Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / 76433ccc5dc830112db945de226838772b3a267e / . / docs / LangRef.html
blob: 7c770cb43762696bc248fc9db2ddfa7397874a4d [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>
Bill Wendling9127adb2008-11-18 22:10:53 +0000[diff] [blame]217 <li><a href="#int_atomics">Atomic intrinsics</a>
218 <ol>
219 <li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
220 <li><a href="#int_atomic_cmp_swap"><tt>llvm.atomic.cmp.swap</tt></a></li>
221 <li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
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>
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">
Bill Wendlinge4164592008-11-19 05:56:17 +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">
Bill Wendlinge4164592008-11-19 05:56:17 +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">
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]241 '<tt>llvm.trap</tt>' Intrinsic</a></li>
242 <li><a href="#int_stackprotector">
243 '<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]244 </ol>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]245 </li>
246 </ol>
247 </li>
248</ol>
249
250<div class="doc_author">
251 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
252 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
253</div>
254
255<!-- *********************************************************************** -->
256<div class="doc_section"> <a name="abstract">Abstract </a></div>
257<!-- *********************************************************************** -->
258
259<div class="doc_text">
260<p>This document is a reference manual for the LLVM assembly language.
Bill Wendlinge7846a52008-08-05 22:29:16 +0000[diff] [blame]261LLVM is a Static Single Assignment (SSA) based representation that provides
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]262type safety, low-level operations, flexibility, and the capability of
263representing 'all' high-level languages cleanly. It is the common code
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]264representation used throughout all phases of the LLVM compilation
265strategy.</p>
266</div>
267
268<!-- *********************************************************************** -->
269<div class="doc_section"> <a name="introduction">Introduction</a> </div>
270<!-- *********************************************************************** -->
271
272<div class="doc_text">
273
274<p>The LLVM code representation is designed to be used in three
275different forms: as an in-memory compiler IR, as an on-disk bitcode
276representation (suitable for fast loading by a Just-In-Time compiler),
277and as a human readable assembly language representation. This allows
278LLVM to provide a powerful intermediate representation for efficient
279compiler transformations and analysis, while providing a natural means
280to debug and visualize the transformations. The three different forms
281of LLVM are all equivalent. This document describes the human readable
282representation and notation.</p>
283
284<p>The LLVM representation aims to be light-weight and low-level
285while being expressive, typed, and extensible at the same time. It
286aims to be a "universal IR" of sorts, by being at a low enough level
287that high-level ideas may be cleanly mapped to it (similar to how
288microprocessors are "universal IR's", allowing many source languages to
289be mapped to them). By providing type information, LLVM can be used as
290the target of optimizations: for example, through pointer analysis, it
291can be proven that a C automatic variable is never accessed outside of
292the current function... allowing it to be promoted to a simple SSA
293value instead of a memory location.</p>
294
295</div>
296
297<!-- _______________________________________________________________________ -->
298<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
299
300<div class="doc_text">
301
302<p>It is important to note that this document describes 'well formed'
303LLVM assembly language. There is a difference between what the parser
304accepts and what is considered 'well formed'. For example, the
305following instruction is syntactically okay, but not well formed:</p>
306
307<div class="doc_code">
308<pre>
309%x = <a href="#i_add">add</a> i32 1, %x
310</pre>
311</div>
312
313<p>...because the definition of <tt>%x</tt> does not dominate all of
314its uses. The LLVM infrastructure provides a verification pass that may
315be used to verify that an LLVM module is well formed. This pass is
316automatically run by the parser after parsing input assembly and by
317the optimizer before it outputs bitcode. The violations pointed out
318by the verifier pass indicate bugs in transformation passes or input to
319the parser.</p>
320</div>
321
Chris Lattnera83fdc02007-10-03 17:34:29 +0000[diff] [blame]322<!-- Describe the typesetting conventions here. -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]323
324<!-- *********************************************************************** -->
325<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
326<!-- *********************************************************************** -->
327
328<div class="doc_text">
329
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]330 <p>LLVM identifiers come in two basic types: global and local. Global
331 identifiers (functions, global variables) begin with the @ character. Local
332 identifiers (register names, types) begin with the % character. Additionally,
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]333 there are three different formats for identifiers, for different purposes:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]334
335<ol>
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]336 <li>Named values are represented as a string of characters with their prefix.
337 For example, %foo, @DivisionByZero, %a.really.long.identifier. The actual
338 regular expression used is '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]339 Identifiers which require other characters in their names can be surrounded
Daniel Dunbar73f9a772008-10-14 23:51:43 +0000[diff] [blame]340 with quotes. Special characters may be escaped using "\xx" where xx is the
341 ASCII code for the character in hexadecimal. In this way, any character can
342 be used in a name value, even quotes themselves.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]343
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]344 <li>Unnamed values are represented as an unsigned numeric value with their
345 prefix. For example, %12, @2, %44.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]346
347 <li>Constants, which are described in a <a href="#constants">section about
348 constants</a>, below.</li>
349</ol>
350
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]351<p>LLVM requires that values start with a prefix for two reasons: Compilers
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]352don't need to worry about name clashes with reserved words, and the set of
353reserved words may be expanded in the future without penalty. Additionally,
354unnamed identifiers allow a compiler to quickly come up with a temporary
355variable without having to avoid symbol table conflicts.</p>
356
357<p>Reserved words in LLVM are very similar to reserved words in other
358languages. There are keywords for different opcodes
359('<tt><a href="#i_add">add</a></tt>',
360 '<tt><a href="#i_bitcast">bitcast</a></tt>',
361 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
362href="#t_void">void</a></tt>', '<tt><a href="#t_primitive">i32</a></tt>', etc...),
363and others. These reserved words cannot conflict with variable names, because
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]364none of them start with a prefix character ('%' or '@').</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]365
366<p>Here is an example of LLVM code to multiply the integer variable
367'<tt>%X</tt>' by 8:</p>
368
369<p>The easy way:</p>
370
371<div class="doc_code">
372<pre>
373%result = <a href="#i_mul">mul</a> i32 %X, 8
374</pre>
375</div>
376
377<p>After strength reduction:</p>
378
379<div class="doc_code">
380<pre>
381%result = <a href="#i_shl">shl</a> i32 %X, i8 3
382</pre>
383</div>
384
385<p>And the hard way:</p>
386
387<div class="doc_code">
388<pre>
389<a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
390<a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
391%result = <a href="#i_add">add</a> i32 %1, %1
392</pre>
393</div>
394
395<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
396important lexical features of LLVM:</p>
397
398<ol>
399
400 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
401 line.</li>
402
403 <li>Unnamed temporaries are created when the result of a computation is not
404 assigned to a named value.</li>
405
406 <li>Unnamed temporaries are numbered sequentially</li>
407
408</ol>
409
410<p>...and it also shows a convention that we follow in this document. When
411demonstrating instructions, we will follow an instruction with a comment that
412defines the type and name of value produced. Comments are shown in italic
413text.</p>
414
415</div>
416
417<!-- *********************************************************************** -->
418<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
419<!-- *********************************************************************** -->
420
421<!-- ======================================================================= -->
422<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
423</div>
424
425<div class="doc_text">
426
427<p>LLVM programs are composed of "Module"s, each of which is a
428translation unit of the input programs. Each module consists of
429functions, global variables, and symbol table entries. Modules may be
430combined together with the LLVM linker, which merges function (and
431global variable) definitions, resolves forward declarations, and merges
432symbol table entries. Here is an example of the "hello world" module:</p>
433
434<div class="doc_code">
435<pre><i>; Declare the string constant as a global constant...</i>
436<a href="#identifiers">@.LC0</a> = <a href="#linkage_internal">internal</a> <a
437 href="#globalvars">constant</a> <a href="#t_array">[13 x i8]</a> c"hello world\0A\00" <i>; [13 x i8]*</i>
438
439<i>; External declaration of the puts function</i>
440<a href="#functionstructure">declare</a> i32 @puts(i8 *) <i>; i32(i8 *)* </i>
441
442<i>; Definition of main function</i>
443define i32 @main() { <i>; i32()* </i>
444 <i>; Convert [13x i8 ]* to i8 *...</i>
445 %cast210 = <a
446 href="#i_getelementptr">getelementptr</a> [13 x i8 ]* @.LC0, i64 0, i64 0 <i>; i8 *</i>
447
448 <i>; Call puts function to write out the string to stdout...</i>
449 <a
450 href="#i_call">call</a> i32 @puts(i8 * %cast210) <i>; i32</i>
451 <a
452 href="#i_ret">ret</a> i32 0<br>}<br>
453</pre>
454</div>
455
456<p>This example is made up of a <a href="#globalvars">global variable</a>
457named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
458function, and a <a href="#functionstructure">function definition</a>
459for "<tt>main</tt>".</p>
460
461<p>In general, a module is made up of a list of global values,
462where both functions and global variables are global values. Global values are
463represented by a pointer to a memory location (in this case, a pointer to an
464array of char, and a pointer to a function), and have one of the following <a
465href="#linkage">linkage types</a>.</p>
466
467</div>
468
469<!-- ======================================================================= -->
470<div class="doc_subsection">
471 <a name="linkage">Linkage Types</a>
472</div>
473
474<div class="doc_text">
475
476<p>
477All Global Variables and Functions have one of the following types of linkage:
478</p>
479
480<dl>
481
Dale Johannesen96e7e092008-05-23 23:13:41 +0000[diff] [blame]482 <dt><tt><b><a name="linkage_internal">internal</a></b></tt>: </dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]483
484 <dd>Global values with internal linkage are only directly accessible by
485 objects in the current module. In particular, linking code into a module with
486 an internal global value may cause the internal to be renamed as necessary to
487 avoid collisions. Because the symbol is internal to the module, all
488 references can be updated. This corresponds to the notion of the
489 '<tt>static</tt>' keyword in C.
490 </dd>
491
492 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
493
494 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
495 the same name when linkage occurs. This is typically used to implement
496 inline functions, templates, or other code which must be generated in each
497 translation unit that uses it. Unreferenced <tt>linkonce</tt> globals are
498 allowed to be discarded.
499 </dd>
500
Dale Johannesen96e7e092008-05-23 23:13:41 +0000[diff] [blame]501 <dt><tt><b><a name="linkage_common">common</a></b></tt>: </dt>
502
503 <dd>"<tt>common</tt>" linkage is exactly the same as <tt>linkonce</tt>
504 linkage, except that unreferenced <tt>common</tt> globals may not be
505 discarded. This is used for globals that may be emitted in multiple
506 translation units, but that are not guaranteed to be emitted into every
507 translation unit that uses them. One example of this is tentative
508 definitions in C, such as "<tt>int X;</tt>" at global scope.
509 </dd>
510
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]511 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
512
Dale Johannesen96e7e092008-05-23 23:13:41 +0000[diff] [blame]513 <dd>"<tt>weak</tt>" linkage is the same as <tt>common</tt> linkage, except
514 that some targets may choose to emit different assembly sequences for them
515 for target-dependent reasons. This is used for globals that are declared
516 "weak" in C source code.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]517 </dd>
518
519 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
520
521 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
522 pointer to array type. When two global variables with appending linkage are
523 linked together, the two global arrays are appended together. This is the
524 LLVM, typesafe, equivalent of having the system linker append together
525 "sections" with identical names when .o files are linked.
526 </dd>
527
528 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt>: </dt>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]529 <dd>The semantics of this linkage follow the ELF object file model: the
530 symbol is weak until linked, if not linked, the symbol becomes null instead
531 of being an undefined reference.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]532 </dd>
533
534 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
535
536 <dd>If none of the above identifiers are used, the global is externally
537 visible, meaning that it participates in linkage and can be used to resolve
538 external symbol references.
539 </dd>
540</dl>
541
542 <p>
543 The next two types of linkage are targeted for Microsoft Windows platform
544 only. They are designed to support importing (exporting) symbols from (to)
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]545 DLLs (Dynamic Link Libraries).
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]546 </p>
547
548 <dl>
549 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt>: </dt>
550
551 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
552 or variable via a global pointer to a pointer that is set up by the DLL
553 exporting the symbol. On Microsoft Windows targets, the pointer name is
554 formed by combining <code>_imp__</code> and the function or variable name.
555 </dd>
556
557 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt>: </dt>
558
559 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
560 pointer to a pointer in a DLL, so that it can be referenced with the
561 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
562 name is formed by combining <code>_imp__</code> and the function or variable
563 name.
564 </dd>
565
566</dl>
567
568<p><a name="linkage_external"></a>For example, since the "<tt>.LC0</tt>"
569variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
570variable and was linked with this one, one of the two would be renamed,
571preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
572external (i.e., lacking any linkage declarations), they are accessible
573outside of the current module.</p>
574<p>It is illegal for a function <i>declaration</i>
575to have any linkage type other than "externally visible", <tt>dllimport</tt>,
576or <tt>extern_weak</tt>.</p>
577<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]578linkages.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]579</div>
580
581<!-- ======================================================================= -->
582<div class="doc_subsection">
583 <a name="callingconv">Calling Conventions</a>
584</div>
585
586<div class="doc_text">
587
588<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
589and <a href="#i_invoke">invokes</a> can all have an optional calling convention
590specified for the call. The calling convention of any pair of dynamic
591caller/callee must match, or the behavior of the program is undefined. The
592following calling conventions are supported by LLVM, and more may be added in
593the future:</p>
594
595<dl>
596 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
597
598 <dd>This calling convention (the default if no other calling convention is
599 specified) matches the target C calling conventions. This calling convention
600 supports varargs function calls and tolerates some mismatch in the declared
601 prototype and implemented declaration of the function (as does normal C).
602 </dd>
603
604 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
605
606 <dd>This calling convention attempts to make calls as fast as possible
607 (e.g. by passing things in registers). This calling convention allows the
608 target to use whatever tricks it wants to produce fast code for the target,
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]609 without having to conform to an externally specified ABI (Application Binary
610 Interface). Implementations of this convention should allow arbitrary
Arnold Schwaighofer07444922008-05-14 09:17:12 +0000[diff] [blame]611 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> to be
612 supported. This calling convention does not support varargs and requires the
613 prototype of all callees to exactly match the prototype of the function
614 definition.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]615 </dd>
616
617 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
618
619 <dd>This calling convention attempts to make code in the caller as efficient
620 as possible under the assumption that the call is not commonly executed. As
621 such, these calls often preserve all registers so that the call does not break
622 any live ranges in the caller side. This calling convention does not support
623 varargs and requires the prototype of all callees to exactly match the
624 prototype of the function definition.
625 </dd>
626
627 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
628
629 <dd>Any calling convention may be specified by number, allowing
630 target-specific calling conventions to be used. Target specific calling
631 conventions start at 64.
632 </dd>
633</dl>
634
635<p>More calling conventions can be added/defined on an as-needed basis, to
636support pascal conventions or any other well-known target-independent
637convention.</p>
638
639</div>
640
641<!-- ======================================================================= -->
642<div class="doc_subsection">
643 <a name="visibility">Visibility Styles</a>
644</div>
645
646<div class="doc_text">
647
648<p>
649All Global Variables and Functions have one of the following visibility styles:
650</p>
651
652<dl>
653 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
654
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]655 <dd>On targets that use the ELF object file format, default visibility means
656 that the declaration is visible to other
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]657 modules and, in shared libraries, means that the declared entity may be
658 overridden. On Darwin, default visibility means that the declaration is
659 visible to other modules. Default visibility corresponds to "external
660 linkage" in the language.
661 </dd>
662
663 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
664
665 <dd>Two declarations of an object with hidden visibility refer to the same
666 object if they are in the same shared object. Usually, hidden visibility
667 indicates that the symbol will not be placed into the dynamic symbol table,
668 so no other module (executable or shared library) can reference it
669 directly.
670 </dd>
671
672 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
673
674 <dd>On ELF, protected visibility indicates that the symbol will be placed in
675 the dynamic symbol table, but that references within the defining module will
676 bind to the local symbol. That is, the symbol cannot be overridden by another
677 module.
678 </dd>
679</dl>
680
681</div>
682
683<!-- ======================================================================= -->
684<div class="doc_subsection">
685 <a name="globalvars">Global Variables</a>
686</div>
687
688<div class="doc_text">
689
690<p>Global variables define regions of memory allocated at compilation time
691instead of run-time. Global variables may optionally be initialized, may have
692an explicit section to be placed in, and may have an optional explicit alignment
693specified. A variable may be defined as "thread_local", which means that it
694will not be shared by threads (each thread will have a separated copy of the
695variable). A variable may be defined as a global "constant," which indicates
696that the contents of the variable will <b>never</b> be modified (enabling better
697optimization, allowing the global data to be placed in the read-only section of
698an executable, etc). Note that variables that need runtime initialization
699cannot be marked "constant" as there is a store to the variable.</p>
700
701<p>
702LLVM explicitly allows <em>declarations</em> of global variables to be marked
703constant, even if the final definition of the global is not. This capability
704can be used to enable slightly better optimization of the program, but requires
705the language definition to guarantee that optimizations based on the
706'constantness' are valid for the translation units that do not include the
707definition.
708</p>
709
710<p>As SSA values, global variables define pointer values that are in
711scope (i.e. they dominate) all basic blocks in the program. Global
712variables always define a pointer to their "content" type because they
713describe a region of memory, and all memory objects in LLVM are
714accessed through pointers.</p>
715
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]716<p>A global variable may be declared to reside in a target-specifc numbered
717address space. For targets that support them, address spaces may affect how
718optimizations are performed and/or what target instructions are used to access
Christopher Lamb20a39e92007-12-12 08:44:39 +0000[diff] [blame]719the variable. The default address space is zero. The address space qualifier
720must precede any other attributes.</p>
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]721
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]722<p>LLVM allows an explicit section to be specified for globals. If the target
723supports it, it will emit globals to the section specified.</p>
724
725<p>An explicit alignment may be specified for a global. If not present, or if
726the alignment is set to zero, the alignment of the global is set by the target
727to whatever it feels convenient. If an explicit alignment is specified, the
728global is forced to have at least that much alignment. All alignments must be
729a power of 2.</p>
730
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]731<p>For example, the following defines a global in a numbered address space with
732an initializer, section, and alignment:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]733
734<div class="doc_code">
735<pre>
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]736@G = constant float 1.0 addrspace(5), section "foo", align 4
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]737</pre>
738</div>
739
740</div>
741
742
743<!-- ======================================================================= -->
744<div class="doc_subsection">
745 <a name="functionstructure">Functions</a>
746</div>
747
748<div class="doc_text">
749
750<p>LLVM function definitions consist of the "<tt>define</tt>" keyord,
751an optional <a href="#linkage">linkage type</a>, an optional
752<a href="#visibility">visibility style</a>, an optional
753<a href="#callingconv">calling convention</a>, a return type, an optional
754<a href="#paramattrs">parameter attribute</a> for the return type, a function
755name, a (possibly empty) argument list (each with optional
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]756<a href="#paramattrs">parameter attributes</a>), optional
757<a href="#fnattrs">function attributes</a>, an optional section,
758an optional alignment, an optional <a href="#gc">garbage collector name</a>,
Chris Lattneree202542008-10-04 18:10:21 +0000[diff] [blame]759an opening curly brace, a list of basic blocks, and a closing curly brace.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]760
761LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
762optional <a href="#linkage">linkage type</a>, an optional
763<a href="#visibility">visibility style</a>, an optional
764<a href="#callingconv">calling convention</a>, a return type, an optional
765<a href="#paramattrs">parameter attribute</a> for the return type, a function
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]766name, a possibly empty list of arguments, an optional alignment, and an optional
Gordon Henriksend606f5b2007-12-10 03:30:21 +0000[diff] [blame]767<a href="#gc">garbage collector name</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]768
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]769<p>A function definition contains a list of basic blocks, forming the CFG
770(Control Flow Graph) for
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]771the function. Each basic block may optionally start with a label (giving the
772basic block a symbol table entry), contains a list of instructions, and ends
773with a <a href="#terminators">terminator</a> instruction (such as a branch or
774function return).</p>
775
776<p>The first basic block in a function is special in two ways: it is immediately
777executed on entrance to the function, and it is not allowed to have predecessor
778basic blocks (i.e. there can not be any branches to the entry block of a
779function). Because the block can have no predecessors, it also cannot have any
780<a href="#i_phi">PHI nodes</a>.</p>
781
782<p>LLVM allows an explicit section to be specified for functions. If the target
783supports it, it will emit functions to the section specified.</p>
784
785<p>An explicit alignment may be specified for a function. If not present, or if
786the alignment is set to zero, the alignment of the function is set by the target
787to whatever it feels convenient. If an explicit alignment is specified, the
788function is forced to have at least that much alignment. All alignments must be
789a power of 2.</p>
790
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]791 <h5>Syntax:</h5>
792
793<div class="doc_code">
Chris Lattner1e5c5cd02008-10-13 16:55:18 +0000[diff] [blame]794<tt>
795define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
796 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
797 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
798 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
799 [<a href="#gc">gc</a>] { ... }
800</tt>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]801</div>
802
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]803</div>
804
805
806<!-- ======================================================================= -->
807<div class="doc_subsection">
808 <a name="aliasstructure">Aliases</a>
809</div>
810<div class="doc_text">
811 <p>Aliases act as "second name" for the aliasee value (which can be either
Anton Korobeynikov96822812008-03-22 08:36:14 +0000[diff] [blame]812 function, global variable, another alias or bitcast of global value). Aliases
813 may have an optional <a href="#linkage">linkage type</a>, and an
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]814 optional <a href="#visibility">visibility style</a>.</p>
815
816 <h5>Syntax:</h5>
817
818<div class="doc_code">
819<pre>
Duncan Sandsd7bfabf2008-09-12 20:48:21 +0000[diff] [blame]820@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]821</pre>
822</div>
823
824</div>
825
826
827
828<!-- ======================================================================= -->
829<div class="doc_subsection"><a name="paramattrs">Parameter Attributes</a></div>
830<div class="doc_text">
831 <p>The return type and each parameter of a function type may have a set of
832 <i>parameter attributes</i> associated with them. Parameter attributes are
833 used to communicate additional information about the result or parameters of
Duncan Sandsf5588dc2007-11-27 13:23:08 +0000[diff] [blame]834 a function. Parameter attributes are considered to be part of the function,
835 not of the function type, so functions with different parameter attributes
836 can have the same function type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]837
838 <p>Parameter attributes are simple keywords that follow the type specified. If
839 multiple parameter attributes are needed, they are space separated. For
840 example:</p>
841
842<div class="doc_code">
843<pre>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]844declare i32 @printf(i8* noalias , ...)
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]845declare i32 @atoi(i8 zeroext)
846declare signext i8 @returns_signed_char()
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]847</pre>
848</div>
849
Duncan Sandsf5588dc2007-11-27 13:23:08 +0000[diff] [blame]850 <p>Note that any attributes for the function result (<tt>nounwind</tt>,
851 <tt>readonly</tt>) come immediately after the argument list.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]852
853 <p>Currently, only the following parameter attributes are defined:</p>
854 <dl>
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]855 <dt><tt>zeroext</tt></dt>
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]856 <dd>This indicates to the code generator that the parameter or return value
857 should be zero-extended to a 32-bit value by the caller (for a parameter)
858 or the callee (for a return value).</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]859
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]860 <dt><tt>signext</tt></dt>
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]861 <dd>This indicates to the code generator that the parameter or return value
862 should be sign-extended to a 32-bit value by the caller (for a parameter)
863 or the callee (for a return value).</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]864
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]865 <dt><tt>inreg</tt></dt>
Dale Johannesenc08a0e22008-09-25 20:47:45 +0000[diff] [blame]866 <dd>This indicates that this parameter or return value should be treated
867 in a special target-dependent fashion during while emitting code for a
868 function call or return (usually, by putting it in a register as opposed
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]869 to memory, though some targets use it to distinguish between two different
870 kinds of registers). Use of this attribute is target-specific.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]871
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]872 <dt><tt><a name="byval">byval</a></tt></dt>
Chris Lattner04c86182008-01-15 04:34:22 +0000[diff] [blame]873 <dd>This indicates that the pointer parameter should really be passed by
874 value to the function. The attribute implies that a hidden copy of the
875 pointee is made between the caller and the callee, so the callee is unable
Chris Lattner6a9f3c42008-08-05 18:21:08 +0000[diff] [blame]876 to modify the value in the callee. This attribute is only valid on LLVM
Chris Lattner04c86182008-01-15 04:34:22 +0000[diff] [blame]877 pointer arguments. It is generally used to pass structs and arrays by
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]878 value, but is also valid on pointers to scalars. The copy is considered to
879 belong to the caller not the callee (for example,
880 <tt><a href="#readonly">readonly</a></tt> functions should not write to
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]881 <tt>byval</tt> parameters). This is not a valid attribute for return
882 values. </dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]883
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]884 <dt><tt>sret</tt></dt>
Duncan Sands616cc032008-02-18 04:19:38 +0000[diff] [blame]885 <dd>This indicates that the pointer parameter specifies the address of a
886 structure that is the return value of the function in the source program.
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]887 This pointer must be guaranteed by the caller to be valid: loads and stores
888 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]889 be applied to the first parameter. This is not a valid attribute for
890 return values. </dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]891
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]892 <dt><tt>noalias</tt></dt>
Nick Lewyckyff384472008-11-24 03:41:24 +0000[diff] [blame]893 <dd>This indicates that the pointer does not alias any global or any other
894 parameter. The caller is responsible for ensuring that this is the
Nick Lewyckya604a942008-11-24 05:00:44 +0000[diff] [blame]895 case. On a function return value, <tt>noalias</tt> additionally indicates
896 that the pointer does not alias any other pointers visible to the
Edwin Török76433cc2008-11-24 08:02:24 +0000[diff] [blame^]897 caller. Note that this applies only to pointers that can be used to actually
898 load/store a value: NULL, unique pointers from malloc(0), and freed pointers
899 are considered to not alias anything.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]900
Duncan Sands4ee46812007-07-27 19:57:41 +0000[diff] [blame]901 <dt><tt>nest</tt></dt>
Duncan Sandsf1a7d4c2008-07-08 09:27:25 +0000[diff] [blame]902 <dd>This indicates that the pointer parameter can be excised using the
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]903 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
904 attribute for return values.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]905 </dl>
906
907</div>
908
909<!-- ======================================================================= -->
910<div class="doc_subsection">
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]911 <a name="gc">Garbage Collector Names</a>
912</div>
913
914<div class="doc_text">
915<p>Each function may specify a garbage collector name, which is simply a
916string.</p>
917
918<div class="doc_code"><pre
919>define void @f() gc "name" { ...</pre></div>
920
921<p>The compiler declares the supported values of <i>name</i>. Specifying a
922collector which will cause the compiler to alter its output in order to support
923the named garbage collection algorithm.</p>
924</div>
925
926<!-- ======================================================================= -->
927<div class="doc_subsection">
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]928 <a name="fnattrs">Function Attributes</a>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]929</div>
930
931<div class="doc_text">
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]932
933<p>Function attributes are set to communicate additional information about
934 a function. Function attributes are considered to be part of the function,
935 not of the function type, so functions with different parameter attributes
936 can have the same function type.</p>
937
938 <p>Function attributes are simple keywords that follow the type specified. If
939 multiple attributes are needed, they are space separated. For
940 example:</p>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]941
942<div class="doc_code">
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]943<pre>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]944define void @f() noinline { ... }
945define void @f() alwaysinline { ... }
946define void @f() alwaysinline optsize { ... }
947define void @f() optsize
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]948</pre>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]949</div>
950
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]951<dl>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]952<dt><tt>alwaysinline</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]953<dd>This attribute indicates that the inliner should attempt to inline this
954function into callers whenever possible, ignoring any active inlining size
955threshold for this caller.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]956
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]957<dt><tt>noinline</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]958<dd>This attribute indicates that the inliner should never inline this function
Chris Lattner377f3ae2008-10-05 17:14:59 +0000[diff] [blame]959in any situation. This attribute may not be used together with the
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]960<tt>alwaysinline</tt> attribute.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]961
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]962<dt><tt>optsize</tt></dt>
Devang Patelc29099b2008-09-29 18:34:44 +0000[diff] [blame]963<dd>This attribute suggests that optimization passes and code generator passes
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]964make choices that keep the code size of this function low, and otherwise do
965optimizations specifically to reduce code size.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]966
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]967<dt><tt>noreturn</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]968<dd>This function attribute indicates that the function never returns normally.
969This produces undefined behavior at runtime if the function ever does
970dynamically return.</dd>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]971
972<dt><tt>nounwind</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]973<dd>This function attribute indicates that the function never returns with an
974unwind or exceptional control flow. If the function does unwind, its runtime
975behavior is undefined.</dd>
976
977<dt><tt>readnone</tt></dt>
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]978<dd>This attribute indicates that the function computes its result (or the
979exception it throws) based strictly on its arguments, without dereferencing any
980pointer arguments or otherwise accessing any mutable state (e.g. memory, control
981registers, etc) visible to caller functions. It does not write through any
982pointer arguments (including <tt><a href="#byval">byval</a></tt> arguments) and
983never changes any state visible to callers.</dd>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]984
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]985<dt><tt><a name="readonly">readonly</a></tt></dt>
986<dd>This attribute indicates that the function does not write through any
987pointer arguments (including <tt><a href="#byval">byval</a></tt> arguments)
988or otherwise modify any state (e.g. memory, control registers, etc) visible to
989caller functions. It may dereference pointer arguments and read state that may
990be set in the caller. A readonly function always returns the same value (or
991throws the same exception) when called with the same set of arguments and global
992state.</dd>
Bill Wendlingdfaabba2008-11-13 01:02:51 +0000[diff] [blame]993
994<dt><tt><a name="ssp">ssp</a></tt></dt>
995<dd>This attribute indicates that the function should emit a stack smashing
996protector. It is in the form of a "canary"&mdash;a random value placed on the
997stack before the local variables that's checked upon return from the function to
998see if it has been overwritten. A heuristic is used to determine if a function
999needs stack protectors or not.</dd>
1000
1001<dt><tt>ssp-req</tt></dt>
1002<dd>This attribute indicates that the function should <em>always</em> emit a
1003stack smashing protector. This overrides the <tt><a href="#ssp">ssp</a></tt>
1004function attribute.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1005</dl>
1006
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1007</div>
1008
1009<!-- ======================================================================= -->
1010<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1011 <a name="moduleasm">Module-Level Inline Assembly</a>
1012</div>
1013
1014<div class="doc_text">
1015<p>
1016Modules may contain "module-level inline asm" blocks, which corresponds to the
1017GCC "file scope inline asm" blocks. These blocks are internally concatenated by
1018LLVM and treated as a single unit, but may be separated in the .ll file if
1019desired. The syntax is very simple:
1020</p>
1021
1022<div class="doc_code">
1023<pre>
1024module asm "inline asm code goes here"
1025module asm "more can go here"
1026</pre>
1027</div>
1028
1029<p>The strings can contain any character by escaping non-printable characters.
1030 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
1031 for the number.
1032</p>
1033
1034<p>
1035 The inline asm code is simply printed to the machine code .s file when
1036 assembly code is generated.
1037</p>
1038</div>
1039
1040<!-- ======================================================================= -->
1041<div class="doc_subsection">
1042 <a name="datalayout">Data Layout</a>
1043</div>
1044
1045<div class="doc_text">
1046<p>A module may specify a target specific data layout string that specifies how
1047data is to be laid out in memory. The syntax for the data layout is simply:</p>
1048<pre> target datalayout = "<i>layout specification</i>"</pre>
1049<p>The <i>layout specification</i> consists of a list of specifications
1050separated by the minus sign character ('-'). Each specification starts with a
1051letter and may include other information after the letter to define some
1052aspect of the data layout. The specifications accepted are as follows: </p>
1053<dl>
1054 <dt><tt>E</tt></dt>
1055 <dd>Specifies that the target lays out data in big-endian form. That is, the
1056 bits with the most significance have the lowest address location.</dd>
1057 <dt><tt>e</tt></dt>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1058 <dd>Specifies that the target lays out data in little-endian form. That is,
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1059 the bits with the least significance have the lowest address location.</dd>
1060 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1061 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
1062 <i>preferred</i> alignments. All sizes are in bits. Specifying the <i>pref</i>
1063 alignment is optional. If omitted, the preceding <tt>:</tt> should be omitted
1064 too.</dd>
1065 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1066 <dd>This specifies the alignment for an integer type of a given bit
1067 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1068 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1069 <dd>This specifies the alignment for a vector type of a given bit
1070 <i>size</i>.</dd>
1071 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1072 <dd>This specifies the alignment for a floating point type of a given bit
1073 <i>size</i>. The value of <i>size</i> must be either 32 (float) or 64
1074 (double).</dd>
1075 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1076 <dd>This specifies the alignment for an aggregate type of a given bit
1077 <i>size</i>.</dd>
1078</dl>
1079<p>When constructing the data layout for a given target, LLVM starts with a
1080default set of specifications which are then (possibly) overriden by the
1081specifications in the <tt>datalayout</tt> keyword. The default specifications
1082are given in this list:</p>
1083<ul>
1084 <li><tt>E</tt> - big endian</li>
1085 <li><tt>p:32:64:64</tt> - 32-bit pointers with 64-bit alignment</li>
1086 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1087 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1088 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1089 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1090 <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]1091 alignment of 64-bits</li>
1092 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1093 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1094 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1095 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1096 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
1097</ul>
Chris Lattner6a9f3c42008-08-05 18:21:08 +0000[diff] [blame]1098<p>When LLVM is determining the alignment for a given type, it uses the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]1099following rules:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1100<ol>
1101 <li>If the type sought is an exact match for one of the specifications, that
1102 specification is used.</li>
1103 <li>If no match is found, and the type sought is an integer type, then the
1104 smallest integer type that is larger than the bitwidth of the sought type is
1105 used. If none of the specifications are larger than the bitwidth then the the
1106 largest integer type is used. For example, given the default specifications
1107 above, the i7 type will use the alignment of i8 (next largest) while both
1108 i65 and i256 will use the alignment of i64 (largest specified).</li>
1109 <li>If no match is found, and the type sought is a vector type, then the
1110 largest vector type that is smaller than the sought vector type will be used
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]1111 as a fall back. This happens because &lt;128 x double&gt; can be implemented
1112 in terms of 64 &lt;2 x double&gt;, for example.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1113</ol>
1114</div>
1115
1116<!-- *********************************************************************** -->
1117<div class="doc_section"> <a name="typesystem">Type System</a> </div>
1118<!-- *********************************************************************** -->
1119
1120<div class="doc_text">
1121
1122<p>The LLVM type system is one of the most important features of the
1123intermediate representation. Being typed enables a number of
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1124optimizations to be performed on the intermediate representation directly,
1125without having to do
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1126extra analyses on the side before the transformation. A strong type
1127system makes it easier to read the generated code and enables novel
1128analyses and transformations that are not feasible to perform on normal
1129three address code representations.</p>
1130
1131</div>
1132
1133<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1134<div class="doc_subsection"> <a name="t_classifications">Type
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1135Classifications</a> </div>
1136<div class="doc_text">
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1137<p>The types fall into a few useful
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1138classifications:</p>
1139
1140<table border="1" cellspacing="0" cellpadding="4">
1141 <tbody>
1142 <tr><th>Classification</th><th>Types</th></tr>
1143 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1144 <td><a href="#t_integer">integer</a></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1145 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
1146 </tr>
1147 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1148 <td><a href="#t_floating">floating point</a></td>
1149 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1150 </tr>
1151 <tr>
1152 <td><a name="t_firstclass">first class</a></td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1153 <td><a href="#t_integer">integer</a>,
1154 <a href="#t_floating">floating point</a>,
1155 <a href="#t_pointer">pointer</a>,
Dan Gohmanf6237db2008-06-18 18:42:13 +0000[diff] [blame]1156 <a href="#t_vector">vector</a>,
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]1157 <a href="#t_struct">structure</a>,
1158 <a href="#t_array">array</a>,
Dan Gohmand13951e2008-05-23 22:50:26 +0000[diff] [blame]1159 <a href="#t_label">label</a>.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1160 </td>
1161 </tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1162 <tr>
1163 <td><a href="#t_primitive">primitive</a></td>
1164 <td><a href="#t_label">label</a>,
1165 <a href="#t_void">void</a>,
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1166 <a href="#t_floating">floating point</a>.</td>
1167 </tr>
1168 <tr>
1169 <td><a href="#t_derived">derived</a></td>
1170 <td><a href="#t_integer">integer</a>,
1171 <a href="#t_array">array</a>,
1172 <a href="#t_function">function</a>,
1173 <a href="#t_pointer">pointer</a>,
1174 <a href="#t_struct">structure</a>,
1175 <a href="#t_pstruct">packed structure</a>,
1176 <a href="#t_vector">vector</a>,
1177 <a href="#t_opaque">opaque</a>.
Dan Gohman032ba852008-10-14 16:32:04 +0000[diff] [blame]1178 </td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1179 </tr>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1180 </tbody>
1181</table>
1182
1183<p>The <a href="#t_firstclass">first class</a> types are perhaps the
1184most important. Values of these types are the only ones which can be
1185produced by instructions, passed as arguments, or used as operands to
Dan Gohman4f29e422008-05-23 21:53:15 +0000[diff] [blame]1186instructions.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1187</div>
1188
1189<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1190<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1191
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1192<div class="doc_text">
1193<p>The primitive types are the fundamental building blocks of the LLVM
1194system.</p>
1195
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1196</div>
1197
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1198<!-- _______________________________________________________________________ -->
1199<div class="doc_subsubsection"> <a name="t_floating">Floating Point Types</a> </div>
1200
1201<div class="doc_text">
1202 <table>
1203 <tbody>
1204 <tr><th>Type</th><th>Description</th></tr>
1205 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1206 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1207 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1208 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1209 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1210 </tbody>
1211 </table>
1212</div>
1213
1214<!-- _______________________________________________________________________ -->
1215<div class="doc_subsubsection"> <a name="t_void">Void Type</a> </div>
1216
1217<div class="doc_text">
1218<h5>Overview:</h5>
1219<p>The void type does not represent any value and has no size.</p>
1220
1221<h5>Syntax:</h5>
1222
1223<pre>
1224 void
1225</pre>
1226</div>
1227
1228<!-- _______________________________________________________________________ -->
1229<div class="doc_subsubsection"> <a name="t_label">Label Type</a> </div>
1230
1231<div class="doc_text">
1232<h5>Overview:</h5>
1233<p>The label type represents code labels.</p>
1234
1235<h5>Syntax:</h5>
1236
1237<pre>
1238 label
1239</pre>
1240</div>
1241
1242
1243<!-- ======================================================================= -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1244<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
1245
1246<div class="doc_text">
1247
1248<p>The real power in LLVM comes from the derived types in the system.
1249This is what allows a programmer to represent arrays, functions,
1250pointers, and other useful types. Note that these derived types may be
1251recursive: For example, it is possible to have a two dimensional array.</p>
1252
1253</div>
1254
1255<!-- _______________________________________________________________________ -->
1256<div class="doc_subsubsection"> <a name="t_integer">Integer Type</a> </div>
1257
1258<div class="doc_text">
1259
1260<h5>Overview:</h5>
1261<p>The integer type is a very simple derived type that simply specifies an
1262arbitrary bit width for the integer type desired. Any bit width from 1 bit to
12632^23-1 (about 8 million) can be specified.</p>
1264
1265<h5>Syntax:</h5>
1266
1267<pre>
1268 iN
1269</pre>
1270
1271<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1272value.</p>
1273
1274<h5>Examples:</h5>
1275<table class="layout">
Chris Lattner251ab812007-12-18 06:18:21 +0000[diff] [blame]1276 <tbody>
1277 <tr>
1278 <td><tt>i1</tt></td>
1279 <td>a single-bit integer.</td>
1280 </tr><tr>
1281 <td><tt>i32</tt></td>
1282 <td>a 32-bit integer.</td>
1283 </tr><tr>
1284 <td><tt>i1942652</tt></td>
1285 <td>a really big integer of over 1 million bits.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1286 </tr>
Chris Lattner251ab812007-12-18 06:18:21 +0000[diff] [blame]1287 </tbody>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1288</table>
1289</div>
1290
1291<!-- _______________________________________________________________________ -->
1292<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
1293
1294<div class="doc_text">
1295
1296<h5>Overview:</h5>
1297
1298<p>The array type is a very simple derived type that arranges elements
1299sequentially in memory. The array type requires a size (number of
1300elements) and an underlying data type.</p>
1301
1302<h5>Syntax:</h5>
1303
1304<pre>
1305 [&lt;# elements&gt; x &lt;elementtype&gt;]
1306</pre>
1307
1308<p>The number of elements is a constant integer value; elementtype may
1309be any type with a size.</p>
1310
1311<h5>Examples:</h5>
1312<table class="layout">
1313 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1314 <td class="left"><tt>[40 x i32]</tt></td>
1315 <td class="left">Array of 40 32-bit integer values.</td>
1316 </tr>
1317 <tr class="layout">
1318 <td class="left"><tt>[41 x i32]</tt></td>
1319 <td class="left">Array of 41 32-bit integer values.</td>
1320 </tr>
1321 <tr class="layout">
1322 <td class="left"><tt>[4 x i8]</tt></td>
1323 <td class="left">Array of 4 8-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1324 </tr>
1325</table>
1326<p>Here are some examples of multidimensional arrays:</p>
1327<table class="layout">
1328 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1329 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1330 <td class="left">3x4 array of 32-bit integer values.</td>
1331 </tr>
1332 <tr class="layout">
1333 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1334 <td class="left">12x10 array of single precision floating point values.</td>
1335 </tr>
1336 <tr class="layout">
1337 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1338 <td class="left">2x3x4 array of 16-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1339 </tr>
1340</table>
1341
1342<p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
1343length array. Normally, accesses past the end of an array are undefined in
1344LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
1345As a special case, however, zero length arrays are recognized to be variable
1346length. This allows implementation of 'pascal style arrays' with the LLVM
1347type "{ i32, [0 x float]}", for example.</p>
1348
1349</div>
1350
1351<!-- _______________________________________________________________________ -->
1352<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
1353<div class="doc_text">
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1354
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1355<h5>Overview:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1356
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1357<p>The function type can be thought of as a function signature. It
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1358consists of a return type and a list of formal parameter types. The
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1359return 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]1360If 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]1361class types, and the struct must have at least one element.</p>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1362
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1363<h5>Syntax:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1364
1365<pre>
1366 &lt;returntype list&gt; (&lt;parameter list&gt;)
1367</pre>
1368
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1369<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
1370specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1371which indicates that the function takes a variable number of arguments.
1372Variable argument functions can access their arguments with the <a
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1373 href="#int_varargs">variable argument handling intrinsic</a> functions.
1374'<tt>&lt;returntype list&gt;</tt>' is a comma-separated list of
1375<a href="#t_firstclass">first class</a> type specifiers.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1376
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1377<h5>Examples:</h5>
1378<table class="layout">
1379 <tr class="layout">
1380 <td class="left"><tt>i32 (i32)</tt></td>
1381 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
1382 </td>
1383 </tr><tr class="layout">
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]1384 <td class="left"><tt>float&nbsp;(i16&nbsp;signext,&nbsp;i32&nbsp;*)&nbsp;*
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1385 </tt></td>
1386 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
1387 an <tt>i16</tt> that should be sign extended and a
1388 <a href="#t_pointer">pointer</a> to <tt>i32</tt>, returning
1389 <tt>float</tt>.
1390 </td>
1391 </tr><tr class="layout">
1392 <td class="left"><tt>i32 (i8*, ...)</tt></td>
1393 <td class="left">A vararg function that takes at least one
1394 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1395 which returns an integer. This is the signature for <tt>printf</tt> in
1396 LLVM.
1397 </td>
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1398 </tr><tr class="layout">
1399 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Devang Patel6dddba22008-03-24 18:10:52 +0000[diff] [blame]1400 <td class="left">A function taking an <tt>i32></tt>, returning two
1401 <tt> i32 </tt> values as an aggregate of type <tt>{ i32, i32 }</tt>
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1402 </td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1403 </tr>
1404</table>
1405
1406</div>
1407<!-- _______________________________________________________________________ -->
1408<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
1409<div class="doc_text">
1410<h5>Overview:</h5>
1411<p>The structure type is used to represent a collection of data members
1412together in memory. The packing of the field types is defined to match
1413the ABI of the underlying processor. The elements of a structure may
1414be any type that has a size.</p>
1415<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1416and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1417field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1418instruction.</p>
1419<h5>Syntax:</h5>
1420<pre> { &lt;type list&gt; }<br></pre>
1421<h5>Examples:</h5>
1422<table class="layout">
1423 <tr class="layout">
1424 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1425 <td class="left">A triple of three <tt>i32</tt> values</td>
1426 </tr><tr class="layout">
1427 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1428 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1429 second element is a <a href="#t_pointer">pointer</a> to a
1430 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1431 an <tt>i32</tt>.</td>
1432 </tr>
1433</table>
1434</div>
1435
1436<!-- _______________________________________________________________________ -->
1437<div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
1438</div>
1439<div class="doc_text">
1440<h5>Overview:</h5>
1441<p>The packed structure type is used to represent a collection of data members
1442together in memory. There is no padding between fields. Further, the alignment
1443of a packed structure is 1 byte. The elements of a packed structure may
1444be any type that has a size.</p>
1445<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1446and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1447field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1448instruction.</p>
1449<h5>Syntax:</h5>
1450<pre> &lt; { &lt;type list&gt; } &gt; <br></pre>
1451<h5>Examples:</h5>
1452<table class="layout">
1453 <tr class="layout">
1454 <td class="left"><tt>&lt; { i32, i32, i32 } &gt;</tt></td>
1455 <td class="left">A triple of three <tt>i32</tt> values</td>
1456 </tr><tr class="layout">
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1457 <td class="left">
1458<tt>&lt;&nbsp;{&nbsp;float,&nbsp;i32&nbsp;(i32)*&nbsp;}&nbsp;&gt;</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1459 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1460 second element is a <a href="#t_pointer">pointer</a> to a
1461 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1462 an <tt>i32</tt>.</td>
1463 </tr>
1464</table>
1465</div>
1466
1467<!-- _______________________________________________________________________ -->
1468<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
1469<div class="doc_text">
1470<h5>Overview:</h5>
1471<p>As in many languages, the pointer type represents a pointer or
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]1472reference to another object, which must live in memory. Pointer types may have
1473an optional address space attribute defining the target-specific numbered
1474address space where the pointed-to object resides. The default address space is
1475zero.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1476<h5>Syntax:</h5>
1477<pre> &lt;type&gt; *<br></pre>
1478<h5>Examples:</h5>
1479<table class="layout">
1480 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1481 <td class="left"><tt>[4x i32]*</tt></td>
1482 <td class="left">A <a href="#t_pointer">pointer</a> to <a
1483 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
1484 </tr>
1485 <tr class="layout">
1486 <td class="left"><tt>i32 (i32 *) *</tt></td>
1487 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1488 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1489 <tt>i32</tt>.</td>
1490 </tr>
1491 <tr class="layout">
1492 <td class="left"><tt>i32 addrspace(5)*</tt></td>
1493 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
1494 that resides in address space #5.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1495 </tr>
1496</table>
1497</div>
1498
1499<!-- _______________________________________________________________________ -->
1500<div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
1501<div class="doc_text">
1502
1503<h5>Overview:</h5>
1504
1505<p>A vector type is a simple derived type that represents a vector
1506of elements. Vector types are used when multiple primitive data
1507are operated in parallel using a single instruction (SIMD).
1508A vector type requires a size (number of
1509elements) and an underlying primitive data type. Vectors must have a power
1510of two length (1, 2, 4, 8, 16 ...). Vector types are
1511considered <a href="#t_firstclass">first class</a>.</p>
1512
1513<h5>Syntax:</h5>
1514
1515<pre>
1516 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1517</pre>
1518
1519<p>The number of elements is a constant integer value; elementtype may
1520be any integer or floating point type.</p>
1521
1522<h5>Examples:</h5>
1523
1524<table class="layout">
1525 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1526 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
1527 <td class="left">Vector of 4 32-bit integer values.</td>
1528 </tr>
1529 <tr class="layout">
1530 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
1531 <td class="left">Vector of 8 32-bit floating-point values.</td>
1532 </tr>
1533 <tr class="layout">
1534 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
1535 <td class="left">Vector of 2 64-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1536 </tr>
1537</table>
1538</div>
1539
1540<!-- _______________________________________________________________________ -->
1541<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1542<div class="doc_text">
1543
1544<h5>Overview:</h5>
1545
1546<p>Opaque types are used to represent unknown types in the system. This
Gordon Henriksenda0706e2007-10-14 00:34:53 +0000[diff] [blame]1547corresponds (for example) to the C notion of a forward declared structure type.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1548In LLVM, opaque types can eventually be resolved to any type (not just a
1549structure type).</p>
1550
1551<h5>Syntax:</h5>
1552
1553<pre>
1554 opaque
1555</pre>
1556
1557<h5>Examples:</h5>
1558
1559<table class="layout">
1560 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1561 <td class="left"><tt>opaque</tt></td>
1562 <td class="left">An opaque type.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1563 </tr>
1564</table>
1565</div>
1566
1567
1568<!-- *********************************************************************** -->
1569<div class="doc_section"> <a name="constants">Constants</a> </div>
1570<!-- *********************************************************************** -->
1571
1572<div class="doc_text">
1573
1574<p>LLVM has several different basic types of constants. This section describes
1575them all and their syntax.</p>
1576
1577</div>
1578
1579<!-- ======================================================================= -->
1580<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
1581
1582<div class="doc_text">
1583
1584<dl>
1585 <dt><b>Boolean constants</b></dt>
1586
1587 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
1588 constants of the <tt><a href="#t_primitive">i1</a></tt> type.
1589 </dd>
1590
1591 <dt><b>Integer constants</b></dt>
1592
1593 <dd>Standard integers (such as '4') are constants of the <a
1594 href="#t_integer">integer</a> type. Negative numbers may be used with
1595 integer types.
1596 </dd>
1597
1598 <dt><b>Floating point constants</b></dt>
1599
1600 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
1601 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]1602 notation (see below). The assembler requires the exact decimal value of
1603 a floating-point constant. For example, the assembler accepts 1.25 but
1604 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
1605 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1606
1607 <dt><b>Null pointer constants</b></dt>
1608
1609 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
1610 and must be of <a href="#t_pointer">pointer type</a>.</dd>
1611
1612</dl>
1613
1614<p>The one non-intuitive notation for constants is the optional hexadecimal form
1615of floating point constants. For example, the form '<tt>double
16160x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
16174.5e+15</tt>'. The only time hexadecimal floating point constants are required
1618(and the only time that they are generated by the disassembler) is when a
1619floating point constant must be emitted but it cannot be represented as a
1620decimal floating point number. For example, NaN's, infinities, and other
1621special values are represented in their IEEE hexadecimal format so that
1622assembly and disassembly do not cause any bits to change in the constants.</p>
1623
1624</div>
1625
1626<!-- ======================================================================= -->
1627<div class="doc_subsection"><a name="aggregateconstants">Aggregate Constants</a>
1628</div>
1629
1630<div class="doc_text">
1631<p>Aggregate constants arise from aggregation of simple constants
1632and smaller aggregate constants.</p>
1633
1634<dl>
1635 <dt><b>Structure constants</b></dt>
1636
1637 <dd>Structure constants are represented with notation similar to structure
1638 type definitions (a comma separated list of elements, surrounded by braces
Chris Lattner6c8de962007-12-25 20:34:52 +0000[diff] [blame]1639 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
1640 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]1641 must have <a href="#t_struct">structure type</a>, and the number and
1642 types of elements must match those specified by the type.
1643 </dd>
1644
1645 <dt><b>Array constants</b></dt>
1646
1647 <dd>Array constants are represented with notation similar to array type
1648 definitions (a comma separated list of elements, surrounded by square brackets
1649 (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74 ]</tt>". Array
1650 constants must have <a href="#t_array">array type</a>, and the number and
1651 types of elements must match those specified by the type.
1652 </dd>
1653
1654 <dt><b>Vector constants</b></dt>
1655
1656 <dd>Vector constants are represented with notation similar to vector type
1657 definitions (a comma separated list of elements, surrounded by
1658 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32 42,
1659 i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must have <a
1660 href="#t_vector">vector type</a>, and the number and types of elements must
1661 match those specified by the type.
1662 </dd>
1663
1664 <dt><b>Zero initialization</b></dt>
1665
1666 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1667 value to zero of <em>any</em> type, including scalar and aggregate types.
1668 This is often used to avoid having to print large zero initializers (e.g. for
1669 large arrays) and is always exactly equivalent to using explicit zero
1670 initializers.
1671 </dd>
1672</dl>
1673
1674</div>
1675
1676<!-- ======================================================================= -->
1677<div class="doc_subsection">
1678 <a name="globalconstants">Global Variable and Function Addresses</a>
1679</div>
1680
1681<div class="doc_text">
1682
1683<p>The addresses of <a href="#globalvars">global variables</a> and <a
1684href="#functionstructure">functions</a> are always implicitly valid (link-time)
1685constants. These constants are explicitly referenced when the <a
1686href="#identifiers">identifier for the global</a> is used and always have <a
1687href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1688file:</p>
1689
1690<div class="doc_code">
1691<pre>
1692@X = global i32 17
1693@Y = global i32 42
1694@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
1695</pre>
1696</div>
1697
1698</div>
1699
1700<!-- ======================================================================= -->
1701<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
1702<div class="doc_text">
1703 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
1704 no specific value. Undefined values may be of any type and be used anywhere
1705 a constant is permitted.</p>
1706
1707 <p>Undefined values indicate to the compiler that the program is well defined
1708 no matter what value is used, giving the compiler more freedom to optimize.
1709 </p>
1710</div>
1711
1712<!-- ======================================================================= -->
1713<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1714</div>
1715
1716<div class="doc_text">
1717
1718<p>Constant expressions are used to allow expressions involving other constants
1719to be used as constants. Constant expressions may be of any <a
1720href="#t_firstclass">first class</a> type and may involve any LLVM operation
1721that does not have side effects (e.g. load and call are not supported). The
1722following is the syntax for constant expressions:</p>
1723
1724<dl>
1725 <dt><b><tt>trunc ( CST to TYPE )</tt></b></dt>
1726 <dd>Truncate a constant to another type. The bit size of CST must be larger
1727 than the bit size of TYPE. Both types must be integers.</dd>
1728
1729 <dt><b><tt>zext ( CST to TYPE )</tt></b></dt>
1730 <dd>Zero extend a constant to another type. The bit size of CST must be
1731 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1732
1733 <dt><b><tt>sext ( CST to TYPE )</tt></b></dt>
1734 <dd>Sign extend a constant to another type. The bit size of CST must be
1735 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1736
1737 <dt><b><tt>fptrunc ( CST to TYPE )</tt></b></dt>
1738 <dd>Truncate a floating point constant to another floating point type. The
1739 size of CST must be larger than the size of TYPE. Both types must be
1740 floating point.</dd>
1741
1742 <dt><b><tt>fpext ( CST to TYPE )</tt></b></dt>
1743 <dd>Floating point extend a constant to another type. The size of CST must be
1744 smaller or equal to the size of TYPE. Both types must be floating point.</dd>
1745
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]1746 <dt><b><tt>fptoui ( CST to TYPE )</tt></b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1747 <dd>Convert a floating point constant to the corresponding unsigned integer
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1748 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1749 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1750 of the same number of elements. If the value won't fit in the integer type,
1751 the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1752
1753 <dt><b><tt>fptosi ( CST to TYPE )</tt></b></dt>
1754 <dd>Convert a floating point constant to the corresponding signed integer
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1755 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1756 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1757 of the same number of elements. If the value won't fit in the integer type,
1758 the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1759
1760 <dt><b><tt>uitofp ( CST to TYPE )</tt></b></dt>
1761 <dd>Convert an unsigned integer constant to the corresponding floating point
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1762 constant. TYPE must be a scalar or vector floating point type. CST must be of
1763 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1764 of the same number of elements. If the value won't fit in the floating point
1765 type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1766
1767 <dt><b><tt>sitofp ( CST to TYPE )</tt></b></dt>
1768 <dd>Convert a signed integer constant to the corresponding floating point
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1769 constant. TYPE must be a scalar or vector floating point type. CST must be of
1770 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1771 of the same number of elements. If the value won't fit in the floating point
1772 type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1773
1774 <dt><b><tt>ptrtoint ( CST to TYPE )</tt></b></dt>
1775 <dd>Convert a pointer typed constant to the corresponding integer constant
1776 TYPE must be an integer type. CST must be of pointer type. The CST value is
1777 zero extended, truncated, or unchanged to make it fit in TYPE.</dd>
1778
1779 <dt><b><tt>inttoptr ( CST to TYPE )</tt></b></dt>
1780 <dd>Convert a integer constant to a pointer constant. TYPE must be a
1781 pointer type. CST must be of integer type. The CST value is zero extended,
1782 truncated, or unchanged to make it fit in a pointer size. This one is
1783 <i>really</i> dangerous!</dd>
1784
1785 <dt><b><tt>bitcast ( CST to TYPE )</tt></b></dt>
1786 <dd>Convert a constant, CST, to another TYPE. The size of CST and TYPE must be
1787 identical (same number of bits). The conversion is done as if the CST value
1788 was stored to memory and read back as TYPE. In other words, no bits change
1789 with this operator, just the type. This can be used for conversion of
1790 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]1791 pointers it is only valid to cast to another pointer type. It is not valid
1792 to bitcast to or from an aggregate type.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1793 </dd>
1794
1795 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
1796
1797 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
1798 constants. As with the <a href="#i_getelementptr">getelementptr</a>
1799 instruction, the index list may have zero or more indexes, which are required
1800 to make sense for the type of "CSTPTR".</dd>
1801
1802 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
1803
1804 <dd>Perform the <a href="#i_select">select operation</a> on
1805 constants.</dd>
1806
1807 <dt><b><tt>icmp COND ( VAL1, VAL2 )</tt></b></dt>
1808 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
1809
1810 <dt><b><tt>fcmp COND ( VAL1, VAL2 )</tt></b></dt>
1811 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
1812
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]1813 <dt><b><tt>vicmp COND ( VAL1, VAL2 )</tt></b></dt>
1814 <dd>Performs the <a href="#i_vicmp">vicmp operation</a> on constants.</dd>
1815
1816 <dt><b><tt>vfcmp COND ( VAL1, VAL2 )</tt></b></dt>
1817 <dd>Performs the <a href="#i_vfcmp">vfcmp operation</a> on constants.</dd>
1818
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1819 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
1820
1821 <dd>Perform the <a href="#i_extractelement">extractelement
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]1822 operation</a> on constants.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1823
1824 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
1825
1826 <dd>Perform the <a href="#i_insertelement">insertelement
1827 operation</a> on constants.</dd>
1828
1829
1830 <dt><b><tt>shufflevector ( VEC1, VEC2, IDXMASK )</tt></b></dt>
1831
1832 <dd>Perform the <a href="#i_shufflevector">shufflevector
1833 operation</a> on constants.</dd>
1834
1835 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
1836
1837 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
1838 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
1839 binary</a> operations. The constraints on operands are the same as those for
1840 the corresponding instruction (e.g. no bitwise operations on floating point
1841 values are allowed).</dd>
1842</dl>
1843</div>
1844
1845<!-- *********************************************************************** -->
1846<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
1847<!-- *********************************************************************** -->
1848
1849<!-- ======================================================================= -->
1850<div class="doc_subsection">
1851<a name="inlineasm">Inline Assembler Expressions</a>
1852</div>
1853
1854<div class="doc_text">
1855
1856<p>
1857LLVM supports inline assembler expressions (as opposed to <a href="#moduleasm">
1858Module-Level Inline Assembly</a>) through the use of a special value. This
1859value represents the inline assembler as a string (containing the instructions
1860to emit), a list of operand constraints (stored as a string), and a flag that
1861indicates whether or not the inline asm expression has side effects. An example
1862inline assembler expression is:
1863</p>
1864
1865<div class="doc_code">
1866<pre>
1867i32 (i32) asm "bswap $0", "=r,r"
1868</pre>
1869</div>
1870
1871<p>
1872Inline assembler expressions may <b>only</b> be used as the callee operand of
1873a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we have:
1874</p>
1875
1876<div class="doc_code">
1877<pre>
1878%X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
1879</pre>
1880</div>
1881
1882<p>
1883Inline asms with side effects not visible in the constraint list must be marked
1884as having side effects. This is done through the use of the
1885'<tt>sideeffect</tt>' keyword, like so:
1886</p>
1887
1888<div class="doc_code">
1889<pre>
1890call void asm sideeffect "eieio", ""()
1891</pre>
1892</div>
1893
1894<p>TODO: The format of the asm and constraints string still need to be
1895documented here. Constraints on what can be done (e.g. duplication, moving, etc
Chris Lattner1cdd64e2008-10-04 18:36:02 +0000[diff] [blame]1896need to be documented). This is probably best done by reference to another
1897document that covers inline asm from a holistic perspective.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1898</p>
1899
1900</div>
1901
1902<!-- *********************************************************************** -->
1903<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
1904<!-- *********************************************************************** -->
1905
1906<div class="doc_text">
1907
1908<p>The LLVM instruction set consists of several different
1909classifications of instructions: <a href="#terminators">terminator
1910instructions</a>, <a href="#binaryops">binary instructions</a>,
1911<a href="#bitwiseops">bitwise binary instructions</a>, <a
1912 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
1913instructions</a>.</p>
1914
1915</div>
1916
1917<!-- ======================================================================= -->
1918<div class="doc_subsection"> <a name="terminators">Terminator
1919Instructions</a> </div>
1920
1921<div class="doc_text">
1922
1923<p>As mentioned <a href="#functionstructure">previously</a>, every
1924basic block in a program ends with a "Terminator" instruction, which
1925indicates which block should be executed after the current block is
1926finished. These terminator instructions typically yield a '<tt>void</tt>'
1927value: they produce control flow, not values (the one exception being
1928the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
1929<p>There are six different terminator instructions: the '<a
1930 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
1931instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
1932the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
1933 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
1934 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
1935
1936</div>
1937
1938<!-- _______________________________________________________________________ -->
1939<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
1940Instruction</a> </div>
1941<div class="doc_text">
1942<h5>Syntax:</h5>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]1943<pre>
1944 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]1945 ret void <i>; Return from void function</i>
1946</pre>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1947
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1948<h5>Overview:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1949
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]1950<p>The '<tt>ret</tt>' instruction is used to return control flow (and
1951optionally a value) from a function back to the caller.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1952<p>There are two forms of the '<tt>ret</tt>' instruction: one that
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]1953returns a value and then causes control flow, and one that just causes
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1954control flow to occur.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1955
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1956<h5>Arguments:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1957
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]1958<p>The '<tt>ret</tt>' instruction optionally accepts a single argument,
1959the return value. The type of the return value must be a
1960'<a href="#t_firstclass">first class</a>' type.</p>
1961
1962<p>A function is not <a href="#wellformed">well formed</a> if
1963it it has a non-void return type and contains a '<tt>ret</tt>'
1964instruction with no return value or a return value with a type that
1965does not match its type, or if it has a void return type and contains
1966a '<tt>ret</tt>' instruction with a return value.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1967
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1968<h5>Semantics:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1969
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1970<p>When the '<tt>ret</tt>' instruction is executed, control flow
1971returns back to the calling function's context. If the caller is a "<a
1972 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
1973the instruction after the call. If the caller was an "<a
1974 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
1975at the beginning of the "normal" destination block. If the instruction
1976returns a value, that value shall set the call or invoke instruction's
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]1977return value.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1978
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1979<h5>Example:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1980
1981<pre>
1982 ret i32 5 <i>; Return an integer value of 5</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1983 ret void <i>; Return from a void function</i>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]1984 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]1985</pre>
1986</div>
1987<!-- _______________________________________________________________________ -->
1988<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
1989<div class="doc_text">
1990<h5>Syntax:</h5>
1991<pre> br i1 &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;<br> br label &lt;dest&gt; <i>; Unconditional branch</i>
1992</pre>
1993<h5>Overview:</h5>
1994<p>The '<tt>br</tt>' instruction is used to cause control flow to
1995transfer to a different basic block in the current function. There are
1996two forms of this instruction, corresponding to a conditional branch
1997and an unconditional branch.</p>
1998<h5>Arguments:</h5>
1999<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
2000single '<tt>i1</tt>' value and two '<tt>label</tt>' values. The
2001unconditional form of the '<tt>br</tt>' instruction takes a single
2002'<tt>label</tt>' value as a target.</p>
2003<h5>Semantics:</h5>
2004<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
2005argument is evaluated. If the value is <tt>true</tt>, control flows
2006to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
2007control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
2008<h5>Example:</h5>
2009<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
2010 href="#i_ret">ret</a> i32 1<br>IfUnequal:<br> <a href="#i_ret">ret</a> i32 0<br></pre>
2011</div>
2012<!-- _______________________________________________________________________ -->
2013<div class="doc_subsubsection">
2014 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
2015</div>
2016
2017<div class="doc_text">
2018<h5>Syntax:</h5>
2019
2020<pre>
2021 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
2022</pre>
2023
2024<h5>Overview:</h5>
2025
2026<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
2027several different places. It is a generalization of the '<tt>br</tt>'
2028instruction, allowing a branch to occur to one of many possible
2029destinations.</p>
2030
2031
2032<h5>Arguments:</h5>
2033
2034<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
2035comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
2036an array of pairs of comparison value constants and '<tt>label</tt>'s. The
2037table is not allowed to contain duplicate constant entries.</p>
2038
2039<h5>Semantics:</h5>
2040
2041<p>The <tt>switch</tt> instruction specifies a table of values and
2042destinations. When the '<tt>switch</tt>' instruction is executed, this
2043table is searched for the given value. If the value is found, control flow is
2044transfered to the corresponding destination; otherwise, control flow is
2045transfered to the default destination.</p>
2046
2047<h5>Implementation:</h5>
2048
2049<p>Depending on properties of the target machine and the particular
2050<tt>switch</tt> instruction, this instruction may be code generated in different
2051ways. For example, it could be generated as a series of chained conditional
2052branches or with a lookup table.</p>
2053
2054<h5>Example:</h5>
2055
2056<pre>
2057 <i>; Emulate a conditional br instruction</i>
2058 %Val = <a href="#i_zext">zext</a> i1 %value to i32
2059 switch i32 %Val, label %truedest [i32 0, label %falsedest ]
2060
2061 <i>; Emulate an unconditional br instruction</i>
2062 switch i32 0, label %dest [ ]
2063
2064 <i>; Implement a jump table:</i>
2065 switch i32 %val, label %otherwise [ i32 0, label %onzero
2066 i32 1, label %onone
2067 i32 2, label %ontwo ]
2068</pre>
2069</div>
2070
2071<!-- _______________________________________________________________________ -->
2072<div class="doc_subsubsection">
2073 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
2074</div>
2075
2076<div class="doc_text">
2077
2078<h5>Syntax:</h5>
2079
2080<pre>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]2081 &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]2082 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
2083</pre>
2084
2085<h5>Overview:</h5>
2086
2087<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
2088function, with the possibility of control flow transfer to either the
2089'<tt>normal</tt>' label or the
2090'<tt>exception</tt>' label. If the callee function returns with the
2091"<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
2092"normal" label. If the callee (or any indirect callees) returns with the "<a
2093href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]2094continued at the dynamically nearest "exception" label.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2095
2096<h5>Arguments:</h5>
2097
2098<p>This instruction requires several arguments:</p>
2099
2100<ol>
2101 <li>
2102 The optional "cconv" marker indicates which <a href="#callingconv">calling
2103 convention</a> the call should use. If none is specified, the call defaults
2104 to using C calling conventions.
2105 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]2106
2107 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
2108 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>',
2109 and '<tt>inreg</tt>' attributes are valid here.</li>
2110
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2111 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
2112 function value being invoked. In most cases, this is a direct function
2113 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
2114 an arbitrary pointer to function value.
2115 </li>
2116
2117 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
2118 function to be invoked. </li>
2119
2120 <li>'<tt>function args</tt>': argument list whose types match the function
2121 signature argument types. If the function signature indicates the function
2122 accepts a variable number of arguments, the extra arguments can be
2123 specified. </li>
2124
2125 <li>'<tt>normal label</tt>': the label reached when the called function
2126 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
2127
2128 <li>'<tt>exception label</tt>': the label reached when a callee returns with
2129 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
2130
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]2131 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]2132 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
2133 '<tt>readnone</tt>' attributes are valid here.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2134</ol>
2135
2136<h5>Semantics:</h5>
2137
2138<p>This instruction is designed to operate as a standard '<tt><a
2139href="#i_call">call</a></tt>' instruction in most regards. The primary
2140difference is that it establishes an association with a label, which is used by
2141the runtime library to unwind the stack.</p>
2142
2143<p>This instruction is used in languages with destructors to ensure that proper
2144cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
2145exception. Additionally, this is important for implementation of
2146'<tt>catch</tt>' clauses in high-level languages that support them.</p>
2147
2148<h5>Example:</h5>
2149<pre>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2150 %retval = invoke i32 @Test(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2151 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2152 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2153 unwind label %TestCleanup <i>; {i32}:retval set</i>
2154</pre>
2155</div>
2156
2157
2158<!-- _______________________________________________________________________ -->
2159
2160<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
2161Instruction</a> </div>
2162
2163<div class="doc_text">
2164
2165<h5>Syntax:</h5>
2166<pre>
2167 unwind
2168</pre>
2169
2170<h5>Overview:</h5>
2171
2172<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
2173at the first callee in the dynamic call stack which used an <a
2174href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
2175primarily used to implement exception handling.</p>
2176
2177<h5>Semantics:</h5>
2178
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]2179<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2180immediately halt. The dynamic call stack is then searched for the first <a
2181href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
2182execution continues at the "exceptional" destination block specified by the
2183<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
2184dynamic call chain, undefined behavior results.</p>
2185</div>
2186
2187<!-- _______________________________________________________________________ -->
2188
2189<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
2190Instruction</a> </div>
2191
2192<div class="doc_text">
2193
2194<h5>Syntax:</h5>
2195<pre>
2196 unreachable
2197</pre>
2198
2199<h5>Overview:</h5>
2200
2201<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
2202instruction is used to inform the optimizer that a particular portion of the
2203code is not reachable. This can be used to indicate that the code after a
2204no-return function cannot be reached, and other facts.</p>
2205
2206<h5>Semantics:</h5>
2207
2208<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
2209</div>
2210
2211
2212
2213<!-- ======================================================================= -->
2214<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
2215<div class="doc_text">
2216<p>Binary operators are used to do most of the computation in a
Chris Lattnerab596d92008-04-01 18:47:32 +0000[diff] [blame]2217program. They require two operands of the same type, execute an operation on them, and
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2218produce a single value. The operands might represent
2219multiple 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]2220The result value has the same type as its operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2221<p>There are several different binary operators:</p>
2222</div>
2223<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2224<div class="doc_subsubsection">
2225 <a name="i_add">'<tt>add</tt>' Instruction</a>
2226</div>
2227
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2228<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2229
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2230<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2231
2232<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2233 &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]2234</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2235
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2236<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2237
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2238<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2239
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2240<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2241
2242<p>The two arguments to the '<tt>add</tt>' instruction must be <a
2243 href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>, or
2244 <a href="#t_vector">vector</a> values. Both arguments must have identical
2245 types.</p>
2246
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2247<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2248
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2249<p>The value produced is the integer or floating point sum of the two
2250operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2251
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2252<p>If an integer sum has unsigned overflow, the result returned is the
2253mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2254the result.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2255
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2256<p>Because LLVM integers use a two's complement representation, this
2257instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2258
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2259<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2260
2261<pre>
2262 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2263</pre>
2264</div>
2265<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2266<div class="doc_subsubsection">
2267 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
2268</div>
2269
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2270<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2271
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2272<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2273
2274<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2275 &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]2276</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2277
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2278<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2279
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2280<p>The '<tt>sub</tt>' instruction returns the difference of its two
2281operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2282
2283<p>Note that the '<tt>sub</tt>' instruction is used to represent the
2284'<tt>neg</tt>' instruction present in most other intermediate
2285representations.</p>
2286
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2287<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2288
2289<p>The two arguments to the '<tt>sub</tt>' instruction must be <a
2290 href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>,
2291 or <a href="#t_vector">vector</a> values. Both arguments must have identical
2292 types.</p>
2293
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2294<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2295
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2296<p>The value produced is the integer or floating point difference of
2297the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2298
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2299<p>If an integer difference has unsigned overflow, the result returned is the
2300mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2301the result.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2302
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2303<p>Because LLVM integers use a two's complement representation, this
2304instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2305
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2306<h5>Example:</h5>
2307<pre>
2308 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
2309 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
2310</pre>
2311</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2312
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2313<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2314<div class="doc_subsubsection">
2315 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
2316</div>
2317
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2318<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2319
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2320<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2321<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]2322</pre>
2323<h5>Overview:</h5>
2324<p>The '<tt>mul</tt>' instruction returns the product of its two
2325operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2326
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2327<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2328
2329<p>The two arguments to the '<tt>mul</tt>' instruction must be <a
2330href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>,
2331or <a href="#t_vector">vector</a> values. Both arguments must have identical
2332types.</p>
2333
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2334<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2335
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2336<p>The value produced is the integer or floating point product of the
2337two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2338
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2339<p>If the result of an integer multiplication has unsigned overflow,
2340the result returned is the mathematical result modulo
23412<sup>n</sup>, where n is the bit width of the result.</p>
2342<p>Because LLVM integers use a two's complement representation, and the
2343result is the same width as the operands, this instruction returns the
2344correct result for both signed and unsigned integers. If a full product
2345(e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands
2346should be sign-extended or zero-extended as appropriate to the
2347width of the full product.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2348<h5>Example:</h5>
2349<pre> &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
2350</pre>
2351</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2352
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2353<!-- _______________________________________________________________________ -->
2354<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
2355</a></div>
2356<div class="doc_text">
2357<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2358<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]2359</pre>
2360<h5>Overview:</h5>
2361<p>The '<tt>udiv</tt>' instruction returns the quotient of its two
2362operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2363
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2364<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2365
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2366<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2367<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2368values. Both arguments must have identical types.</p>
2369
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2370<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2371
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2372<p>The value produced is the unsigned integer quotient of the two operands.</p>
2373<p>Note that unsigned integer division and signed integer division are distinct
2374operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
2375<p>Division by zero leads to undefined behavior.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2376<h5>Example:</h5>
2377<pre> &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2378</pre>
2379</div>
2380<!-- _______________________________________________________________________ -->
2381<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
2382</a> </div>
2383<div class="doc_text">
2384<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2385<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2386 &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]2387</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2388
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2389<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2390
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2391<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two
2392operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2393
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2394<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2395
2396<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
2397<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2398values. Both arguments must have identical types.</p>
2399
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2400<h5>Semantics:</h5>
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2401<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]2402<p>Note that signed integer division and unsigned integer division are distinct
2403operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
2404<p>Division by zero leads to undefined behavior. Overflow also leads to
2405undefined behavior; this is a rare case, but can occur, for example,
2406by doing a 32-bit division of -2147483648 by -1.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2407<h5>Example:</h5>
2408<pre> &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2409</pre>
2410</div>
2411<!-- _______________________________________________________________________ -->
2412<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
2413Instruction</a> </div>
2414<div class="doc_text">
2415<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2416<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2417 &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]2418</pre>
2419<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2420
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2421<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two
2422operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2423
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2424<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2425
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2426<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2427<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2428of floating point values. Both arguments must have identical types.</p>
2429
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2430<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2431
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2432<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2433
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2434<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2435
2436<pre>
2437 &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]2438</pre>
2439</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2440
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2441<!-- _______________________________________________________________________ -->
2442<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
2443</div>
2444<div class="doc_text">
2445<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2446<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]2447</pre>
2448<h5>Overview:</h5>
2449<p>The '<tt>urem</tt>' instruction returns the remainder from the
2450unsigned division of its two arguments.</p>
2451<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2452<p>The two arguments to the '<tt>urem</tt>' instruction must be
2453<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2454values. Both arguments must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2455<h5>Semantics:</h5>
2456<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2457This instruction always performs an unsigned division to get the remainder.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2458<p>Note that unsigned integer remainder and signed integer remainder are
2459distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
2460<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2461<h5>Example:</h5>
2462<pre> &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2463</pre>
2464
2465</div>
2466<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2467<div class="doc_subsubsection">
2468 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
2469</div>
2470
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2471<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2472
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2473<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2474
2475<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2476 &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]2477</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2478
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2479<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2480
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2481<p>The '<tt>srem</tt>' instruction returns the remainder from the
Dan Gohman3e3fd8c2007-11-05 23:35:22 +0000[diff] [blame]2482signed division of its two operands. This instruction can also take
2483<a href="#t_vector">vector</a> versions of the values in which case
2484the elements must be integers.</p>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]2485
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2486<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2487
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2488<p>The two arguments to the '<tt>srem</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2489<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2490values. Both arguments must have identical types.</p>
2491
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2492<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2493
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2494<p>This instruction returns the <i>remainder</i> of a division (where the result
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2495has the same sign as the dividend, <tt>op1</tt>), not the <i>modulo</i>
2496operator (where the result has the same sign as the divisor, <tt>op2</tt>) of
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2497a value. For more information about the difference, see <a
2498 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
2499Math Forum</a>. For a table of how this is implemented in various languages,
2500please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
2501Wikipedia: modulo operation</a>.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2502<p>Note that signed integer remainder and unsigned integer remainder are
2503distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
2504<p>Taking the remainder of a division by zero leads to undefined behavior.
2505Overflow also leads to undefined behavior; this is a rare case, but can occur,
2506for example, by taking the remainder of a 32-bit division of -2147483648 by -1.
2507(The remainder doesn't actually overflow, but this rule lets srem be
2508implemented using instructions that return both the result of the division
2509and the remainder.)</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2510<h5>Example:</h5>
2511<pre> &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2512</pre>
2513
2514</div>
2515<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2516<div class="doc_subsubsection">
2517 <a name="i_frem">'<tt>frem</tt>' Instruction</a> </div>
2518
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2519<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2520
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2521<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2522<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]2523</pre>
2524<h5>Overview:</h5>
2525<p>The '<tt>frem</tt>' instruction returns the remainder from the
2526division of its two operands.</p>
2527<h5>Arguments:</h5>
2528<p>The two arguments to the '<tt>frem</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2529<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2530of floating point values. Both arguments must have identical types.</p>
2531
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2532<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2533
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2534<p>This instruction returns the <i>remainder</i> of a division.
2535The remainder has the same sign as the dividend.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2536
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2537<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2538
2539<pre>
2540 &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]2541</pre>
2542</div>
2543
2544<!-- ======================================================================= -->
2545<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
2546Operations</a> </div>
2547<div class="doc_text">
2548<p>Bitwise binary operators are used to do various forms of
2549bit-twiddling in a program. They are generally very efficient
2550instructions and can commonly be strength reduced from other
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2551instructions. They require two operands of the same type, execute an operation on them,
2552and 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]2553</div>
2554
2555<!-- _______________________________________________________________________ -->
2556<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
2557Instruction</a> </div>
2558<div class="doc_text">
2559<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2560<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]2561</pre>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2562
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2563<h5>Overview:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2564
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2565<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
2566the left a specified number of bits.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2567
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2568<h5>Arguments:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2569
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2570<p>Both arguments to the '<tt>shl</tt>' instruction must be the same <a
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]2571 href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2572type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2573
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2574<h5>Semantics:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2575
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2576<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod 2<sup>n</sup>,
2577where n is the width of the result. If <tt>op2</tt> is (statically or dynamically) negative or
2578equal 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]2579
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2580<h5>Example:</h5><pre>
2581 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
2582 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
2583 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2584 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2585</pre>
2586</div>
2587<!-- _______________________________________________________________________ -->
2588<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
2589Instruction</a> </div>
2590<div class="doc_text">
2591<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2592<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]2593</pre>
2594
2595<h5>Overview:</h5>
2596<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
2597operand shifted to the right a specified number of bits with zero fill.</p>
2598
2599<h5>Arguments:</h5>
2600<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]2601<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]2602type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2603
2604<h5>Semantics:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2605
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2606<p>This instruction always performs a logical shift right operation. The most
2607significant bits of the result will be filled with zero bits after the
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2608shift. If <tt>op2</tt> is (statically or dynamically) equal to or larger than
2609the number of bits in <tt>op1</tt>, the result is undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2610
2611<h5>Example:</h5>
2612<pre>
2613 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
2614 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
2615 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
2616 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2617 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2618</pre>
2619</div>
2620
2621<!-- _______________________________________________________________________ -->
2622<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
2623Instruction</a> </div>
2624<div class="doc_text">
2625
2626<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2627<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]2628</pre>
2629
2630<h5>Overview:</h5>
2631<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
2632operand shifted to the right a specified number of bits with sign extension.</p>
2633
2634<h5>Arguments:</h5>
2635<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]2636<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]2637type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2638
2639<h5>Semantics:</h5>
2640<p>This instruction always performs an arithmetic shift right operation,
2641The most significant bits of the result will be filled with the sign bit
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2642of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
2643larger than the number of bits in <tt>op1</tt>, the result is undefined.
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2644</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2645
2646<h5>Example:</h5>
2647<pre>
2648 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
2649 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
2650 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
2651 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2652 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2653</pre>
2654</div>
2655
2656<!-- _______________________________________________________________________ -->
2657<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
2658Instruction</a> </div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2659
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2660<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2661
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2662<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2663
2664<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2665 &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]2666</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2667
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2668<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2669
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2670<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
2671its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2672
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2673<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2674
2675<p>The two arguments to the '<tt>and</tt>' instruction must be
2676<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2677values. Both arguments must have identical types.</p>
2678
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2679<h5>Semantics:</h5>
2680<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
2681<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]2682<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2683<table border="1" cellspacing="0" cellpadding="4">
2684 <tbody>
2685 <tr>
2686 <td>In0</td>
2687 <td>In1</td>
2688 <td>Out</td>
2689 </tr>
2690 <tr>
2691 <td>0</td>
2692 <td>0</td>
2693 <td>0</td>
2694 </tr>
2695 <tr>
2696 <td>0</td>
2697 <td>1</td>
2698 <td>0</td>
2699 </tr>
2700 <tr>
2701 <td>1</td>
2702 <td>0</td>
2703 <td>0</td>
2704 </tr>
2705 <tr>
2706 <td>1</td>
2707 <td>1</td>
2708 <td>1</td>
2709 </tr>
2710 </tbody>
2711</table>
2712</div>
2713<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2714<pre>
2715 &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]2716 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
2717 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
2718</pre>
2719</div>
2720<!-- _______________________________________________________________________ -->
2721<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
2722<div class="doc_text">
2723<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2724<pre> &lt;result&gt; = 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]2725</pre>
2726<h5>Overview:</h5>
2727<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
2728or of its two operands.</p>
2729<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2730
2731<p>The two arguments to the '<tt>or</tt>' instruction must be
2732<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2733values. Both arguments must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2734<h5>Semantics:</h5>
2735<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
2736<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]2737<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2738<table border="1" cellspacing="0" cellpadding="4">
2739 <tbody>
2740 <tr>
2741 <td>In0</td>
2742 <td>In1</td>
2743 <td>Out</td>
2744 </tr>
2745 <tr>
2746 <td>0</td>
2747 <td>0</td>
2748 <td>0</td>
2749 </tr>
2750 <tr>
2751 <td>0</td>
2752 <td>1</td>
2753 <td>1</td>
2754 </tr>
2755 <tr>
2756 <td>1</td>
2757 <td>0</td>
2758 <td>1</td>
2759 </tr>
2760 <tr>
2761 <td>1</td>
2762 <td>1</td>
2763 <td>1</td>
2764 </tr>
2765 </tbody>
2766</table>
2767</div>
2768<h5>Example:</h5>
2769<pre> &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
2770 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
2771 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
2772</pre>
2773</div>
2774<!-- _______________________________________________________________________ -->
2775<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
2776Instruction</a> </div>
2777<div class="doc_text">
2778<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2779<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]2780</pre>
2781<h5>Overview:</h5>
2782<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
2783or of its two operands. The <tt>xor</tt> is used to implement the
2784"one's complement" operation, which is the "~" operator in C.</p>
2785<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2786<p>The two arguments to the '<tt>xor</tt>' instruction must be
2787<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2788values. Both arguments must have identical types.</p>
2789
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2790<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2791
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2792<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
2793<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]2794<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2795<table border="1" cellspacing="0" cellpadding="4">
2796 <tbody>
2797 <tr>
2798 <td>In0</td>
2799 <td>In1</td>
2800 <td>Out</td>
2801 </tr>
2802 <tr>
2803 <td>0</td>
2804 <td>0</td>
2805 <td>0</td>
2806 </tr>
2807 <tr>
2808 <td>0</td>
2809 <td>1</td>
2810 <td>1</td>
2811 </tr>
2812 <tr>
2813 <td>1</td>
2814 <td>0</td>
2815 <td>1</td>
2816 </tr>
2817 <tr>
2818 <td>1</td>
2819 <td>1</td>
2820 <td>0</td>
2821 </tr>
2822 </tbody>
2823</table>
2824</div>
2825<p> </p>
2826<h5>Example:</h5>
2827<pre> &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
2828 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
2829 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
2830 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
2831</pre>
2832</div>
2833
2834<!-- ======================================================================= -->
2835<div class="doc_subsection">
2836 <a name="vectorops">Vector Operations</a>
2837</div>
2838
2839<div class="doc_text">
2840
2841<p>LLVM supports several instructions to represent vector operations in a
2842target-independent manner. These instructions cover the element-access and
2843vector-specific operations needed to process vectors effectively. While LLVM
2844does directly support these vector operations, many sophisticated algorithms
2845will want to use target-specific intrinsics to take full advantage of a specific
2846target.</p>
2847
2848</div>
2849
2850<!-- _______________________________________________________________________ -->
2851<div class="doc_subsubsection">
2852 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
2853</div>
2854
2855<div class="doc_text">
2856
2857<h5>Syntax:</h5>
2858
2859<pre>
2860 &lt;result&gt; = extractelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, i32 &lt;idx&gt; <i>; yields &lt;ty&gt;</i>
2861</pre>
2862
2863<h5>Overview:</h5>
2864
2865<p>
2866The '<tt>extractelement</tt>' instruction extracts a single scalar
2867element from a vector at a specified index.
2868</p>
2869
2870
2871<h5>Arguments:</h5>
2872
2873<p>
2874The first operand of an '<tt>extractelement</tt>' instruction is a
2875value of <a href="#t_vector">vector</a> type. The second operand is
2876an index indicating the position from which to extract the element.
2877The index may be a variable.</p>
2878
2879<h5>Semantics:</h5>
2880
2881<p>
2882The result is a scalar of the same type as the element type of
2883<tt>val</tt>. Its value is the value at position <tt>idx</tt> of
2884<tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
2885results are undefined.
2886</p>
2887
2888<h5>Example:</h5>
2889
2890<pre>
2891 %result = extractelement &lt;4 x i32&gt; %vec, i32 0 <i>; yields i32</i>
2892</pre>
2893</div>
2894
2895
2896<!-- _______________________________________________________________________ -->
2897<div class="doc_subsubsection">
2898 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
2899</div>
2900
2901<div class="doc_text">
2902
2903<h5>Syntax:</h5>
2904
2905<pre>
Dan Gohmanbcc3c502008-05-12 23:38:42 +0000[diff] [blame]2906 &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]2907</pre>
2908
2909<h5>Overview:</h5>
2910
2911<p>
2912The '<tt>insertelement</tt>' instruction inserts a scalar
2913element into a vector at a specified index.
2914</p>
2915
2916
2917<h5>Arguments:</h5>
2918
2919<p>
2920The first operand of an '<tt>insertelement</tt>' instruction is a
2921value of <a href="#t_vector">vector</a> type. The second operand is a
2922scalar value whose type must equal the element type of the first
2923operand. The third operand is an index indicating the position at
2924which to insert the value. The index may be a variable.</p>
2925
2926<h5>Semantics:</h5>
2927
2928<p>
2929The result is a vector of the same type as <tt>val</tt>. Its
2930element values are those of <tt>val</tt> except at position
2931<tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
2932exceeds the length of <tt>val</tt>, the results are undefined.
2933</p>
2934
2935<h5>Example:</h5>
2936
2937<pre>
2938 %result = insertelement &lt;4 x i32&gt; %vec, i32 1, i32 0 <i>; yields &lt;4 x i32&gt;</i>
2939</pre>
2940</div>
2941
2942<!-- _______________________________________________________________________ -->
2943<div class="doc_subsubsection">
2944 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
2945</div>
2946
2947<div class="doc_text">
2948
2949<h5>Syntax:</h5>
2950
2951<pre>
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]2952 &lt;result&gt; = shufflevector &lt;n x &lt;ty&gt;&gt; &lt;v1&gt;, &lt;n x &lt;ty&gt;&gt; &lt;v2&gt;, &lt;m x i32&gt; &lt;mask&gt; <i>; yields &lt;m x &lt;ty&gt;&gt;</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2953</pre>
2954
2955<h5>Overview:</h5>
2956
2957<p>
2958The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]2959from two input vectors, returning a vector with the same element type as
2960the input and length that is the same as the shuffle mask.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2961</p>
2962
2963<h5>Arguments:</h5>
2964
2965<p>
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]2966The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
2967with types that match each other. The third argument is a shuffle mask whose
2968element type is always 'i32'. The result of the instruction is a vector whose
2969length is the same as the shuffle mask and whose element type is the same as
2970the element type of the first two operands.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2971</p>
2972
2973<p>
2974The shuffle mask operand is required to be a constant vector with either
2975constant integer or undef values.
2976</p>
2977
2978<h5>Semantics:</h5>
2979
2980<p>
2981The elements of the two input vectors are numbered from left to right across
2982both of the vectors. The shuffle mask operand specifies, for each element of
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]2983the result vector, which element of the two input vectors the result element
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2984gets. The element selector may be undef (meaning "don't care") and the second
2985operand may be undef if performing a shuffle from only one vector.
2986</p>
2987
2988<h5>Example:</h5>
2989
2990<pre>
2991 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
2992 &lt;4 x i32&gt; &lt;i32 0, i32 4, i32 1, i32 5&gt; <i>; yields &lt;4 x i32&gt;</i>
2993 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; undef,
2994 &lt;4 x i32&gt; &lt;i32 0, i32 1, i32 2, i32 3&gt; <i>; yields &lt;4 x i32&gt;</i> - Identity shuffle.
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]2995 %result = shufflevector &lt;8 x i32&gt; %v1, &lt;8 x i32&gt; undef,
2996 &lt;4 x i32&gt; &lt;i32 0, i32 1, i32 2, i32 3&gt; <i>; yields &lt;4 x i32&gt;</i>
2997 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
2998 &lt;8 x i32&gt; &lt;i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7 &gt; <i>; yields &lt;8 x i32&gt;</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2999</pre>
3000</div>
3001
3002
3003<!-- ======================================================================= -->
3004<div class="doc_subsection">
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3005 <a name="aggregateops">Aggregate Operations</a>
3006</div>
3007
3008<div class="doc_text">
3009
3010<p>LLVM supports several instructions for working with aggregate values.
3011</p>
3012
3013</div>
3014
3015<!-- _______________________________________________________________________ -->
3016<div class="doc_subsubsection">
3017 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
3018</div>
3019
3020<div class="doc_text">
3021
3022<h5>Syntax:</h5>
3023
3024<pre>
3025 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
3026</pre>
3027
3028<h5>Overview:</h5>
3029
3030<p>
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3031The '<tt>extractvalue</tt>' instruction extracts the value of a struct field
3032or array element from an aggregate value.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3033</p>
3034
3035
3036<h5>Arguments:</h5>
3037
3038<p>
3039The first operand of an '<tt>extractvalue</tt>' instruction is a
3040value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a>
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3041type. The operands are constant indices to specify which value to extract
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3042in a similar manner as indices in a
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3043'<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
3044</p>
3045
3046<h5>Semantics:</h5>
3047
3048<p>
3049The result is the value at the position in the aggregate specified by
3050the index operands.
3051</p>
3052
3053<h5>Example:</h5>
3054
3055<pre>
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3056 %result = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3057</pre>
3058</div>
3059
3060
3061<!-- _______________________________________________________________________ -->
3062<div class="doc_subsubsection">
3063 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
3064</div>
3065
3066<div class="doc_text">
3067
3068<h5>Syntax:</h5>
3069
3070<pre>
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3071 &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]3072</pre>
3073
3074<h5>Overview:</h5>
3075
3076<p>
3077The '<tt>insertvalue</tt>' instruction inserts a value
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3078into a struct field or array element in an aggregate.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3079</p>
3080
3081
3082<h5>Arguments:</h5>
3083
3084<p>
3085The first operand of an '<tt>insertvalue</tt>' instruction is a
3086value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a> type.
3087The second operand is a first-class value to insert.
Dan Gohman4f29e422008-05-23 21:53:15 +0000[diff] [blame]3088The following operands are constant indices
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3089indicating the position at which to insert the value in a similar manner as
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3090indices in a
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3091'<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
3092The value to insert must have the same type as the value identified
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3093by the indices.
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3094</p>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3095
3096<h5>Semantics:</h5>
3097
3098<p>
3099The result is an aggregate of the same type as <tt>val</tt>. Its
3100value is that of <tt>val</tt> except that the value at the position
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3101specified by the indices is that of <tt>elt</tt>.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3102</p>
3103
3104<h5>Example:</h5>
3105
3106<pre>
Dan Gohmanb1aab4e2008-06-23 15:26:37 +0000[diff] [blame]3107 %result = insertvalue {i32, float} %agg, i32 1, 0 <i>; yields {i32, float}</i>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3108</pre>
3109</div>
3110
3111
3112<!-- ======================================================================= -->
3113<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3114 <a name="memoryops">Memory Access and Addressing Operations</a>
3115</div>
3116
3117<div class="doc_text">
3118
3119<p>A key design point of an SSA-based representation is how it
3120represents memory. In LLVM, no memory locations are in SSA form, which
3121makes things very simple. This section describes how to read, write,
3122allocate, and free memory in LLVM.</p>
3123
3124</div>
3125
3126<!-- _______________________________________________________________________ -->
3127<div class="doc_subsubsection">
3128 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
3129</div>
3130
3131<div class="doc_text">
3132
3133<h5>Syntax:</h5>
3134
3135<pre>
3136 &lt;result&gt; = malloc &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
3137</pre>
3138
3139<h5>Overview:</h5>
3140
3141<p>The '<tt>malloc</tt>' instruction allocates memory from the system
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]3142heap and returns a pointer to it. The object is always allocated in the generic
3143address space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3144
3145<h5>Arguments:</h5>
3146
3147<p>The '<tt>malloc</tt>' instruction allocates
3148<tt>sizeof(&lt;type&gt;)*NumElements</tt>
3149bytes of memory from the operating system and returns a pointer of the
3150appropriate type to the program. If "NumElements" is specified, it is the
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3151number of elements allocated, otherwise "NumElements" is defaulted to be one.
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3152If a constant alignment is specified, the value result of the allocation is guaranteed to
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3153be aligned to at least that boundary. If not specified, or if zero, the target can
3154choose to align the allocation on any convenient boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3155
3156<p>'<tt>type</tt>' must be a sized type.</p>
3157
3158<h5>Semantics:</h5>
3159
3160<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
Nick Lewyckyff384472008-11-24 03:41:24 +0000[diff] [blame]3161a pointer is returned. The result of a zero byte allocation is undefined. The
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]3162result is null if there is insufficient memory available.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3163
3164<h5>Example:</h5>
3165
3166<pre>
3167 %array = malloc [4 x i8 ] <i>; yields {[%4 x i8]*}:array</i>
3168
3169 %size = <a href="#i_add">add</a> i32 2, 2 <i>; yields {i32}:size = i32 4</i>
3170 %array1 = malloc i8, i32 4 <i>; yields {i8*}:array1</i>
3171 %array2 = malloc [12 x i8], i32 %size <i>; yields {[12 x i8]*}:array2</i>
3172 %array3 = malloc i32, i32 4, align 1024 <i>; yields {i32*}:array3</i>
3173 %array4 = malloc i32, align 1024 <i>; yields {i32*}:array4</i>
3174</pre>
3175</div>
3176
3177<!-- _______________________________________________________________________ -->
3178<div class="doc_subsubsection">
3179 <a name="i_free">'<tt>free</tt>' Instruction</a>
3180</div>
3181
3182<div class="doc_text">
3183
3184<h5>Syntax:</h5>
3185
3186<pre>
3187 free &lt;type&gt; &lt;value&gt; <i>; yields {void}</i>
3188</pre>
3189
3190<h5>Overview:</h5>
3191
3192<p>The '<tt>free</tt>' instruction returns memory back to the unused
3193memory heap to be reallocated in the future.</p>
3194
3195<h5>Arguments:</h5>
3196
3197<p>'<tt>value</tt>' shall be a pointer value that points to a value
3198that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
3199instruction.</p>
3200
3201<h5>Semantics:</h5>
3202
3203<p>Access to the memory pointed to by the pointer is no longer defined
Chris Lattner329d6532008-04-19 22:41:32 +0000[diff] [blame]3204after this instruction executes. If the pointer is null, the operation
3205is a noop.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3206
3207<h5>Example:</h5>
3208
3209<pre>
3210 %array = <a href="#i_malloc">malloc</a> [4 x i8] <i>; yields {[4 x i8]*}:array</i>
3211 free [4 x i8]* %array
3212</pre>
3213</div>
3214
3215<!-- _______________________________________________________________________ -->
3216<div class="doc_subsubsection">
3217 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
3218</div>
3219
3220<div class="doc_text">
3221
3222<h5>Syntax:</h5>
3223
3224<pre>
3225 &lt;result&gt; = alloca &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
3226</pre>
3227
3228<h5>Overview:</h5>
3229
3230<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
3231currently executing function, to be automatically released when this function
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]3232returns to its caller. The object is always allocated in the generic address
3233space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3234
3235<h5>Arguments:</h5>
3236
3237<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
3238bytes of memory on the runtime stack, returning a pointer of the
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3239appropriate type to the program. If "NumElements" is specified, it is the
3240number of elements allocated, otherwise "NumElements" is defaulted to be one.
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3241If a constant alignment is specified, the value result of the allocation is guaranteed
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3242to be aligned to at least that boundary. If not specified, or if zero, the target
3243can choose to align the allocation on any convenient boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3244
3245<p>'<tt>type</tt>' may be any sized type.</p>
3246
3247<h5>Semantics:</h5>
3248
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]3249<p>Memory is allocated; a pointer is returned. The operation is undefiend if
3250there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3251memory is automatically released when the function returns. The '<tt>alloca</tt>'
3252instruction is commonly used to represent automatic variables that must
3253have an address available. When the function returns (either with the <tt><a
3254 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]3255instructions), the memory is reclaimed. Allocating zero bytes
3256is legal, but the result is undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3257
3258<h5>Example:</h5>
3259
3260<pre>
3261 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
3262 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
3263 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
3264 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
3265</pre>
3266</div>
3267
3268<!-- _______________________________________________________________________ -->
3269<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
3270Instruction</a> </div>
3271<div class="doc_text">
3272<h5>Syntax:</h5>
3273<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>
3274<h5>Overview:</h5>
3275<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
3276<h5>Arguments:</h5>
3277<p>The argument to the '<tt>load</tt>' instruction specifies the memory
3278address from which to load. The pointer must point to a <a
3279 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
3280marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
3281the number or order of execution of this <tt>load</tt> with other
3282volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
3283instructions. </p>
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3284<p>
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3285The optional constant "align" argument specifies the alignment of the operation
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3286(that is, the alignment of the memory address). A value of 0 or an
3287omitted "align" argument means that the operation has the preferential
3288alignment for the target. It is the responsibility of the code emitter
3289to ensure that the alignment information is correct. Overestimating
3290the alignment results in an undefined behavior. Underestimating the
3291alignment may produce less efficient code. An alignment of 1 is always
3292safe.
3293</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3294<h5>Semantics:</h5>
3295<p>The location of memory pointed to is loaded.</p>
3296<h5>Examples:</h5>
3297<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
3298 <a
3299 href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
3300 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
3301</pre>
3302</div>
3303<!-- _______________________________________________________________________ -->
3304<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
3305Instruction</a> </div>
3306<div class="doc_text">
3307<h5>Syntax:</h5>
3308<pre> store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
3309 volatile store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
3310</pre>
3311<h5>Overview:</h5>
3312<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
3313<h5>Arguments:</h5>
3314<p>There are two arguments to the '<tt>store</tt>' instruction: a value
3315to 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]3316operand must be a pointer to the <a href="#t_firstclass">first class</a> type
3317of the '<tt>&lt;value&gt;</tt>'
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3318operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
3319optimizer is not allowed to modify the number or order of execution of
3320this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
3321 href="#i_store">store</a></tt> instructions.</p>
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3322<p>
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3323The optional constant "align" argument specifies the alignment of the operation
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3324(that is, the alignment of the memory address). A value of 0 or an
3325omitted "align" argument means that the operation has the preferential
3326alignment for the target. It is the responsibility of the code emitter
3327to ensure that the alignment information is correct. Overestimating
3328the alignment results in an undefined behavior. Underestimating the
3329alignment may produce less efficient code. An alignment of 1 is always
3330safe.
3331</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3332<h5>Semantics:</h5>
3333<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>'
3334at the location specified by the '<tt>&lt;pointer&gt;</tt>' operand.</p>
3335<h5>Example:</h5>
3336<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling63ffa142007-10-22 05:10:05 +0000[diff] [blame]3337 store i32 3, i32* %ptr <i>; yields {void}</i>
3338 %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]3339</pre>
3340</div>
3341
3342<!-- _______________________________________________________________________ -->
3343<div class="doc_subsubsection">
3344 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
3345</div>
3346
3347<div class="doc_text">
3348<h5>Syntax:</h5>
3349<pre>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3350 &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]3351</pre>
3352
3353<h5>Overview:</h5>
3354
3355<p>
3356The '<tt>getelementptr</tt>' instruction is used to get the address of a
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3357subelement of an aggregate data structure. It performs address calculation only
3358and does not access memory.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3359
3360<h5>Arguments:</h5>
3361
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3362<p>The first argument is always a pointer, and forms the basis of the
3363calculation. The remaining arguments are indices, that indicate which of the
3364elements of the aggregate object are indexed. The interpretation of each index
3365is dependent on the type being indexed into. The first index always indexes the
3366pointer value given as the first argument, the second index indexes a value of
3367the type pointed to (not necessarily the value directly pointed to, since the
3368first index can be non-zero), etc. The first type indexed into must be a pointer
3369value, subsequent types can be arrays, vectors and structs. Note that subsequent
3370types being indexed into can never be pointers, since that would require loading
3371the pointer before continuing calculation.</p>
3372
3373<p>The type of each index argument depends on the type it is indexing into.
3374When indexing into a (packed) structure, only <tt>i32</tt> integer
3375<b>constants</b> are allowed. When indexing into an array, pointer or vector,
3376only integers of 32 or 64 bits are allowed (also non-constants). 32-bit values
3377will be sign extended to 64-bits if required.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3378
3379<p>For example, let's consider a C code fragment and how it gets
3380compiled to LLVM:</p>
3381
3382<div class="doc_code">
3383<pre>
3384struct RT {
3385 char A;
3386 int B[10][20];
3387 char C;
3388};
3389struct ST {
3390 int X;
3391 double Y;
3392 struct RT Z;
3393};
3394
3395int *foo(struct ST *s) {
3396 return &amp;s[1].Z.B[5][13];
3397}
3398</pre>
3399</div>
3400
3401<p>The LLVM code generated by the GCC frontend is:</p>
3402
3403<div class="doc_code">
3404<pre>
3405%RT = type { i8 , [10 x [20 x i32]], i8 }
3406%ST = type { i32, double, %RT }
3407
3408define i32* %foo(%ST* %s) {
3409entry:
3410 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
3411 ret i32* %reg
3412}
3413</pre>
3414</div>
3415
3416<h5>Semantics:</h5>
3417
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3418<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
3419type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
3420}</tt>' type, a structure. The second index indexes into the third element of
3421the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
3422i8 }</tt>' type, another structure. The third index indexes into the second
3423element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
3424array. The two dimensions of the array are subscripted into, yielding an
3425'<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
3426to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
3427
3428<p>Note that it is perfectly legal to index partially through a
3429structure, returning a pointer to an inner element. Because of this,
3430the LLVM code for the given testcase is equivalent to:</p>
3431
3432<pre>
3433 define i32* %foo(%ST* %s) {
3434 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
3435 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
3436 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
3437 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
3438 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
3439 ret i32* %t5
3440 }
3441</pre>
3442
3443<p>Note that it is undefined to access an array out of bounds: array and
3444pointer indexes must always be within the defined bounds of the array type.
Chris Lattnera7d94ba2008-04-24 05:59:56 +0000[diff] [blame]3445The one exception for this rule is zero length arrays. These arrays are
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3446defined to be accessible as variable length arrays, which requires access
3447beyond the zero'th element.</p>
3448
3449<p>The getelementptr instruction is often confusing. For some more insight
3450into how it works, see <a href="GetElementPtr.html">the getelementptr
3451FAQ</a>.</p>
3452
3453<h5>Example:</h5>
3454
3455<pre>
3456 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3457 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
3458 <i>; yields i8*:vptr</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3459 %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]3460 <i>; yields i8*:eptr</i>
3461 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3462</pre>
3463</div>
3464
3465<!-- ======================================================================= -->
3466<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
3467</div>
3468<div class="doc_text">
3469<p>The instructions in this category are the conversion instructions (casting)
3470which all take a single operand and a type. They perform various bit conversions
3471on the operand.</p>
3472</div>
3473
3474<!-- _______________________________________________________________________ -->
3475<div class="doc_subsubsection">
3476 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
3477</div>
3478<div class="doc_text">
3479
3480<h5>Syntax:</h5>
3481<pre>
3482 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3483</pre>
3484
3485<h5>Overview:</h5>
3486<p>
3487The '<tt>trunc</tt>' instruction truncates its operand to the type <tt>ty2</tt>.
3488</p>
3489
3490<h5>Arguments:</h5>
3491<p>
3492The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
3493be an <a href="#t_integer">integer</a> type, and a type that specifies the size
3494and type of the result, which must be an <a href="#t_integer">integer</a>
3495type. The bit size of <tt>value</tt> must be larger than the bit size of
3496<tt>ty2</tt>. Equal sized types are not allowed.</p>
3497
3498<h5>Semantics:</h5>
3499<p>
3500The '<tt>trunc</tt>' instruction truncates the high order bits in <tt>value</tt>
3501and converts the remaining bits to <tt>ty2</tt>. Since the source size must be
3502larger than the destination size, <tt>trunc</tt> cannot be a <i>no-op cast</i>.
3503It will always truncate bits.</p>
3504
3505<h5>Example:</h5>
3506<pre>
3507 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
3508 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
3509 %Y = trunc i32 122 to i1 <i>; yields i1:false</i>
3510</pre>
3511</div>
3512
3513<!-- _______________________________________________________________________ -->
3514<div class="doc_subsubsection">
3515 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
3516</div>
3517<div class="doc_text">
3518
3519<h5>Syntax:</h5>
3520<pre>
3521 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3522</pre>
3523
3524<h5>Overview:</h5>
3525<p>The '<tt>zext</tt>' instruction zero extends its operand to type
3526<tt>ty2</tt>.</p>
3527
3528
3529<h5>Arguments:</h5>
3530<p>The '<tt>zext</tt>' instruction takes a value to cast, which must be of
3531<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3532also be of <a href="#t_integer">integer</a> type. The bit size of the
3533<tt>value</tt> must be smaller than the bit size of the destination type,
3534<tt>ty2</tt>.</p>
3535
3536<h5>Semantics:</h5>
3537<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
3538bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
3539
3540<p>When zero extending from i1, the result will always be either 0 or 1.</p>
3541
3542<h5>Example:</h5>
3543<pre>
3544 %X = zext i32 257 to i64 <i>; yields i64:257</i>
3545 %Y = zext i1 true to i32 <i>; yields i32:1</i>
3546</pre>
3547</div>
3548
3549<!-- _______________________________________________________________________ -->
3550<div class="doc_subsubsection">
3551 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
3552</div>
3553<div class="doc_text">
3554
3555<h5>Syntax:</h5>
3556<pre>
3557 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3558</pre>
3559
3560<h5>Overview:</h5>
3561<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
3562
3563<h5>Arguments:</h5>
3564<p>
3565The '<tt>sext</tt>' instruction takes a value to cast, which must be of
3566<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3567also be of <a href="#t_integer">integer</a> type. The bit size of the
3568<tt>value</tt> must be smaller than the bit size of the destination type,
3569<tt>ty2</tt>.</p>
3570
3571<h5>Semantics:</h5>
3572<p>
3573The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
3574bit (highest order bit) of the <tt>value</tt> until it reaches the bit size of
3575the type <tt>ty2</tt>.</p>
3576
3577<p>When sign extending from i1, the extension always results in -1 or 0.</p>
3578
3579<h5>Example:</h5>
3580<pre>
3581 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
3582 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
3583</pre>
3584</div>
3585
3586<!-- _______________________________________________________________________ -->
3587<div class="doc_subsubsection">
3588 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
3589</div>
3590
3591<div class="doc_text">
3592
3593<h5>Syntax:</h5>
3594
3595<pre>
3596 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3597</pre>
3598
3599<h5>Overview:</h5>
3600<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
3601<tt>ty2</tt>.</p>
3602
3603
3604<h5>Arguments:</h5>
3605<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
3606 point</a> value to cast and a <a href="#t_floating">floating point</a> type to
3607cast it to. The size of <tt>value</tt> must be larger than the size of
3608<tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
3609<i>no-op cast</i>.</p>
3610
3611<h5>Semantics:</h5>
3612<p> The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
3613<a href="#t_floating">floating point</a> type to a smaller
3614<a href="#t_floating">floating point</a> type. If the value cannot fit within
3615the destination type, <tt>ty2</tt>, then the results are undefined.</p>
3616
3617<h5>Example:</h5>
3618<pre>
3619 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
3620 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
3621</pre>
3622</div>
3623
3624<!-- _______________________________________________________________________ -->
3625<div class="doc_subsubsection">
3626 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
3627</div>
3628<div class="doc_text">
3629
3630<h5>Syntax:</h5>
3631<pre>
3632 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3633</pre>
3634
3635<h5>Overview:</h5>
3636<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
3637floating point value.</p>
3638
3639<h5>Arguments:</h5>
3640<p>The '<tt>fpext</tt>' instruction takes a
3641<a href="#t_floating">floating point</a> <tt>value</tt> to cast,
3642and a <a href="#t_floating">floating point</a> type to cast it to. The source
3643type must be smaller than the destination type.</p>
3644
3645<h5>Semantics:</h5>
3646<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
3647<a href="#t_floating">floating point</a> type to a larger
3648<a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
3649used to make a <i>no-op cast</i> because it always changes bits. Use
3650<tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
3651
3652<h5>Example:</h5>
3653<pre>
3654 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
3655 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
3656</pre>
3657</div>
3658
3659<!-- _______________________________________________________________________ -->
3660<div class="doc_subsubsection">
3661 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
3662</div>
3663<div class="doc_text">
3664
3665<h5>Syntax:</h5>
3666<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3667 &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]3668</pre>
3669
3670<h5>Overview:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3671<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3672unsigned integer equivalent of type <tt>ty2</tt>.
3673</p>
3674
3675<h5>Arguments:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3676<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]3677scalar or vector <a href="#t_floating">floating point</a> value, and a type
3678to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
3679type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
3680vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3681
3682<h5>Semantics:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3683<p> The '<tt>fptoui</tt>' instruction converts its
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3684<a href="#t_floating">floating point</a> operand into the nearest (rounding
3685towards zero) unsigned integer value. If the value cannot fit in <tt>ty2</tt>,
3686the results are undefined.</p>
3687
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3688<h5>Example:</h5>
3689<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3690 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]3691 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3692 %X = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3693</pre>
3694</div>
3695
3696<!-- _______________________________________________________________________ -->
3697<div class="doc_subsubsection">
3698 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
3699</div>
3700<div class="doc_text">
3701
3702<h5>Syntax:</h5>
3703<pre>
3704 &lt;result&gt; = fptosi &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3705</pre>
3706
3707<h5>Overview:</h5>
3708<p>The '<tt>fptosi</tt>' instruction converts
3709<a href="#t_floating">floating point</a> <tt>value</tt> to type <tt>ty2</tt>.
3710</p>
3711
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3712<h5>Arguments:</h5>
3713<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]3714scalar or vector <a href="#t_floating">floating point</a> value, and a type
3715to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
3716type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
3717vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3718
3719<h5>Semantics:</h5>
3720<p>The '<tt>fptosi</tt>' instruction converts its
3721<a href="#t_floating">floating point</a> operand into the nearest (rounding
3722towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
3723the results are undefined.</p>
3724
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3725<h5>Example:</h5>
3726<pre>
3727 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]3728 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3729 %X = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
3730</pre>
3731</div>
3732
3733<!-- _______________________________________________________________________ -->
3734<div class="doc_subsubsection">
3735 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
3736</div>
3737<div class="doc_text">
3738
3739<h5>Syntax:</h5>
3740<pre>
3741 &lt;result&gt; = uitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3742</pre>
3743
3744<h5>Overview:</h5>
3745<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
3746integer and converts that value to the <tt>ty2</tt> type.</p>
3747
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3748<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]3749<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be a
3750scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
3751to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
3752type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
3753floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3754
3755<h5>Semantics:</h5>
3756<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
3757integer quantity and converts it to the corresponding floating point value. If
3758the value cannot fit in the floating point value, the results are undefined.</p>
3759
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3760<h5>Example:</h5>
3761<pre>
3762 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3763 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3764</pre>
3765</div>
3766
3767<!-- _______________________________________________________________________ -->
3768<div class="doc_subsubsection">
3769 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
3770</div>
3771<div class="doc_text">
3772
3773<h5>Syntax:</h5>
3774<pre>
3775 &lt;result&gt; = sitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3776</pre>
3777
3778<h5>Overview:</h5>
3779<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed
3780integer and converts that value to the <tt>ty2</tt> type.</p>
3781
3782<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]3783<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be a
3784scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
3785to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
3786type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
3787floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3788
3789<h5>Semantics:</h5>
3790<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed
3791integer quantity and converts it to the corresponding floating point value. If
3792the value cannot fit in the floating point value, the results are undefined.</p>
3793
3794<h5>Example:</h5>
3795<pre>
3796 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3797 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3798</pre>
3799</div>
3800
3801<!-- _______________________________________________________________________ -->
3802<div class="doc_subsubsection">
3803 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
3804</div>
3805<div class="doc_text">
3806
3807<h5>Syntax:</h5>
3808<pre>
3809 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3810</pre>
3811
3812<h5>Overview:</h5>
3813<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
3814the integer type <tt>ty2</tt>.</p>
3815
3816<h5>Arguments:</h5>
3817<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
3818must 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]3819<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]3820
3821<h5>Semantics:</h5>
3822<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
3823<tt>ty2</tt> by interpreting the pointer value as an integer and either
3824truncating or zero extending that value to the size of the integer type. If
3825<tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
3826<tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
3827are the same size, then nothing is done (<i>no-op cast</i>) other than a type
3828change.</p>
3829
3830<h5>Example:</h5>
3831<pre>
3832 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
3833 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
3834</pre>
3835</div>
3836
3837<!-- _______________________________________________________________________ -->
3838<div class="doc_subsubsection">
3839 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
3840</div>
3841<div class="doc_text">
3842
3843<h5>Syntax:</h5>
3844<pre>
3845 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3846</pre>
3847
3848<h5>Overview:</h5>
3849<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to
3850a pointer type, <tt>ty2</tt>.</p>
3851
3852<h5>Arguments:</h5>
3853<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
3854value to cast, and a type to cast it to, which must be a
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3855<a href="#t_pointer">pointer</a> type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3856
3857<h5>Semantics:</h5>
3858<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
3859<tt>ty2</tt> by applying either a zero extension or a truncation depending on
3860the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
3861size of a pointer then a truncation is done. If <tt>value</tt> is smaller than
3862the size of a pointer then a zero extension is done. If they are the same size,
3863nothing is done (<i>no-op cast</i>).</p>
3864
3865<h5>Example:</h5>
3866<pre>
3867 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
3868 %X = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
3869 %Y = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
3870</pre>
3871</div>
3872
3873<!-- _______________________________________________________________________ -->
3874<div class="doc_subsubsection">
3875 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
3876</div>
3877<div class="doc_text">
3878
3879<h5>Syntax:</h5>
3880<pre>
3881 &lt;result&gt; = bitcast &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3882</pre>
3883
3884<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3885
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3886<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
3887<tt>ty2</tt> without changing any bits.</p>
3888
3889<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3890
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3891<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be
Dan Gohman7305fa02008-09-08 16:45:59 +0000[diff] [blame]3892a non-aggregate first class value, and a type to cast it to, which must also be
3893a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes of
3894<tt>value</tt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3895and the destination type, <tt>ty2</tt>, must be identical. If the source
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3896type is a pointer, the destination type must also be a pointer. This
3897instruction supports bitwise conversion of vectors to integers and to vectors
3898of other types (as long as they have the same size).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3899
3900<h5>Semantics:</h5>
3901<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
3902<tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
3903this conversion. The conversion is done as if the <tt>value</tt> had been
3904stored to memory and read back as type <tt>ty2</tt>. Pointer types may only be
3905converted to other pointer types with this instruction. To convert pointers to
3906other types, use the <a href="#i_inttoptr">inttoptr</a> or
3907<a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
3908
3909<h5>Example:</h5>
3910<pre>
3911 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
3912 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3913 %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]3914</pre>
3915</div>
3916
3917<!-- ======================================================================= -->
3918<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
3919<div class="doc_text">
3920<p>The instructions in this category are the "miscellaneous"
3921instructions, which defy better classification.</p>
3922</div>
3923
3924<!-- _______________________________________________________________________ -->
3925<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
3926</div>
3927<div class="doc_text">
3928<h5>Syntax:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3929<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]3930</pre>
3931<h5>Overview:</h5>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]3932<p>The '<tt>icmp</tt>' instruction returns a boolean value or
3933a vector of boolean values based on comparison
3934of its two integer, integer vector, or pointer operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3935<h5>Arguments:</h5>
3936<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
3937the condition code indicating the kind of comparison to perform. It is not
3938a value, just a keyword. The possible condition code are:
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3939</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3940<ol>
3941 <li><tt>eq</tt>: equal</li>
3942 <li><tt>ne</tt>: not equal </li>
3943 <li><tt>ugt</tt>: unsigned greater than</li>
3944 <li><tt>uge</tt>: unsigned greater or equal</li>
3945 <li><tt>ult</tt>: unsigned less than</li>
3946 <li><tt>ule</tt>: unsigned less or equal</li>
3947 <li><tt>sgt</tt>: signed greater than</li>
3948 <li><tt>sge</tt>: signed greater or equal</li>
3949 <li><tt>slt</tt>: signed less than</li>
3950 <li><tt>sle</tt>: signed less or equal</li>
3951</ol>
3952<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]3953<a href="#t_pointer">pointer</a>
3954or integer <a href="#t_vector">vector</a> typed.
3955They must also be identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3956<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3957<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]3958the condition code given as <tt>cond</tt>. The comparison performed always
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]3959yields 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]3960</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3961<ol>
3962 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
3963 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
3964 </li>
3965 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3966 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3967 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3968 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3969 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3970 <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]3971 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3972 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3973 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3974 <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]3975 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3976 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3977 <li><tt>sge</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3978 <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]3979 <li><tt>slt</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3980 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3981 <li><tt>sle</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3982 <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]3983</ol>
3984<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
3985values are compared as if they were integers.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]3986<p>If the operands are integer vectors, then they are compared
3987element by element. The result is an <tt>i1</tt> vector with
3988the same number of elements as the values being compared.
3989Otherwise, the result is an <tt>i1</tt>.
3990</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3991
3992<h5>Example:</h5>
3993<pre> &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
3994 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
3995 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
3996 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
3997 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
3998 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
3999</pre>
4000</div>
4001
4002<!-- _______________________________________________________________________ -->
4003<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
4004</div>
4005<div class="doc_text">
4006<h5>Syntax:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4007<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]4008</pre>
4009<h5>Overview:</h5>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4010<p>The '<tt>fcmp</tt>' instruction returns a boolean value
4011or vector of boolean values based on comparison
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4012of its operands.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4013<p>
4014If the operands are floating point scalars, then the result
4015type is a boolean (<a href="#t_primitive"><tt>i1</tt></a>).
4016</p>
4017<p>If the operands are floating point vectors, then the result type
4018is a vector of boolean with the same number of elements as the
4019operands being compared.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4020<h5>Arguments:</h5>
4021<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
4022the condition code indicating the kind of comparison to perform. It is not
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4023a value, just a keyword. The possible condition code are:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4024<ol>
4025 <li><tt>false</tt>: no comparison, always returns false</li>
4026 <li><tt>oeq</tt>: ordered and equal</li>
4027 <li><tt>ogt</tt>: ordered and greater than </li>
4028 <li><tt>oge</tt>: ordered and greater than or equal</li>
4029 <li><tt>olt</tt>: ordered and less than </li>
4030 <li><tt>ole</tt>: ordered and less than or equal</li>
4031 <li><tt>one</tt>: ordered and not equal</li>
4032 <li><tt>ord</tt>: ordered (no nans)</li>
4033 <li><tt>ueq</tt>: unordered or equal</li>
4034 <li><tt>ugt</tt>: unordered or greater than </li>
4035 <li><tt>uge</tt>: unordered or greater than or equal</li>
4036 <li><tt>ult</tt>: unordered or less than </li>
4037 <li><tt>ule</tt>: unordered or less than or equal</li>
4038 <li><tt>une</tt>: unordered or not equal</li>
4039 <li><tt>uno</tt>: unordered (either nans)</li>
4040 <li><tt>true</tt>: no comparison, always returns true</li>
4041</ol>
4042<p><i>Ordered</i> means that neither operand is a QNAN while
4043<i>unordered</i> means that either operand may be a QNAN.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4044<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be
4045either a <a href="#t_floating">floating point</a> type
4046or a <a href="#t_vector">vector</a> of floating point type.
4047They must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4048<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4049<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4050according to the condition code given as <tt>cond</tt>.
4051If the operands are vectors, then the vectors are compared
4052element by element.
4053Each comparison performed
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4054always yields an <a href="#t_primitive">i1</a> result, as follows:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4055<ol>
4056 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
4057 <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]4058 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4059 <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]4060 <tt>op1</tt> is greather than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4061 <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]4062 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4063 <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]4064 <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4065 <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]4066 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4067 <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]4068 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4069 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
4070 <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]4071 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4072 <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]4073 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4074 <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]4075 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4076 <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]4077 <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4078 <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]4079 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4080 <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]4081 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4082 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
4083 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
4084</ol>
4085
4086<h5>Example:</h5>
4087<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]4088 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
4089 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
4090 &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]4091</pre>
4092</div>
4093
4094<!-- _______________________________________________________________________ -->
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4095<div class="doc_subsubsection">
4096 <a name="i_vicmp">'<tt>vicmp</tt>' Instruction</a>
4097</div>
4098<div class="doc_text">
4099<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4100<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]4101</pre>
4102<h5>Overview:</h5>
4103<p>The '<tt>vicmp</tt>' instruction returns an integer vector value based on
4104element-wise comparison of its two integer vector operands.</p>
4105<h5>Arguments:</h5>
4106<p>The '<tt>vicmp</tt>' instruction takes three operands. The first operand is
4107the condition code indicating the kind of comparison to perform. It is not
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4108a value, just a keyword. The possible condition code are:</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4109<ol>
4110 <li><tt>eq</tt>: equal</li>
4111 <li><tt>ne</tt>: not equal </li>
4112 <li><tt>ugt</tt>: unsigned greater than</li>
4113 <li><tt>uge</tt>: unsigned greater or equal</li>
4114 <li><tt>ult</tt>: unsigned less than</li>
4115 <li><tt>ule</tt>: unsigned less or equal</li>
4116 <li><tt>sgt</tt>: signed greater than</li>
4117 <li><tt>sge</tt>: signed greater or equal</li>
4118 <li><tt>slt</tt>: signed less than</li>
4119 <li><tt>sle</tt>: signed less or equal</li>
4120</ol>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4121<p>The remaining two arguments must be <a href="#t_vector">vector</a> or
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4122<a href="#t_integer">integer</a> typed. They must also be identical types.</p>
4123<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4124<p>The '<tt>vicmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4125according to the condition code given as <tt>cond</tt>. The comparison yields a
4126<a href="#t_vector">vector</a> of <a href="#t_integer">integer</a> result, of
4127identical type as the values being compared. The most significant bit in each
4128element is 1 if the element-wise comparison evaluates to true, and is 0
4129otherwise. All other bits of the result are undefined. The condition codes
4130are evaluated identically to the <a href="#i_icmp">'<tt>icmp</tt>'
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4131instruction</a>.</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4132
4133<h5>Example:</h5>
4134<pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4135 &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>
4136 &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]4137</pre>
4138</div>
4139
4140<!-- _______________________________________________________________________ -->
4141<div class="doc_subsubsection">
4142 <a name="i_vfcmp">'<tt>vfcmp</tt>' Instruction</a>
4143</div>
4144<div class="doc_text">
4145<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4146<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]4147<h5>Overview:</h5>
4148<p>The '<tt>vfcmp</tt>' instruction returns an integer vector value based on
4149element-wise comparison of its two floating point vector operands. The output
4150elements have the same width as the input elements.</p>
4151<h5>Arguments:</h5>
4152<p>The '<tt>vfcmp</tt>' instruction takes three operands. The first operand is
4153the condition code indicating the kind of comparison to perform. It is not
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4154a value, just a keyword. The possible condition code are:</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4155<ol>
4156 <li><tt>false</tt>: no comparison, always returns false</li>
4157 <li><tt>oeq</tt>: ordered and equal</li>
4158 <li><tt>ogt</tt>: ordered and greater than </li>
4159 <li><tt>oge</tt>: ordered and greater than or equal</li>
4160 <li><tt>olt</tt>: ordered and less than </li>
4161 <li><tt>ole</tt>: ordered and less than or equal</li>
4162 <li><tt>one</tt>: ordered and not equal</li>
4163 <li><tt>ord</tt>: ordered (no nans)</li>
4164 <li><tt>ueq</tt>: unordered or equal</li>
4165 <li><tt>ugt</tt>: unordered or greater than </li>
4166 <li><tt>uge</tt>: unordered or greater than or equal</li>
4167 <li><tt>ult</tt>: unordered or less than </li>
4168 <li><tt>ule</tt>: unordered or less than or equal</li>
4169 <li><tt>une</tt>: unordered or not equal</li>
4170 <li><tt>uno</tt>: unordered (either nans)</li>
4171 <li><tt>true</tt>: no comparison, always returns true</li>
4172</ol>
4173<p>The remaining two arguments must be <a href="#t_vector">vector</a> of
4174<a href="#t_floating">floating point</a> typed. They must also be identical
4175types.</p>
4176<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4177<p>The '<tt>vfcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4178according to the condition code given as <tt>cond</tt>. The comparison yields a
4179<a href="#t_vector">vector</a> of <a href="#t_integer">integer</a> result, with
4180an identical number of elements as the values being compared, and each element
4181having identical with to the width of the floating point elements. The most
4182significant bit in each element is 1 if the element-wise comparison evaluates to
4183true, and is 0 otherwise. All other bits of the result are undefined. The
4184condition codes are evaluated identically to the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4185<a href="#i_fcmp">'<tt>fcmp</tt>' instruction</a>.</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4186
4187<h5>Example:</h5>
4188<pre>
Chris Lattner1e5c5cd02008-10-13 16:55:18 +0000[diff] [blame]4189 <i>; yields: result=&lt;2 x i32&gt; &lt; i32 0, i32 -1 &gt;</i>
4190 &lt;result&gt; = vfcmp oeq &lt;2 x float&gt; &lt; float 4, float 0 &gt;, &lt; float 5, float 0 &gt;
4191
4192 <i>; yields: result=&lt;2 x i64&gt; &lt; i64 -1, i64 0 &gt;</i>
4193 &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]4194</pre>
4195</div>
4196
4197<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4198<div class="doc_subsubsection">
4199 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
4200</div>
4201
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4202<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4203
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4204<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4205
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4206<pre> &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...<br></pre>
4207<h5>Overview:</h5>
4208<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in
4209the SSA graph representing the function.</p>
4210<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4211
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4212<p>The type of the incoming values is specified with the first type
4213field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
4214as arguments, with one pair for each predecessor basic block of the
4215current block. Only values of <a href="#t_firstclass">first class</a>
4216type may be used as the value arguments to the PHI node. Only labels
4217may be used as the label arguments.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4218
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4219<p>There must be no non-phi instructions between the start of a basic
4220block and the PHI instructions: i.e. PHI instructions must be first in
4221a basic block.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4222
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4223<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4224
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4225<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
4226specified by the pair corresponding to the predecessor basic block that executed
4227just prior to the current block.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4228
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4229<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4230<pre>
4231Loop: ; Infinite loop that counts from 0 on up...
4232 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
4233 %nextindvar = add i32 %indvar, 1
4234 br label %Loop
4235</pre>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4236</div>
4237
4238<!-- _______________________________________________________________________ -->
4239<div class="doc_subsubsection">
4240 <a name="i_select">'<tt>select</tt>' Instruction</a>
4241</div>
4242
4243<div class="doc_text">
4244
4245<h5>Syntax:</h5>
4246
4247<pre>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4248 &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>
4249
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4250 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4251</pre>
4252
4253<h5>Overview:</h5>
4254
4255<p>
4256The '<tt>select</tt>' instruction is used to choose one value based on a
4257condition, without branching.
4258</p>
4259
4260
4261<h5>Arguments:</h5>
4262
4263<p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4264The '<tt>select</tt>' instruction requires an 'i1' value or
4265a vector of 'i1' values indicating the
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4266condition, and two values of the same <a href="#t_firstclass">first class</a>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4267type. If the val1/val2 are vectors and
4268the condition is a scalar, then entire vectors are selected, not
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4269individual elements.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4270</p>
4271
4272<h5>Semantics:</h5>
4273
4274<p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4275If 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]4276value argument; otherwise, it returns the second value argument.
4277</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4278<p>
4279If the condition is a vector of i1, then the value arguments must
4280be vectors of the same size, and the selection is done element
4281by element.
4282</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4283
4284<h5>Example:</h5>
4285
4286<pre>
4287 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
4288</pre>
4289</div>
4290
4291
4292<!-- _______________________________________________________________________ -->
4293<div class="doc_subsubsection">
4294 <a name="i_call">'<tt>call</tt>' Instruction</a>
4295</div>
4296
4297<div class="doc_text">
4298
4299<h5>Syntax:</h5>
4300<pre>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]4301 &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]4302</pre>
4303
4304<h5>Overview:</h5>
4305
4306<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
4307
4308<h5>Arguments:</h5>
4309
4310<p>This instruction requires several arguments:</p>
4311
4312<ol>
4313 <li>
4314 <p>The optional "tail" marker indicates whether the callee function accesses
4315 any allocas or varargs in the caller. If the "tail" marker is present, the
4316 function call is eligible for tail call optimization. Note that calls may
4317 be marked "tail" even if they do not occur before a <a
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4318 href="#i_ret"><tt>ret</tt></a> instruction.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4319 </li>
4320 <li>
4321 <p>The optional "cconv" marker indicates which <a href="#callingconv">calling
4322 convention</a> the call should use. If none is specified, the call defaults
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4323 to using C calling conventions.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4324 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4325
4326 <li>
4327 <p>The optional <a href="#paramattrs">Parameter Attributes</a> list for
4328 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>',
4329 and '<tt>inreg</tt>' attributes are valid here.</p>
4330 </li>
4331
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4332 <li>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]4333 <p>'<tt>ty</tt>': the type of the call instruction itself which is also
4334 the type of the return value. Functions that return no value are marked
4335 <tt><a href="#t_void">void</a></tt>.</p>
4336 </li>
4337 <li>
4338 <p>'<tt>fnty</tt>': shall be the signature of the pointer to function
4339 value being invoked. The argument types must match the types implied by
4340 this signature. This type can be omitted if the function is not varargs
4341 and if the function type does not return a pointer to a function.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4342 </li>
4343 <li>
4344 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
4345 be invoked. In most cases, this is a direct function invocation, but
4346 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
4347 to function value.</p>
4348 </li>
4349 <li>
4350 <p>'<tt>function args</tt>': argument list whose types match the
4351 function signature argument types. All arguments must be of
4352 <a href="#t_firstclass">first class</a> type. If the function signature
4353 indicates the function accepts a variable number of arguments, the extra
4354 arguments can be specified.</p>
4355 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4356 <li>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]4357 <p>The optional <a href="#fnattrs">function attributes</a> list. Only
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4358 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
4359 '<tt>readnone</tt>' attributes are valid here.</p>
4360 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4361</ol>
4362
4363<h5>Semantics:</h5>
4364
4365<p>The '<tt>call</tt>' instruction is used to cause control flow to
4366transfer to a specified function, with its incoming arguments bound to
4367the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
4368instruction in the called function, control flow continues with the
4369instruction after the function call, and the return value of the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4370function is bound to the result argument.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4371
4372<h5>Example:</h5>
4373
4374<pre>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]4375 %retval = call i32 @test(i32 %argc)
Chris Lattner5e893ef2008-03-21 17:24:17 +0000[diff] [blame]4376 call i32 (i8 *, ...)* @printf(i8 * %msg, i32 12, i8 42) <i>; yields i32</i>
4377 %X = tail call i32 @foo() <i>; yields i32</i>
4378 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
4379 call void %foo(i8 97 signext)
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]4380
4381 %struct.A = type { i32, i8 }
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4382 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]4383 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
4384 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattnerac454b32008-10-08 06:26:11 +0000[diff] [blame]4385 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijman2c4e05a2008-10-07 10:03:45 +0000[diff] [blame]4386 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4387</pre>
4388
4389</div>
4390
4391<!-- _______________________________________________________________________ -->
4392<div class="doc_subsubsection">
4393 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
4394</div>
4395
4396<div class="doc_text">
4397
4398<h5>Syntax:</h5>
4399
4400<pre>
4401 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
4402</pre>
4403
4404<h5>Overview:</h5>
4405
4406<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
4407the "variable argument" area of a function call. It is used to implement the
4408<tt>va_arg</tt> macro in C.</p>
4409
4410<h5>Arguments:</h5>
4411
4412<p>This instruction takes a <tt>va_list*</tt> value and the type of
4413the argument. It returns a value of the specified argument type and
4414increments the <tt>va_list</tt> to point to the next argument. The
4415actual type of <tt>va_list</tt> is target specific.</p>
4416
4417<h5>Semantics:</h5>
4418
4419<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
4420type from the specified <tt>va_list</tt> and causes the
4421<tt>va_list</tt> to point to the next argument. For more information,
4422see the variable argument handling <a href="#int_varargs">Intrinsic
4423Functions</a>.</p>
4424
4425<p>It is legal for this instruction to be called in a function which does not
4426take a variable number of arguments, for example, the <tt>vfprintf</tt>
4427function.</p>
4428
4429<p><tt>va_arg</tt> is an LLVM instruction instead of an <a
4430href="#intrinsics">intrinsic function</a> because it takes a type as an
4431argument.</p>
4432
4433<h5>Example:</h5>
4434
4435<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
4436
4437</div>
4438
4439<!-- *********************************************************************** -->
4440<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
4441<!-- *********************************************************************** -->
4442
4443<div class="doc_text">
4444
4445<p>LLVM supports the notion of an "intrinsic function". These functions have
4446well known names and semantics and are required to follow certain restrictions.
4447Overall, these intrinsics represent an extension mechanism for the LLVM
4448language that does not require changing all of the transformations in LLVM when
4449adding to the language (or the bitcode reader/writer, the parser, etc...).</p>
4450
4451<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
4452prefix is reserved in LLVM for intrinsic names; thus, function names may not
4453begin with this prefix. Intrinsic functions must always be external functions:
4454you cannot define the body of intrinsic functions. Intrinsic functions may
4455only be used in call or invoke instructions: it is illegal to take the address
4456of an intrinsic function. Additionally, because intrinsic functions are part
4457of the LLVM language, it is required if any are added that they be documented
4458here.</p>
4459
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4460<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents
4461a family of functions that perform the same operation but on different data
4462types. Because LLVM can represent over 8 million different integer types,
4463overloading is used commonly to allow an intrinsic function to operate on any
4464integer type. One or more of the argument types or the result type can be
4465overloaded to accept any integer type. Argument types may also be defined as
4466exactly matching a previous argument's type or the result type. This allows an
4467intrinsic function which accepts multiple arguments, but needs all of them to
4468be of the same type, to only be overloaded with respect to a single argument or
4469the result.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4470
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4471<p>Overloaded intrinsics will have the names of its overloaded argument types
4472encoded into its function name, each preceded by a period. Only those types
4473which are overloaded result in a name suffix. Arguments whose type is matched
4474against another type do not. For example, the <tt>llvm.ctpop</tt> function can
4475take an integer of any width and returns an integer of exactly the same integer
4476width. This leads to a family of functions such as
4477<tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29 %val)</tt>.
4478Only one type, the return type, is overloaded, and only one type suffix is
4479required. Because the argument's type is matched against the return type, it
4480does not require its own name suffix.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4481
4482<p>To learn how to add an intrinsic function, please see the
4483<a href="ExtendingLLVM.html">Extending LLVM Guide</a>.
4484</p>
4485
4486</div>
4487
4488<!-- ======================================================================= -->
4489<div class="doc_subsection">
4490 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
4491</div>
4492
4493<div class="doc_text">
4494
4495<p>Variable argument support is defined in LLVM with the <a
4496 href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
4497intrinsic functions. These functions are related to the similarly
4498named macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
4499
4500<p>All of these functions operate on arguments that use a
4501target-specific value type "<tt>va_list</tt>". The LLVM assembly
4502language reference manual does not define what this type is, so all
4503transformations should be prepared to handle these functions regardless of
4504the type used.</p>
4505
4506<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
4507instruction and the variable argument handling intrinsic functions are
4508used.</p>
4509
4510<div class="doc_code">
4511<pre>
4512define i32 @test(i32 %X, ...) {
4513 ; Initialize variable argument processing
4514 %ap = alloca i8*
4515 %ap2 = bitcast i8** %ap to i8*
4516 call void @llvm.va_start(i8* %ap2)
4517
4518 ; Read a single integer argument
4519 %tmp = va_arg i8** %ap, i32
4520
4521 ; Demonstrate usage of llvm.va_copy and llvm.va_end
4522 %aq = alloca i8*
4523 %aq2 = bitcast i8** %aq to i8*
4524 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
4525 call void @llvm.va_end(i8* %aq2)
4526
4527 ; Stop processing of arguments.
4528 call void @llvm.va_end(i8* %ap2)
4529 ret i32 %tmp
4530}
4531
4532declare void @llvm.va_start(i8*)
4533declare void @llvm.va_copy(i8*, i8*)
4534declare void @llvm.va_end(i8*)
4535</pre>
4536</div>
4537
4538</div>
4539
4540<!-- _______________________________________________________________________ -->
4541<div class="doc_subsubsection">
4542 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
4543</div>
4544
4545
4546<div class="doc_text">
4547<h5>Syntax:</h5>
4548<pre> declare void %llvm.va_start(i8* &lt;arglist&gt;)<br></pre>
4549<h5>Overview:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4550<p>The '<tt>llvm.va_start</tt>' intrinsic initializes
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4551<tt>*&lt;arglist&gt;</tt> for subsequent use by <tt><a
4552href="#i_va_arg">va_arg</a></tt>.</p>
4553
4554<h5>Arguments:</h5>
4555
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4556<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]4557
4558<h5>Semantics:</h5>
4559
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4560<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]4561macro available in C. In a target-dependent way, it initializes the
4562<tt>va_list</tt> element to which the argument points, so that the next call to
4563<tt>va_arg</tt> will produce the first variable argument passed to the function.
4564Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
4565last argument of the function as the compiler can figure that out.</p>
4566
4567</div>
4568
4569<!-- _______________________________________________________________________ -->
4570<div class="doc_subsubsection">
4571 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
4572</div>
4573
4574<div class="doc_text">
4575<h5>Syntax:</h5>
4576<pre> declare void @llvm.va_end(i8* &lt;arglist&gt;)<br></pre>
4577<h5>Overview:</h5>
4578
4579<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
4580which has been initialized previously with <tt><a href="#int_va_start">llvm.va_start</a></tt>
4581or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
4582
4583<h5>Arguments:</h5>
4584
4585<p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
4586
4587<h5>Semantics:</h5>
4588
4589<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
4590macro available in C. In a target-dependent way, it destroys the
4591<tt>va_list</tt> element to which the argument points. Calls to <a
4592href="#int_va_start"><tt>llvm.va_start</tt></a> and <a href="#int_va_copy">
4593<tt>llvm.va_copy</tt></a> must be matched exactly with calls to
4594<tt>llvm.va_end</tt>.</p>
4595
4596</div>
4597
4598<!-- _______________________________________________________________________ -->
4599<div class="doc_subsubsection">
4600 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
4601</div>
4602
4603<div class="doc_text">
4604
4605<h5>Syntax:</h5>
4606
4607<pre>
4608 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
4609</pre>
4610
4611<h5>Overview:</h5>
4612
4613<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
4614from the source argument list to the destination argument list.</p>
4615
4616<h5>Arguments:</h5>
4617
4618<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
4619The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
4620
4621
4622<h5>Semantics:</h5>
4623
4624<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
4625macro available in C. In a target-dependent way, it copies the source
4626<tt>va_list</tt> element into the destination <tt>va_list</tt> element. This
4627intrinsic is necessary because the <tt><a href="#int_va_start">
4628llvm.va_start</a></tt> intrinsic may be arbitrarily complex and require, for
4629example, memory allocation.</p>
4630
4631</div>
4632
4633<!-- ======================================================================= -->
4634<div class="doc_subsection">
4635 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
4636</div>
4637
4638<div class="doc_text">
4639
4640<p>
4641LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]4642Collection</a> (GC) requires the implementation and generation of these
4643intrinsics.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4644These intrinsics allow identification of <a href="#int_gcroot">GC roots on the
4645stack</a>, as well as garbage collector implementations that require <a
4646href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a> barriers.
4647Front-ends for type-safe garbage collected languages should generate these
4648intrinsics to make use of the LLVM garbage collectors. For more details, see <a
4649href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
4650</p>
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]4651
4652<p>The garbage collection intrinsics only operate on objects in the generic
4653 address space (address space zero).</p>
4654
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4655</div>
4656
4657<!-- _______________________________________________________________________ -->
4658<div class="doc_subsubsection">
4659 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
4660</div>
4661
4662<div class="doc_text">
4663
4664<h5>Syntax:</h5>
4665
4666<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4667 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4668</pre>
4669
4670<h5>Overview:</h5>
4671
4672<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
4673the code generator, and allows some metadata to be associated with it.</p>
4674
4675<h5>Arguments:</h5>
4676
4677<p>The first argument specifies the address of a stack object that contains the
4678root pointer. The second pointer (which must be either a constant or a global
4679value address) contains the meta-data to be associated with the root.</p>
4680
4681<h5>Semantics:</h5>
4682
Chris Lattnera7d94ba2008-04-24 05:59:56 +0000[diff] [blame]4683<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]4684location. At compile-time, the code generator generates information to allow
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4685the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
4686intrinsic may only be used in a function which <a href="#gc">specifies a GC
4687algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4688
4689</div>
4690
4691
4692<!-- _______________________________________________________________________ -->
4693<div class="doc_subsubsection">
4694 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
4695</div>
4696
4697<div class="doc_text">
4698
4699<h5>Syntax:</h5>
4700
4701<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4702 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4703</pre>
4704
4705<h5>Overview:</h5>
4706
4707<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
4708locations, allowing garbage collector implementations that require read
4709barriers.</p>
4710
4711<h5>Arguments:</h5>
4712
4713<p>The second argument is the address to read from, which should be an address
4714allocated from the garbage collector. The first object is a pointer to the
4715start of the referenced object, if needed by the language runtime (otherwise
4716null).</p>
4717
4718<h5>Semantics:</h5>
4719
4720<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
4721instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4722garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
4723may only be used in a function which <a href="#gc">specifies a GC
4724algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4725
4726</div>
4727
4728
4729<!-- _______________________________________________________________________ -->
4730<div class="doc_subsubsection">
4731 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
4732</div>
4733
4734<div class="doc_text">
4735
4736<h5>Syntax:</h5>
4737
4738<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4739 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4740</pre>
4741
4742<h5>Overview:</h5>
4743
4744<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
4745locations, allowing garbage collector implementations that require write
4746barriers (such as generational or reference counting collectors).</p>
4747
4748<h5>Arguments:</h5>
4749
4750<p>The first argument is the reference to store, the second is the start of the
4751object to store it to, and the third is the address of the field of Obj to
4752store to. If the runtime does not require a pointer to the object, Obj may be
4753null.</p>
4754
4755<h5>Semantics:</h5>
4756
4757<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
4758instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4759garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
4760may only be used in a function which <a href="#gc">specifies a GC
4761algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4762
4763</div>
4764
4765
4766
4767<!-- ======================================================================= -->
4768<div class="doc_subsection">
4769 <a name="int_codegen">Code Generator Intrinsics</a>
4770</div>
4771
4772<div class="doc_text">
4773<p>
4774These intrinsics are provided by LLVM to expose special features that may only
4775be implemented with code generator support.
4776</p>
4777
4778</div>
4779
4780<!-- _______________________________________________________________________ -->
4781<div class="doc_subsubsection">
4782 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
4783</div>
4784
4785<div class="doc_text">
4786
4787<h5>Syntax:</h5>
4788<pre>
4789 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
4790</pre>
4791
4792<h5>Overview:</h5>
4793
4794<p>
4795The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
4796target-specific value indicating the return address of the current function
4797or one of its callers.
4798</p>
4799
4800<h5>Arguments:</h5>
4801
4802<p>
4803The argument to this intrinsic indicates which function to return the address
4804for. Zero indicates the calling function, one indicates its caller, etc. The
4805argument is <b>required</b> to be a constant integer value.
4806</p>
4807
4808<h5>Semantics:</h5>
4809
4810<p>
4811The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
4812the return address of the specified call frame, or zero if it cannot be
4813identified. The value returned by this intrinsic is likely to be incorrect or 0
4814for arguments other than zero, so it should only be used for debugging purposes.
4815</p>
4816
4817<p>
4818Note that calling this intrinsic does not prevent function inlining or other
4819aggressive transformations, so the value returned may not be that of the obvious
4820source-language caller.
4821</p>
4822</div>
4823
4824
4825<!-- _______________________________________________________________________ -->
4826<div class="doc_subsubsection">
4827 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
4828</div>
4829
4830<div class="doc_text">
4831
4832<h5>Syntax:</h5>
4833<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4834 declare i8 *@llvm.frameaddress(i32 &lt;level&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4835</pre>
4836
4837<h5>Overview:</h5>
4838
4839<p>
4840The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
4841target-specific frame pointer value for the specified stack frame.
4842</p>
4843
4844<h5>Arguments:</h5>
4845
4846<p>
4847The argument to this intrinsic indicates which function to return the frame
4848pointer for. Zero indicates the calling function, one indicates its caller,
4849etc. The argument is <b>required</b> to be a constant integer value.
4850</p>
4851
4852<h5>Semantics:</h5>
4853
4854<p>
4855The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
4856the frame address of the specified call frame, or zero if it cannot be
4857identified. The value returned by this intrinsic is likely to be incorrect or 0
4858for arguments other than zero, so it should only be used for debugging purposes.
4859</p>
4860
4861<p>
4862Note that calling this intrinsic does not prevent function inlining or other
4863aggressive transformations, so the value returned may not be that of the obvious
4864source-language caller.
4865</p>
4866</div>
4867
4868<!-- _______________________________________________________________________ -->
4869<div class="doc_subsubsection">
4870 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
4871</div>
4872
4873<div class="doc_text">
4874
4875<h5>Syntax:</h5>
4876<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4877 declare i8 *@llvm.stacksave()
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4878</pre>
4879
4880<h5>Overview:</h5>
4881
4882<p>
4883The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
4884the function stack, for use with <a href="#int_stackrestore">
4885<tt>llvm.stackrestore</tt></a>. This is useful for implementing language
4886features like scoped automatic variable sized arrays in C99.
4887</p>
4888
4889<h5>Semantics:</h5>
4890
4891<p>
4892This intrinsic returns a opaque pointer value that can be passed to <a
4893href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
4894<tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
4895<tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
4896state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
4897practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
4898that were allocated after the <tt>llvm.stacksave</tt> was executed.
4899</p>
4900
4901</div>
4902
4903<!-- _______________________________________________________________________ -->
4904<div class="doc_subsubsection">
4905 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
4906</div>
4907
4908<div class="doc_text">
4909
4910<h5>Syntax:</h5>
4911<pre>
4912 declare void @llvm.stackrestore(i8 * %ptr)
4913</pre>
4914
4915<h5>Overview:</h5>
4916
4917<p>
4918The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
4919the function stack to the state it was in when the corresponding <a
4920href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
4921useful for implementing language features like scoped automatic variable sized
4922arrays in C99.
4923</p>
4924
4925<h5>Semantics:</h5>
4926
4927<p>
4928See the description for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.
4929</p>
4930
4931</div>
4932
4933
4934<!-- _______________________________________________________________________ -->
4935<div class="doc_subsubsection">
4936 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
4937</div>
4938
4939<div class="doc_text">
4940
4941<h5>Syntax:</h5>
4942<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4943 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]4944</pre>
4945
4946<h5>Overview:</h5>
4947
4948
4949<p>
4950The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
4951a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
4952no
4953effect on the behavior of the program but can change its performance
4954characteristics.
4955</p>
4956
4957<h5>Arguments:</h5>
4958
4959<p>
4960<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
4961determining if the fetch should be for a read (0) or write (1), and
4962<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
4963locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
4964<tt>locality</tt> arguments must be constant integers.
4965</p>
4966
4967<h5>Semantics:</h5>
4968
4969<p>
4970This intrinsic does not modify the behavior of the program. In particular,
4971prefetches cannot trap and do not produce a value. On targets that support this
4972intrinsic, the prefetch can provide hints to the processor cache for better
4973performance.
4974</p>
4975
4976</div>
4977
4978<!-- _______________________________________________________________________ -->
4979<div class="doc_subsubsection">
4980 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
4981</div>
4982
4983<div class="doc_text">
4984
4985<h5>Syntax:</h5>
4986<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4987 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4988</pre>
4989
4990<h5>Overview:</h5>
4991
4992
4993<p>
4994The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]4995(PC) in a region of
4996code to simulators and other tools. The method is target specific, but it is
4997expected that the marker will use exported symbols to transmit the PC of the
4998marker.
4999The marker makes no guarantees that it will remain with any specific instruction
5000after optimizations. It is possible that the presence of a marker will inhibit
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5001optimizations. The intended use is to be inserted after optimizations to allow
5002correlations of simulation runs.
5003</p>
5004
5005<h5>Arguments:</h5>
5006
5007<p>
5008<tt>id</tt> is a numerical id identifying the marker.
5009</p>
5010
5011<h5>Semantics:</h5>
5012
5013<p>
5014This intrinsic does not modify the behavior of the program. Backends that do not
5015support this intrinisic may ignore it.
5016</p>
5017
5018</div>
5019
5020<!-- _______________________________________________________________________ -->
5021<div class="doc_subsubsection">
5022 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
5023</div>
5024
5025<div class="doc_text">
5026
5027<h5>Syntax:</h5>
5028<pre>
5029 declare i64 @llvm.readcyclecounter( )
5030</pre>
5031
5032<h5>Overview:</h5>
5033
5034
5035<p>
5036The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
5037counter register (or similar low latency, high accuracy clocks) on those targets
5038that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
5039As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
5040should only be used for small timings.
5041</p>
5042
5043<h5>Semantics:</h5>
5044
5045<p>
5046When directly supported, reading the cycle counter should not modify any memory.
5047Implementations are allowed to either return a application specific value or a
5048system wide value. On backends without support, this is lowered to a constant 0.
5049</p>
5050
5051</div>
5052
5053<!-- ======================================================================= -->
5054<div class="doc_subsection">
5055 <a name="int_libc">Standard C Library Intrinsics</a>
5056</div>
5057
5058<div class="doc_text">
5059<p>
5060LLVM provides intrinsics for a few important standard C library functions.
5061These intrinsics allow source-language front-ends to pass information about the
5062alignment of the pointer arguments to the code generator, providing opportunity
5063for more efficient code generation.
5064</p>
5065
5066</div>
5067
5068<!-- _______________________________________________________________________ -->
5069<div class="doc_subsubsection">
5070 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
5071</div>
5072
5073<div class="doc_text">
5074
5075<h5>Syntax:</h5>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5076<p>This is an overloaded intrinsic. You can use llvm.memcpy on any integer bit
5077width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5078<pre>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5079 declare void @llvm.memcpy.i8(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5080 i8 &lt;len&gt;, i32 &lt;align&gt;)
5081 declare void @llvm.memcpy.i16(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5082 i16 &lt;len&gt;, i32 &lt;align&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5083 declare void @llvm.memcpy.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5084 i32 &lt;len&gt;, i32 &lt;align&gt;)
5085 declare void @llvm.memcpy.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5086 i64 &lt;len&gt;, i32 &lt;align&gt;)
5087</pre>
5088
5089<h5>Overview:</h5>
5090
5091<p>
5092The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
5093location to the destination location.
5094</p>
5095
5096<p>
5097Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
5098intrinsics do not return a value, and takes an extra alignment argument.
5099</p>
5100
5101<h5>Arguments:</h5>
5102
5103<p>
5104The first argument is a pointer to the destination, the second is a pointer to
5105the source. The third argument is an integer argument
5106specifying the number of bytes to copy, and the fourth argument is the alignment
5107of the source and destination locations.
5108</p>
5109
5110<p>
5111If the call to this intrinisic has an alignment value that is not 0 or 1, then
5112the caller guarantees that both the source and destination pointers are aligned
5113to that boundary.
5114</p>
5115
5116<h5>Semantics:</h5>
5117
5118<p>
5119The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
5120location to the destination location, which are not allowed to overlap. It
5121copies "len" bytes of memory over. If the argument is known to be aligned to
5122some boundary, this can be specified as the fourth argument, otherwise it should
5123be set to 0 or 1.
5124</p>
5125</div>
5126
5127
5128<!-- _______________________________________________________________________ -->
5129<div class="doc_subsubsection">
5130 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
5131</div>
5132
5133<div class="doc_text">
5134
5135<h5>Syntax:</h5>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5136<p>This is an overloaded intrinsic. You can use llvm.memmove on any integer bit
5137width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5138<pre>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5139 declare void @llvm.memmove.i8(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5140 i8 &lt;len&gt;, i32 &lt;align&gt;)
5141 declare void @llvm.memmove.i16(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5142 i16 &lt;len&gt;, i32 &lt;align&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5143 declare void @llvm.memmove.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5144 i32 &lt;len&gt;, i32 &lt;align&gt;)
5145 declare void @llvm.memmove.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5146 i64 &lt;len&gt;, i32 &lt;align&gt;)
5147</pre>
5148
5149<h5>Overview:</h5>
5150
5151<p>
5152The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the source
5153location to the destination location. It is similar to the
Chris Lattnerdba16ea2008-01-06 19:51:52 +0000[diff] [blame]5154'<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to overlap.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5155</p>
5156
5157<p>
5158Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
5159intrinsics do not return a value, and takes an extra alignment argument.
5160</p>
5161
5162<h5>Arguments:</h5>
5163
5164<p>
5165The first argument is a pointer to the destination, the second is a pointer to
5166the source. The third argument is an integer argument
5167specifying the number of bytes to copy, and the fourth argument is the alignment
5168of the source and destination locations.
5169</p>
5170
5171<p>
5172If the call to this intrinisic has an alignment value that is not 0 or 1, then
5173the caller guarantees that the source and destination pointers are aligned to
5174that boundary.
5175</p>
5176
5177<h5>Semantics:</h5>
5178
5179<p>
5180The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the source
5181location to the destination location, which may overlap. It
5182copies "len" bytes of memory over. If the argument is known to be aligned to
5183some boundary, this can be specified as the fourth argument, otherwise it should
5184be set to 0 or 1.
5185</p>
5186</div>
5187
5188
5189<!-- _______________________________________________________________________ -->
5190<div class="doc_subsubsection">
5191 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
5192</div>
5193
5194<div class="doc_text">
5195
5196<h5>Syntax:</h5>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5197<p>This is an overloaded intrinsic. You can use llvm.memset on any integer bit
5198width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5199<pre>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5200 declare void @llvm.memset.i8(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5201 i8 &lt;len&gt;, i32 &lt;align&gt;)
5202 declare void @llvm.memset.i16(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5203 i16 &lt;len&gt;, i32 &lt;align&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5204 declare void @llvm.memset.i32(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5205 i32 &lt;len&gt;, i32 &lt;align&gt;)
5206 declare void @llvm.memset.i64(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5207 i64 &lt;len&gt;, i32 &lt;align&gt;)
5208</pre>
5209
5210<h5>Overview:</h5>
5211
5212<p>
5213The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a particular
5214byte value.
5215</p>
5216
5217<p>
5218Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
5219does not return a value, and takes an extra alignment argument.
5220</p>
5221
5222<h5>Arguments:</h5>
5223
5224<p>
5225The first argument is a pointer to the destination to fill, the second is the
5226byte value to fill it with, the third argument is an integer
5227argument specifying the number of bytes to fill, and the fourth argument is the
5228known alignment of destination location.
5229</p>
5230
5231<p>
5232If the call to this intrinisic has an alignment value that is not 0 or 1, then
5233the caller guarantees that the destination pointer is aligned to that boundary.
5234</p>
5235
5236<h5>Semantics:</h5>
5237
5238<p>
5239The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting at
5240the
5241destination location. If the argument is known to be aligned to some boundary,
5242this can be specified as the fourth argument, otherwise it should be set to 0 or
52431.
5244</p>
5245</div>
5246
5247
5248<!-- _______________________________________________________________________ -->
5249<div class="doc_subsubsection">
5250 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
5251</div>
5252
5253<div class="doc_text">
5254
5255<h5>Syntax:</h5>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5256<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]5257floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5258types however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5259<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5260 declare float @llvm.sqrt.f32(float %Val)
5261 declare double @llvm.sqrt.f64(double %Val)
5262 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
5263 declare fp128 @llvm.sqrt.f128(fp128 %Val)
5264 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5265</pre>
5266
5267<h5>Overview:</h5>
5268
5269<p>
5270The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5271returning the same value as the libm '<tt>sqrt</tt>' functions would. Unlike
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5272<tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
Chris Lattnerf914a002008-01-29 07:00:44 +0000[diff] [blame]5273negative numbers other than -0.0 (which allows for better optimization, because
5274there is no need to worry about errno being set). <tt>llvm.sqrt(-0.0)</tt> is
5275defined to return -0.0 like IEEE sqrt.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5276</p>
5277
5278<h5>Arguments:</h5>
5279
5280<p>
5281The argument and return value are floating point numbers of the same type.
5282</p>
5283
5284<h5>Semantics:</h5>
5285
5286<p>
5287This function returns the sqrt of the specified operand if it is a nonnegative
5288floating point number.
5289</p>
5290</div>
5291
5292<!-- _______________________________________________________________________ -->
5293<div class="doc_subsubsection">
5294 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
5295</div>
5296
5297<div class="doc_text">
5298
5299<h5>Syntax:</h5>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5300<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]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 Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5303<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5304 declare float @llvm.powi.f32(float %Val, i32 %power)
5305 declare double @llvm.powi.f64(double %Val, i32 %power)
5306 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
5307 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
5308 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5309</pre>
5310
5311<h5>Overview:</h5>
5312
5313<p>
5314The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
5315specified (positive or negative) power. The order of evaluation of
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5316multiplications is not defined. When a vector of floating point type is
5317used, the second argument remains a scalar integer value.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5318</p>
5319
5320<h5>Arguments:</h5>
5321
5322<p>
5323The second argument is an integer power, and the first is a value to raise to
5324that power.
5325</p>
5326
5327<h5>Semantics:</h5>
5328
5329<p>
5330This function returns the first value raised to the second power with an
5331unspecified sequence of rounding operations.</p>
5332</div>
5333
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5334<!-- _______________________________________________________________________ -->
5335<div class="doc_subsubsection">
5336 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
5337</div>
5338
5339<div class="doc_text">
5340
5341<h5>Syntax:</h5>
5342<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
5343floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5344types however.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5345<pre>
5346 declare float @llvm.sin.f32(float %Val)
5347 declare double @llvm.sin.f64(double %Val)
5348 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
5349 declare fp128 @llvm.sin.f128(fp128 %Val)
5350 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
5351</pre>
5352
5353<h5>Overview:</h5>
5354
5355<p>
5356The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.
5357</p>
5358
5359<h5>Arguments:</h5>
5360
5361<p>
5362The argument and return value are floating point numbers of the same type.
5363</p>
5364
5365<h5>Semantics:</h5>
5366
5367<p>
5368This function returns the sine of the specified operand, returning the
5369same values as the libm <tt>sin</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5370conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5371</div>
5372
5373<!-- _______________________________________________________________________ -->
5374<div class="doc_subsubsection">
5375 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
5376</div>
5377
5378<div class="doc_text">
5379
5380<h5>Syntax:</h5>
5381<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
5382floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5383types however.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5384<pre>
5385 declare float @llvm.cos.f32(float %Val)
5386 declare double @llvm.cos.f64(double %Val)
5387 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
5388 declare fp128 @llvm.cos.f128(fp128 %Val)
5389 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
5390</pre>
5391
5392<h5>Overview:</h5>
5393
5394<p>
5395The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.
5396</p>
5397
5398<h5>Arguments:</h5>
5399
5400<p>
5401The argument and return value are floating point numbers of the same type.
5402</p>
5403
5404<h5>Semantics:</h5>
5405
5406<p>
5407This function returns the cosine of the specified operand, returning the
5408same values as the libm <tt>cos</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
5412<!-- _______________________________________________________________________ -->
5413<div class="doc_subsubsection">
5414 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
5415</div>
5416
5417<div class="doc_text">
5418
5419<h5>Syntax:</h5>
5420<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
5421floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5422types however.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5423<pre>
5424 declare float @llvm.pow.f32(float %Val, float %Power)
5425 declare double @llvm.pow.f64(double %Val, double %Power)
5426 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
5427 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
5428 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
5429</pre>
5430
5431<h5>Overview:</h5>
5432
5433<p>
5434The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
5435specified (positive or negative) power.
5436</p>
5437
5438<h5>Arguments:</h5>
5439
5440<p>
5441The second argument is a floating point power, and the first is a value to
5442raise to that power.
5443</p>
5444
5445<h5>Semantics:</h5>
5446
5447<p>
5448This function returns the first value raised to the second power,
5449returning the
5450same values as the libm <tt>pow</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5451conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5452</div>
5453
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5454
5455<!-- ======================================================================= -->
5456<div class="doc_subsection">
5457 <a name="int_manip">Bit Manipulation Intrinsics</a>
5458</div>
5459
5460<div class="doc_text">
5461<p>
5462LLVM provides intrinsics for a few important bit manipulation operations.
5463These allow efficient code generation for some algorithms.
5464</p>
5465
5466</div>
5467
5468<!-- _______________________________________________________________________ -->
5469<div class="doc_subsubsection">
5470 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
5471</div>
5472
5473<div class="doc_text">
5474
5475<h5>Syntax:</h5>
5476<p>This is an overloaded intrinsic function. You can use bswap on any integer
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5477type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5478<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5479 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
5480 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
5481 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5482</pre>
5483
5484<h5>Overview:</h5>
5485
5486<p>
5487The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
5488values with an even number of bytes (positive multiple of 16 bits). These are
5489useful for performing operations on data that is not in the target's native
5490byte order.
5491</p>
5492
5493<h5>Semantics:</h5>
5494
5495<p>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5496The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5497and low byte of the input i16 swapped. Similarly, the <tt>llvm.bswap.i32</tt>
5498intrinsic returns an i32 value that has the four bytes of the input i32
5499swapped, 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]5500i32 will have its bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i48</tt>,
5501<tt>llvm.bswap.i64</tt> and other intrinsics extend this concept to
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5502additional even-byte lengths (6 bytes, 8 bytes and more, respectively).
5503</p>
5504
5505</div>
5506
5507<!-- _______________________________________________________________________ -->
5508<div class="doc_subsubsection">
5509 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
5510</div>
5511
5512<div class="doc_text">
5513
5514<h5>Syntax:</h5>
5515<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]5516width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5517<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5518 declare i8 @llvm.ctpop.i8 (i8 &lt;src&gt;)
5519 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5520 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5521 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
5522 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5523</pre>
5524
5525<h5>Overview:</h5>
5526
5527<p>
5528The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
5529value.
5530</p>
5531
5532<h5>Arguments:</h5>
5533
5534<p>
5535The only argument is the value to be counted. The argument may be of any
5536integer type. The return type must match the argument type.
5537</p>
5538
5539<h5>Semantics:</h5>
5540
5541<p>
5542The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
5543</p>
5544</div>
5545
5546<!-- _______________________________________________________________________ -->
5547<div class="doc_subsubsection">
5548 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
5549</div>
5550
5551<div class="doc_text">
5552
5553<h5>Syntax:</h5>
5554<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]5555integer bit width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5556<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5557 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
5558 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5559 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5560 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
5561 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5562</pre>
5563
5564<h5>Overview:</h5>
5565
5566<p>
5567The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
5568leading zeros in a variable.
5569</p>
5570
5571<h5>Arguments:</h5>
5572
5573<p>
5574The only argument is the value to be counted. The argument may be of any
5575integer type. The return type must match the argument type.
5576</p>
5577
5578<h5>Semantics:</h5>
5579
5580<p>
5581The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
5582in a variable. If the src == 0 then the result is the size in bits of the type
5583of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.
5584</p>
5585</div>
5586
5587
5588
5589<!-- _______________________________________________________________________ -->
5590<div class="doc_subsubsection">
5591 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
5592</div>
5593
5594<div class="doc_text">
5595
5596<h5>Syntax:</h5>
5597<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5598integer bit width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5599<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5600 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
5601 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5602 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5603 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
5604 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5605</pre>
5606
5607<h5>Overview:</h5>
5608
5609<p>
5610The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
5611trailing zeros.
5612</p>
5613
5614<h5>Arguments:</h5>
5615
5616<p>
5617The only argument is the value to be counted. The argument may be of any
5618integer type. The return type must match the argument type.
5619</p>
5620
5621<h5>Semantics:</h5>
5622
5623<p>
5624The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
5625in a variable. If the src == 0 then the result is the size in bits of the type
5626of src. For example, <tt>llvm.cttz(2) = 1</tt>.
5627</p>
5628</div>
5629
5630<!-- _______________________________________________________________________ -->
5631<div class="doc_subsubsection">
5632 <a name="int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic</a>
5633</div>
5634
5635<div class="doc_text">
5636
5637<h5>Syntax:</h5>
5638<p>This is an overloaded intrinsic. You can use <tt>llvm.part.select</tt>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5639on any integer bit width.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5640<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5641 declare i17 @llvm.part.select.i17 (i17 %val, i32 %loBit, i32 %hiBit)
5642 declare i29 @llvm.part.select.i29 (i29 %val, i32 %loBit, i32 %hiBit)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5643</pre>
5644
5645<h5>Overview:</h5>
5646<p>The '<tt>llvm.part.select</tt>' family of intrinsic functions selects a
5647range of bits from an integer value and returns them in the same bit width as
5648the original value.</p>
5649
5650<h5>Arguments:</h5>
5651<p>The first argument, <tt>%val</tt> and the result may be integer types of
5652any bit width but they must have the same bit width. The second and third
5653arguments must be <tt>i32</tt> type since they specify only a bit index.</p>
5654
5655<h5>Semantics:</h5>
5656<p>The operation of the '<tt>llvm.part.select</tt>' intrinsic has two modes
5657of operation: forwards and reverse. If <tt>%loBit</tt> is greater than
5658<tt>%hiBits</tt> then the intrinsic operates in reverse mode. Otherwise it
5659operates in forward mode.</p>
5660<p>In forward mode, this intrinsic is the equivalent of shifting <tt>%val</tt>
5661right by <tt>%loBit</tt> bits and then ANDing it with a mask with
5662only the <tt>%hiBit - %loBit</tt> bits set, as follows:</p>
5663<ol>
5664 <li>The <tt>%val</tt> is shifted right (LSHR) by the number of bits specified
5665 by <tt>%loBits</tt>. This normalizes the value to the low order bits.</li>
5666 <li>The <tt>%loBits</tt> value is subtracted from the <tt>%hiBits</tt> value
5667 to determine the number of bits to retain.</li>
5668 <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]5669 <li>The mask is ANDed with <tt>%val</tt> to produce the result.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5670</ol>
5671<p>In reverse mode, a similar computation is made except that the bits are
5672returned in the reverse order. So, for example, if <tt>X</tt> has the value
5673<tt>i16 0x0ACF (101011001111)</tt> and we apply
5674<tt>part.select(i16 X, 8, 3)</tt> to it, we get back the value
5675<tt>i16 0x0026 (000000100110)</tt>.</p>
5676</div>
5677
5678<div class="doc_subsubsection">
5679 <a name="int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic</a>
5680</div>
5681
5682<div class="doc_text">
5683
5684<h5>Syntax:</h5>
5685<p>This is an overloaded intrinsic. You can use <tt>llvm.part.set</tt>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5686on any integer bit width.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5687<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5688 declare i17 @llvm.part.set.i17.i9 (i17 %val, i9 %repl, i32 %lo, i32 %hi)
5689 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]5690</pre>
5691
5692<h5>Overview:</h5>
5693<p>The '<tt>llvm.part.set</tt>' family of intrinsic functions replaces a range
5694of bits in an integer value with another integer value. It returns the integer
5695with the replaced bits.</p>
5696
5697<h5>Arguments:</h5>
5698<p>The first argument, <tt>%val</tt> and the result may be integer types of
5699any bit width but they must have the same bit width. <tt>%val</tt> is the value
5700whose bits will be replaced. The second argument, <tt>%repl</tt> may be an
5701integer of any bit width. The third and fourth arguments must be <tt>i32</tt>
5702type since they specify only a bit index.</p>
5703
5704<h5>Semantics:</h5>
5705<p>The operation of the '<tt>llvm.part.set</tt>' intrinsic has two modes
5706of operation: forwards and reverse. If <tt>%lo</tt> is greater than
5707<tt>%hi</tt> then the intrinsic operates in reverse mode. Otherwise it
5708operates in forward mode.</p>
5709<p>For both modes, the <tt>%repl</tt> value is prepared for use by either
5710truncating it down to the size of the replacement area or zero extending it
5711up to that size.</p>
5712<p>In forward mode, the bits between <tt>%lo</tt> and <tt>%hi</tt> (inclusive)
5713are replaced with corresponding bits from <tt>%repl</tt>. That is the 0th bit
5714in <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]5715to the <tt>%hi</tt>th bit.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5716<p>In reverse mode, a similar computation is made except that the bits are
5717reversed. That is, the <tt>0</tt>th bit in <tt>%repl</tt> replaces the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5718<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]5719<h5>Examples:</h5>
5720<pre>
5721 llvm.part.set(0xFFFF, 0, 4, 7) -&gt; 0xFF0F
5722 llvm.part.set(0xFFFF, 0, 7, 4) -&gt; 0xFF0F
5723 llvm.part.set(0xFFFF, 1, 7, 4) -&gt; 0xFF8F
5724 llvm.part.set(0xFFFF, F, 8, 3) -&gt; 0xFFE7
5725 llvm.part.set(0xFFFF, 0, 3, 8) -&gt; 0xFE07
5726</pre>
5727</div>
5728
5729<!-- ======================================================================= -->
5730<div class="doc_subsection">
5731 <a name="int_debugger">Debugger Intrinsics</a>
5732</div>
5733
5734<div class="doc_text">
5735<p>
5736The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
5737are described in the <a
5738href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
5739Debugging</a> document.
5740</p>
5741</div>
5742
5743
5744<!-- ======================================================================= -->
5745<div class="doc_subsection">
5746 <a name="int_eh">Exception Handling Intrinsics</a>
5747</div>
5748
5749<div class="doc_text">
5750<p> The LLVM exception handling intrinsics (which all start with
5751<tt>llvm.eh.</tt> prefix), are described in the <a
5752href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
5753Handling</a> document. </p>
5754</div>
5755
5756<!-- ======================================================================= -->
5757<div class="doc_subsection">
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5758 <a name="int_trampoline">Trampoline Intrinsic</a>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5759</div>
5760
5761<div class="doc_text">
5762<p>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5763 This intrinsic makes it possible to excise one parameter, marked with
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5764 the <tt>nest</tt> attribute, from a function. The result is a callable
5765 function pointer lacking the nest parameter - the caller does not need
5766 to provide a value for it. Instead, the value to use is stored in
5767 advance in a "trampoline", a block of memory usually allocated
5768 on the stack, which also contains code to splice the nest value into the
5769 argument list. This is used to implement the GCC nested function address
5770 extension.
5771</p>
5772<p>
5773 For example, if the function is
5774 <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]5775 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]5776<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5777 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
5778 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
5779 %p = call i8* @llvm.init.trampoline( i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8*), i8* %nval )
5780 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5781</pre>
Bill Wendlinge262b4c2007-09-22 09:23:55 +0000[diff] [blame]5782 <p>The call <tt>%val = call i32 %fp( i32 %x, i32 %y )</tt> is then equivalent
5783 to <tt>%val = call i32 %f( i8* %nval, i32 %x, i32 %y )</tt>.</p>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5784</div>
5785
5786<!-- _______________________________________________________________________ -->
5787<div class="doc_subsubsection">
5788 <a name="int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a>
5789</div>
5790<div class="doc_text">
5791<h5>Syntax:</h5>
5792<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5793declare 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]5794</pre>
5795<h5>Overview:</h5>
5796<p>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5797 This fills the memory pointed to by <tt>tramp</tt> with code
5798 and returns a function pointer suitable for executing it.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5799</p>
5800<h5>Arguments:</h5>
5801<p>
5802 The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
5803 pointers. The <tt>tramp</tt> argument must point to a sufficiently large
5804 and sufficiently aligned block of memory; this memory is written to by the
Duncan Sands35012212007-08-22 23:39:54 +0000[diff] [blame]5805 intrinsic. Note that the size and the alignment are target-specific - LLVM
5806 currently provides no portable way of determining them, so a front-end that
5807 generates this intrinsic needs to have some target-specific knowledge.
5808 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]5809</p>
5810<h5>Semantics:</h5>
5811<p>
5812 The block of memory pointed to by <tt>tramp</tt> is filled with target
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5813 dependent code, turning it into a function. A pointer to this function is
5814 returned, but needs to be bitcast to an
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5815 <a href="#int_trampoline">appropriate function pointer type</a>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5816 before being called. The new function's signature is the same as that of
5817 <tt>func</tt> with any arguments marked with the <tt>nest</tt> attribute
5818 removed. At most one such <tt>nest</tt> argument is allowed, and it must be
5819 of pointer type. Calling the new function is equivalent to calling
5820 <tt>func</tt> with the same argument list, but with <tt>nval</tt> used for the
5821 missing <tt>nest</tt> argument. If, after calling
5822 <tt>llvm.init.trampoline</tt>, the memory pointed to by <tt>tramp</tt> is
5823 modified, then the effect of any later call to the returned function pointer is
5824 undefined.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5825</p>
5826</div>
5827
5828<!-- ======================================================================= -->
5829<div class="doc_subsection">
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]5830 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
5831</div>
5832
5833<div class="doc_text">
5834<p>
5835 These intrinsic functions expand the "universal IR" of LLVM to represent
5836 hardware constructs for atomic operations and memory synchronization. This
5837 provides an interface to the hardware, not an interface to the programmer. It
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]5838 is aimed at a low enough level to allow any programming models or APIs
5839 (Application Programming Interfaces) which
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]5840 need atomic behaviors to map cleanly onto it. It is also modeled primarily on
5841 hardware behavior. Just as hardware provides a "universal IR" for source
5842 languages, it also provides a starting point for developing a "universal"
5843 atomic operation and synchronization IR.
5844</p>
5845<p>
5846 These do <em>not</em> form an API such as high-level threading libraries,
5847 software transaction memory systems, atomic primitives, and intrinsic
5848 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
5849 application libraries. The hardware interface provided by LLVM should allow
5850 a clean implementation of all of these APIs and parallel programming models.
5851 No one model or paradigm should be selected above others unless the hardware
5852 itself ubiquitously does so.
5853
5854</p>
5855</div>
5856
5857<!-- _______________________________________________________________________ -->
5858<div class="doc_subsubsection">
5859 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
5860</div>
5861<div class="doc_text">
5862<h5>Syntax:</h5>
5863<pre>
5864declare void @llvm.memory.barrier( i1 &lt;ll&gt;, i1 &lt;ls&gt;, i1 &lt;sl&gt;, i1 &lt;ss&gt;,
5865i1 &lt;device&gt; )
5866
5867</pre>
5868<h5>Overview:</h5>
5869<p>
5870 The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
5871 specific pairs of memory access types.
5872</p>
5873<h5>Arguments:</h5>
5874<p>
5875 The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
5876 The first four arguments enables a specific barrier as listed below. The fith
5877 argument specifies that the barrier applies to io or device or uncached memory.
5878
5879</p>
5880 <ul>
5881 <li><tt>ll</tt>: load-load barrier</li>
5882 <li><tt>ls</tt>: load-store barrier</li>
5883 <li><tt>sl</tt>: store-load barrier</li>
5884 <li><tt>ss</tt>: store-store barrier</li>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5885 <li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]5886 </ul>
5887<h5>Semantics:</h5>
5888<p>
5889 This intrinsic causes the system to enforce some ordering constraints upon
5890 the loads and stores of the program. This barrier does not indicate
5891 <em>when</em> any events will occur, it only enforces an <em>order</em> in
5892 which they occur. For any of the specified pairs of load and store operations
5893 (f.ex. load-load, or store-load), all of the first operations preceding the
5894 barrier will complete before any of the second operations succeeding the
5895 barrier begin. Specifically the semantics for each pairing is as follows:
5896</p>
5897 <ul>
5898 <li><tt>ll</tt>: All loads before the barrier must complete before any load
5899 after the barrier begins.</li>
5900
5901 <li><tt>ls</tt>: All loads before the barrier must complete before any
5902 store after the barrier begins.</li>
5903 <li><tt>ss</tt>: All stores before the barrier must complete before any
5904 store after the barrier begins.</li>
5905 <li><tt>sl</tt>: All stores before the barrier must complete before any
5906 load after the barrier begins.</li>
5907 </ul>
5908<p>
5909 These semantics are applied with a logical "and" behavior when more than one
5910 is enabled in a single memory barrier intrinsic.
5911</p>
5912<p>
5913 Backends may implement stronger barriers than those requested when they do not
5914 support as fine grained a barrier as requested. Some architectures do not
5915 need all types of barriers and on such architectures, these become noops.
5916</p>
5917<h5>Example:</h5>
5918<pre>
5919%ptr = malloc i32
5920 store i32 4, %ptr
5921
5922%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
5923 call void @llvm.memory.barrier( i1 false, i1 true, i1 false, i1 false )
5924 <i>; guarantee the above finishes</i>
5925 store i32 8, %ptr <i>; before this begins</i>
5926</pre>
5927</div>
5928
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5929<!-- _______________________________________________________________________ -->
5930<div class="doc_subsubsection">
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]5931 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5932</div>
5933<div class="doc_text">
5934<h5>Syntax:</h5>
5935<p>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]5936 This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
5937 any integer bit width and for different address spaces. Not all targets
5938 support all bit widths however.</p>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5939
5940<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]5941declare i8 @llvm.atomic.cmp.swap.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt; )
5942declare i16 @llvm.atomic.cmp.swap.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt; )
5943declare i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt; )
5944declare 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]5945
5946</pre>
5947<h5>Overview:</h5>
5948<p>
5949 This loads a value in memory and compares it to a given value. If they are
5950 equal, it stores a new value into the memory.
5951</p>
5952<h5>Arguments:</h5>
5953<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]5954 The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result as
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5955 well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
5956 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
5957 this integer type. While any bit width integer may be used, targets may only
5958 lower representations they support in hardware.
5959
5960</p>
5961<h5>Semantics:</h5>
5962<p>
5963 This entire intrinsic must be executed atomically. It first loads the value
5964 in memory pointed to by <tt>ptr</tt> and compares it with the value
5965 <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the memory. The
5966 loaded value is yielded in all cases. This provides the equivalent of an
5967 atomic compare-and-swap operation within the SSA framework.
5968</p>
5969<h5>Examples:</h5>
5970
5971<pre>
5972%ptr = malloc i32
5973 store i32 4, %ptr
5974
5975%val1 = add i32 4, 4
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]5976%result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 4, %val1 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5977 <i>; yields {i32}:result1 = 4</i>
5978%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
5979%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
5980
5981%val2 = add i32 1, 1
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]5982%result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 5, %val2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5983 <i>; yields {i32}:result2 = 8</i>
5984%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
5985
5986%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
5987</pre>
5988</div>
5989
5990<!-- _______________________________________________________________________ -->
5991<div class="doc_subsubsection">
5992 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
5993</div>
5994<div class="doc_text">
5995<h5>Syntax:</h5>
5996
5997<p>
5998 This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
5999 integer bit width. Not all targets support all bit widths however.</p>
6000<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6001declare i8 @llvm.atomic.swap.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;val&gt; )
6002declare i16 @llvm.atomic.swap.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;val&gt; )
6003declare i32 @llvm.atomic.swap.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;val&gt; )
6004declare i64 @llvm.atomic.swap.i64.p0i64( i64* &lt;ptr&gt;, i64 &lt;val&gt; )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6005
6006</pre>
6007<h5>Overview:</h5>
6008<p>
6009 This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
6010 the value from memory. It then stores the value in <tt>val</tt> in the memory
6011 at <tt>ptr</tt>.
6012</p>
6013<h5>Arguments:</h5>
6014
6015<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6016 The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both the
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6017 <tt>val</tt> argument and the result must be integers of the same bit width.
6018 The first argument, <tt>ptr</tt>, must be a pointer to a value of this
6019 integer type. The targets may only lower integer representations they
6020 support.
6021</p>
6022<h5>Semantics:</h5>
6023<p>
6024 This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
6025 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
6026 equivalent of an atomic swap operation within the SSA framework.
6027
6028</p>
6029<h5>Examples:</h5>
6030<pre>
6031%ptr = malloc i32
6032 store i32 4, %ptr
6033
6034%val1 = add i32 4, 4
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6035%result1 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val1 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6036 <i>; yields {i32}:result1 = 4</i>
6037%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
6038%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
6039
6040%val2 = add i32 1, 1
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6041%result2 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6042 <i>; yields {i32}:result2 = 8</i>
6043
6044%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
6045%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
6046</pre>
6047</div>
6048
6049<!-- _______________________________________________________________________ -->
6050<div class="doc_subsubsection">
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6051 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6052
6053</div>
6054<div class="doc_text">
6055<h5>Syntax:</h5>
6056<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6057 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]6058 integer bit width. Not all targets support all bit widths however.</p>
6059<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6060declare i8 @llvm.atomic.load.add.i8..p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6061declare i16 @llvm.atomic.load.add.i16..p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6062declare i32 @llvm.atomic.load.add.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6063declare 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]6064
6065</pre>
6066<h5>Overview:</h5>
6067<p>
6068 This intrinsic adds <tt>delta</tt> to the value stored in memory at
6069 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
6070</p>
6071<h5>Arguments:</h5>
6072<p>
6073
6074 The intrinsic takes two arguments, the first a pointer to an integer value
6075 and the second an integer value. The result is also an integer value. These
6076 integer types can have any bit width, but they must all have the same bit
6077 width. The targets may only lower integer representations they support.
6078</p>
6079<h5>Semantics:</h5>
6080<p>
6081 This intrinsic does a series of operations atomically. It first loads the
6082 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
6083 to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
6084</p>
6085
6086<h5>Examples:</h5>
6087<pre>
6088%ptr = malloc i32
6089 store i32 4, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6090%result1 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 4 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6091 <i>; yields {i32}:result1 = 4</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6092%result2 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6093 <i>; yields {i32}:result2 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6094%result3 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 5 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6095 <i>; yields {i32}:result3 = 10</i>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6096%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6097</pre>
6098</div>
6099
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6100<!-- _______________________________________________________________________ -->
6101<div class="doc_subsubsection">
6102 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
6103
6104</div>
6105<div class="doc_text">
6106<h5>Syntax:</h5>
6107<p>
6108 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]6109 any integer bit width and for different address spaces. Not all targets
6110 support all bit widths however.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6111<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6112declare i8 @llvm.atomic.load.sub.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6113declare i16 @llvm.atomic.load.sub.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6114declare i32 @llvm.atomic.load.sub.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6115declare 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]6116
6117</pre>
6118<h5>Overview:</h5>
6119<p>
6120 This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
6121 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
6122</p>
6123<h5>Arguments:</h5>
6124<p>
6125
6126 The intrinsic takes two arguments, the first a pointer to an integer value
6127 and the second an integer value. The result is also an integer value. These
6128 integer types can have any bit width, but they must all have the same bit
6129 width. The targets may only lower integer representations they support.
6130</p>
6131<h5>Semantics:</h5>
6132<p>
6133 This intrinsic does a series of operations atomically. It first loads the
6134 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
6135 result to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
6136</p>
6137
6138<h5>Examples:</h5>
6139<pre>
6140%ptr = malloc i32
6141 store i32 8, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6142%result1 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 4 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6143 <i>; yields {i32}:result1 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6144%result2 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 2 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6145 <i>; yields {i32}:result2 = 4</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6146%result3 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 5 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6147 <i>; yields {i32}:result3 = 2</i>
6148%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
6149</pre>
6150</div>
6151
6152<!-- _______________________________________________________________________ -->
6153<div class="doc_subsubsection">
6154 <a name="int_atomic_load_and">'<tt>llvm.atomic.load.and.*</tt>' Intrinsic</a><br>
6155 <a name="int_atomic_load_nand">'<tt>llvm.atomic.load.nand.*</tt>' Intrinsic</a><br>
6156 <a name="int_atomic_load_or">'<tt>llvm.atomic.load.or.*</tt>' Intrinsic</a><br>
6157 <a name="int_atomic_load_xor">'<tt>llvm.atomic.load.xor.*</tt>' Intrinsic</a><br>
6158
6159</div>
6160<div class="doc_text">
6161<h5>Syntax:</h5>
6162<p>
6163 These are overloaded intrinsics. You can use <tt>llvm.atomic.load_and</tt>,
6164 <tt>llvm.atomic.load_nand</tt>, <tt>llvm.atomic.load_or</tt>, and
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6165 <tt>llvm.atomic.load_xor</tt> on any integer bit width and for different
6166 address spaces. Not all targets support all bit widths however.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6167<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6168declare i8 @llvm.atomic.load.and.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6169declare i16 @llvm.atomic.load.and.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6170declare i32 @llvm.atomic.load.and.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6171declare 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]6172
6173</pre>
6174
6175<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6176declare i8 @llvm.atomic.load.or.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6177declare i16 @llvm.atomic.load.or.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6178declare i32 @llvm.atomic.load.or.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6179declare 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]6180
6181</pre>
6182
6183<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6184declare i8 @llvm.atomic.load.nand.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6185declare i16 @llvm.atomic.load.nand.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6186declare i32 @llvm.atomic.load.nand.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6187declare 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]6188
6189</pre>
6190
6191<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6192declare i8 @llvm.atomic.load.xor.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6193declare i16 @llvm.atomic.load.xor.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6194declare i32 @llvm.atomic.load.xor.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6195declare 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]6196
6197</pre>
6198<h5>Overview:</h5>
6199<p>
6200 These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
6201 the value stored in memory at <tt>ptr</tt>. It yields the original value
6202 at <tt>ptr</tt>.
6203</p>
6204<h5>Arguments:</h5>
6205<p>
6206
6207 These intrinsics take two arguments, the first a pointer to an integer value
6208 and the second an integer value. The result is also an integer value. These
6209 integer types can have any bit width, but they must all have the same bit
6210 width. The targets may only lower integer representations they support.
6211</p>
6212<h5>Semantics:</h5>
6213<p>
6214 These intrinsics does a series of operations atomically. They first load the
6215 value stored at <tt>ptr</tt>. They then do the bitwise operation
6216 <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the original
6217 value stored at <tt>ptr</tt>.
6218</p>
6219
6220<h5>Examples:</h5>
6221<pre>
6222%ptr = malloc i32
6223 store i32 0x0F0F, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6224%result0 = call i32 @llvm.atomic.load.nand.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6225 <i>; yields {i32}:result0 = 0x0F0F</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6226%result1 = call i32 @llvm.atomic.load.and.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6227 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6228%result2 = call i32 @llvm.atomic.load.or.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6229 <i>; yields {i32}:result2 = 0xF0</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6230%result3 = call i32 @llvm.atomic.load.xor.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6231 <i>; yields {i32}:result3 = FF</i>
6232%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
6233</pre>
6234</div>
6235
6236
6237<!-- _______________________________________________________________________ -->
6238<div class="doc_subsubsection">
6239 <a name="int_atomic_load_max">'<tt>llvm.atomic.load.max.*</tt>' Intrinsic</a><br>
6240 <a name="int_atomic_load_min">'<tt>llvm.atomic.load.min.*</tt>' Intrinsic</a><br>
6241 <a name="int_atomic_load_umax">'<tt>llvm.atomic.load.umax.*</tt>' Intrinsic</a><br>
6242 <a name="int_atomic_load_umin">'<tt>llvm.atomic.load.umin.*</tt>' Intrinsic</a><br>
6243
6244</div>
6245<div class="doc_text">
6246<h5>Syntax:</h5>
6247<p>
6248 These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
6249 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6250 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
6251 address spaces. Not all targets
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6252 support all bit widths however.</p>
6253<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6254declare i8 @llvm.atomic.load.max.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6255declare i16 @llvm.atomic.load.max.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6256declare i32 @llvm.atomic.load.max.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6257declare 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]6258
6259</pre>
6260
6261<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6262declare i8 @llvm.atomic.load.min.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6263declare i16 @llvm.atomic.load.min.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6264declare i32 @llvm.atomic.load.min.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6265declare 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]6266
6267</pre>
6268
6269<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6270declare i8 @llvm.atomic.load.umax.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6271declare i16 @llvm.atomic.load.umax.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6272declare i32 @llvm.atomic.load.umax.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6273declare 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]6274
6275</pre>
6276
6277<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6278declare i8 @llvm.atomic.load.umin.i8..p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6279declare i16 @llvm.atomic.load.umin.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6280declare i32 @llvm.atomic.load.umin.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6281declare 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]6282
6283</pre>
6284<h5>Overview:</h5>
6285<p>
6286 These intrinsics takes the signed or unsigned minimum or maximum of
6287 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
6288 original value at <tt>ptr</tt>.
6289</p>
6290<h5>Arguments:</h5>
6291<p>
6292
6293 These intrinsics take two arguments, the first a pointer to an integer value
6294 and the second an integer value. The result is also an integer value. These
6295 integer types can have any bit width, but they must all have the same bit
6296 width. The targets may only lower integer representations they support.
6297</p>
6298<h5>Semantics:</h5>
6299<p>
6300 These intrinsics does a series of operations atomically. They first load the
6301 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or max
6302 <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They yield
6303 the original value stored at <tt>ptr</tt>.
6304</p>
6305
6306<h5>Examples:</h5>
6307<pre>
6308%ptr = malloc i32
6309 store i32 7, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6310%result0 = call i32 @llvm.atomic.load.min.i32.p0i32( i32* %ptr, i32 -2 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6311 <i>; yields {i32}:result0 = 7</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6312%result1 = call i32 @llvm.atomic.load.max.i32.p0i32( i32* %ptr, i32 8 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6313 <i>; yields {i32}:result1 = -2</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6314%result2 = call i32 @llvm.atomic.load.umin.i32.p0i32( i32* %ptr, i32 10 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6315 <i>; yields {i32}:result2 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6316%result3 = call i32 @llvm.atomic.load.umax.i32.p0i32( i32* %ptr, i32 30 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6317 <i>; yields {i32}:result3 = 8</i>
6318%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
6319</pre>
6320</div>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6321
6322<!-- ======================================================================= -->
6323<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6324 <a name="int_general">General Intrinsics</a>
6325</div>
6326
6327<div class="doc_text">
6328<p> This class of intrinsics is designed to be generic and has
6329no specific purpose. </p>
6330</div>
6331
6332<!-- _______________________________________________________________________ -->
6333<div class="doc_subsubsection">
6334 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
6335</div>
6336
6337<div class="doc_text">
6338
6339<h5>Syntax:</h5>
6340<pre>
6341 declare void @llvm.var.annotation(i8* &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6342</pre>
6343
6344<h5>Overview:</h5>
6345
6346<p>
6347The '<tt>llvm.var.annotation</tt>' intrinsic
6348</p>
6349
6350<h5>Arguments:</h5>
6351
6352<p>
6353The first argument is a pointer to a value, the second is a pointer to a
6354global string, the third is a pointer to a global string which is the source
6355file name, and the last argument is the line number.
6356</p>
6357
6358<h5>Semantics:</h5>
6359
6360<p>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]6361This intrinsic allows annotation of local variables with arbitrary strings.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6362This can be useful for special purpose optimizations that want to look for these
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]6363annotations. These have no other defined use, they are ignored by code
6364generation and optimization.
6365</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6366</div>
6367
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6368<!-- _______________________________________________________________________ -->
6369<div class="doc_subsubsection">
Tanya Lattnerc9869b12007-09-21 23:57:59 +0000[diff] [blame]6370 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6371</div>
6372
6373<div class="doc_text">
6374
6375<h5>Syntax:</h5>
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]6376<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
6377any integer bit width.
6378</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6379<pre>
Tanya Lattner09161fe2007-09-22 00:03:01 +0000[diff] [blame]6380 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6381 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6382 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6383 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6384 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]6385</pre>
6386
6387<h5>Overview:</h5>
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]6388
6389<p>
6390The '<tt>llvm.annotation</tt>' intrinsic.
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6391</p>
6392
6393<h5>Arguments:</h5>
6394
6395<p>
6396The first argument is an integer value (result of some expression),
6397the second is a pointer to a global string, the third is a pointer to a global
6398string which is the source file name, and the last argument is the line number.
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]6399It returns the value of the first argument.
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6400</p>
6401
6402<h5>Semantics:</h5>
6403
6404<p>
6405This intrinsic allows annotations to be put on arbitrary expressions
6406with arbitrary strings. This can be useful for special purpose optimizations
6407that want to look for these annotations. These have no other defined use, they
6408are ignored by code generation and optimization.
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]6409</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6410</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6411
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]6412<!-- _______________________________________________________________________ -->
6413<div class="doc_subsubsection">
6414 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
6415</div>
6416
6417<div class="doc_text">
6418
6419<h5>Syntax:</h5>
6420<pre>
6421 declare void @llvm.trap()
6422</pre>
6423
6424<h5>Overview:</h5>
6425
6426<p>
6427The '<tt>llvm.trap</tt>' intrinsic
6428</p>
6429
6430<h5>Arguments:</h5>
6431
6432<p>
6433None
6434</p>
6435
6436<h5>Semantics:</h5>
6437
6438<p>
6439This intrinsics is lowered to the target dependent trap instruction. If the
6440target does not have a trap instruction, this intrinsic will be lowered to the
6441call of the abort() function.
6442</p>
6443</div>
6444
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]6445<!-- _______________________________________________________________________ -->
6446<div class="doc_subsubsection">
Misha Brukman5dd7f4d2008-11-22 23:55:29 +0000[diff] [blame]6447 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]6448</div>
6449<div class="doc_text">
6450<h5>Syntax:</h5>
6451<pre>
6452declare void @llvm.stackprotector( i8* &lt;guard&gt;, i8** &lt;slot&gt; )
6453
6454</pre>
6455<h5>Overview:</h5>
6456<p>
6457 The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and stores
6458 it onto the stack at <tt>slot</tt>. The stack slot is adjusted to ensure that
6459 it is placed on the stack before local variables.
6460</p>
6461<h5>Arguments:</h5>
6462<p>
6463 The <tt>llvm.stackprotector</tt> intrinsic requires two pointer arguments. The
6464 first argument is the value loaded from the stack guard
6465 <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt> that
6466 has enough space to hold the value of the guard.
6467</p>
6468<h5>Semantics:</h5>
6469<p>
6470 This intrinsic causes the prologue/epilogue inserter to force the position of
6471 the <tt>AllocaInst</tt> stack slot to be before local variables on the
6472 stack. This is to ensure that if a local variable on the stack is overwritten,
6473 it will destroy the value of the guard. When the function exits, the guard on
6474 the stack is checked against the original guard. If they're different, then
6475 the program aborts by calling the <tt>__stack_chk_fail()</tt> function.
6476</p>
6477</div>
6478
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6479<!-- *********************************************************************** -->
6480<hr>
6481<address>
6482 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
6483 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
6484 <a href="http://validator.w3.org/check/referer"><img
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]6485 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!"></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6486
6487 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
6488 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
6489 Last modified: $Date$
6490</address>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]6491
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6492</body>
6493</html>