ( CST to TYPE )</tt></b></dt>
1736 <dd>Convert a floating point constant to the corresponding signed integer
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1737 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1738 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1739 of the same number of elements. If the value won't fit in the integer type,
1740 the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1741
1742 <dt><b><tt>uitofp ( CST to TYPE )</tt></b></dt>
1743 <dd>Convert an unsigned integer constant to the corresponding floating point
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1744 constant. TYPE must be a scalar or vector floating point type. CST must be of
1745 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1746 of the same number of elements. If the value won't fit in the floating point
1747 type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1748
1749 <dt><b><tt>sitofp ( CST to TYPE )</tt></b></dt>
1750 <dd>Convert a signed integer constant to the corresponding floating point
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1751 constant. TYPE must be a scalar or vector floating point type. CST must be of
1752 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1753 of the same number of elements. If the value won't fit in the floating point
1754 type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1755
1756 <dt><b><tt>ptrtoint ( CST to TYPE )</tt></b></dt>
1757 <dd>Convert a pointer typed constant to the corresponding integer constant
1758 TYPE must be an integer type. CST must be of pointer type. The CST value is
1759 zero extended, truncated, or unchanged to make it fit in TYPE.</dd>
1760
1761 <dt><b><tt>inttoptr ( CST to TYPE )</tt></b></dt>
1762 <dd>Convert a integer constant to a pointer constant. TYPE must be a
1763 pointer type. CST must be of integer type. The CST value is zero extended,
1764 truncated, or unchanged to make it fit in a pointer size. This one is
1765 <i>really</i> dangerous!</dd>
1766
1767 <dt><b><tt>bitcast ( CST to TYPE )</tt></b></dt>
1768 <dd>Convert a constant, CST, to another TYPE. The size of CST and TYPE must be
1769 identical (same number of bits). The conversion is done as if the CST value
1770 was stored to memory and read back as TYPE. In other words, no bits change
1771 with this operator, just the type. This can be used for conversion of
1772 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]1773 pointers it is only valid to cast to another pointer type. It is not valid
1774 to bitcast to or from an aggregate type.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1775 </dd>
1776
1777 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
1778
1779 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
1780 constants. As with the <a href="#i_getelementptr">getelementptr</a>
1781 instruction, the index list may have zero or more indexes, which are required
1782 to make sense for the type of "CSTPTR".</dd>
1783
1784 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
1785
1786 <dd>Perform the <a href="#i_select">select operation</a> on
1787 constants.</dd>
1788
1789 <dt><b><tt>icmp COND ( VAL1, VAL2 )</tt></b></dt>
1790 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
1791
1792 <dt><b><tt>fcmp COND ( VAL1, VAL2 )</tt></b></dt>
1793 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
1794
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]1795 <dt><b><tt>vicmp COND ( VAL1, VAL2 )</tt></b></dt>
1796 <dd>Performs the <a href="#i_vicmp">vicmp operation</a> on constants.</dd>
1797
1798 <dt><b><tt>vfcmp COND ( VAL1, VAL2 )</tt></b></dt>
1799 <dd>Performs the <a href="#i_vfcmp">vfcmp operation</a> on constants.</dd>
1800
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1801 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
1802
1803 <dd>Perform the <a href="#i_extractelement">extractelement
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]1804 operation</a> on constants.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1805
1806 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
1807
1808 <dd>Perform the <a href="#i_insertelement">insertelement
1809 operation</a> on constants.</dd>
1810
1811
1812 <dt><b><tt>shufflevector ( VEC1, VEC2, IDXMASK )</tt></b></dt>
1813
1814 <dd>Perform the <a href="#i_shufflevector">shufflevector
1815 operation</a> on constants.</dd>
1816
1817 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
1818
1819 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
1820 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
1821 binary</a> operations. The constraints on operands are the same as those for
1822 the corresponding instruction (e.g. no bitwise operations on floating point
1823 values are allowed).</dd>
1824</dl>
1825</div>
1826
1827<!-- *********************************************************************** -->
1828<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
1829<!-- *********************************************************************** -->
1830
1831<!-- ======================================================================= -->
1832<div class="doc_subsection">
1833<a name="inlineasm">Inline Assembler Expressions</a>
1834</div>
1835
1836<div class="doc_text">
1837
1838<p>
1839LLVM supports inline assembler expressions (as opposed to <a href="#moduleasm">
1840Module-Level Inline Assembly</a>) through the use of a special value. This
1841value represents the inline assembler as a string (containing the instructions
1842to emit), a list of operand constraints (stored as a string), and a flag that
1843indicates whether or not the inline asm expression has side effects. An example
1844inline assembler expression is:
1845</p>
1846
1847<div class="doc_code">
1848<pre>
1849i32 (i32) asm "bswap $0", "=r,r"
1850</pre>
1851</div>
1852
1853<p>
1854Inline assembler expressions may <b>only</b> be used as the callee operand of
1855a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we have:
1856</p>
1857
1858<div class="doc_code">
1859<pre>
1860%X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
1861</pre>
1862</div>
1863
1864<p>
1865Inline asms with side effects not visible in the constraint list must be marked
1866as having side effects. This is done through the use of the
1867'<tt>sideeffect</tt>' keyword, like so:
1868</p>
1869
1870<div class="doc_code">
1871<pre>
1872call void asm sideeffect "eieio", ""()
1873</pre>
1874</div>
1875
1876<p>TODO: The format of the asm and constraints string still need to be
1877documented here. Constraints on what can be done (e.g. duplication, moving, etc
Chris Lattner1cdd64e2008-10-04 18:36:02 +0000[diff] [blame]1878need to be documented). This is probably best done by reference to another
1879document that covers inline asm from a holistic perspective.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1880</p>
1881
1882</div>
1883
1884<!-- *********************************************************************** -->
1885<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
1886<!-- *********************************************************************** -->
1887
1888<div class="doc_text">
1889
1890<p>The LLVM instruction set consists of several different
1891classifications of instructions: <a href="#terminators">terminator
1892instructions</a>, <a href="#binaryops">binary instructions</a>,
1893<a href="#bitwiseops">bitwise binary instructions</a>, <a
1894 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
1895instructions</a>.</p>
1896
1897</div>
1898
1899<!-- ======================================================================= -->
1900<div class="doc_subsection"> <a name="terminators">Terminator
1901Instructions</a> </div>
1902
1903<div class="doc_text">
1904
1905<p>As mentioned <a href="#functionstructure">previously</a>, every
1906basic block in a program ends with a "Terminator" instruction, which
1907indicates which block should be executed after the current block is
1908finished. These terminator instructions typically yield a '<tt>void</tt>'
1909value: they produce control flow, not values (the one exception being
1910the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
1911<p>There are six different terminator instructions: the '<a
1912 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
1913instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
1914the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
1915 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
1916 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
1917
1918</div>
1919
1920<!-- _______________________________________________________________________ -->
1921<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
1922Instruction</a> </div>
1923<div class="doc_text">
1924<h5>Syntax:</h5>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]1925<pre>
1926 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]1927 ret void <i>; Return from void function</i>
1928</pre>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1929
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1930<h5>Overview:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1931
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]1932<p>The '<tt>ret</tt>' instruction is used to return control flow (and
1933optionally a value) from a function back to the caller.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1934<p>There are two forms of the '<tt>ret</tt>' instruction: one that
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]1935returns a value and then causes control flow, and one that just causes
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1936control flow to occur.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1937
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1938<h5>Arguments:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1939
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]1940<p>The '<tt>ret</tt>' instruction optionally accepts a single argument,
1941the return value. The type of the return value must be a
1942'<a href="#t_firstclass">first class</a>' type.</p>
1943
1944<p>A function is not <a href="#wellformed">well formed</a> if
1945it it has a non-void return type and contains a '<tt>ret</tt>'
1946instruction with no return value or a return value with a type that
1947does not match its type, or if it has a void return type and contains
1948a '<tt>ret</tt>' instruction with a return value.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1949
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1950<h5>Semantics:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1951
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1952<p>When the '<tt>ret</tt>' instruction is executed, control flow
1953returns back to the calling function's context. If the caller is a "<a
1954 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
1955the instruction after the call. If the caller was an "<a
1956 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
1957at the beginning of the "normal" destination block. If the instruction
1958returns a value, that value shall set the call or invoke instruction's
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]1959return value.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1960
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1961<h5>Example:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1962
1963<pre>
1964 ret i32 5 <i>; Return an integer value of 5</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1965 ret void <i>; Return from a void function</i>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]1966 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]1967</pre>
1968</div>
1969<!-- _______________________________________________________________________ -->
1970<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
1971<div class="doc_text">
1972<h5>Syntax:</h5>
1973<pre> br i1 &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;<br> br label &lt;dest&gt; <i>; Unconditional branch</i>
1974</pre>
1975<h5>Overview:</h5>
1976<p>The '<tt>br</tt>' instruction is used to cause control flow to
1977transfer to a different basic block in the current function. There are
1978two forms of this instruction, corresponding to a conditional branch
1979and an unconditional branch.</p>
1980<h5>Arguments:</h5>
1981<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
1982single '<tt>i1</tt>' value and two '<tt>label</tt>' values. The
1983unconditional form of the '<tt>br</tt>' instruction takes a single
1984'<tt>label</tt>' value as a target.</p>
1985<h5>Semantics:</h5>
1986<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
1987argument is evaluated. If the value is <tt>true</tt>, control flows
1988to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
1989control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
1990<h5>Example:</h5>
1991<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
1992 href="#i_ret">ret</a> i32 1<br>IfUnequal:<br> <a href="#i_ret">ret</a> i32 0<br></pre>
1993</div>
1994<!-- _______________________________________________________________________ -->
1995<div class="doc_subsubsection">
1996 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
1997</div>
1998
1999<div class="doc_text">
2000<h5>Syntax:</h5>
2001
2002<pre>
2003 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
2004</pre>
2005
2006<h5>Overview:</h5>
2007
2008<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
2009several different places. It is a generalization of the '<tt>br</tt>'
2010instruction, allowing a branch to occur to one of many possible
2011destinations.</p>
2012
2013
2014<h5>Arguments:</h5>
2015
2016<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
2017comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
2018an array of pairs of comparison value constants and '<tt>label</tt>'s. The
2019table is not allowed to contain duplicate constant entries.</p>
2020
2021<h5>Semantics:</h5>
2022
2023<p>The <tt>switch</tt> instruction specifies a table of values and
2024destinations. When the '<tt>switch</tt>' instruction is executed, this
2025table is searched for the given value. If the value is found, control flow is
2026transfered to the corresponding destination; otherwise, control flow is
2027transfered to the default destination.</p>
2028
2029<h5>Implementation:</h5>
2030
2031<p>Depending on properties of the target machine and the particular
2032<tt>switch</tt> instruction, this instruction may be code generated in different
2033ways. For example, it could be generated as a series of chained conditional
2034branches or with a lookup table.</p>
2035
2036<h5>Example:</h5>
2037
2038<pre>
2039 <i>; Emulate a conditional br instruction</i>
2040 %Val = <a href="#i_zext">zext</a> i1 %value to i32
2041 switch i32 %Val, label %truedest [i32 0, label %falsedest ]
2042
2043 <i>; Emulate an unconditional br instruction</i>
2044 switch i32 0, label %dest [ ]
2045
2046 <i>; Implement a jump table:</i>
2047 switch i32 %val, label %otherwise [ i32 0, label %onzero
2048 i32 1, label %onone
2049 i32 2, label %ontwo ]
2050</pre>
2051</div>
2052
2053<!-- _______________________________________________________________________ -->
2054<div class="doc_subsubsection">
2055 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
2056</div>
2057
2058<div class="doc_text">
2059
2060<h5>Syntax:</h5>
2061
2062<pre>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]2063 &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]2064 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
2065</pre>
2066
2067<h5>Overview:</h5>
2068
2069<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
2070function, with the possibility of control flow transfer to either the
2071'<tt>normal</tt>' label or the
2072'<tt>exception</tt>' label. If the callee function returns with the
2073"<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
2074"normal" label. If the callee (or any indirect callees) returns with the "<a
2075href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]2076continued at the dynamically nearest "exception" label.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2077
2078<h5>Arguments:</h5>
2079
2080<p>This instruction requires several arguments:</p>
2081
2082<ol>
2083 <li>
2084 The optional "cconv" marker indicates which <a href="#callingconv">calling
2085 convention</a> the call should use. If none is specified, the call defaults
2086 to using C calling conventions.
2087 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]2088
2089 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
2090 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>',
2091 and '<tt>inreg</tt>' attributes are valid here.</li>
2092
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2093 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
2094 function value being invoked. In most cases, this is a direct function
2095 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
2096 an arbitrary pointer to function value.
2097 </li>
2098
2099 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
2100 function to be invoked. </li>
2101
2102 <li>'<tt>function args</tt>': argument list whose types match the function
2103 signature argument types. If the function signature indicates the function
2104 accepts a variable number of arguments, the extra arguments can be
2105 specified. </li>
2106
2107 <li>'<tt>normal label</tt>': the label reached when the called function
2108 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
2109
2110 <li>'<tt>exception label</tt>': the label reached when a callee returns with
2111 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
2112
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]2113 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]2114 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
2115 '<tt>readnone</tt>' attributes are valid here.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2116</ol>
2117
2118<h5>Semantics:</h5>
2119
2120<p>This instruction is designed to operate as a standard '<tt><a
2121href="#i_call">call</a></tt>' instruction in most regards. The primary
2122difference is that it establishes an association with a label, which is used by
2123the runtime library to unwind the stack.</p>
2124
2125<p>This instruction is used in languages with destructors to ensure that proper
2126cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
2127exception. Additionally, this is important for implementation of
2128'<tt>catch</tt>' clauses in high-level languages that support them.</p>
2129
2130<h5>Example:</h5>
2131<pre>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2132 %retval = invoke i32 @Test(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2133 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2134 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2135 unwind label %TestCleanup <i>; {i32}:retval set</i>
2136</pre>
2137</div>
2138
2139
2140<!-- _______________________________________________________________________ -->
2141
2142<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
2143Instruction</a> </div>
2144
2145<div class="doc_text">
2146
2147<h5>Syntax:</h5>
2148<pre>
2149 unwind
2150</pre>
2151
2152<h5>Overview:</h5>
2153
2154<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
2155at the first callee in the dynamic call stack which used an <a
2156href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
2157primarily used to implement exception handling.</p>
2158
2159<h5>Semantics:</h5>
2160
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]2161<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2162immediately halt. The dynamic call stack is then searched for the first <a
2163href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
2164execution continues at the "exceptional" destination block specified by the
2165<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
2166dynamic call chain, undefined behavior results.</p>
2167</div>
2168
2169<!-- _______________________________________________________________________ -->
2170
2171<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
2172Instruction</a> </div>
2173
2174<div class="doc_text">
2175
2176<h5>Syntax:</h5>
2177<pre>
2178 unreachable
2179</pre>
2180
2181<h5>Overview:</h5>
2182
2183<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
2184instruction is used to inform the optimizer that a particular portion of the
2185code is not reachable. This can be used to indicate that the code after a
2186no-return function cannot be reached, and other facts.</p>
2187
2188<h5>Semantics:</h5>
2189
2190<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
2191</div>
2192
2193
2194
2195<!-- ======================================================================= -->
2196<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
2197<div class="doc_text">
2198<p>Binary operators are used to do most of the computation in a
Chris Lattnerab596d92008-04-01 18:47:32 +0000[diff] [blame]2199program. They require two operands of the same type, execute an operation on them, and
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2200produce a single value. The operands might represent
2201multiple 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]2202The result value has the same type as its operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2203<p>There are several different binary operators:</p>
2204</div>
2205<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2206<div class="doc_subsubsection">
2207 <a name="i_add">'<tt>add</tt>' Instruction</a>
2208</div>
2209
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2210<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2211
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2212<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2213
2214<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2215 &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]2216</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2217
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2218<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2219
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2220<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2221
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2222<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2223
2224<p>The two arguments to the '<tt>add</tt>' instruction must be <a
2225 href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>, or
2226 <a href="#t_vector">vector</a> values. Both arguments must have identical
2227 types.</p>
2228
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2229<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2230
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2231<p>The value produced is the integer or floating point sum of the two
2232operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2233
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2234<p>If an integer sum has unsigned overflow, the result returned is the
2235mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2236the result.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2237
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2238<p>Because LLVM integers use a two's complement representation, this
2239instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2240
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2241<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2242
2243<pre>
2244 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2245</pre>
2246</div>
2247<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2248<div class="doc_subsubsection">
2249 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
2250</div>
2251
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2252<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2253
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2254<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2255
2256<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2257 &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]2258</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2259
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2260<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2261
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2262<p>The '<tt>sub</tt>' instruction returns the difference of its two
2263operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2264
2265<p>Note that the '<tt>sub</tt>' instruction is used to represent the
2266'<tt>neg</tt>' instruction present in most other intermediate
2267representations.</p>
2268
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2269<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2270
2271<p>The two arguments to the '<tt>sub</tt>' instruction must be <a
2272 href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>,
2273 or <a href="#t_vector">vector</a> values. Both arguments must have identical
2274 types.</p>
2275
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2276<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2277
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2278<p>The value produced is the integer or floating point difference of
2279the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2280
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2281<p>If an integer difference has unsigned overflow, the result returned is the
2282mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2283the result.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2284
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2285<p>Because LLVM integers use a two's complement representation, this
2286instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2287
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2288<h5>Example:</h5>
2289<pre>
2290 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
2291 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
2292</pre>
2293</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2294
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2295<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2296<div class="doc_subsubsection">
2297 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
2298</div>
2299
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2300<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2301
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2302<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2303<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]2304</pre>
2305<h5>Overview:</h5>
2306<p>The '<tt>mul</tt>' instruction returns the product of its two
2307operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2308
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2309<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2310
2311<p>The two arguments to the '<tt>mul</tt>' instruction must be <a
2312href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>,
2313or <a href="#t_vector">vector</a> values. Both arguments must have identical
2314types.</p>
2315
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2316<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2317
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2318<p>The value produced is the integer or floating point product of the
2319two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2320
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2321<p>If the result of an integer multiplication has unsigned overflow,
2322the result returned is the mathematical result modulo
23232<sup>n</sup>, where n is the bit width of the result.</p>
2324<p>Because LLVM integers use a two's complement representation, and the
2325result is the same width as the operands, this instruction returns the
2326correct result for both signed and unsigned integers. If a full product
2327(e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands
2328should be sign-extended or zero-extended as appropriate to the
2329width of the full product.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2330<h5>Example:</h5>
2331<pre> &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
2332</pre>
2333</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2334
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2335<!-- _______________________________________________________________________ -->
2336<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
2337</a></div>
2338<div class="doc_text">
2339<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2340<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]2341</pre>
2342<h5>Overview:</h5>
2343<p>The '<tt>udiv</tt>' instruction returns the quotient of its two
2344operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2345
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2346<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2347
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2348<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2349<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2350values. Both arguments must have identical types.</p>
2351
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2352<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2353
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2354<p>The value produced is the unsigned integer quotient of the two operands.</p>
2355<p>Note that unsigned integer division and signed integer division are distinct
2356operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
2357<p>Division by zero leads to undefined behavior.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2358<h5>Example:</h5>
2359<pre> &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2360</pre>
2361</div>
2362<!-- _______________________________________________________________________ -->
2363<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
2364</a> </div>
2365<div class="doc_text">
2366<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2367<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2368 &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]2369</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2370
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2371<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2372
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2373<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two
2374operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2375
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2376<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2377
2378<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
2379<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2380values. Both arguments must have identical types.</p>
2381
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2382<h5>Semantics:</h5>
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2383<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]2384<p>Note that signed integer division and unsigned integer division are distinct
2385operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
2386<p>Division by zero leads to undefined behavior. Overflow also leads to
2387undefined behavior; this is a rare case, but can occur, for example,
2388by doing a 32-bit division of -2147483648 by -1.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2389<h5>Example:</h5>
2390<pre> &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2391</pre>
2392</div>
2393<!-- _______________________________________________________________________ -->
2394<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
2395Instruction</a> </div>
2396<div class="doc_text">
2397<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2398<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2399 &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]2400</pre>
2401<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2402
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2403<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two
2404operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2405
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2406<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2407
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2408<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2409<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2410of floating point values. Both arguments must have identical types.</p>
2411
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2412<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2413
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2414<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2415
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2416<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2417
2418<pre>
2419 &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]2420</pre>
2421</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2422
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2423<!-- _______________________________________________________________________ -->
2424<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
2425</div>
2426<div class="doc_text">
2427<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2428<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]2429</pre>
2430<h5>Overview:</h5>
2431<p>The '<tt>urem</tt>' instruction returns the remainder from the
2432unsigned division of its two arguments.</p>
2433<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2434<p>The two arguments to the '<tt>urem</tt>' instruction must be
2435<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2436values. Both arguments must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2437<h5>Semantics:</h5>
2438<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2439This instruction always performs an unsigned division to get the remainder.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2440<p>Note that unsigned integer remainder and signed integer remainder are
2441distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
2442<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2443<h5>Example:</h5>
2444<pre> &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2445</pre>
2446
2447</div>
2448<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2449<div class="doc_subsubsection">
2450 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
2451</div>
2452
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2453<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2454
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2455<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2456
2457<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2458 &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]2459</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2460
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2461<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2462
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2463<p>The '<tt>srem</tt>' instruction returns the remainder from the
Dan Gohman3e3fd8c2007-11-05 23:35:22 +0000[diff] [blame]2464signed division of its two operands. This instruction can also take
2465<a href="#t_vector">vector</a> versions of the values in which case
2466the elements must be integers.</p>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]2467
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2468<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2469
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2470<p>The two arguments to the '<tt>srem</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2471<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2472values. Both arguments must have identical types.</p>
2473
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2474<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2475
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2476<p>This instruction returns the <i>remainder</i> of a division (where the result
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2477has the same sign as the dividend, <tt>op1</tt>), not the <i>modulo</i>
2478operator (where the result has the same sign as the divisor, <tt>op2</tt>) of
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2479a value. For more information about the difference, see <a
2480 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
2481Math Forum</a>. For a table of how this is implemented in various languages,
2482please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
2483Wikipedia: modulo operation</a>.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2484<p>Note that signed integer remainder and unsigned integer remainder are
2485distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
2486<p>Taking the remainder of a division by zero leads to undefined behavior.
2487Overflow also leads to undefined behavior; this is a rare case, but can occur,
2488for example, by taking the remainder of a 32-bit division of -2147483648 by -1.
2489(The remainder doesn't actually overflow, but this rule lets srem be
2490implemented using instructions that return both the result of the division
2491and the remainder.)</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2492<h5>Example:</h5>
2493<pre> &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2494</pre>
2495
2496</div>
2497<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2498<div class="doc_subsubsection">
2499 <a name="i_frem">'<tt>frem</tt>' Instruction</a> </div>
2500
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2501<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2502
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2503<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2504<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]2505</pre>
2506<h5>Overview:</h5>
2507<p>The '<tt>frem</tt>' instruction returns the remainder from the
2508division of its two operands.</p>
2509<h5>Arguments:</h5>
2510<p>The two arguments to the '<tt>frem</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2511<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2512of floating point values. Both arguments must have identical types.</p>
2513
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2514<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2515
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2516<p>This instruction returns the <i>remainder</i> of a division.
2517The remainder has the same sign as the dividend.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2518
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2519<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2520
2521<pre>
2522 &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]2523</pre>
2524</div>
2525
2526<!-- ======================================================================= -->
2527<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
2528Operations</a> </div>
2529<div class="doc_text">
2530<p>Bitwise binary operators are used to do various forms of
2531bit-twiddling in a program. They are generally very efficient
2532instructions and can commonly be strength reduced from other
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2533instructions. They require two operands of the same type, execute an operation on them,
2534and 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]2535</div>
2536
2537<!-- _______________________________________________________________________ -->
2538<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
2539Instruction</a> </div>
2540<div class="doc_text">
2541<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2542<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]2543</pre>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2544
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2545<h5>Overview:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2546
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2547<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
2548the left a specified number of bits.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2549
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2550<h5>Arguments:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2551
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2552<p>Both arguments to the '<tt>shl</tt>' instruction must be the same <a
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]2553 href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2554type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2555
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2556<h5>Semantics:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2557
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2558<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod 2<sup>n</sup>,
2559where n is the width of the result. If <tt>op2</tt> is (statically or dynamically) negative or
2560equal 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]2561
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2562<h5>Example:</h5><pre>
2563 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
2564 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
2565 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2566 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2567</pre>
2568</div>
2569<!-- _______________________________________________________________________ -->
2570<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
2571Instruction</a> </div>
2572<div class="doc_text">
2573<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2574<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]2575</pre>
2576
2577<h5>Overview:</h5>
2578<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
2579operand shifted to the right a specified number of bits with zero fill.</p>
2580
2581<h5>Arguments:</h5>
2582<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]2583<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]2584type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2585
2586<h5>Semantics:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2587
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2588<p>This instruction always performs a logical shift right operation. The most
2589significant bits of the result will be filled with zero bits after the
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2590shift. If <tt>op2</tt> is (statically or dynamically) equal to or larger than
2591the number of bits in <tt>op1</tt>, the result is undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2592
2593<h5>Example:</h5>
2594<pre>
2595 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
2596 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
2597 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
2598 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2599 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2600</pre>
2601</div>
2602
2603<!-- _______________________________________________________________________ -->
2604<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
2605Instruction</a> </div>
2606<div class="doc_text">
2607
2608<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2609<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]2610</pre>
2611
2612<h5>Overview:</h5>
2613<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
2614operand shifted to the right a specified number of bits with sign extension.</p>
2615
2616<h5>Arguments:</h5>
2617<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]2618<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]2619type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2620
2621<h5>Semantics:</h5>
2622<p>This instruction always performs an arithmetic shift right operation,
2623The most significant bits of the result will be filled with the sign bit
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2624of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
2625larger than the number of bits in <tt>op1</tt>, the result is undefined.
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2626</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2627
2628<h5>Example:</h5>
2629<pre>
2630 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
2631 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
2632 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
2633 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2634 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2635</pre>
2636</div>
2637
2638<!-- _______________________________________________________________________ -->
2639<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
2640Instruction</a> </div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2641
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2642<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2643
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2644<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2645
2646<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2647 &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]2648</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2649
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2650<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2651
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2652<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
2653its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2654
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2655<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2656
2657<p>The two arguments to the '<tt>and</tt>' instruction must be
2658<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2659values. Both arguments must have identical types.</p>
2660
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2661<h5>Semantics:</h5>
2662<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
2663<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]2664<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2665<table border="1" cellspacing="0" cellpadding="4">
2666 <tbody>
2667 <tr>
2668 <td>In0</td>
2669 <td>In1</td>
2670 <td>Out</td>
2671 </tr>
2672 <tr>
2673 <td>0</td>
2674 <td>0</td>
2675 <td>0</td>
2676 </tr>
2677 <tr>
2678 <td>0</td>
2679 <td>1</td>
2680 <td>0</td>
2681 </tr>
2682 <tr>
2683 <td>1</td>
2684 <td>0</td>
2685 <td>0</td>
2686 </tr>
2687 <tr>
2688 <td>1</td>
2689 <td>1</td>
2690 <td>1</td>
2691 </tr>
2692 </tbody>
2693</table>
2694</div>
2695<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2696<pre>
2697 &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]2698 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
2699 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
2700</pre>
2701</div>
2702<!-- _______________________________________________________________________ -->
2703<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
2704<div class="doc_text">
2705<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2706<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]2707</pre>
2708<h5>Overview:</h5>
2709<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
2710or of its two operands.</p>
2711<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2712
2713<p>The two arguments to the '<tt>or</tt>' instruction must be
2714<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2715values. Both arguments must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2716<h5>Semantics:</h5>
2717<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
2718<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]2719<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2720<table border="1" cellspacing="0" cellpadding="4">
2721 <tbody>
2722 <tr>
2723 <td>In0</td>
2724 <td>In1</td>
2725 <td>Out</td>
2726 </tr>
2727 <tr>
2728 <td>0</td>
2729 <td>0</td>
2730 <td>0</td>
2731 </tr>
2732 <tr>
2733 <td>0</td>
2734 <td>1</td>
2735 <td>1</td>
2736 </tr>
2737 <tr>
2738 <td>1</td>
2739 <td>0</td>
2740 <td>1</td>
2741 </tr>
2742 <tr>
2743 <td>1</td>
2744 <td>1</td>
2745 <td>1</td>
2746 </tr>
2747 </tbody>
2748</table>
2749</div>
2750<h5>Example:</h5>
2751<pre> &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
2752 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
2753 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
2754</pre>
2755</div>
2756<!-- _______________________________________________________________________ -->
2757<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
2758Instruction</a> </div>
2759<div class="doc_text">
2760<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2761<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]2762</pre>
2763<h5>Overview:</h5>
2764<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
2765or of its two operands. The <tt>xor</tt> is used to implement the
2766"one's complement" operation, which is the "~" operator in C.</p>
2767<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2768<p>The two arguments to the '<tt>xor</tt>' instruction must be
2769<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2770values. Both arguments must have identical types.</p>
2771
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2772<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2773
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2774<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
2775<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]2776<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2777<table border="1" cellspacing="0" cellpadding="4">
2778 <tbody>
2779 <tr>
2780 <td>In0</td>
2781 <td>In1</td>
2782 <td>Out</td>
2783 </tr>
2784 <tr>
2785 <td>0</td>
2786 <td>0</td>
2787 <td>0</td>
2788 </tr>
2789 <tr>
2790 <td>0</td>
2791 <td>1</td>
2792 <td>1</td>
2793 </tr>
2794 <tr>
2795 <td>1</td>
2796 <td>0</td>
2797 <td>1</td>
2798 </tr>
2799 <tr>
2800 <td>1</td>
2801 <td>1</td>
2802 <td>0</td>
2803 </tr>
2804 </tbody>
2805</table>
2806</div>
2807<p> </p>
2808<h5>Example:</h5>
2809<pre> &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
2810 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
2811 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
2812 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
2813</pre>
2814</div>
2815
2816<!-- ======================================================================= -->
2817<div class="doc_subsection">
2818 <a name="vectorops">Vector Operations</a>
2819</div>
2820
2821<div class="doc_text">
2822
2823<p>LLVM supports several instructions to represent vector operations in a
2824target-independent manner. These instructions cover the element-access and
2825vector-specific operations needed to process vectors effectively. While LLVM
2826does directly support these vector operations, many sophisticated algorithms
2827will want to use target-specific intrinsics to take full advantage of a specific
2828target.</p>
2829
2830</div>
2831
2832<!-- _______________________________________________________________________ -->
2833<div class="doc_subsubsection">
2834 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
2835</div>
2836
2837<div class="doc_text">
2838
2839<h5>Syntax:</h5>
2840
2841<pre>
2842 &lt;result&gt; = extractelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, i32 &lt;idx&gt; <i>; yields &lt;ty&gt;</i>
2843</pre>
2844
2845<h5>Overview:</h5>
2846
2847<p>
2848The '<tt>extractelement</tt>' instruction extracts a single scalar
2849element from a vector at a specified index.
2850</p>
2851
2852
2853<h5>Arguments:</h5>
2854
2855<p>
2856The first operand of an '<tt>extractelement</tt>' instruction is a
2857value of <a href="#t_vector">vector</a> type. The second operand is
2858an index indicating the position from which to extract the element.
2859The index may be a variable.</p>
2860
2861<h5>Semantics:</h5>
2862
2863<p>
2864The result is a scalar of the same type as the element type of
2865<tt>val</tt>. Its value is the value at position <tt>idx</tt> of
2866<tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
2867results are undefined.
2868</p>
2869
2870<h5>Example:</h5>
2871
2872<pre>
2873 %result = extractelement &lt;4 x i32&gt; %vec, i32 0 <i>; yields i32</i>
2874</pre>
2875</div>
2876
2877
2878<!-- _______________________________________________________________________ -->
2879<div class="doc_subsubsection">
2880 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
2881</div>
2882
2883<div class="doc_text">
2884
2885<h5>Syntax:</h5>
2886
2887<pre>
Dan Gohmanbcc3c502008-05-12 23:38:42 +0000[diff] [blame]2888 &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]2889</pre>
2890
2891<h5>Overview:</h5>
2892
2893<p>
2894The '<tt>insertelement</tt>' instruction inserts a scalar
2895element into a vector at a specified index.
2896</p>
2897
2898
2899<h5>Arguments:</h5>
2900
2901<p>
2902The first operand of an '<tt>insertelement</tt>' instruction is a
2903value of <a href="#t_vector">vector</a> type. The second operand is a
2904scalar value whose type must equal the element type of the first
2905operand. The third operand is an index indicating the position at
2906which to insert the value. The index may be a variable.</p>
2907
2908<h5>Semantics:</h5>
2909
2910<p>
2911The result is a vector of the same type as <tt>val</tt>. Its
2912element values are those of <tt>val</tt> except at position
2913<tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
2914exceeds the length of <tt>val</tt>, the results are undefined.
2915</p>
2916
2917<h5>Example:</h5>
2918
2919<pre>
2920 %result = insertelement &lt;4 x i32&gt; %vec, i32 1, i32 0 <i>; yields &lt;4 x i32&gt;</i>
2921</pre>
2922</div>
2923
2924<!-- _______________________________________________________________________ -->
2925<div class="doc_subsubsection">
2926 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
2927</div>
2928
2929<div class="doc_text">
2930
2931<h5>Syntax:</h5>
2932
2933<pre>
2934 &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>
2935</pre>
2936
2937<h5>Overview:</h5>
2938
2939<p>
2940The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
2941from two input vectors, returning a vector of the same type.
2942</p>
2943
2944<h5>Arguments:</h5>
2945
2946<p>
2947The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
2948with types that match each other and types that match the result of the
2949instruction. The third argument is a shuffle mask, which has the same number
2950of elements as the other vector type, but whose element type is always 'i32'.
2951</p>
2952
2953<p>
2954The shuffle mask operand is required to be a constant vector with either
2955constant integer or undef values.
2956</p>
2957
2958<h5>Semantics:</h5>
2959
2960<p>
2961The elements of the two input vectors are numbered from left to right across
2962both of the vectors. The shuffle mask operand specifies, for each element of
2963the result vector, which element of the two input registers the result element
2964gets. The element selector may be undef (meaning "don't care") and the second
2965operand may be undef if performing a shuffle from only one vector.
2966</p>
2967
2968<h5>Example:</h5>
2969
2970<pre>
2971 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
2972 &lt;4 x i32&gt; &lt;i32 0, i32 4, i32 1, i32 5&gt; <i>; yields &lt;4 x i32&gt;</i>
2973 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; undef,
2974 &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.
2975</pre>
2976</div>
2977
2978
2979<!-- ======================================================================= -->
2980<div class="doc_subsection">
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]2981 <a name="aggregateops">Aggregate Operations</a>
2982</div>
2983
2984<div class="doc_text">
2985
2986<p>LLVM supports several instructions for working with aggregate values.
2987</p>
2988
2989</div>
2990
2991<!-- _______________________________________________________________________ -->
2992<div class="doc_subsubsection">
2993 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
2994</div>
2995
2996<div class="doc_text">
2997
2998<h5>Syntax:</h5>
2999
3000<pre>
3001 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
3002</pre>
3003
3004<h5>Overview:</h5>
3005
3006<p>
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3007The '<tt>extractvalue</tt>' instruction extracts the value of a struct field
3008or array element from an aggregate value.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3009</p>
3010
3011
3012<h5>Arguments:</h5>
3013
3014<p>
3015The first operand of an '<tt>extractvalue</tt>' instruction is a
3016value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a>
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3017type. The operands are constant indices to specify which value to extract
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3018in a similar manner as indices in a
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3019'<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
3020</p>
3021
3022<h5>Semantics:</h5>
3023
3024<p>
3025The result is the value at the position in the aggregate specified by
3026the index operands.
3027</p>
3028
3029<h5>Example:</h5>
3030
3031<pre>
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3032 %result = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3033</pre>
3034</div>
3035
3036
3037<!-- _______________________________________________________________________ -->
3038<div class="doc_subsubsection">
3039 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
3040</div>
3041
3042<div class="doc_text">
3043
3044<h5>Syntax:</h5>
3045
3046<pre>
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3047 &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]3048</pre>
3049
3050<h5>Overview:</h5>
3051
3052<p>
3053The '<tt>insertvalue</tt>' instruction inserts a value
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3054into a struct field or array element in an aggregate.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3055</p>
3056
3057
3058<h5>Arguments:</h5>
3059
3060<p>
3061The first operand of an '<tt>insertvalue</tt>' instruction is a
3062value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a> type.
3063The second operand is a first-class value to insert.
Dan Gohman4f29e422008-05-23 21:53:15 +0000[diff] [blame]3064The following operands are constant indices
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3065indicating the position at which to insert the value in a similar manner as
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3066indices in a
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3067'<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
3068The value to insert must have the same type as the value identified
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3069by the indices.
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3070</p>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3071
3072<h5>Semantics:</h5>
3073
3074<p>
3075The result is an aggregate of the same type as <tt>val</tt>. Its
3076value is that of <tt>val</tt> except that the value at the position
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3077specified by the indices is that of <tt>elt</tt>.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3078</p>
3079
3080<h5>Example:</h5>
3081
3082<pre>
Dan Gohmanb1aab4e2008-06-23 15:26:37 +0000[diff] [blame]3083 %result = insertvalue {i32, float} %agg, i32 1, 0 <i>; yields {i32, float}</i>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3084</pre>
3085</div>
3086
3087
3088<!-- ======================================================================= -->
3089<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3090 <a name="memoryops">Memory Access and Addressing Operations</a>
3091</div>
3092
3093<div class="doc_text">
3094
3095<p>A key design point of an SSA-based representation is how it
3096represents memory. In LLVM, no memory locations are in SSA form, which
3097makes things very simple. This section describes how to read, write,
3098allocate, and free memory in LLVM.</p>
3099
3100</div>
3101
3102<!-- _______________________________________________________________________ -->
3103<div class="doc_subsubsection">
3104 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
3105</div>
3106
3107<div class="doc_text">
3108
3109<h5>Syntax:</h5>
3110
3111<pre>
3112 &lt;result&gt; = malloc &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
3113</pre>
3114
3115<h5>Overview:</h5>
3116
3117<p>The '<tt>malloc</tt>' instruction allocates memory from the system
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]3118heap and returns a pointer to it. The object is always allocated in the generic
3119address space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3120
3121<h5>Arguments:</h5>
3122
3123<p>The '<tt>malloc</tt>' instruction allocates
3124<tt>sizeof(&lt;type&gt;)*NumElements</tt>
3125bytes of memory from the operating system and returns a pointer of the
3126appropriate type to the program. If "NumElements" is specified, it is the
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3127number of elements allocated, otherwise "NumElements" is defaulted to be one.
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3128If a constant alignment is specified, the value result of the allocation is guaranteed to
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3129be aligned to at least that boundary. If not specified, or if zero, the target can
3130choose to align the allocation on any convenient boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3131
3132<p>'<tt>type</tt>' must be a sized type.</p>
3133
3134<h5>Semantics:</h5>
3135
3136<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]3137a pointer is returned. The result of a zero byte allocattion is undefined. The
3138result is null if there is insufficient memory available.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3139
3140<h5>Example:</h5>
3141
3142<pre>
3143 %array = malloc [4 x i8 ] <i>; yields {[%4 x i8]*}:array</i>
3144
3145 %size = <a href="#i_add">add</a> i32 2, 2 <i>; yields {i32}:size = i32 4</i>
3146 %array1 = malloc i8, i32 4 <i>; yields {i8*}:array1</i>
3147 %array2 = malloc [12 x i8], i32 %size <i>; yields {[12 x i8]*}:array2</i>
3148 %array3 = malloc i32, i32 4, align 1024 <i>; yields {i32*}:array3</i>
3149 %array4 = malloc i32, align 1024 <i>; yields {i32*}:array4</i>
3150</pre>
3151</div>
3152
3153<!-- _______________________________________________________________________ -->
3154<div class="doc_subsubsection">
3155 <a name="i_free">'<tt>free</tt>' Instruction</a>
3156</div>
3157
3158<div class="doc_text">
3159
3160<h5>Syntax:</h5>
3161
3162<pre>
3163 free &lt;type&gt; &lt;value&gt; <i>; yields {void}</i>
3164</pre>
3165
3166<h5>Overview:</h5>
3167
3168<p>The '<tt>free</tt>' instruction returns memory back to the unused
3169memory heap to be reallocated in the future.</p>
3170
3171<h5>Arguments:</h5>
3172
3173<p>'<tt>value</tt>' shall be a pointer value that points to a value
3174that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
3175instruction.</p>
3176
3177<h5>Semantics:</h5>
3178
3179<p>Access to the memory pointed to by the pointer is no longer defined
Chris Lattner329d6532008-04-19 22:41:32 +0000[diff] [blame]3180after this instruction executes. If the pointer is null, the operation
3181is a noop.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3182
3183<h5>Example:</h5>
3184
3185<pre>
3186 %array = <a href="#i_malloc">malloc</a> [4 x i8] <i>; yields {[4 x i8]*}:array</i>
3187 free [4 x i8]* %array
3188</pre>
3189</div>
3190
3191<!-- _______________________________________________________________________ -->
3192<div class="doc_subsubsection">
3193 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
3194</div>
3195
3196<div class="doc_text">
3197
3198<h5>Syntax:</h5>
3199
3200<pre>
3201 &lt;result&gt; = alloca &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
3202</pre>
3203
3204<h5>Overview:</h5>
3205
3206<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
3207currently executing function, to be automatically released when this function
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]3208returns to its caller. The object is always allocated in the generic address
3209space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3210
3211<h5>Arguments:</h5>
3212
3213<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
3214bytes of memory on the runtime stack, returning a pointer of the
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3215appropriate type to the program. If "NumElements" is specified, it is the
3216number of elements allocated, otherwise "NumElements" is defaulted to be one.
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3217If a constant alignment is specified, the value result of the allocation is guaranteed
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3218to be aligned to at least that boundary. If not specified, or if zero, the target
3219can choose to align the allocation on any convenient boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3220
3221<p>'<tt>type</tt>' may be any sized type.</p>
3222
3223<h5>Semantics:</h5>
3224
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]3225<p>Memory is allocated; a pointer is returned. The operation is undefiend if
3226there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3227memory is automatically released when the function returns. The '<tt>alloca</tt>'
3228instruction is commonly used to represent automatic variables that must
3229have an address available. When the function returns (either with the <tt><a
3230 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]3231instructions), the memory is reclaimed. Allocating zero bytes
3232is legal, but the result is undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3233
3234<h5>Example:</h5>
3235
3236<pre>
3237 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
3238 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
3239 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
3240 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
3241</pre>
3242</div>
3243
3244<!-- _______________________________________________________________________ -->
3245<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
3246Instruction</a> </div>
3247<div class="doc_text">
3248<h5>Syntax:</h5>
3249<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>
3250<h5>Overview:</h5>
3251<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
3252<h5>Arguments:</h5>
3253<p>The argument to the '<tt>load</tt>' instruction specifies the memory
3254address from which to load. The pointer must point to a <a
3255 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
3256marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
3257the number or order of execution of this <tt>load</tt> with other
3258volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
3259instructions. </p>
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3260<p>
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3261The optional constant "align" argument specifies the alignment of the operation
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3262(that is, the alignment of the memory address). A value of 0 or an
3263omitted "align" argument means that the operation has the preferential
3264alignment for the target. It is the responsibility of the code emitter
3265to ensure that the alignment information is correct. Overestimating
3266the alignment results in an undefined behavior. Underestimating the
3267alignment may produce less efficient code. An alignment of 1 is always
3268safe.
3269</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3270<h5>Semantics:</h5>
3271<p>The location of memory pointed to is loaded.</p>
3272<h5>Examples:</h5>
3273<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
3274 <a
3275 href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
3276 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
3277</pre>
3278</div>
3279<!-- _______________________________________________________________________ -->
3280<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
3281Instruction</a> </div>
3282<div class="doc_text">
3283<h5>Syntax:</h5>
3284<pre> store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
3285 volatile store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
3286</pre>
3287<h5>Overview:</h5>
3288<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
3289<h5>Arguments:</h5>
3290<p>There are two arguments to the '<tt>store</tt>' instruction: a value
3291to 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]3292operand must be a pointer to the <a href="#t_firstclass">first class</a> type
3293of the '<tt>&lt;value&gt;</tt>'
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3294operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
3295optimizer is not allowed to modify the number or order of execution of
3296this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
3297 href="#i_store">store</a></tt> instructions.</p>
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3298<p>
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3299The optional constant "align" argument specifies the alignment of the operation
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3300(that is, the alignment of the memory address). A value of 0 or an
3301omitted "align" argument means that the operation has the preferential
3302alignment for the target. It is the responsibility of the code emitter
3303to ensure that the alignment information is correct. Overestimating
3304the alignment results in an undefined behavior. Underestimating the
3305alignment may produce less efficient code. An alignment of 1 is always
3306safe.
3307</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3308<h5>Semantics:</h5>
3309<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>'
3310at the location specified by the '<tt>&lt;pointer&gt;</tt>' operand.</p>
3311<h5>Example:</h5>
3312<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling63ffa142007-10-22 05:10:05 +0000[diff] [blame]3313 store i32 3, i32* %ptr <i>; yields {void}</i>
3314 %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]3315</pre>
3316</div>
3317
3318<!-- _______________________________________________________________________ -->
3319<div class="doc_subsubsection">
3320 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
3321</div>
3322
3323<div class="doc_text">
3324<h5>Syntax:</h5>
3325<pre>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3326 &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]3327</pre>
3328
3329<h5>Overview:</h5>
3330
3331<p>
3332The '<tt>getelementptr</tt>' instruction is used to get the address of a
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3333subelement of an aggregate data structure. It performs address calculation only
3334and does not access memory.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3335
3336<h5>Arguments:</h5>
3337
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3338<p>The first argument is always a pointer, and forms the basis of the
3339calculation. The remaining arguments are indices, that indicate which of the
3340elements of the aggregate object are indexed. The interpretation of each index
3341is dependent on the type being indexed into. The first index always indexes the
3342pointer value given as the first argument, the second index indexes a value of
3343the type pointed to (not necessarily the value directly pointed to, since the
3344first index can be non-zero), etc. The first type indexed into must be a pointer
3345value, subsequent types can be arrays, vectors and structs. Note that subsequent
3346types being indexed into can never be pointers, since that would require loading
3347the pointer before continuing calculation.</p>
3348
3349<p>The type of each index argument depends on the type it is indexing into.
3350When indexing into a (packed) structure, only <tt>i32</tt> integer
3351<b>constants</b> are allowed. When indexing into an array, pointer or vector,
3352only integers of 32 or 64 bits are allowed (also non-constants). 32-bit values
3353will be sign extended to 64-bits if required.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3354
3355<p>For example, let's consider a C code fragment and how it gets
3356compiled to LLVM:</p>
3357
3358<div class="doc_code">
3359<pre>
3360struct RT {
3361 char A;
3362 int B[10][20];
3363 char C;
3364};
3365struct ST {
3366 int X;
3367 double Y;
3368 struct RT Z;
3369};
3370
3371int *foo(struct ST *s) {
3372 return &amp;s[1].Z.B[5][13];
3373}
3374</pre>
3375</div>
3376
3377<p>The LLVM code generated by the GCC frontend is:</p>
3378
3379<div class="doc_code">
3380<pre>
3381%RT = type { i8 , [10 x [20 x i32]], i8 }
3382%ST = type { i32, double, %RT }
3383
3384define i32* %foo(%ST* %s) {
3385entry:
3386 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
3387 ret i32* %reg
3388}
3389</pre>
3390</div>
3391
3392<h5>Semantics:</h5>
3393
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3394<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
3395type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
3396}</tt>' type, a structure. The second index indexes into the third element of
3397the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
3398i8 }</tt>' type, another structure. The third index indexes into the second
3399element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
3400array. The two dimensions of the array are subscripted into, yielding an
3401'<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
3402to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
3403
3404<p>Note that it is perfectly legal to index partially through a
3405structure, returning a pointer to an inner element. Because of this,
3406the LLVM code for the given testcase is equivalent to:</p>
3407
3408<pre>
3409 define i32* %foo(%ST* %s) {
3410 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
3411 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
3412 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
3413 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
3414 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
3415 ret i32* %t5
3416 }
3417</pre>
3418
3419<p>Note that it is undefined to access an array out of bounds: array and
3420pointer indexes must always be within the defined bounds of the array type.
Chris Lattnera7d94ba2008-04-24 05:59:56 +0000[diff] [blame]3421The one exception for this rule is zero length arrays. These arrays are
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3422defined to be accessible as variable length arrays, which requires access
3423beyond the zero'th element.</p>
3424
3425<p>The getelementptr instruction is often confusing. For some more insight
3426into how it works, see <a href="GetElementPtr.html">the getelementptr
3427FAQ</a>.</p>
3428
3429<h5>Example:</h5>
3430
3431<pre>
3432 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3433 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
3434 <i>; yields i8*:vptr</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3435 %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]3436 <i>; yields i8*:eptr</i>
3437 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3438</pre>
3439</div>
3440
3441<!-- ======================================================================= -->
3442<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
3443</div>
3444<div class="doc_text">
3445<p>The instructions in this category are the conversion instructions (casting)
3446which all take a single operand and a type. They perform various bit conversions
3447on the operand.</p>
3448</div>
3449
3450<!-- _______________________________________________________________________ -->
3451<div class="doc_subsubsection">
3452 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
3453</div>
3454<div class="doc_text">
3455
3456<h5>Syntax:</h5>
3457<pre>
3458 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3459</pre>
3460
3461<h5>Overview:</h5>
3462<p>
3463The '<tt>trunc</tt>' instruction truncates its operand to the type <tt>ty2</tt>.
3464</p>
3465
3466<h5>Arguments:</h5>
3467<p>
3468The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
3469be an <a href="#t_integer">integer</a> type, and a type that specifies the size
3470and type of the result, which must be an <a href="#t_integer">integer</a>
3471type. The bit size of <tt>value</tt> must be larger than the bit size of
3472<tt>ty2</tt>. Equal sized types are not allowed.</p>
3473
3474<h5>Semantics:</h5>
3475<p>
3476The '<tt>trunc</tt>' instruction truncates the high order bits in <tt>value</tt>
3477and converts the remaining bits to <tt>ty2</tt>. Since the source size must be
3478larger than the destination size, <tt>trunc</tt> cannot be a <i>no-op cast</i>.
3479It will always truncate bits.</p>
3480
3481<h5>Example:</h5>
3482<pre>
3483 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
3484 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
3485 %Y = trunc i32 122 to i1 <i>; yields i1:false</i>
3486</pre>
3487</div>
3488
3489<!-- _______________________________________________________________________ -->
3490<div class="doc_subsubsection">
3491 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
3492</div>
3493<div class="doc_text">
3494
3495<h5>Syntax:</h5>
3496<pre>
3497 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3498</pre>
3499
3500<h5>Overview:</h5>
3501<p>The '<tt>zext</tt>' instruction zero extends its operand to type
3502<tt>ty2</tt>.</p>
3503
3504
3505<h5>Arguments:</h5>
3506<p>The '<tt>zext</tt>' instruction takes a value to cast, which must be of
3507<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3508also be of <a href="#t_integer">integer</a> type. The bit size of the
3509<tt>value</tt> must be smaller than the bit size of the destination type,
3510<tt>ty2</tt>.</p>
3511
3512<h5>Semantics:</h5>
3513<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
3514bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
3515
3516<p>When zero extending from i1, the result will always be either 0 or 1.</p>
3517
3518<h5>Example:</h5>
3519<pre>
3520 %X = zext i32 257 to i64 <i>; yields i64:257</i>
3521 %Y = zext i1 true to i32 <i>; yields i32:1</i>
3522</pre>
3523</div>
3524
3525<!-- _______________________________________________________________________ -->
3526<div class="doc_subsubsection">
3527 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
3528</div>
3529<div class="doc_text">
3530
3531<h5>Syntax:</h5>
3532<pre>
3533 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3534</pre>
3535
3536<h5>Overview:</h5>
3537<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
3538
3539<h5>Arguments:</h5>
3540<p>
3541The '<tt>sext</tt>' instruction takes a value to cast, which must be of
3542<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3543also be of <a href="#t_integer">integer</a> type. The bit size of the
3544<tt>value</tt> must be smaller than the bit size of the destination type,
3545<tt>ty2</tt>.</p>
3546
3547<h5>Semantics:</h5>
3548<p>
3549The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
3550bit (highest order bit) of the <tt>value</tt> until it reaches the bit size of
3551the type <tt>ty2</tt>.</p>
3552
3553<p>When sign extending from i1, the extension always results in -1 or 0.</p>
3554
3555<h5>Example:</h5>
3556<pre>
3557 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
3558 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
3559</pre>
3560</div>
3561
3562<!-- _______________________________________________________________________ -->
3563<div class="doc_subsubsection">
3564 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
3565</div>
3566
3567<div class="doc_text">
3568
3569<h5>Syntax:</h5>
3570
3571<pre>
3572 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3573</pre>
3574
3575<h5>Overview:</h5>
3576<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
3577<tt>ty2</tt>.</p>
3578
3579
3580<h5>Arguments:</h5>
3581<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
3582 point</a> value to cast and a <a href="#t_floating">floating point</a> type to
3583cast it to. The size of <tt>value</tt> must be larger than the size of
3584<tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
3585<i>no-op cast</i>.</p>
3586
3587<h5>Semantics:</h5>
3588<p> The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
3589<a href="#t_floating">floating point</a> type to a smaller
3590<a href="#t_floating">floating point</a> type. If the value cannot fit within
3591the destination type, <tt>ty2</tt>, then the results are undefined.</p>
3592
3593<h5>Example:</h5>
3594<pre>
3595 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
3596 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
3597</pre>
3598</div>
3599
3600<!-- _______________________________________________________________________ -->
3601<div class="doc_subsubsection">
3602 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
3603</div>
3604<div class="doc_text">
3605
3606<h5>Syntax:</h5>
3607<pre>
3608 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3609</pre>
3610
3611<h5>Overview:</h5>
3612<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
3613floating point value.</p>
3614
3615<h5>Arguments:</h5>
3616<p>The '<tt>fpext</tt>' instruction takes a
3617<a href="#t_floating">floating point</a> <tt>value</tt> to cast,
3618and a <a href="#t_floating">floating point</a> type to cast it to. The source
3619type must be smaller than the destination type.</p>
3620
3621<h5>Semantics:</h5>
3622<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
3623<a href="#t_floating">floating point</a> type to a larger
3624<a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
3625used to make a <i>no-op cast</i> because it always changes bits. Use
3626<tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
3627
3628<h5>Example:</h5>
3629<pre>
3630 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
3631 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
3632</pre>
3633</div>
3634
3635<!-- _______________________________________________________________________ -->
3636<div class="doc_subsubsection">
3637 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
3638</div>
3639<div class="doc_text">
3640
3641<h5>Syntax:</h5>
3642<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3643 &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]3644</pre>
3645
3646<h5>Overview:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3647<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3648unsigned integer equivalent of type <tt>ty2</tt>.
3649</p>
3650
3651<h5>Arguments:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3652<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]3653scalar or vector <a href="#t_floating">floating point</a> value, and a type
3654to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
3655type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
3656vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3657
3658<h5>Semantics:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3659<p> The '<tt>fptoui</tt>' instruction converts its
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3660<a href="#t_floating">floating point</a> operand into the nearest (rounding
3661towards zero) unsigned integer value. If the value cannot fit in <tt>ty2</tt>,
3662the results are undefined.</p>
3663
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3664<h5>Example:</h5>
3665<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3666 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]3667 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3668 %X = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3669</pre>
3670</div>
3671
3672<!-- _______________________________________________________________________ -->
3673<div class="doc_subsubsection">
3674 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
3675</div>
3676<div class="doc_text">
3677
3678<h5>Syntax:</h5>
3679<pre>
3680 &lt;result&gt; = fptosi &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3681</pre>
3682
3683<h5>Overview:</h5>
3684<p>The '<tt>fptosi</tt>' instruction converts
3685<a href="#t_floating">floating point</a> <tt>value</tt> to type <tt>ty2</tt>.
3686</p>
3687
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3688<h5>Arguments:</h5>
3689<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]3690scalar or vector <a href="#t_floating">floating point</a> value, and a type
3691to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
3692type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
3693vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3694
3695<h5>Semantics:</h5>
3696<p>The '<tt>fptosi</tt>' instruction converts its
3697<a href="#t_floating">floating point</a> operand into the nearest (rounding
3698towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
3699the results are undefined.</p>
3700
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3701<h5>Example:</h5>
3702<pre>
3703 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]3704 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3705 %X = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
3706</pre>
3707</div>
3708
3709<!-- _______________________________________________________________________ -->
3710<div class="doc_subsubsection">
3711 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
3712</div>
3713<div class="doc_text">
3714
3715<h5>Syntax:</h5>
3716<pre>
3717 &lt;result&gt; = uitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3718</pre>
3719
3720<h5>Overview:</h5>
3721<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
3722integer and converts that value to the <tt>ty2</tt> type.</p>
3723
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3724<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]3725<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be a
3726scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
3727to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
3728type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
3729floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3730
3731<h5>Semantics:</h5>
3732<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
3733integer quantity and converts it to the corresponding floating point value. If
3734the value cannot fit in the floating point value, the results are undefined.</p>
3735
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3736<h5>Example:</h5>
3737<pre>
3738 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3739 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3740</pre>
3741</div>
3742
3743<!-- _______________________________________________________________________ -->
3744<div class="doc_subsubsection">
3745 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
3746</div>
3747<div class="doc_text">
3748
3749<h5>Syntax:</h5>
3750<pre>
3751 &lt;result&gt; = sitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3752</pre>
3753
3754<h5>Overview:</h5>
3755<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed
3756integer and converts that value to the <tt>ty2</tt> type.</p>
3757
3758<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]3759<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be a
3760scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
3761to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
3762type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
3763floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3764
3765<h5>Semantics:</h5>
3766<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed
3767integer quantity and converts it to the corresponding floating point value. If
3768the value cannot fit in the floating point value, the results are undefined.</p>
3769
3770<h5>Example:</h5>
3771<pre>
3772 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3773 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3774</pre>
3775</div>
3776
3777<!-- _______________________________________________________________________ -->
3778<div class="doc_subsubsection">
3779 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
3780</div>
3781<div class="doc_text">
3782
3783<h5>Syntax:</h5>
3784<pre>
3785 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3786</pre>
3787
3788<h5>Overview:</h5>
3789<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
3790the integer type <tt>ty2</tt>.</p>
3791
3792<h5>Arguments:</h5>
3793<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
3794must 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]3795<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]3796
3797<h5>Semantics:</h5>
3798<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
3799<tt>ty2</tt> by interpreting the pointer value as an integer and either
3800truncating or zero extending that value to the size of the integer type. If
3801<tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
3802<tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
3803are the same size, then nothing is done (<i>no-op cast</i>) other than a type
3804change.</p>
3805
3806<h5>Example:</h5>
3807<pre>
3808 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
3809 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
3810</pre>
3811</div>
3812
3813<!-- _______________________________________________________________________ -->
3814<div class="doc_subsubsection">
3815 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
3816</div>
3817<div class="doc_text">
3818
3819<h5>Syntax:</h5>
3820<pre>
3821 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3822</pre>
3823
3824<h5>Overview:</h5>
3825<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to
3826a pointer type, <tt>ty2</tt>.</p>
3827
3828<h5>Arguments:</h5>
3829<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
3830value to cast, and a type to cast it to, which must be a
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3831<a href="#t_pointer">pointer</a> type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3832
3833<h5>Semantics:</h5>
3834<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
3835<tt>ty2</tt> by applying either a zero extension or a truncation depending on
3836the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
3837size of a pointer then a truncation is done. If <tt>value</tt> is smaller than
3838the size of a pointer then a zero extension is done. If they are the same size,
3839nothing is done (<i>no-op cast</i>).</p>
3840
3841<h5>Example:</h5>
3842<pre>
3843 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
3844 %X = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
3845 %Y = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
3846</pre>
3847</div>
3848
3849<!-- _______________________________________________________________________ -->
3850<div class="doc_subsubsection">
3851 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
3852</div>
3853<div class="doc_text">
3854
3855<h5>Syntax:</h5>
3856<pre>
3857 &lt;result&gt; = bitcast &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3858</pre>
3859
3860<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3861
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3862<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
3863<tt>ty2</tt> without changing any bits.</p>
3864
3865<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3866
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3867<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be
Dan Gohman7305fa02008-09-08 16:45:59 +0000[diff] [blame]3868a non-aggregate first class value, and a type to cast it to, which must also be
3869a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes of
3870<tt>value</tt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3871and the destination type, <tt>ty2</tt>, must be identical. If the source
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3872type is a pointer, the destination type must also be a pointer. This
3873instruction supports bitwise conversion of vectors to integers and to vectors
3874of other types (as long as they have the same size).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3875
3876<h5>Semantics:</h5>
3877<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
3878<tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
3879this conversion. The conversion is done as if the <tt>value</tt> had been
3880stored to memory and read back as type <tt>ty2</tt>. Pointer types may only be
3881converted to other pointer types with this instruction. To convert pointers to
3882other types, use the <a href="#i_inttoptr">inttoptr</a> or
3883<a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
3884
3885<h5>Example:</h5>
3886<pre>
3887 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
3888 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3889 %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]3890</pre>
3891</div>
3892
3893<!-- ======================================================================= -->
3894<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
3895<div class="doc_text">
3896<p>The instructions in this category are the "miscellaneous"
3897instructions, which defy better classification.</p>
3898</div>
3899
3900<!-- _______________________________________________________________________ -->
3901<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
3902</div>
3903<div class="doc_text">
3904<h5>Syntax:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3905<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]3906</pre>
3907<h5>Overview:</h5>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]3908<p>The '<tt>icmp</tt>' instruction returns a boolean value or
3909a vector of boolean values based on comparison
3910of its two integer, integer vector, or pointer operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3911<h5>Arguments:</h5>
3912<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
3913the condition code indicating the kind of comparison to perform. It is not
3914a value, just a keyword. The possible condition code are:
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3915</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3916<ol>
3917 <li><tt>eq</tt>: equal</li>
3918 <li><tt>ne</tt>: not equal </li>
3919 <li><tt>ugt</tt>: unsigned greater than</li>
3920 <li><tt>uge</tt>: unsigned greater or equal</li>
3921 <li><tt>ult</tt>: unsigned less than</li>
3922 <li><tt>ule</tt>: unsigned less or equal</li>
3923 <li><tt>sgt</tt>: signed greater than</li>
3924 <li><tt>sge</tt>: signed greater or equal</li>
3925 <li><tt>slt</tt>: signed less than</li>
3926 <li><tt>sle</tt>: signed less or equal</li>
3927</ol>
3928<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]3929<a href="#t_pointer">pointer</a>
3930or integer <a href="#t_vector">vector</a> typed.
3931They must also be identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3932<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3933<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]3934the condition code given as <tt>cond</tt>. The comparison performed always
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]3935yields 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]3936</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3937<ol>
3938 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
3939 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
3940 </li>
3941 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3942 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3943 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3944 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3945 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3946 <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]3947 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3948 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3949 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3950 <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]3951 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3952 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3953 <li><tt>sge</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3954 <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]3955 <li><tt>slt</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3956 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3957 <li><tt>sle</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3958 <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]3959</ol>
3960<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
3961values are compared as if they were integers.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]3962<p>If the operands are integer vectors, then they are compared
3963element by element. The result is an <tt>i1</tt> vector with
3964the same number of elements as the values being compared.
3965Otherwise, the result is an <tt>i1</tt>.
3966</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3967
3968<h5>Example:</h5>
3969<pre> &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
3970 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
3971 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
3972 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
3973 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
3974 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
3975</pre>
3976</div>
3977
3978<!-- _______________________________________________________________________ -->
3979<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
3980</div>
3981<div class="doc_text">
3982<h5>Syntax:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3983<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]3984</pre>
3985<h5>Overview:</h5>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]3986<p>The '<tt>fcmp</tt>' instruction returns a boolean value
3987or vector of boolean values based on comparison
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3988of its operands.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]3989<p>
3990If the operands are floating point scalars, then the result
3991type is a boolean (<a href="#t_primitive"><tt>i1</tt></a>).
3992</p>
3993<p>If the operands are floating point vectors, then the result type
3994is a vector of boolean with the same number of elements as the
3995operands being compared.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3996<h5>Arguments:</h5>
3997<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
3998the condition code indicating the kind of comparison to perform. It is not
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3999a value, just a keyword. The possible condition code are:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4000<ol>
4001 <li><tt>false</tt>: no comparison, always returns false</li>
4002 <li><tt>oeq</tt>: ordered and equal</li>
4003 <li><tt>ogt</tt>: ordered and greater than </li>
4004 <li><tt>oge</tt>: ordered and greater than or equal</li>
4005 <li><tt>olt</tt>: ordered and less than </li>
4006 <li><tt>ole</tt>: ordered and less than or equal</li>
4007 <li><tt>one</tt>: ordered and not equal</li>
4008 <li><tt>ord</tt>: ordered (no nans)</li>
4009 <li><tt>ueq</tt>: unordered or equal</li>
4010 <li><tt>ugt</tt>: unordered or greater than </li>
4011 <li><tt>uge</tt>: unordered or greater than or equal</li>
4012 <li><tt>ult</tt>: unordered or less than </li>
4013 <li><tt>ule</tt>: unordered or less than or equal</li>
4014 <li><tt>une</tt>: unordered or not equal</li>
4015 <li><tt>uno</tt>: unordered (either nans)</li>
4016 <li><tt>true</tt>: no comparison, always returns true</li>
4017</ol>
4018<p><i>Ordered</i> means that neither operand is a QNAN while
4019<i>unordered</i> means that either operand may be a QNAN.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4020<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be
4021either a <a href="#t_floating">floating point</a> type
4022or a <a href="#t_vector">vector</a> of floating point type.
4023They must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4024<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4025<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4026according to the condition code given as <tt>cond</tt>.
4027If the operands are vectors, then the vectors are compared
4028element by element.
4029Each comparison performed
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4030always yields an <a href="#t_primitive">i1</a> result, as follows:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4031<ol>
4032 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
4033 <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]4034 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4035 <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]4036 <tt>op1</tt> is greather than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4037 <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]4038 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4039 <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]4040 <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4041 <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]4042 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4043 <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]4044 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4045 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
4046 <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]4047 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4048 <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]4049 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4050 <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]4051 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4052 <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]4053 <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4054 <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]4055 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4056 <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]4057 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4058 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
4059 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
4060</ol>
4061
4062<h5>Example:</h5>
4063<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]4064 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
4065 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
4066 &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]4067</pre>
4068</div>
4069
4070<!-- _______________________________________________________________________ -->
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4071<div class="doc_subsubsection">
4072 <a name="i_vicmp">'<tt>vicmp</tt>' Instruction</a>
4073</div>
4074<div class="doc_text">
4075<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4076<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]4077</pre>
4078<h5>Overview:</h5>
4079<p>The '<tt>vicmp</tt>' instruction returns an integer vector value based on
4080element-wise comparison of its two integer vector operands.</p>
4081<h5>Arguments:</h5>
4082<p>The '<tt>vicmp</tt>' instruction takes three operands. The first operand is
4083the condition code indicating the kind of comparison to perform. It is not
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4084a value, just a keyword. The possible condition code are:</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4085<ol>
4086 <li><tt>eq</tt>: equal</li>
4087 <li><tt>ne</tt>: not equal </li>
4088 <li><tt>ugt</tt>: unsigned greater than</li>
4089 <li><tt>uge</tt>: unsigned greater or equal</li>
4090 <li><tt>ult</tt>: unsigned less than</li>
4091 <li><tt>ule</tt>: unsigned less or equal</li>
4092 <li><tt>sgt</tt>: signed greater than</li>
4093 <li><tt>sge</tt>: signed greater or equal</li>
4094 <li><tt>slt</tt>: signed less than</li>
4095 <li><tt>sle</tt>: signed less or equal</li>
4096</ol>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4097<p>The remaining two arguments must be <a href="#t_vector">vector</a> or
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4098<a href="#t_integer">integer</a> typed. They must also be identical types.</p>
4099<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4100<p>The '<tt>vicmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4101according to the condition code given as <tt>cond</tt>. The comparison yields a
4102<a href="#t_vector">vector</a> of <a href="#t_integer">integer</a> result, of
4103identical type as the values being compared. The most significant bit in each
4104element is 1 if the element-wise comparison evaluates to true, and is 0
4105otherwise. All other bits of the result are undefined. The condition codes
4106are evaluated identically to the <a href="#i_icmp">'<tt>icmp</tt>'
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4107instruction</a>.</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4108
4109<h5>Example:</h5>
4110<pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4111 &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>
4112 &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]4113</pre>
4114</div>
4115
4116<!-- _______________________________________________________________________ -->
4117<div class="doc_subsubsection">
4118 <a name="i_vfcmp">'<tt>vfcmp</tt>' Instruction</a>
4119</div>
4120<div class="doc_text">
4121<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4122<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]4123<h5>Overview:</h5>
4124<p>The '<tt>vfcmp</tt>' instruction returns an integer vector value based on
4125element-wise comparison of its two floating point vector operands. The output
4126elements have the same width as the input elements.</p>
4127<h5>Arguments:</h5>
4128<p>The '<tt>vfcmp</tt>' instruction takes three operands. The first operand is
4129the condition code indicating the kind of comparison to perform. It is not
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4130a value, just a keyword. The possible condition code are:</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4131<ol>
4132 <li><tt>false</tt>: no comparison, always returns false</li>
4133 <li><tt>oeq</tt>: ordered and equal</li>
4134 <li><tt>ogt</tt>: ordered and greater than </li>
4135 <li><tt>oge</tt>: ordered and greater than or equal</li>
4136 <li><tt>olt</tt>: ordered and less than </li>
4137 <li><tt>ole</tt>: ordered and less than or equal</li>
4138 <li><tt>one</tt>: ordered and not equal</li>
4139 <li><tt>ord</tt>: ordered (no nans)</li>
4140 <li><tt>ueq</tt>: unordered or equal</li>
4141 <li><tt>ugt</tt>: unordered or greater than </li>
4142 <li><tt>uge</tt>: unordered or greater than or equal</li>
4143 <li><tt>ult</tt>: unordered or less than </li>
4144 <li><tt>ule</tt>: unordered or less than or equal</li>
4145 <li><tt>une</tt>: unordered or not equal</li>
4146 <li><tt>uno</tt>: unordered (either nans)</li>
4147 <li><tt>true</tt>: no comparison, always returns true</li>
4148</ol>
4149<p>The remaining two arguments must be <a href="#t_vector">vector</a> of
4150<a href="#t_floating">floating point</a> typed. They must also be identical
4151types.</p>
4152<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4153<p>The '<tt>vfcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4154according to the condition code given as <tt>cond</tt>. The comparison yields a
4155<a href="#t_vector">vector</a> of <a href="#t_integer">integer</a> result, with
4156an identical number of elements as the values being compared, and each element
4157having identical with to the width of the floating point elements. The most
4158significant bit in each element is 1 if the element-wise comparison evaluates to
4159true, and is 0 otherwise. All other bits of the result are undefined. The
4160condition codes are evaluated identically to the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4161<a href="#i_fcmp">'<tt>fcmp</tt>' instruction</a>.</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4162
4163<h5>Example:</h5>
4164<pre>
Chris Lattner1e5c5cd02008-10-13 16:55:18 +0000[diff] [blame]4165 <i>; yields: result=&lt;2 x i32&gt; &lt; i32 0, i32 -1 &gt;</i>
4166 &lt;result&gt; = vfcmp oeq &lt;2 x float&gt; &lt; float 4, float 0 &gt;, &lt; float 5, float 0 &gt;
4167
4168 <i>; yields: result=&lt;2 x i64&gt; &lt; i64 -1, i64 0 &gt;</i>
4169 &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]4170</pre>
4171</div>
4172
4173<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4174<div class="doc_subsubsection">
4175 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
4176</div>
4177
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4178<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4179
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4180<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4181
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4182<pre> &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...<br></pre>
4183<h5>Overview:</h5>
4184<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in
4185the SSA graph representing the function.</p>
4186<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4187
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4188<p>The type of the incoming values is specified with the first type
4189field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
4190as arguments, with one pair for each predecessor basic block of the
4191current block. Only values of <a href="#t_firstclass">first class</a>
4192type may be used as the value arguments to the PHI node. Only labels
4193may be used as the label arguments.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4194
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4195<p>There must be no non-phi instructions between the start of a basic
4196block and the PHI instructions: i.e. PHI instructions must be first in
4197a basic block.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4198
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4199<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4200
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4201<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
4202specified by the pair corresponding to the predecessor basic block that executed
4203just prior to the current block.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4204
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4205<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4206<pre>
4207Loop: ; Infinite loop that counts from 0 on up...
4208 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
4209 %nextindvar = add i32 %indvar, 1
4210 br label %Loop
4211</pre>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4212</div>
4213
4214<!-- _______________________________________________________________________ -->
4215<div class="doc_subsubsection">
4216 <a name="i_select">'<tt>select</tt>' Instruction</a>
4217</div>
4218
4219<div class="doc_text">
4220
4221<h5>Syntax:</h5>
4222
4223<pre>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4224 &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>
4225
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4226 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4227</pre>
4228
4229<h5>Overview:</h5>
4230
4231<p>
4232The '<tt>select</tt>' instruction is used to choose one value based on a
4233condition, without branching.
4234</p>
4235
4236
4237<h5>Arguments:</h5>
4238
4239<p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4240The '<tt>select</tt>' instruction requires an 'i1' value or
4241a vector of 'i1' values indicating the
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4242condition, and two values of the same <a href="#t_firstclass">first class</a>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4243type. If the val1/val2 are vectors and
4244the condition is a scalar, then entire vectors are selected, not
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4245individual elements.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4246</p>
4247
4248<h5>Semantics:</h5>
4249
4250<p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4251If 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]4252value argument; otherwise, it returns the second value argument.
4253</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4254<p>
4255If the condition is a vector of i1, then the value arguments must
4256be vectors of the same size, and the selection is done element
4257by element.
4258</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4259
4260<h5>Example:</h5>
4261
4262<pre>
4263 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
4264</pre>
4265</div>
4266
4267
4268<!-- _______________________________________________________________________ -->
4269<div class="doc_subsubsection">
4270 <a name="i_call">'<tt>call</tt>' Instruction</a>
4271</div>
4272
4273<div class="doc_text">
4274
4275<h5>Syntax:</h5>
4276<pre>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]4277 &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]4278</pre>
4279
4280<h5>Overview:</h5>
4281
4282<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
4283
4284<h5>Arguments:</h5>
4285
4286<p>This instruction requires several arguments:</p>
4287
4288<ol>
4289 <li>
4290 <p>The optional "tail" marker indicates whether the callee function accesses
4291 any allocas or varargs in the caller. If the "tail" marker is present, the
4292 function call is eligible for tail call optimization. Note that calls may
4293 be marked "tail" even if they do not occur before a <a
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4294 href="#i_ret"><tt>ret</tt></a> instruction.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4295 </li>
4296 <li>
4297 <p>The optional "cconv" marker indicates which <a href="#callingconv">calling
4298 convention</a> the call should use. If none is specified, the call defaults
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4299 to using C calling conventions.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4300 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4301
4302 <li>
4303 <p>The optional <a href="#paramattrs">Parameter Attributes</a> list for
4304 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>',
4305 and '<tt>inreg</tt>' attributes are valid here.</p>
4306 </li>
4307
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4308 <li>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]4309 <p>'<tt>ty</tt>': the type of the call instruction itself which is also
4310 the type of the return value. Functions that return no value are marked
4311 <tt><a href="#t_void">void</a></tt>.</p>
4312 </li>
4313 <li>
4314 <p>'<tt>fnty</tt>': shall be the signature of the pointer to function
4315 value being invoked. The argument types must match the types implied by
4316 this signature. This type can be omitted if the function is not varargs
4317 and if the function type does not return a pointer to a function.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4318 </li>
4319 <li>
4320 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
4321 be invoked. In most cases, this is a direct function invocation, but
4322 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
4323 to function value.</p>
4324 </li>
4325 <li>
4326 <p>'<tt>function args</tt>': argument list whose types match the
4327 function signature argument types. All arguments must be of
4328 <a href="#t_firstclass">first class</a> type. If the function signature
4329 indicates the function accepts a variable number of arguments, the extra
4330 arguments can be specified.</p>
4331 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4332 <li>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]4333 <p>The optional <a href="#fnattrs">function attributes</a> list. Only
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4334 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
4335 '<tt>readnone</tt>' attributes are valid here.</p>
4336 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4337</ol>
4338
4339<h5>Semantics:</h5>
4340
4341<p>The '<tt>call</tt>' instruction is used to cause control flow to
4342transfer to a specified function, with its incoming arguments bound to
4343the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
4344instruction in the called function, control flow continues with the
4345instruction after the function call, and the return value of the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4346function is bound to the result argument.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4347
4348<h5>Example:</h5>
4349
4350<pre>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]4351 %retval = call i32 @test(i32 %argc)
Chris Lattner5e893ef2008-03-21 17:24:17 +0000[diff] [blame]4352 call i32 (i8 *, ...)* @printf(i8 * %msg, i32 12, i8 42) <i>; yields i32</i>
4353 %X = tail call i32 @foo() <i>; yields i32</i>
4354 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
4355 call void %foo(i8 97 signext)
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]4356
4357 %struct.A = type { i32, i8 }
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4358 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]4359 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
4360 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattnerac454b32008-10-08 06:26:11 +0000[diff] [blame]4361 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijman2c4e05a2008-10-07 10:03:45 +0000[diff] [blame]4362 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4363</pre>
4364
4365</div>
4366
4367<!-- _______________________________________________________________________ -->
4368<div class="doc_subsubsection">
4369 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
4370</div>
4371
4372<div class="doc_text">
4373
4374<h5>Syntax:</h5>
4375
4376<pre>
4377 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
4378</pre>
4379
4380<h5>Overview:</h5>
4381
4382<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
4383the "variable argument" area of a function call. It is used to implement the
4384<tt>va_arg</tt> macro in C.</p>
4385
4386<h5>Arguments:</h5>
4387
4388<p>This instruction takes a <tt>va_list*</tt> value and the type of
4389the argument. It returns a value of the specified argument type and
4390increments the <tt>va_list</tt> to point to the next argument. The
4391actual type of <tt>va_list</tt> is target specific.</p>
4392
4393<h5>Semantics:</h5>
4394
4395<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
4396type from the specified <tt>va_list</tt> and causes the
4397<tt>va_list</tt> to point to the next argument. For more information,
4398see the variable argument handling <a href="#int_varargs">Intrinsic
4399Functions</a>.</p>
4400
4401<p>It is legal for this instruction to be called in a function which does not
4402take a variable number of arguments, for example, the <tt>vfprintf</tt>
4403function.</p>
4404
4405<p><tt>va_arg</tt> is an LLVM instruction instead of an <a
4406href="#intrinsics">intrinsic function</a> because it takes a type as an
4407argument.</p>
4408
4409<h5>Example:</h5>
4410
4411<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
4412
4413</div>
4414
4415<!-- *********************************************************************** -->
4416<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
4417<!-- *********************************************************************** -->
4418
4419<div class="doc_text">
4420
4421<p>LLVM supports the notion of an "intrinsic function". These functions have
4422well known names and semantics and are required to follow certain restrictions.
4423Overall, these intrinsics represent an extension mechanism for the LLVM
4424language that does not require changing all of the transformations in LLVM when
4425adding to the language (or the bitcode reader/writer, the parser, etc...).</p>
4426
4427<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
4428prefix is reserved in LLVM for intrinsic names; thus, function names may not
4429begin with this prefix. Intrinsic functions must always be external functions:
4430you cannot define the body of intrinsic functions. Intrinsic functions may
4431only be used in call or invoke instructions: it is illegal to take the address
4432of an intrinsic function. Additionally, because intrinsic functions are part
4433of the LLVM language, it is required if any are added that they be documented
4434here.</p>
4435
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4436<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents
4437a family of functions that perform the same operation but on different data
4438types. Because LLVM can represent over 8 million different integer types,
4439overloading is used commonly to allow an intrinsic function to operate on any
4440integer type. One or more of the argument types or the result type can be
4441overloaded to accept any integer type. Argument types may also be defined as
4442exactly matching a previous argument's type or the result type. This allows an
4443intrinsic function which accepts multiple arguments, but needs all of them to
4444be of the same type, to only be overloaded with respect to a single argument or
4445the result.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4446
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4447<p>Overloaded intrinsics will have the names of its overloaded argument types
4448encoded into its function name, each preceded by a period. Only those types
4449which are overloaded result in a name suffix. Arguments whose type is matched
4450against another type do not. For example, the <tt>llvm.ctpop</tt> function can
4451take an integer of any width and returns an integer of exactly the same integer
4452width. This leads to a family of functions such as
4453<tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29 %val)</tt>.
4454Only one type, the return type, is overloaded, and only one type suffix is
4455required. Because the argument's type is matched against the return type, it
4456does not require its own name suffix.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4457
4458<p>To learn how to add an intrinsic function, please see the
4459<a href="ExtendingLLVM.html">Extending LLVM Guide</a>.
4460</p>
4461
4462</div>
4463
4464<!-- ======================================================================= -->
4465<div class="doc_subsection">
4466 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
4467</div>
4468
4469<div class="doc_text">
4470
4471<p>Variable argument support is defined in LLVM with the <a
4472 href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
4473intrinsic functions. These functions are related to the similarly
4474named macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
4475
4476<p>All of these functions operate on arguments that use a
4477target-specific value type "<tt>va_list</tt>". The LLVM assembly
4478language reference manual does not define what this type is, so all
4479transformations should be prepared to handle these functions regardless of
4480the type used.</p>
4481
4482<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
4483instruction and the variable argument handling intrinsic functions are
4484used.</p>
4485
4486<div class="doc_code">
4487<pre>
4488define i32 @test(i32 %X, ...) {
4489 ; Initialize variable argument processing
4490 %ap = alloca i8*
4491 %ap2 = bitcast i8** %ap to i8*
4492 call void @llvm.va_start(i8* %ap2)
4493
4494 ; Read a single integer argument
4495 %tmp = va_arg i8** %ap, i32
4496
4497 ; Demonstrate usage of llvm.va_copy and llvm.va_end
4498 %aq = alloca i8*
4499 %aq2 = bitcast i8** %aq to i8*
4500 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
4501 call void @llvm.va_end(i8* %aq2)
4502
4503 ; Stop processing of arguments.
4504 call void @llvm.va_end(i8* %ap2)
4505 ret i32 %tmp
4506}
4507
4508declare void @llvm.va_start(i8*)
4509declare void @llvm.va_copy(i8*, i8*)
4510declare void @llvm.va_end(i8*)
4511</pre>
4512</div>
4513
4514</div>
4515
4516<!-- _______________________________________________________________________ -->
4517<div class="doc_subsubsection">
4518 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
4519</div>
4520
4521
4522<div class="doc_text">
4523<h5>Syntax:</h5>
4524<pre> declare void %llvm.va_start(i8* &lt;arglist&gt;)<br></pre>
4525<h5>Overview:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4526<p>The '<tt>llvm.va_start</tt>' intrinsic initializes
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4527<tt>*&lt;arglist&gt;</tt> for subsequent use by <tt><a
4528href="#i_va_arg">va_arg</a></tt>.</p>
4529
4530<h5>Arguments:</h5>
4531
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4532<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]4533
4534<h5>Semantics:</h5>
4535
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4536<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]4537macro available in C. In a target-dependent way, it initializes the
4538<tt>va_list</tt> element to which the argument points, so that the next call to
4539<tt>va_arg</tt> will produce the first variable argument passed to the function.
4540Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
4541last argument of the function as the compiler can figure that out.</p>
4542
4543</div>
4544
4545<!-- _______________________________________________________________________ -->
4546<div class="doc_subsubsection">
4547 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
4548</div>
4549
4550<div class="doc_text">
4551<h5>Syntax:</h5>
4552<pre> declare void @llvm.va_end(i8* &lt;arglist&gt;)<br></pre>
4553<h5>Overview:</h5>
4554
4555<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
4556which has been initialized previously with <tt><a href="#int_va_start">llvm.va_start</a></tt>
4557or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
4558
4559<h5>Arguments:</h5>
4560
4561<p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
4562
4563<h5>Semantics:</h5>
4564
4565<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
4566macro available in C. In a target-dependent way, it destroys the
4567<tt>va_list</tt> element to which the argument points. Calls to <a
4568href="#int_va_start"><tt>llvm.va_start</tt></a> and <a href="#int_va_copy">
4569<tt>llvm.va_copy</tt></a> must be matched exactly with calls to
4570<tt>llvm.va_end</tt>.</p>
4571
4572</div>
4573
4574<!-- _______________________________________________________________________ -->
4575<div class="doc_subsubsection">
4576 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
4577</div>
4578
4579<div class="doc_text">
4580
4581<h5>Syntax:</h5>
4582
4583<pre>
4584 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
4585</pre>
4586
4587<h5>Overview:</h5>
4588
4589<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
4590from the source argument list to the destination argument list.</p>
4591
4592<h5>Arguments:</h5>
4593
4594<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
4595The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
4596
4597
4598<h5>Semantics:</h5>
4599
4600<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
4601macro available in C. In a target-dependent way, it copies the source
4602<tt>va_list</tt> element into the destination <tt>va_list</tt> element. This
4603intrinsic is necessary because the <tt><a href="#int_va_start">
4604llvm.va_start</a></tt> intrinsic may be arbitrarily complex and require, for
4605example, memory allocation.</p>
4606
4607</div>
4608
4609<!-- ======================================================================= -->
4610<div class="doc_subsection">
4611 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
4612</div>
4613
4614<div class="doc_text">
4615
4616<p>
4617LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]4618Collection</a> (GC) requires the implementation and generation of these
4619intrinsics.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4620These intrinsics allow identification of <a href="#int_gcroot">GC roots on the
4621stack</a>, as well as garbage collector implementations that require <a
4622href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a> barriers.
4623Front-ends for type-safe garbage collected languages should generate these
4624intrinsics to make use of the LLVM garbage collectors. For more details, see <a
4625href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
4626</p>
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]4627
4628<p>The garbage collection intrinsics only operate on objects in the generic
4629 address space (address space zero).</p>
4630
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4631</div>
4632
4633<!-- _______________________________________________________________________ -->
4634<div class="doc_subsubsection">
4635 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
4636</div>
4637
4638<div class="doc_text">
4639
4640<h5>Syntax:</h5>
4641
4642<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4643 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4644</pre>
4645
4646<h5>Overview:</h5>
4647
4648<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
4649the code generator, and allows some metadata to be associated with it.</p>
4650
4651<h5>Arguments:</h5>
4652
4653<p>The first argument specifies the address of a stack object that contains the
4654root pointer. The second pointer (which must be either a constant or a global
4655value address) contains the meta-data to be associated with the root.</p>
4656
4657<h5>Semantics:</h5>
4658
Chris Lattnera7d94ba2008-04-24 05:59:56 +0000[diff] [blame]4659<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]4660location. At compile-time, the code generator generates information to allow
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4661the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
4662intrinsic may only be used in a function which <a href="#gc">specifies a GC
4663algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4664
4665</div>
4666
4667
4668<!-- _______________________________________________________________________ -->
4669<div class="doc_subsubsection">
4670 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
4671</div>
4672
4673<div class="doc_text">
4674
4675<h5>Syntax:</h5>
4676
4677<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4678 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4679</pre>
4680
4681<h5>Overview:</h5>
4682
4683<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
4684locations, allowing garbage collector implementations that require read
4685barriers.</p>
4686
4687<h5>Arguments:</h5>
4688
4689<p>The second argument is the address to read from, which should be an address
4690allocated from the garbage collector. The first object is a pointer to the
4691start of the referenced object, if needed by the language runtime (otherwise
4692null).</p>
4693
4694<h5>Semantics:</h5>
4695
4696<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
4697instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4698garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
4699may only be used in a function which <a href="#gc">specifies a GC
4700algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4701
4702</div>
4703
4704
4705<!-- _______________________________________________________________________ -->
4706<div class="doc_subsubsection">
4707 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
4708</div>
4709
4710<div class="doc_text">
4711
4712<h5>Syntax:</h5>
4713
4714<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4715 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4716</pre>
4717
4718<h5>Overview:</h5>
4719
4720<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
4721locations, allowing garbage collector implementations that require write
4722barriers (such as generational or reference counting collectors).</p>
4723
4724<h5>Arguments:</h5>
4725
4726<p>The first argument is the reference to store, the second is the start of the
4727object to store it to, and the third is the address of the field of Obj to
4728store to. If the runtime does not require a pointer to the object, Obj may be
4729null.</p>
4730
4731<h5>Semantics:</h5>
4732
4733<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
4734instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4735garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
4736may only be used in a function which <a href="#gc">specifies a GC
4737algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4738
4739</div>
4740
4741
4742
4743<!-- ======================================================================= -->
4744<div class="doc_subsection">
4745 <a name="int_codegen">Code Generator Intrinsics</a>
4746</div>
4747
4748<div class="doc_text">
4749<p>
4750These intrinsics are provided by LLVM to expose special features that may only
4751be implemented with code generator support.
4752</p>
4753
4754</div>
4755
4756<!-- _______________________________________________________________________ -->
4757<div class="doc_subsubsection">
4758 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
4759</div>
4760
4761<div class="doc_text">
4762
4763<h5>Syntax:</h5>
4764<pre>
4765 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
4766</pre>
4767
4768<h5>Overview:</h5>
4769
4770<p>
4771The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
4772target-specific value indicating the return address of the current function
4773or one of its callers.
4774</p>
4775
4776<h5>Arguments:</h5>
4777
4778<p>
4779The argument to this intrinsic indicates which function to return the address
4780for. Zero indicates the calling function, one indicates its caller, etc. The
4781argument is <b>required</b> to be a constant integer value.
4782</p>
4783
4784<h5>Semantics:</h5>
4785
4786<p>
4787The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
4788the return address of the specified call frame, or zero if it cannot be
4789identified. The value returned by this intrinsic is likely to be incorrect or 0
4790for arguments other than zero, so it should only be used for debugging purposes.
4791</p>
4792
4793<p>
4794Note that calling this intrinsic does not prevent function inlining or other
4795aggressive transformations, so the value returned may not be that of the obvious
4796source-language caller.
4797</p>
4798</div>
4799
4800
4801<!-- _______________________________________________________________________ -->
4802<div class="doc_subsubsection">
4803 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
4804</div>
4805
4806<div class="doc_text">
4807
4808<h5>Syntax:</h5>
4809<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4810 declare i8 *@llvm.frameaddress(i32 &lt;level&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4811</pre>
4812
4813<h5>Overview:</h5>
4814
4815<p>
4816The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
4817target-specific frame pointer value for the specified stack frame.
4818</p>
4819
4820<h5>Arguments:</h5>
4821
4822<p>
4823The argument to this intrinsic indicates which function to return the frame
4824pointer for. Zero indicates the calling function, one indicates its caller,
4825etc. The argument is <b>required</b> to be a constant integer value.
4826</p>
4827
4828<h5>Semantics:</h5>
4829
4830<p>
4831The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
4832the frame address of the specified call frame, or zero if it cannot be
4833identified. The value returned by this intrinsic is likely to be incorrect or 0
4834for arguments other than zero, so it should only be used for debugging purposes.
4835</p>
4836
4837<p>
4838Note that calling this intrinsic does not prevent function inlining or other
4839aggressive transformations, so the value returned may not be that of the obvious
4840source-language caller.
4841</p>
4842</div>
4843
4844<!-- _______________________________________________________________________ -->
4845<div class="doc_subsubsection">
4846 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
4847</div>
4848
4849<div class="doc_text">
4850
4851<h5>Syntax:</h5>
4852<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4853 declare i8 *@llvm.stacksave()
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4854</pre>
4855
4856<h5>Overview:</h5>
4857
4858<p>
4859The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
4860the function stack, for use with <a href="#int_stackrestore">
4861<tt>llvm.stackrestore</tt></a>. This is useful for implementing language
4862features like scoped automatic variable sized arrays in C99.
4863</p>
4864
4865<h5>Semantics:</h5>
4866
4867<p>
4868This intrinsic returns a opaque pointer value that can be passed to <a
4869href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
4870<tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
4871<tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
4872state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
4873practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
4874that were allocated after the <tt>llvm.stacksave</tt> was executed.
4875</p>
4876
4877</div>
4878
4879<!-- _______________________________________________________________________ -->
4880<div class="doc_subsubsection">
4881 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
4882</div>
4883
4884<div class="doc_text">
4885
4886<h5>Syntax:</h5>
4887<pre>
4888 declare void @llvm.stackrestore(i8 * %ptr)
4889</pre>
4890
4891<h5>Overview:</h5>
4892
4893<p>
4894The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
4895the function stack to the state it was in when the corresponding <a
4896href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
4897useful for implementing language features like scoped automatic variable sized
4898arrays in C99.
4899</p>
4900
4901<h5>Semantics:</h5>
4902
4903<p>
4904See the description for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.
4905</p>
4906
4907</div>
4908
4909
4910<!-- _______________________________________________________________________ -->
4911<div class="doc_subsubsection">
4912 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
4913</div>
4914
4915<div class="doc_text">
4916
4917<h5>Syntax:</h5>
4918<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4919 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]4920</pre>
4921
4922<h5>Overview:</h5>
4923
4924
4925<p>
4926The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
4927a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
4928no
4929effect on the behavior of the program but can change its performance
4930characteristics.
4931</p>
4932
4933<h5>Arguments:</h5>
4934
4935<p>
4936<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
4937determining if the fetch should be for a read (0) or write (1), and
4938<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
4939locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
4940<tt>locality</tt> arguments must be constant integers.
4941</p>
4942
4943<h5>Semantics:</h5>
4944
4945<p>
4946This intrinsic does not modify the behavior of the program. In particular,
4947prefetches cannot trap and do not produce a value. On targets that support this
4948intrinsic, the prefetch can provide hints to the processor cache for better
4949performance.
4950</p>
4951
4952</div>
4953
4954<!-- _______________________________________________________________________ -->
4955<div class="doc_subsubsection">
4956 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
4957</div>
4958
4959<div class="doc_text">
4960
4961<h5>Syntax:</h5>
4962<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4963 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4964</pre>
4965
4966<h5>Overview:</h5>
4967
4968
4969<p>
4970The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]4971(PC) in a region of
4972code to simulators and other tools. The method is target specific, but it is
4973expected that the marker will use exported symbols to transmit the PC of the
4974marker.
4975The marker makes no guarantees that it will remain with any specific instruction
4976after optimizations. It is possible that the presence of a marker will inhibit
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4977optimizations. The intended use is to be inserted after optimizations to allow
4978correlations of simulation runs.
4979</p>
4980
4981<h5>Arguments:</h5>
4982
4983<p>
4984<tt>id</tt> is a numerical id identifying the marker.
4985</p>
4986
4987<h5>Semantics:</h5>
4988
4989<p>
4990This intrinsic does not modify the behavior of the program. Backends that do not
4991support this intrinisic may ignore it.
4992</p>
4993
4994</div>
4995
4996<!-- _______________________________________________________________________ -->
4997<div class="doc_subsubsection">
4998 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
4999</div>
5000
5001<div class="doc_text">
5002
5003<h5>Syntax:</h5>
5004<pre>
5005 declare i64 @llvm.readcyclecounter( )
5006</pre>
5007
5008<h5>Overview:</h5>
5009
5010
5011<p>
5012The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
5013counter register (or similar low latency, high accuracy clocks) on those targets
5014that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
5015As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
5016should only be used for small timings.
5017</p>
5018
5019<h5>Semantics:</h5>
5020
5021<p>
5022When directly supported, reading the cycle counter should not modify any memory.
5023Implementations are allowed to either return a application specific value or a
5024system wide value. On backends without support, this is lowered to a constant 0.
5025</p>
5026
5027</div>
5028
5029<!-- ======================================================================= -->
5030<div class="doc_subsection">
5031 <a name="int_libc">Standard C Library Intrinsics</a>
5032</div>
5033
5034<div class="doc_text">
5035<p>
5036LLVM provides intrinsics for a few important standard C library functions.
5037These intrinsics allow source-language front-ends to pass information about the
5038alignment of the pointer arguments to the code generator, providing opportunity
5039for more efficient code generation.
5040</p>
5041
5042</div>
5043
5044<!-- _______________________________________________________________________ -->
5045<div class="doc_subsubsection">
5046 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
5047</div>
5048
5049<div class="doc_text">
5050
5051<h5>Syntax:</h5>
5052<pre>
5053 declare void @llvm.memcpy.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5054 i32 &lt;len&gt;, i32 &lt;align&gt;)
5055 declare void @llvm.memcpy.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5056 i64 &lt;len&gt;, i32 &lt;align&gt;)
5057</pre>
5058
5059<h5>Overview:</h5>
5060
5061<p>
5062The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
5063location to the destination location.
5064</p>
5065
5066<p>
5067Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
5068intrinsics do not return a value, and takes an extra alignment argument.
5069</p>
5070
5071<h5>Arguments:</h5>
5072
5073<p>
5074The first argument is a pointer to the destination, the second is a pointer to
5075the source. The third argument is an integer argument
5076specifying the number of bytes to copy, and the fourth argument is the alignment
5077of the source and destination locations.
5078</p>
5079
5080<p>
5081If the call to this intrinisic has an alignment value that is not 0 or 1, then
5082the caller guarantees that both the source and destination pointers are aligned
5083to that boundary.
5084</p>
5085
5086<h5>Semantics:</h5>
5087
5088<p>
5089The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
5090location to the destination location, which are not allowed to overlap. It
5091copies "len" bytes of memory over. If the argument is known to be aligned to
5092some boundary, this can be specified as the fourth argument, otherwise it should
5093be set to 0 or 1.
5094</p>
5095</div>
5096
5097
5098<!-- _______________________________________________________________________ -->
5099<div class="doc_subsubsection">
5100 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
5101</div>
5102
5103<div class="doc_text">
5104
5105<h5>Syntax:</h5>
5106<pre>
5107 declare void @llvm.memmove.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5108 i32 &lt;len&gt;, i32 &lt;align&gt;)
5109 declare void @llvm.memmove.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5110 i64 &lt;len&gt;, i32 &lt;align&gt;)
5111</pre>
5112
5113<h5>Overview:</h5>
5114
5115<p>
5116The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the source
5117location to the destination location. It is similar to the
Chris Lattnerdba16ea2008-01-06 19:51:52 +0000[diff] [blame]5118'<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to overlap.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5119</p>
5120
5121<p>
5122Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
5123intrinsics do not return a value, and takes an extra alignment argument.
5124</p>
5125
5126<h5>Arguments:</h5>
5127
5128<p>
5129The first argument is a pointer to the destination, the second is a pointer to
5130the source. The third argument is an integer argument
5131specifying the number of bytes to copy, and the fourth argument is the alignment
5132of the source and destination locations.
5133</p>
5134
5135<p>
5136If the call to this intrinisic has an alignment value that is not 0 or 1, then
5137the caller guarantees that the source and destination pointers are aligned to
5138that boundary.
5139</p>
5140
5141<h5>Semantics:</h5>
5142
5143<p>
5144The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the source
5145location to the destination location, which may overlap. It
5146copies "len" bytes of memory over. If the argument is known to be aligned to
5147some boundary, this can be specified as the fourth argument, otherwise it should
5148be set to 0 or 1.
5149</p>
5150</div>
5151
5152
5153<!-- _______________________________________________________________________ -->
5154<div class="doc_subsubsection">
5155 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
5156</div>
5157
5158<div class="doc_text">
5159
5160<h5>Syntax:</h5>
5161<pre>
5162 declare void @llvm.memset.i32(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5163 i32 &lt;len&gt;, i32 &lt;align&gt;)
5164 declare void @llvm.memset.i64(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5165 i64 &lt;len&gt;, i32 &lt;align&gt;)
5166</pre>
5167
5168<h5>Overview:</h5>
5169
5170<p>
5171The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a particular
5172byte value.
5173</p>
5174
5175<p>
5176Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
5177does not return a value, and takes an extra alignment argument.
5178</p>
5179
5180<h5>Arguments:</h5>
5181
5182<p>
5183The first argument is a pointer to the destination to fill, the second is the
5184byte value to fill it with, the third argument is an integer
5185argument specifying the number of bytes to fill, and the fourth argument is the
5186known alignment of destination location.
5187</p>
5188
5189<p>
5190If the call to this intrinisic has an alignment value that is not 0 or 1, then
5191the caller guarantees that the destination pointer is aligned to that boundary.
5192</p>
5193
5194<h5>Semantics:</h5>
5195
5196<p>
5197The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting at
5198the
5199destination location. If the argument is known to be aligned to some boundary,
5200this can be specified as the fourth argument, otherwise it should be set to 0 or
52011.
5202</p>
5203</div>
5204
5205
5206<!-- _______________________________________________________________________ -->
5207<div class="doc_subsubsection">
5208 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
5209</div>
5210
5211<div class="doc_text">
5212
5213<h5>Syntax:</h5>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5214<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]5215floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5216types however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5217<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5218 declare float @llvm.sqrt.f32(float %Val)
5219 declare double @llvm.sqrt.f64(double %Val)
5220 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
5221 declare fp128 @llvm.sqrt.f128(fp128 %Val)
5222 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5223</pre>
5224
5225<h5>Overview:</h5>
5226
5227<p>
5228The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5229returning the same value as the libm '<tt>sqrt</tt>' functions would. Unlike
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5230<tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
Chris Lattnerf914a002008-01-29 07:00:44 +0000[diff] [blame]5231negative numbers other than -0.0 (which allows for better optimization, because
5232there is no need to worry about errno being set). <tt>llvm.sqrt(-0.0)</tt> is
5233defined to return -0.0 like IEEE sqrt.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5234</p>
5235
5236<h5>Arguments:</h5>
5237
5238<p>
5239The argument and return value are floating point numbers of the same type.
5240</p>
5241
5242<h5>Semantics:</h5>
5243
5244<p>
5245This function returns the sqrt of the specified operand if it is a nonnegative
5246floating point number.
5247</p>
5248</div>
5249
5250<!-- _______________________________________________________________________ -->
5251<div class="doc_subsubsection">
5252 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
5253</div>
5254
5255<div class="doc_text">
5256
5257<h5>Syntax:</h5>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5258<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]5259floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5260types however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5261<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5262 declare float @llvm.powi.f32(float %Val, i32 %power)
5263 declare double @llvm.powi.f64(double %Val, i32 %power)
5264 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
5265 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
5266 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5267</pre>
5268
5269<h5>Overview:</h5>
5270
5271<p>
5272The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
5273specified (positive or negative) power. The order of evaluation of
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5274multiplications is not defined. When a vector of floating point type is
5275used, the second argument remains a scalar integer value.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5276</p>
5277
5278<h5>Arguments:</h5>
5279
5280<p>
5281The second argument is an integer power, and the first is a value to raise to
5282that power.
5283</p>
5284
5285<h5>Semantics:</h5>
5286
5287<p>
5288This function returns the first value raised to the second power with an
5289unspecified sequence of rounding operations.</p>
5290</div>
5291
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5292<!-- _______________________________________________________________________ -->
5293<div class="doc_subsubsection">
5294 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
5295</div>
5296
5297<div class="doc_text">
5298
5299<h5>Syntax:</h5>
5300<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
5301floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5302types however.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5303<pre>
5304 declare float @llvm.sin.f32(float %Val)
5305 declare double @llvm.sin.f64(double %Val)
5306 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
5307 declare fp128 @llvm.sin.f128(fp128 %Val)
5308 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
5309</pre>
5310
5311<h5>Overview:</h5>
5312
5313<p>
5314The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.
5315</p>
5316
5317<h5>Arguments:</h5>
5318
5319<p>
5320The argument and return value are floating point numbers of the same type.
5321</p>
5322
5323<h5>Semantics:</h5>
5324
5325<p>
5326This function returns the sine of the specified operand, returning the
5327same values as the libm <tt>sin</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5328conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5329</div>
5330
5331<!-- _______________________________________________________________________ -->
5332<div class="doc_subsubsection">
5333 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
5334</div>
5335
5336<div class="doc_text">
5337
5338<h5>Syntax:</h5>
5339<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
5340floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5341types however.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5342<pre>
5343 declare float @llvm.cos.f32(float %Val)
5344 declare double @llvm.cos.f64(double %Val)
5345 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
5346 declare fp128 @llvm.cos.f128(fp128 %Val)
5347 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
5348</pre>
5349
5350<h5>Overview:</h5>
5351
5352<p>
5353The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.
5354</p>
5355
5356<h5>Arguments:</h5>
5357
5358<p>
5359The argument and return value are floating point numbers of the same type.
5360</p>
5361
5362<h5>Semantics:</h5>
5363
5364<p>
5365This function returns the cosine of the specified operand, returning the
5366same values as the libm <tt>cos</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5367conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5368</div>
5369
5370<!-- _______________________________________________________________________ -->
5371<div class="doc_subsubsection">
5372 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
5373</div>
5374
5375<div class="doc_text">
5376
5377<h5>Syntax:</h5>
5378<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
5379floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5380types however.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5381<pre>
5382 declare float @llvm.pow.f32(float %Val, float %Power)
5383 declare double @llvm.pow.f64(double %Val, double %Power)
5384 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
5385 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
5386 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
5387</pre>
5388
5389<h5>Overview:</h5>
5390
5391<p>
5392The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
5393specified (positive or negative) power.
5394</p>
5395
5396<h5>Arguments:</h5>
5397
5398<p>
5399The second argument is a floating point power, and the first is a value to
5400raise to that power.
5401</p>
5402
5403<h5>Semantics:</h5>
5404
5405<p>
5406This function returns the first value raised to the second power,
5407returning the
5408same values as the libm <tt>pow</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5409conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5410</div>
5411
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5412
5413<!-- ======================================================================= -->
5414<div class="doc_subsection">
5415 <a name="int_manip">Bit Manipulation Intrinsics</a>
5416</div>
5417
5418<div class="doc_text">
5419<p>
5420LLVM provides intrinsics for a few important bit manipulation operations.
5421These allow efficient code generation for some algorithms.
5422</p>
5423
5424</div>
5425
5426<!-- _______________________________________________________________________ -->
5427<div class="doc_subsubsection">
5428 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
5429</div>
5430
5431<div class="doc_text">
5432
5433<h5>Syntax:</h5>
5434<p>This is an overloaded intrinsic function. You can use bswap on any integer
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5435type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5436<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5437 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
5438 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
5439 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5440</pre>
5441
5442<h5>Overview:</h5>
5443
5444<p>
5445The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
5446values with an even number of bytes (positive multiple of 16 bits). These are
5447useful for performing operations on data that is not in the target's native
5448byte order.
5449</p>
5450
5451<h5>Semantics:</h5>
5452
5453<p>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5454The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5455and low byte of the input i16 swapped. Similarly, the <tt>llvm.bswap.i32</tt>
5456intrinsic returns an i32 value that has the four bytes of the input i32
5457swapped, 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]5458i32 will have its bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i48</tt>,
5459<tt>llvm.bswap.i64</tt> and other intrinsics extend this concept to
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5460additional even-byte lengths (6 bytes, 8 bytes and more, respectively).
5461</p>
5462
5463</div>
5464
5465<!-- _______________________________________________________________________ -->
5466<div class="doc_subsubsection">
5467 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
5468</div>
5469
5470<div class="doc_text">
5471
5472<h5>Syntax:</h5>
5473<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]5474width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5475<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5476 declare i8 @llvm.ctpop.i8 (i8 &lt;src&gt;)
5477 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5478 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5479 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
5480 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5481</pre>
5482
5483<h5>Overview:</h5>
5484
5485<p>
5486The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
5487value.
5488</p>
5489
5490<h5>Arguments:</h5>
5491
5492<p>
5493The only argument is the value to be counted. The argument may be of any
5494integer type. The return type must match the argument type.
5495</p>
5496
5497<h5>Semantics:</h5>
5498
5499<p>
5500The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
5501</p>
5502</div>
5503
5504<!-- _______________________________________________________________________ -->
5505<div class="doc_subsubsection">
5506 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
5507</div>
5508
5509<div class="doc_text">
5510
5511<h5>Syntax:</h5>
5512<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]5513integer bit width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5514<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5515 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
5516 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5517 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5518 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
5519 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5520</pre>
5521
5522<h5>Overview:</h5>
5523
5524<p>
5525The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
5526leading zeros in a variable.
5527</p>
5528
5529<h5>Arguments:</h5>
5530
5531<p>
5532The only argument is the value to be counted. The argument may be of any
5533integer type. The return type must match the argument type.
5534</p>
5535
5536<h5>Semantics:</h5>
5537
5538<p>
5539The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
5540in a variable. If the src == 0 then the result is the size in bits of the type
5541of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.
5542</p>
5543</div>
5544
5545
5546
5547<!-- _______________________________________________________________________ -->
5548<div class="doc_subsubsection">
5549 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
5550</div>
5551
5552<div class="doc_text">
5553
5554<h5>Syntax:</h5>
5555<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]5556integer bit width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5557<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5558 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
5559 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5560 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5561 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
5562 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5563</pre>
5564
5565<h5>Overview:</h5>
5566
5567<p>
5568The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
5569trailing zeros.
5570</p>
5571
5572<h5>Arguments:</h5>
5573
5574<p>
5575The only argument is the value to be counted. The argument may be of any
5576integer type. The return type must match the argument type.
5577</p>
5578
5579<h5>Semantics:</h5>
5580
5581<p>
5582The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
5583in a variable. If the src == 0 then the result is the size in bits of the type
5584of src. For example, <tt>llvm.cttz(2) = 1</tt>.
5585</p>
5586</div>
5587
5588<!-- _______________________________________________________________________ -->
5589<div class="doc_subsubsection">
5590 <a name="int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic</a>
5591</div>
5592
5593<div class="doc_text">
5594
5595<h5>Syntax:</h5>
5596<p>This is an overloaded intrinsic. You can use <tt>llvm.part.select</tt>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5597on any integer bit width.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5598<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5599 declare i17 @llvm.part.select.i17 (i17 %val, i32 %loBit, i32 %hiBit)
5600 declare i29 @llvm.part.select.i29 (i29 %val, i32 %loBit, i32 %hiBit)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5601</pre>
5602
5603<h5>Overview:</h5>
5604<p>The '<tt>llvm.part.select</tt>' family of intrinsic functions selects a
5605range of bits from an integer value and returns them in the same bit width as
5606the original value.</p>
5607
5608<h5>Arguments:</h5>
5609<p>The first argument, <tt>%val</tt> and the result may be integer types of
5610any bit width but they must have the same bit width. The second and third
5611arguments must be <tt>i32</tt> type since they specify only a bit index.</p>
5612
5613<h5>Semantics:</h5>
5614<p>The operation of the '<tt>llvm.part.select</tt>' intrinsic has two modes
5615of operation: forwards and reverse. If <tt>%loBit</tt> is greater than
5616<tt>%hiBits</tt> then the intrinsic operates in reverse mode. Otherwise it
5617operates in forward mode.</p>
5618<p>In forward mode, this intrinsic is the equivalent of shifting <tt>%val</tt>
5619right by <tt>%loBit</tt> bits and then ANDing it with a mask with
5620only the <tt>%hiBit - %loBit</tt> bits set, as follows:</p>
5621<ol>
5622 <li>The <tt>%val</tt> is shifted right (LSHR) by the number of bits specified
5623 by <tt>%loBits</tt>. This normalizes the value to the low order bits.</li>
5624 <li>The <tt>%loBits</tt> value is subtracted from the <tt>%hiBits</tt> value
5625 to determine the number of bits to retain.</li>
5626 <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]5627 <li>The mask is ANDed with <tt>%val</tt> to produce the result.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5628</ol>
5629<p>In reverse mode, a similar computation is made except that the bits are
5630returned in the reverse order. So, for example, if <tt>X</tt> has the value
5631<tt>i16 0x0ACF (101011001111)</tt> and we apply
5632<tt>part.select(i16 X, 8, 3)</tt> to it, we get back the value
5633<tt>i16 0x0026 (000000100110)</tt>.</p>
5634</div>
5635
5636<div class="doc_subsubsection">
5637 <a name="int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic</a>
5638</div>
5639
5640<div class="doc_text">
5641
5642<h5>Syntax:</h5>
5643<p>This is an overloaded intrinsic. You can use <tt>llvm.part.set</tt>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5644on any integer bit width.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5645<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5646 declare i17 @llvm.part.set.i17.i9 (i17 %val, i9 %repl, i32 %lo, i32 %hi)
5647 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]5648</pre>
5649
5650<h5>Overview:</h5>
5651<p>The '<tt>llvm.part.set</tt>' family of intrinsic functions replaces a range
5652of bits in an integer value with another integer value. It returns the integer
5653with the replaced bits.</p>
5654
5655<h5>Arguments:</h5>
5656<p>The first argument, <tt>%val</tt> and the result may be integer types of
5657any bit width but they must have the same bit width. <tt>%val</tt> is the value
5658whose bits will be replaced. The second argument, <tt>%repl</tt> may be an
5659integer of any bit width. The third and fourth arguments must be <tt>i32</tt>
5660type since they specify only a bit index.</p>
5661
5662<h5>Semantics:</h5>
5663<p>The operation of the '<tt>llvm.part.set</tt>' intrinsic has two modes
5664of operation: forwards and reverse. If <tt>%lo</tt> is greater than
5665<tt>%hi</tt> then the intrinsic operates in reverse mode. Otherwise it
5666operates in forward mode.</p>
5667<p>For both modes, the <tt>%repl</tt> value is prepared for use by either
5668truncating it down to the size of the replacement area or zero extending it
5669up to that size.</p>
5670<p>In forward mode, the bits between <tt>%lo</tt> and <tt>%hi</tt> (inclusive)
5671are replaced with corresponding bits from <tt>%repl</tt>. That is the 0th bit
5672in <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]5673to the <tt>%hi</tt>th bit.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5674<p>In reverse mode, a similar computation is made except that the bits are
5675reversed. That is, the <tt>0</tt>th bit in <tt>%repl</tt> replaces the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5676<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]5677<h5>Examples:</h5>
5678<pre>
5679 llvm.part.set(0xFFFF, 0, 4, 7) -&gt; 0xFF0F
5680 llvm.part.set(0xFFFF, 0, 7, 4) -&gt; 0xFF0F
5681 llvm.part.set(0xFFFF, 1, 7, 4) -&gt; 0xFF8F
5682 llvm.part.set(0xFFFF, F, 8, 3) -&gt; 0xFFE7
5683 llvm.part.set(0xFFFF, 0, 3, 8) -&gt; 0xFE07
5684</pre>
5685</div>
5686
5687<!-- ======================================================================= -->
5688<div class="doc_subsection">
5689 <a name="int_debugger">Debugger Intrinsics</a>
5690</div>
5691
5692<div class="doc_text">
5693<p>
5694The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
5695are described in the <a
5696href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
5697Debugging</a> document.
5698</p>
5699</div>
5700
5701
5702<!-- ======================================================================= -->
5703<div class="doc_subsection">
5704 <a name="int_eh">Exception Handling Intrinsics</a>
5705</div>
5706
5707<div class="doc_text">
5708<p> The LLVM exception handling intrinsics (which all start with
5709<tt>llvm.eh.</tt> prefix), are described in the <a
5710href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
5711Handling</a> document. </p>
5712</div>
5713
5714<!-- ======================================================================= -->
5715<div class="doc_subsection">
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5716 <a name="int_trampoline">Trampoline Intrinsic</a>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5717</div>
5718
5719<div class="doc_text">
5720<p>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5721 This intrinsic makes it possible to excise one parameter, marked with
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5722 the <tt>nest</tt> attribute, from a function. The result is a callable
5723 function pointer lacking the nest parameter - the caller does not need
5724 to provide a value for it. Instead, the value to use is stored in
5725 advance in a "trampoline", a block of memory usually allocated
5726 on the stack, which also contains code to splice the nest value into the
5727 argument list. This is used to implement the GCC nested function address
5728 extension.
5729</p>
5730<p>
5731 For example, if the function is
5732 <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]5733 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]5734<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5735 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
5736 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
5737 %p = call i8* @llvm.init.trampoline( i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8*), i8* %nval )
5738 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5739</pre>
Bill Wendlinge262b4c2007-09-22 09:23:55 +0000[diff] [blame]5740 <p>The call <tt>%val = call i32 %fp( i32 %x, i32 %y )</tt> is then equivalent
5741 to <tt>%val = call i32 %f( i8* %nval, i32 %x, i32 %y )</tt>.</p>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5742</div>
5743
5744<!-- _______________________________________________________________________ -->
5745<div class="doc_subsubsection">
5746 <a name="int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a>
5747</div>
5748<div class="doc_text">
5749<h5>Syntax:</h5>
5750<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5751declare 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]5752</pre>
5753<h5>Overview:</h5>
5754<p>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5755 This fills the memory pointed to by <tt>tramp</tt> with code
5756 and returns a function pointer suitable for executing it.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5757</p>
5758<h5>Arguments:</h5>
5759<p>
5760 The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
5761 pointers. The <tt>tramp</tt> argument must point to a sufficiently large
5762 and sufficiently aligned block of memory; this memory is written to by the
Duncan Sands35012212007-08-22 23:39:54 +0000[diff] [blame]5763 intrinsic. Note that the size and the alignment are target-specific - LLVM
5764 currently provides no portable way of determining them, so a front-end that
5765 generates this intrinsic needs to have some target-specific knowledge.
5766 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]5767</p>
5768<h5>Semantics:</h5>
5769<p>
5770 The block of memory pointed to by <tt>tramp</tt> is filled with target
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5771 dependent code, turning it into a function. A pointer to this function is
5772 returned, but needs to be bitcast to an
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5773 <a href="#int_trampoline">appropriate function pointer type</a>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5774 before being called. The new function's signature is the same as that of
5775 <tt>func</tt> with any arguments marked with the <tt>nest</tt> attribute
5776 removed. At most one such <tt>nest</tt> argument is allowed, and it must be
5777 of pointer type. Calling the new function is equivalent to calling
5778 <tt>func</tt> with the same argument list, but with <tt>nval</tt> used for the
5779 missing <tt>nest</tt> argument. If, after calling
5780 <tt>llvm.init.trampoline</tt>, the memory pointed to by <tt>tramp</tt> is
5781 modified, then the effect of any later call to the returned function pointer is
5782 undefined.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5783</p>
5784</div>
5785
5786<!-- ======================================================================= -->
5787<div class="doc_subsection">
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]5788 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
5789</div>
5790
5791<div class="doc_text">
5792<p>
5793 These intrinsic functions expand the "universal IR" of LLVM to represent
5794 hardware constructs for atomic operations and memory synchronization. This
5795 provides an interface to the hardware, not an interface to the programmer. It
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]5796 is aimed at a low enough level to allow any programming models or APIs
5797 (Application Programming Interfaces) which
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]5798 need atomic behaviors to map cleanly onto it. It is also modeled primarily on
5799 hardware behavior. Just as hardware provides a "universal IR" for source
5800 languages, it also provides a starting point for developing a "universal"
5801 atomic operation and synchronization IR.
5802</p>
5803<p>
5804 These do <em>not</em> form an API such as high-level threading libraries,
5805 software transaction memory systems, atomic primitives, and intrinsic
5806 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
5807 application libraries. The hardware interface provided by LLVM should allow
5808 a clean implementation of all of these APIs and parallel programming models.
5809 No one model or paradigm should be selected above others unless the hardware
5810 itself ubiquitously does so.
5811
5812</p>
5813</div>
5814
5815<!-- _______________________________________________________________________ -->
5816<div class="doc_subsubsection">
5817 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
5818</div>
5819<div class="doc_text">
5820<h5>Syntax:</h5>
5821<pre>
5822declare void @llvm.memory.barrier( i1 &lt;ll&gt;, i1 &lt;ls&gt;, i1 &lt;sl&gt;, i1 &lt;ss&gt;,
5823i1 &lt;device&gt; )
5824
5825</pre>
5826<h5>Overview:</h5>
5827<p>
5828 The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
5829 specific pairs of memory access types.
5830</p>
5831<h5>Arguments:</h5>
5832<p>
5833 The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
5834 The first four arguments enables a specific barrier as listed below. The fith
5835 argument specifies that the barrier applies to io or device or uncached memory.
5836
5837</p>
5838 <ul>
5839 <li><tt>ll</tt>: load-load barrier</li>
5840 <li><tt>ls</tt>: load-store barrier</li>
5841 <li><tt>sl</tt>: store-load barrier</li>
5842 <li><tt>ss</tt>: store-store barrier</li>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5843 <li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]5844 </ul>
5845<h5>Semantics:</h5>
5846<p>
5847 This intrinsic causes the system to enforce some ordering constraints upon
5848 the loads and stores of the program. This barrier does not indicate
5849 <em>when</em> any events will occur, it only enforces an <em>order</em> in
5850 which they occur. For any of the specified pairs of load and store operations
5851 (f.ex. load-load, or store-load), all of the first operations preceding the
5852 barrier will complete before any of the second operations succeeding the
5853 barrier begin. Specifically the semantics for each pairing is as follows:
5854</p>
5855 <ul>
5856 <li><tt>ll</tt>: All loads before the barrier must complete before any load
5857 after the barrier begins.</li>
5858
5859 <li><tt>ls</tt>: All loads before the barrier must complete before any
5860 store after the barrier begins.</li>
5861 <li><tt>ss</tt>: All stores before the barrier must complete before any
5862 store after the barrier begins.</li>
5863 <li><tt>sl</tt>: All stores before the barrier must complete before any
5864 load after the barrier begins.</li>
5865 </ul>
5866<p>
5867 These semantics are applied with a logical "and" behavior when more than one
5868 is enabled in a single memory barrier intrinsic.
5869</p>
5870<p>
5871 Backends may implement stronger barriers than those requested when they do not
5872 support as fine grained a barrier as requested. Some architectures do not
5873 need all types of barriers and on such architectures, these become noops.
5874</p>
5875<h5>Example:</h5>
5876<pre>
5877%ptr = malloc i32
5878 store i32 4, %ptr
5879
5880%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
5881 call void @llvm.memory.barrier( i1 false, i1 true, i1 false, i1 false )
5882 <i>; guarantee the above finishes</i>
5883 store i32 8, %ptr <i>; before this begins</i>
5884</pre>
5885</div>
5886
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5887<!-- _______________________________________________________________________ -->
5888<div class="doc_subsubsection">
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]5889 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5890</div>
5891<div class="doc_text">
5892<h5>Syntax:</h5>
5893<p>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]5894 This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
5895 any integer bit width and for different address spaces. Not all targets
5896 support all bit widths however.</p>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5897
5898<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]5899declare i8 @llvm.atomic.cmp.swap.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt; )
5900declare i16 @llvm.atomic.cmp.swap.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt; )
5901declare i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt; )
5902declare 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]5903
5904</pre>
5905<h5>Overview:</h5>
5906<p>
5907 This loads a value in memory and compares it to a given value. If they are
5908 equal, it stores a new value into the memory.
5909</p>
5910<h5>Arguments:</h5>
5911<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]5912 The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result as
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5913 well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
5914 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
5915 this integer type. While any bit width integer may be used, targets may only
5916 lower representations they support in hardware.
5917
5918</p>
5919<h5>Semantics:</h5>
5920<p>
5921 This entire intrinsic must be executed atomically. It first loads the value
5922 in memory pointed to by <tt>ptr</tt> and compares it with the value
5923 <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the memory. The
5924 loaded value is yielded in all cases. This provides the equivalent of an
5925 atomic compare-and-swap operation within the SSA framework.
5926</p>
5927<h5>Examples:</h5>
5928
5929<pre>
5930%ptr = malloc i32
5931 store i32 4, %ptr
5932
5933%val1 = add i32 4, 4
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]5934%result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 4, %val1 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5935 <i>; yields {i32}:result1 = 4</i>
5936%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
5937%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
5938
5939%val2 = add i32 1, 1
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]5940%result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 5, %val2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5941 <i>; yields {i32}:result2 = 8</i>
5942%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
5943
5944%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
5945</pre>
5946</div>
5947
5948<!-- _______________________________________________________________________ -->
5949<div class="doc_subsubsection">
5950 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
5951</div>
5952<div class="doc_text">
5953<h5>Syntax:</h5>
5954
5955<p>
5956 This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
5957 integer bit width. Not all targets support all bit widths however.</p>
5958<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]5959declare i8 @llvm.atomic.swap.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;val&gt; )
5960declare i16 @llvm.atomic.swap.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;val&gt; )
5961declare i32 @llvm.atomic.swap.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;val&gt; )
5962declare i64 @llvm.atomic.swap.i64.p0i64( i64* &lt;ptr&gt;, i64 &lt;val&gt; )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5963
5964</pre>
5965<h5>Overview:</h5>
5966<p>
5967 This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
5968 the value from memory. It then stores the value in <tt>val</tt> in the memory
5969 at <tt>ptr</tt>.
5970</p>
5971<h5>Arguments:</h5>
5972
5973<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]5974 The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both the
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5975 <tt>val</tt> argument and the result must be integers of the same bit width.
5976 The first argument, <tt>ptr</tt>, must be a pointer to a value of this
5977 integer type. The targets may only lower integer representations they
5978 support.
5979</p>
5980<h5>Semantics:</h5>
5981<p>
5982 This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
5983 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
5984 equivalent of an atomic swap operation within the SSA framework.
5985
5986</p>
5987<h5>Examples:</h5>
5988<pre>
5989%ptr = malloc i32
5990 store i32 4, %ptr
5991
5992%val1 = add i32 4, 4
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]5993%result1 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val1 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5994 <i>; yields {i32}:result1 = 4</i>
5995%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
5996%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
5997
5998%val2 = add i32 1, 1
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]5999%result2 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6000 <i>; yields {i32}:result2 = 8</i>
6001
6002%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
6003%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
6004</pre>
6005</div>
6006
6007<!-- _______________________________________________________________________ -->
6008<div class="doc_subsubsection">
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6009 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6010
6011</div>
6012<div class="doc_text">
6013<h5>Syntax:</h5>
6014<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6015 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]6016 integer bit width. Not all targets support all bit widths however.</p>
6017<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6018declare i8 @llvm.atomic.load.add.i8..p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6019declare i16 @llvm.atomic.load.add.i16..p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6020declare i32 @llvm.atomic.load.add.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6021declare 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]6022
6023</pre>
6024<h5>Overview:</h5>
6025<p>
6026 This intrinsic adds <tt>delta</tt> to the value stored in memory at
6027 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
6028</p>
6029<h5>Arguments:</h5>
6030<p>
6031
6032 The intrinsic takes two arguments, the first a pointer to an integer value
6033 and the second an integer value. The result is also an integer value. These
6034 integer types can have any bit width, but they must all have the same bit
6035 width. The targets may only lower integer representations they support.
6036</p>
6037<h5>Semantics:</h5>
6038<p>
6039 This intrinsic does a series of operations atomically. It first loads the
6040 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
6041 to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
6042</p>
6043
6044<h5>Examples:</h5>
6045<pre>
6046%ptr = malloc i32
6047 store i32 4, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6048%result1 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 4 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6049 <i>; yields {i32}:result1 = 4</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6050%result2 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6051 <i>; yields {i32}:result2 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6052%result3 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 5 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6053 <i>; yields {i32}:result3 = 10</i>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6054%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6055</pre>
6056</div>
6057
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6058<!-- _______________________________________________________________________ -->
6059<div class="doc_subsubsection">
6060 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
6061
6062</div>
6063<div class="doc_text">
6064<h5>Syntax:</h5>
6065<p>
6066 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]6067 any integer bit width and for different address spaces. Not all targets
6068 support all bit widths however.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6069<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6070declare i8 @llvm.atomic.load.sub.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6071declare i16 @llvm.atomic.load.sub.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6072declare i32 @llvm.atomic.load.sub.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6073declare 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]6074
6075</pre>
6076<h5>Overview:</h5>
6077<p>
6078 This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
6079 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
6080</p>
6081<h5>Arguments:</h5>
6082<p>
6083
6084 The intrinsic takes two arguments, the first a pointer to an integer value
6085 and the second an integer value. The result is also an integer value. These
6086 integer types can have any bit width, but they must all have the same bit
6087 width. The targets may only lower integer representations they support.
6088</p>
6089<h5>Semantics:</h5>
6090<p>
6091 This intrinsic does a series of operations atomically. It first loads the
6092 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
6093 result to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
6094</p>
6095
6096<h5>Examples:</h5>
6097<pre>
6098%ptr = malloc i32
6099 store i32 8, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6100%result1 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 4 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6101 <i>; yields {i32}:result1 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6102%result2 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 2 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6103 <i>; yields {i32}:result2 = 4</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6104%result3 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 5 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6105 <i>; yields {i32}:result3 = 2</i>
6106%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
6107</pre>
6108</div>
6109
6110<!-- _______________________________________________________________________ -->
6111<div class="doc_subsubsection">
6112 <a name="int_atomic_load_and">'<tt>llvm.atomic.load.and.*</tt>' Intrinsic</a><br>
6113 <a name="int_atomic_load_nand">'<tt>llvm.atomic.load.nand.*</tt>' Intrinsic</a><br>
6114 <a name="int_atomic_load_or">'<tt>llvm.atomic.load.or.*</tt>' Intrinsic</a><br>
6115 <a name="int_atomic_load_xor">'<tt>llvm.atomic.load.xor.*</tt>' Intrinsic</a><br>
6116
6117</div>
6118<div class="doc_text">
6119<h5>Syntax:</h5>
6120<p>
6121 These are overloaded intrinsics. You can use <tt>llvm.atomic.load_and</tt>,
6122 <tt>llvm.atomic.load_nand</tt>, <tt>llvm.atomic.load_or</tt>, and
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6123 <tt>llvm.atomic.load_xor</tt> on any integer bit width and for different
6124 address spaces. Not all targets support all bit widths however.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6125<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6126declare i8 @llvm.atomic.load.and.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6127declare i16 @llvm.atomic.load.and.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6128declare i32 @llvm.atomic.load.and.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6129declare 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]6130
6131</pre>
6132
6133<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6134declare i8 @llvm.atomic.load.or.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6135declare i16 @llvm.atomic.load.or.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6136declare i32 @llvm.atomic.load.or.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6137declare 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]6138
6139</pre>
6140
6141<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6142declare i8 @llvm.atomic.load.nand.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6143declare i16 @llvm.atomic.load.nand.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6144declare i32 @llvm.atomic.load.nand.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6145declare 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]6146
6147</pre>
6148
6149<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6150declare i8 @llvm.atomic.load.xor.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6151declare i16 @llvm.atomic.load.xor.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6152declare i32 @llvm.atomic.load.xor.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6153declare 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]6154
6155</pre>
6156<h5>Overview:</h5>
6157<p>
6158 These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
6159 the value stored in memory at <tt>ptr</tt>. It yields the original value
6160 at <tt>ptr</tt>.
6161</p>
6162<h5>Arguments:</h5>
6163<p>
6164
6165 These intrinsics take two arguments, the first a pointer to an integer value
6166 and the second an integer value. The result is also an integer value. These
6167 integer types can have any bit width, but they must all have the same bit
6168 width. The targets may only lower integer representations they support.
6169</p>
6170<h5>Semantics:</h5>
6171<p>
6172 These intrinsics does a series of operations atomically. They first load the
6173 value stored at <tt>ptr</tt>. They then do the bitwise operation
6174 <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the original
6175 value stored at <tt>ptr</tt>.
6176</p>
6177
6178<h5>Examples:</h5>
6179<pre>
6180%ptr = malloc i32
6181 store i32 0x0F0F, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6182%result0 = call i32 @llvm.atomic.load.nand.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6183 <i>; yields {i32}:result0 = 0x0F0F</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6184%result1 = call i32 @llvm.atomic.load.and.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6185 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6186%result2 = call i32 @llvm.atomic.load.or.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6187 <i>; yields {i32}:result2 = 0xF0</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6188%result3 = call i32 @llvm.atomic.load.xor.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6189 <i>; yields {i32}:result3 = FF</i>
6190%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
6191</pre>
6192</div>
6193
6194
6195<!-- _______________________________________________________________________ -->
6196<div class="doc_subsubsection">
6197 <a name="int_atomic_load_max">'<tt>llvm.atomic.load.max.*</tt>' Intrinsic</a><br>
6198 <a name="int_atomic_load_min">'<tt>llvm.atomic.load.min.*</tt>' Intrinsic</a><br>
6199 <a name="int_atomic_load_umax">'<tt>llvm.atomic.load.umax.*</tt>' Intrinsic</a><br>
6200 <a name="int_atomic_load_umin">'<tt>llvm.atomic.load.umin.*</tt>' Intrinsic</a><br>
6201
6202</div>
6203<div class="doc_text">
6204<h5>Syntax:</h5>
6205<p>
6206 These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
6207 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6208 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
6209 address spaces. Not all targets
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6210 support all bit widths however.</p>
6211<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6212declare i8 @llvm.atomic.load.max.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6213declare i16 @llvm.atomic.load.max.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6214declare i32 @llvm.atomic.load.max.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6215declare 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]6216
6217</pre>
6218
6219<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6220declare i8 @llvm.atomic.load.min.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6221declare i16 @llvm.atomic.load.min.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6222declare i32 @llvm.atomic.load.min.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6223declare 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]6224
6225</pre>
6226
6227<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6228declare i8 @llvm.atomic.load.umax.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6229declare i16 @llvm.atomic.load.umax.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6230declare i32 @llvm.atomic.load.umax.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6231declare 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]6232
6233</pre>
6234
6235<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6236declare i8 @llvm.atomic.load.umin.i8..p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6237declare i16 @llvm.atomic.load.umin.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6238declare i32 @llvm.atomic.load.umin.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6239declare 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]6240
6241</pre>
6242<h5>Overview:</h5>
6243<p>
6244 These intrinsics takes the signed or unsigned minimum or maximum of
6245 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
6246 original value at <tt>ptr</tt>.
6247</p>
6248<h5>Arguments:</h5>
6249<p>
6250
6251 These intrinsics take two arguments, the first a pointer to an integer value
6252 and the second an integer value. The result is also an integer value. These
6253 integer types can have any bit width, but they must all have the same bit
6254 width. The targets may only lower integer representations they support.
6255</p>
6256<h5>Semantics:</h5>
6257<p>
6258 These intrinsics does a series of operations atomically. They first load the
6259 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or max
6260 <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They yield
6261 the original value stored at <tt>ptr</tt>.
6262</p>
6263
6264<h5>Examples:</h5>
6265<pre>
6266%ptr = malloc i32
6267 store i32 7, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6268%result0 = call i32 @llvm.atomic.load.min.i32.p0i32( i32* %ptr, i32 -2 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6269 <i>; yields {i32}:result0 = 7</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6270%result1 = call i32 @llvm.atomic.load.max.i32.p0i32( i32* %ptr, i32 8 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6271 <i>; yields {i32}:result1 = -2</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6272%result2 = call i32 @llvm.atomic.load.umin.i32.p0i32( i32* %ptr, i32 10 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6273 <i>; yields {i32}:result2 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6274%result3 = call i32 @llvm.atomic.load.umax.i32.p0i32( i32* %ptr, i32 30 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6275 <i>; yields {i32}:result3 = 8</i>
6276%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
6277</pre>
6278</div>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6279
6280<!-- ======================================================================= -->
6281<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6282 <a name="int_general">General Intrinsics</a>
6283</div>
6284
6285<div class="doc_text">
6286<p> This class of intrinsics is designed to be generic and has
6287no specific purpose. </p>
6288</div>
6289
6290<!-- _______________________________________________________________________ -->
6291<div class="doc_subsubsection">
6292 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
6293</div>
6294
6295<div class="doc_text">
6296
6297<h5>Syntax:</h5>
6298<pre>
6299 declare void @llvm.var.annotation(i8* &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6300</pre>
6301
6302<h5>Overview:</h5>
6303
6304<p>
6305The '<tt>llvm.var.annotation</tt>' intrinsic
6306</p>
6307
6308<h5>Arguments:</h5>
6309
6310<p>
6311The first argument is a pointer to a value, the second is a pointer to a
6312global string, the third is a pointer to a global string which is the source
6313file name, and the last argument is the line number.
6314</p>
6315
6316<h5>Semantics:</h5>
6317
6318<p>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]6319This intrinsic allows annotation of local variables with arbitrary strings.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6320This can be useful for special purpose optimizations that want to look for these
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]6321annotations. These have no other defined use, they are ignored by code
6322generation and optimization.
6323</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6324</div>
6325
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6326<!-- _______________________________________________________________________ -->
6327<div class="doc_subsubsection">
Tanya Lattnerc9869b12007-09-21 23:57:59 +0000[diff] [blame]6328 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6329</div>
6330
6331<div class="doc_text">
6332
6333<h5>Syntax:</h5>
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]6334<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
6335any integer bit width.
6336</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6337<pre>
Tanya Lattner09161fe2007-09-22 00:03:01 +0000[diff] [blame]6338 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6339 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6340 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6341 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6342 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]6343</pre>
6344
6345<h5>Overview:</h5>
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]6346
6347<p>
6348The '<tt>llvm.annotation</tt>' intrinsic.
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6349</p>
6350
6351<h5>Arguments:</h5>
6352
6353<p>
6354The first argument is an integer value (result of some expression),
6355the second is a pointer to a global string, the third is a pointer to a global
6356string which is the source file name, and the last argument is the line number.
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]6357It returns the value of the first argument.
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6358</p>
6359
6360<h5>Semantics:</h5>
6361
6362<p>
6363This intrinsic allows annotations to be put on arbitrary expressions
6364with arbitrary strings. This can be useful for special purpose optimizations
6365that want to look for these annotations. These have no other defined use, they
6366are ignored by code generation and optimization.
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]6367</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6368</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6369
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]6370<!-- _______________________________________________________________________ -->
6371<div class="doc_subsubsection">
6372 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
6373</div>
6374
6375<div class="doc_text">
6376
6377<h5>Syntax:</h5>
6378<pre>
6379 declare void @llvm.trap()
6380</pre>
6381
6382<h5>Overview:</h5>
6383
6384<p>
6385The '<tt>llvm.trap</tt>' intrinsic
6386</p>
6387
6388<h5>Arguments:</h5>
6389
6390<p>
6391None
6392</p>
6393
6394<h5>Semantics:</h5>
6395
6396<p>
6397This intrinsics is lowered to the target dependent trap instruction. If the
6398target does not have a trap instruction, this intrinsic will be lowered to the
6399call of the abort() function.
6400</p>
6401</div>
6402
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6403<!-- *********************************************************************** -->
6404<hr>
6405<address>
6406 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
6407 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
6408 <a href="http://validator.w3.org/check/referer"><img
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]6409 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!"></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6410
6411 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
6412 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
6413 Last modified: $Date$
6414</address>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]6415
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6416</body>
6417</html>