Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / e7f855e87b5e3a5f255737a85a731c1688ff1c24 / . / docs / LangRef.html
blob: 15f95e2a11c477052022048040e426d6bb0ed6f2 [file] [log] [blame]
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
3<html>
4<head>
5 <title>LLVM Assembly Language Reference Manual</title>
6 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
7 <meta name="author" content="Chris Lattner">
8 <meta name="description"
9 content="LLVM Assembly Language Reference Manual.">
10 <link rel="stylesheet" href="llvm.css" type="text/css">
11</head>
12
13<body>
14
15<div class="doc_title"> LLVM Language Reference Manual </div>
16<ol>
17 <li><a href="#abstract">Abstract</a></li>
18 <li><a href="#introduction">Introduction</a></li>
19 <li><a href="#identifiers">Identifiers</a></li>
20 <li><a href="#highlevel">High Level Structure</a>
21 <ol>
22 <li><a href="#modulestructure">Module Structure</a></li>
23 <li><a href="#linkage">Linkage Types</a></li>
24 <li><a href="#callingconv">Calling Conventions</a></li>
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]25 <li><a href="#namedtypes">Named Types</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]26 <li><a href="#globalvars">Global Variables</a></li>
27 <li><a href="#functionstructure">Functions</a></li>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]28 <li><a href="#aliasstructure">Aliases</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]29 <li><a href="#paramattrs">Parameter Attributes</a></li>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]30 <li><a href="#fnattrs">Function Attributes</a></li>
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]31 <li><a href="#gc">Garbage Collector Names</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]32 <li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
33 <li><a href="#datalayout">Data Layout</a></li>
34 </ol>
35 </li>
36 <li><a href="#typesystem">Type System</a>
37 <ol>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]38 <li><a href="#t_classifications">Type Classifications</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]39 <li><a href="#t_primitive">Primitive Types</a>
40 <ol>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]41 <li><a href="#t_floating">Floating Point Types</a></li>
42 <li><a href="#t_void">Void Type</a></li>
43 <li><a href="#t_label">Label Type</a></li>
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]44 <li><a href="#t_metadata">Metadata Type</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]45 </ol>
46 </li>
47 <li><a href="#t_derived">Derived Types</a>
48 <ol>
Chris Lattner251ab812007-12-18 06:18:21 +0000[diff] [blame]49 <li><a href="#t_integer">Integer Type</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]50 <li><a href="#t_array">Array Type</a></li>
51 <li><a href="#t_function">Function Type</a></li>
52 <li><a href="#t_pointer">Pointer Type</a></li>
53 <li><a href="#t_struct">Structure Type</a></li>
54 <li><a href="#t_pstruct">Packed Structure Type</a></li>
55 <li><a href="#t_vector">Vector Type</a></li>
56 <li><a href="#t_opaque">Opaque Type</a></li>
57 </ol>
58 </li>
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]59 <li><a href="#t_uprefs">Type Up-references</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]60 </ol>
61 </li>
62 <li><a href="#constants">Constants</a>
63 <ol>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]64 <li><a href="#simpleconstants">Simple Constants</a></li>
Chris Lattner97063852009-02-28 18:32:25 +0000[diff] [blame]65 <li><a href="#complexconstants">Complex Constants</a></li>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]66 <li><a href="#globalconstants">Global Variable and Function Addresses</a></li>
67 <li><a href="#undefvalues">Undefined Values</a></li>
68 <li><a href="#constantexprs">Constant Expressions</a></li>
Nick Lewycky4dcf8102009-04-04 07:22:01 +0000[diff] [blame]69 <li><a href="#metadata">Embedded Metadata</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]70 </ol>
71 </li>
72 <li><a href="#othervalues">Other Values</a>
73 <ol>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]74 <li><a href="#inlineasm">Inline Assembler Expressions</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]75 </ol>
76 </li>
77 <li><a href="#instref">Instruction Reference</a>
78 <ol>
79 <li><a href="#terminators">Terminator Instructions</a>
80 <ol>
81 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
82 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
83 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
84 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
85 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
86 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
87 </ol>
88 </li>
89 <li><a href="#binaryops">Binary Operations</a>
90 <ol>
91 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]92 <li><a href="#i_fadd">'<tt>fadd</tt>' Instruction</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]93 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]94 <li><a href="#i_fsub">'<tt>fsub</tt>' Instruction</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]95 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]96 <li><a href="#i_fmul">'<tt>fmul</tt>' Instruction</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]97 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
98 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
99 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
100 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
101 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
102 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
103 </ol>
104 </li>
105 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
106 <ol>
107 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
108 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
109 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
110 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
111 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
112 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
113 </ol>
114 </li>
115 <li><a href="#vectorops">Vector Operations</a>
116 <ol>
117 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
118 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
119 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
120 </ol>
121 </li>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]122 <li><a href="#aggregateops">Aggregate Operations</a>
123 <ol>
124 <li><a href="#i_extractvalue">'<tt>extractvalue</tt>' Instruction</a></li>
125 <li><a href="#i_insertvalue">'<tt>insertvalue</tt>' Instruction</a></li>
126 </ol>
127 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]128 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
129 <ol>
130 <li><a href="#i_malloc">'<tt>malloc</tt>' Instruction</a></li>
131 <li><a href="#i_free">'<tt>free</tt>' Instruction</a></li>
132 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
133 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
134 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
135 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
136 </ol>
137 </li>
138 <li><a href="#convertops">Conversion Operations</a>
139 <ol>
140 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
141 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
142 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
143 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
144 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
145 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
146 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
147 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
148 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
149 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
150 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
151 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
152 </ol>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]153 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]154 <li><a href="#otherops">Other Operations</a>
155 <ol>
156 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
157 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
158 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
159 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
160 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
161 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
162 </ol>
163 </li>
164 </ol>
165 </li>
166 <li><a href="#intrinsics">Intrinsic Functions</a>
167 <ol>
168 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
169 <ol>
170 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
171 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
172 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
173 </ol>
174 </li>
175 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
176 <ol>
177 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
178 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
179 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
180 </ol>
181 </li>
182 <li><a href="#int_codegen">Code Generator Intrinsics</a>
183 <ol>
184 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
185 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
186 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
187 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
188 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
189 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
190 <li><a href="#int_readcyclecounter"><tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
191 </ol>
192 </li>
193 <li><a href="#int_libc">Standard C Library Intrinsics</a>
194 <ol>
195 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
196 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
197 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
198 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
199 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]200 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
201 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
202 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]203 </ol>
204 </li>
205 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
206 <ol>
207 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
208 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
209 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
210 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]211 </ol>
212 </li>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]213 <li><a href="#int_overflow">Arithmetic with Overflow Intrinsics</a>
214 <ol>
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]215 <li><a href="#int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt> Intrinsics</a></li>
216 <li><a href="#int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt> Intrinsics</a></li>
217 <li><a href="#int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt> Intrinsics</a></li>
218 <li><a href="#int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt> Intrinsics</a></li>
219 <li><a href="#int_smul_overflow">'<tt>llvm.smul.with.overflow.*</tt> Intrinsics</a></li>
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]220 <li><a href="#int_umul_overflow">'<tt>llvm.umul.with.overflow.*</tt> Intrinsics</a></li>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]221 </ol>
222 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]223 <li><a href="#int_debugger">Debugger intrinsics</a></li>
224 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]225 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]226 <ol>
227 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]228 </ol>
229 </li>
Bill Wendling9127adb2008-11-18 22:10:53 +0000[diff] [blame]230 <li><a href="#int_atomics">Atomic intrinsics</a>
231 <ol>
232 <li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
233 <li><a href="#int_atomic_cmp_swap"><tt>llvm.atomic.cmp.swap</tt></a></li>
234 <li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
235 <li><a href="#int_atomic_load_add"><tt>llvm.atomic.load.add</tt></a></li>
236 <li><a href="#int_atomic_load_sub"><tt>llvm.atomic.load.sub</tt></a></li>
237 <li><a href="#int_atomic_load_and"><tt>llvm.atomic.load.and</tt></a></li>
238 <li><a href="#int_atomic_load_nand"><tt>llvm.atomic.load.nand</tt></a></li>
239 <li><a href="#int_atomic_load_or"><tt>llvm.atomic.load.or</tt></a></li>
240 <li><a href="#int_atomic_load_xor"><tt>llvm.atomic.load.xor</tt></a></li>
241 <li><a href="#int_atomic_load_max"><tt>llvm.atomic.load.max</tt></a></li>
242 <li><a href="#int_atomic_load_min"><tt>llvm.atomic.load.min</tt></a></li>
243 <li><a href="#int_atomic_load_umax"><tt>llvm.atomic.load.umax</tt></a></li>
244 <li><a href="#int_atomic_load_umin"><tt>llvm.atomic.load.umin</tt></a></li>
245 </ol>
246 </li>
Reid Spencerb043f672007-07-20 19:59:11 +0000[diff] [blame]247 <li><a href="#int_general">General intrinsics</a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]248 <ol>
Reid Spencerb043f672007-07-20 19:59:11 +0000[diff] [blame]249 <li><a href="#int_var_annotation">
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]250 '<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]251 <li><a href="#int_annotation">
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]252 '<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]253 <li><a href="#int_trap">
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]254 '<tt>llvm.trap</tt>' Intrinsic</a></li>
255 <li><a href="#int_stackprotector">
256 '<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]257 </ol>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]258 </li>
259 </ol>
260 </li>
261</ol>
262
263<div class="doc_author">
264 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
265 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
266</div>
267
268<!-- *********************************************************************** -->
269<div class="doc_section"> <a name="abstract">Abstract </a></div>
270<!-- *********************************************************************** -->
271
272<div class="doc_text">
273<p>This document is a reference manual for the LLVM assembly language.
Bill Wendlinge7846a52008-08-05 22:29:16 +0000[diff] [blame]274LLVM is a Static Single Assignment (SSA) based representation that provides
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]275type safety, low-level operations, flexibility, and the capability of
276representing 'all' high-level languages cleanly. It is the common code
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]277representation used throughout all phases of the LLVM compilation
278strategy.</p>
279</div>
280
281<!-- *********************************************************************** -->
282<div class="doc_section"> <a name="introduction">Introduction</a> </div>
283<!-- *********************************************************************** -->
284
285<div class="doc_text">
286
287<p>The LLVM code representation is designed to be used in three
288different forms: as an in-memory compiler IR, as an on-disk bitcode
289representation (suitable for fast loading by a Just-In-Time compiler),
290and as a human readable assembly language representation. This allows
291LLVM to provide a powerful intermediate representation for efficient
292compiler transformations and analysis, while providing a natural means
293to debug and visualize the transformations. The three different forms
294of LLVM are all equivalent. This document describes the human readable
295representation and notation.</p>
296
297<p>The LLVM representation aims to be light-weight and low-level
298while being expressive, typed, and extensible at the same time. It
299aims to be a "universal IR" of sorts, by being at a low enough level
300that high-level ideas may be cleanly mapped to it (similar to how
301microprocessors are "universal IR's", allowing many source languages to
302be mapped to them). By providing type information, LLVM can be used as
303the target of optimizations: for example, through pointer analysis, it
304can be proven that a C automatic variable is never accessed outside of
305the current function... allowing it to be promoted to a simple SSA
306value instead of a memory location.</p>
307
308</div>
309
310<!-- _______________________________________________________________________ -->
311<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
312
313<div class="doc_text">
314
315<p>It is important to note that this document describes 'well formed'
316LLVM assembly language. There is a difference between what the parser
317accepts and what is considered 'well formed'. For example, the
318following instruction is syntactically okay, but not well formed:</p>
319
320<div class="doc_code">
321<pre>
322%x = <a href="#i_add">add</a> i32 1, %x
323</pre>
324</div>
325
326<p>...because the definition of <tt>%x</tt> does not dominate all of
327its uses. The LLVM infrastructure provides a verification pass that may
328be used to verify that an LLVM module is well formed. This pass is
329automatically run by the parser after parsing input assembly and by
330the optimizer before it outputs bitcode. The violations pointed out
331by the verifier pass indicate bugs in transformation passes or input to
332the parser.</p>
333</div>
334
Chris Lattnera83fdc02007-10-03 17:34:29 +0000[diff] [blame]335<!-- Describe the typesetting conventions here. -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]336
337<!-- *********************************************************************** -->
338<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
339<!-- *********************************************************************** -->
340
341<div class="doc_text">
342
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]343 <p>LLVM identifiers come in two basic types: global and local. Global
344 identifiers (functions, global variables) begin with the @ character. Local
345 identifiers (register names, types) begin with the % character. Additionally,
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]346 there are three different formats for identifiers, for different purposes:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]347
348<ol>
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]349 <li>Named values are represented as a string of characters with their prefix.
350 For example, %foo, @DivisionByZero, %a.really.long.identifier. The actual
351 regular expression used is '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]352 Identifiers which require other characters in their names can be surrounded
Daniel Dunbar73f9a772008-10-14 23:51:43 +0000[diff] [blame]353 with quotes. Special characters may be escaped using "\xx" where xx is the
354 ASCII code for the character in hexadecimal. In this way, any character can
355 be used in a name value, even quotes themselves.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]356
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]357 <li>Unnamed values are represented as an unsigned numeric value with their
358 prefix. For example, %12, @2, %44.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]359
360 <li>Constants, which are described in a <a href="#constants">section about
361 constants</a>, below.</li>
362</ol>
363
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]364<p>LLVM requires that values start with a prefix for two reasons: Compilers
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]365don't need to worry about name clashes with reserved words, and the set of
366reserved words may be expanded in the future without penalty. Additionally,
367unnamed identifiers allow a compiler to quickly come up with a temporary
368variable without having to avoid symbol table conflicts.</p>
369
370<p>Reserved words in LLVM are very similar to reserved words in other
371languages. There are keywords for different opcodes
372('<tt><a href="#i_add">add</a></tt>',
373 '<tt><a href="#i_bitcast">bitcast</a></tt>',
374 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
375href="#t_void">void</a></tt>', '<tt><a href="#t_primitive">i32</a></tt>', etc...),
376and others. These reserved words cannot conflict with variable names, because
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]377none of them start with a prefix character ('%' or '@').</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]378
379<p>Here is an example of LLVM code to multiply the integer variable
380'<tt>%X</tt>' by 8:</p>
381
382<p>The easy way:</p>
383
384<div class="doc_code">
385<pre>
386%result = <a href="#i_mul">mul</a> i32 %X, 8
387</pre>
388</div>
389
390<p>After strength reduction:</p>
391
392<div class="doc_code">
393<pre>
394%result = <a href="#i_shl">shl</a> i32 %X, i8 3
395</pre>
396</div>
397
398<p>And the hard way:</p>
399
400<div class="doc_code">
401<pre>
402<a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
403<a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
404%result = <a href="#i_add">add</a> i32 %1, %1
405</pre>
406</div>
407
408<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
409important lexical features of LLVM:</p>
410
411<ol>
412
413 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
414 line.</li>
415
416 <li>Unnamed temporaries are created when the result of a computation is not
417 assigned to a named value.</li>
418
419 <li>Unnamed temporaries are numbered sequentially</li>
420
421</ol>
422
423<p>...and it also shows a convention that we follow in this document. When
424demonstrating instructions, we will follow an instruction with a comment that
425defines the type and name of value produced. Comments are shown in italic
426text.</p>
427
428</div>
429
430<!-- *********************************************************************** -->
431<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
432<!-- *********************************************************************** -->
433
434<!-- ======================================================================= -->
435<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
436</div>
437
438<div class="doc_text">
439
440<p>LLVM programs are composed of "Module"s, each of which is a
441translation unit of the input programs. Each module consists of
442functions, global variables, and symbol table entries. Modules may be
443combined together with the LLVM linker, which merges function (and
444global variable) definitions, resolves forward declarations, and merges
445symbol table entries. Here is an example of the "hello world" module:</p>
446
447<div class="doc_code">
448<pre><i>; Declare the string constant as a global constant...</i>
449<a href="#identifiers">@.LC0</a> = <a href="#linkage_internal">internal</a> <a
450 href="#globalvars">constant</a> <a href="#t_array">[13 x i8]</a> c"hello world\0A\00" <i>; [13 x i8]*</i>
451
452<i>; External declaration of the puts function</i>
453<a href="#functionstructure">declare</a> i32 @puts(i8 *) <i>; i32(i8 *)* </i>
454
455<i>; Definition of main function</i>
456define i32 @main() { <i>; i32()* </i>
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]457 <i>; Convert [13 x i8]* to i8 *...</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]458 %cast210 = <a
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]459 href="#i_getelementptr">getelementptr</a> [13 x i8]* @.LC0, i64 0, i64 0 <i>; i8 *</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]460
461 <i>; Call puts function to write out the string to stdout...</i>
462 <a
463 href="#i_call">call</a> i32 @puts(i8 * %cast210) <i>; i32</i>
464 <a
465 href="#i_ret">ret</a> i32 0<br>}<br>
466</pre>
467</div>
468
469<p>This example is made up of a <a href="#globalvars">global variable</a>
470named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
471function, and a <a href="#functionstructure">function definition</a>
472for "<tt>main</tt>".</p>
473
474<p>In general, a module is made up of a list of global values,
475where both functions and global variables are global values. Global values are
476represented by a pointer to a memory location (in this case, a pointer to an
477array of char, and a pointer to a function), and have one of the following <a
478href="#linkage">linkage types</a>.</p>
479
480</div>
481
482<!-- ======================================================================= -->
483<div class="doc_subsection">
484 <a name="linkage">Linkage Types</a>
485</div>
486
487<div class="doc_text">
488
489<p>
490All Global Variables and Functions have one of the following types of linkage:
491</p>
492
493<dl>
494
Rafael Espindolaa168fc92009-01-15 20:18:42 +0000[diff] [blame]495 <dt><tt><b><a name="linkage_private">private</a></b></tt>: </dt>
496
497 <dd>Global values with private linkage are only directly accessible by
498 objects in the current module. In particular, linking code into a module with
499 an private global value may cause the private to be renamed as necessary to
500 avoid collisions. Because the symbol is private to the module, all
501 references can be updated. This doesn't show up in any symbol table in the
502 object file.
503 </dd>
504
Dale Johannesen96e7e092008-05-23 23:13:41 +0000[diff] [blame]505 <dt><tt><b><a name="linkage_internal">internal</a></b></tt>: </dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]506
Duncan Sandsa75223a2009-01-16 09:29:46 +0000[diff] [blame]507 <dd> Similar to private, but the value shows as a local symbol (STB_LOCAL in
Rafael Espindolaa168fc92009-01-15 20:18:42 +0000[diff] [blame]508 the case of ELF) in the object file. This corresponds to the notion of the
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]509 '<tt>static</tt>' keyword in C.
510 </dd>
511
Chris Lattner68433442009-04-13 05:44:34 +0000[diff] [blame]512 <dt><tt><b><a name="available_externally">available_externally</a></b></tt>:
513 </dt>
514
515 <dd>Globals with "<tt>available_externally</tt>" linkage are never emitted
516 into the object file corresponding to the LLVM module. They exist to
517 allow inlining and other optimizations to take place given knowledge of the
518 definition of the global, which is known to be somewhere outside the module.
519 Globals with <tt>available_externally</tt> linkage are allowed to be discarded
520 at will, and are otherwise the same as <tt>linkonce_odr</tt>. This linkage
521 type is only allowed on definitions, not declarations.</dd>
522
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]523 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
524
525 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
526 the same name when linkage occurs. This is typically used to implement
527 inline functions, templates, or other code which must be generated in each
528 translation unit that uses it. Unreferenced <tt>linkonce</tt> globals are
529 allowed to be discarded.
530 </dd>
531
Dale Johannesen96e7e092008-05-23 23:13:41 +0000[diff] [blame]532 <dt><tt><b><a name="linkage_common">common</a></b></tt>: </dt>
533
534 <dd>"<tt>common</tt>" linkage is exactly the same as <tt>linkonce</tt>
535 linkage, except that unreferenced <tt>common</tt> globals may not be
536 discarded. This is used for globals that may be emitted in multiple
537 translation units, but that are not guaranteed to be emitted into every
538 translation unit that uses them. One example of this is tentative
539 definitions in C, such as "<tt>int X;</tt>" at global scope.
540 </dd>
541
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]542 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
543
Dale Johannesen96e7e092008-05-23 23:13:41 +0000[diff] [blame]544 <dd>"<tt>weak</tt>" linkage is the same as <tt>common</tt> linkage, except
545 that some targets may choose to emit different assembly sequences for them
546 for target-dependent reasons. This is used for globals that are declared
547 "weak" in C source code.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]548 </dd>
549
550 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
551
552 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
553 pointer to array type. When two global variables with appending linkage are
554 linked together, the two global arrays are appended together. This is the
555 LLVM, typesafe, equivalent of having the system linker append together
556 "sections" with identical names when .o files are linked.
557 </dd>
558
559 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt>: </dt>
Duncan Sands19d161f2009-03-07 15:45:40 +0000[diff] [blame]560
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]561 <dd>The semantics of this linkage follow the ELF object file model: the
562 symbol is weak until linked, if not linked, the symbol becomes null instead
563 of being an undefined reference.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]564 </dd>
565
Duncan Sands19d161f2009-03-07 15:45:40 +0000[diff] [blame]566 <dt><tt><b><a name="linkage_linkonce">linkonce_odr</a></b></tt>: </dt>
Duncan Sands19d161f2009-03-07 15:45:40 +0000[diff] [blame]567 <dt><tt><b><a name="linkage_weak">weak_odr</a></b></tt>: </dt>
Chris Lattner68433442009-04-13 05:44:34 +0000[diff] [blame]568 <dd>Some languages allow differing globals to be merged, such as two
Duncan Sands19d161f2009-03-07 15:45:40 +0000[diff] [blame]569 functions with different semantics. Other languages, such as <tt>C++</tt>,
570 ensure that only equivalent globals are ever merged (the "one definition
Chris Lattner68433442009-04-13 05:44:34 +0000[diff] [blame]571 rule" - "ODR"). Such languages can use the <tt>linkonce_odr</tt>
Duncan Sandsb95df792009-03-11 20:14:15 +0000[diff] [blame]572 and <tt>weak_odr</tt> linkage types to indicate that the global will only
573 be merged with equivalent globals. These linkage types are otherwise the
574 same as their non-<tt>odr</tt> versions.
Duncan Sands19d161f2009-03-07 15:45:40 +0000[diff] [blame]575 </dd>
576
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]577 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
578
579 <dd>If none of the above identifiers are used, the global is externally
580 visible, meaning that it participates in linkage and can be used to resolve
581 external symbol references.
582 </dd>
583</dl>
584
585 <p>
586 The next two types of linkage are targeted for Microsoft Windows platform
587 only. They are designed to support importing (exporting) symbols from (to)
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]588 DLLs (Dynamic Link Libraries).
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]589 </p>
590
591 <dl>
592 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt>: </dt>
593
594 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
595 or variable via a global pointer to a pointer that is set up by the DLL
596 exporting the symbol. On Microsoft Windows targets, the pointer name is
Dan Gohman5ec99832009-01-12 21:35:55 +0000[diff] [blame]597 formed by combining <code>__imp_</code> and the function or variable name.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]598 </dd>
599
600 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt>: </dt>
601
602 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
603 pointer to a pointer in a DLL, so that it can be referenced with the
604 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
Dan Gohman5ec99832009-01-12 21:35:55 +0000[diff] [blame]605 name is formed by combining <code>__imp_</code> and the function or variable
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]606 name.
607 </dd>
608
609</dl>
610
Dan Gohman4dfac702008-11-24 17:18:39 +0000[diff] [blame]611<p>For example, since the "<tt>.LC0</tt>"
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]612variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
613variable and was linked with this one, one of the two would be renamed,
614preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
615external (i.e., lacking any linkage declarations), they are accessible
616outside of the current module.</p>
617<p>It is illegal for a function <i>declaration</i>
Duncan Sands565f65d2009-03-11 08:08:06 +0000[diff] [blame]618to have any linkage type other than "externally visible", <tt>dllimport</tt>
619or <tt>extern_weak</tt>.</p>
Duncan Sands19d161f2009-03-07 15:45:40 +0000[diff] [blame]620<p>Aliases can have only <tt>external</tt>, <tt>internal</tt>, <tt>weak</tt>
621or <tt>weak_odr</tt> linkages.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]622</div>
623
624<!-- ======================================================================= -->
625<div class="doc_subsection">
626 <a name="callingconv">Calling Conventions</a>
627</div>
628
629<div class="doc_text">
630
631<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
632and <a href="#i_invoke">invokes</a> can all have an optional calling convention
633specified for the call. The calling convention of any pair of dynamic
634caller/callee must match, or the behavior of the program is undefined. The
635following calling conventions are supported by LLVM, and more may be added in
636the future:</p>
637
638<dl>
639 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
640
641 <dd>This calling convention (the default if no other calling convention is
642 specified) matches the target C calling conventions. This calling convention
643 supports varargs function calls and tolerates some mismatch in the declared
644 prototype and implemented declaration of the function (as does normal C).
645 </dd>
646
647 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
648
649 <dd>This calling convention attempts to make calls as fast as possible
650 (e.g. by passing things in registers). This calling convention allows the
651 target to use whatever tricks it wants to produce fast code for the target,
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]652 without having to conform to an externally specified ABI (Application Binary
653 Interface). Implementations of this convention should allow arbitrary
Arnold Schwaighofer07444922008-05-14 09:17:12 +0000[diff] [blame]654 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> to be
655 supported. This calling convention does not support varargs and requires the
656 prototype of all callees to exactly match the prototype of the function
657 definition.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]658 </dd>
659
660 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
661
662 <dd>This calling convention attempts to make code in the caller as efficient
663 as possible under the assumption that the call is not commonly executed. As
664 such, these calls often preserve all registers so that the call does not break
665 any live ranges in the caller side. This calling convention does not support
666 varargs and requires the prototype of all callees to exactly match the
667 prototype of the function definition.
668 </dd>
669
670 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
671
672 <dd>Any calling convention may be specified by number, allowing
673 target-specific calling conventions to be used. Target specific calling
674 conventions start at 64.
675 </dd>
676</dl>
677
678<p>More calling conventions can be added/defined on an as-needed basis, to
679support pascal conventions or any other well-known target-independent
680convention.</p>
681
682</div>
683
684<!-- ======================================================================= -->
685<div class="doc_subsection">
686 <a name="visibility">Visibility Styles</a>
687</div>
688
689<div class="doc_text">
690
691<p>
692All Global Variables and Functions have one of the following visibility styles:
693</p>
694
695<dl>
696 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
697
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]698 <dd>On targets that use the ELF object file format, default visibility means
699 that the declaration is visible to other
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]700 modules and, in shared libraries, means that the declared entity may be
701 overridden. On Darwin, default visibility means that the declaration is
702 visible to other modules. Default visibility corresponds to "external
703 linkage" in the language.
704 </dd>
705
706 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
707
708 <dd>Two declarations of an object with hidden visibility refer to the same
709 object if they are in the same shared object. Usually, hidden visibility
710 indicates that the symbol will not be placed into the dynamic symbol table,
711 so no other module (executable or shared library) can reference it
712 directly.
713 </dd>
714
715 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
716
717 <dd>On ELF, protected visibility indicates that the symbol will be placed in
718 the dynamic symbol table, but that references within the defining module will
719 bind to the local symbol. That is, the symbol cannot be overridden by another
720 module.
721 </dd>
722</dl>
723
724</div>
725
726<!-- ======================================================================= -->
727<div class="doc_subsection">
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]728 <a name="namedtypes">Named Types</a>
729</div>
730
731<div class="doc_text">
732
733<p>LLVM IR allows you to specify name aliases for certain types. This can make
734it easier to read the IR and make the IR more condensed (particularly when
735recursive types are involved). An example of a name specification is:
736</p>
737
738<div class="doc_code">
739<pre>
740%mytype = type { %mytype*, i32 }
741</pre>
742</div>
743
744<p>You may give a name to any <a href="#typesystem">type</a> except "<a
745href="t_void">void</a>". Type name aliases may be used anywhere a type is
746expected with the syntax "%mytype".</p>
747
748<p>Note that type names are aliases for the structural type that they indicate,
749and that you can therefore specify multiple names for the same type. This often
750leads to confusing behavior when dumping out a .ll file. Since LLVM IR uses
751structural typing, the name is not part of the type. When printing out LLVM IR,
752the printer will pick <em>one name</em> to render all types of a particular
753shape. This means that if you have code where two different source types end up
754having the same LLVM type, that the dumper will sometimes print the "wrong" or
755unexpected type. This is an important design point and isn't going to
756change.</p>
757
758</div>
759
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]760<!-- ======================================================================= -->
761<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]762 <a name="globalvars">Global Variables</a>
763</div>
764
765<div class="doc_text">
766
767<p>Global variables define regions of memory allocated at compilation time
768instead of run-time. Global variables may optionally be initialized, may have
769an explicit section to be placed in, and may have an optional explicit alignment
770specified. A variable may be defined as "thread_local", which means that it
771will not be shared by threads (each thread will have a separated copy of the
772variable). A variable may be defined as a global "constant," which indicates
773that the contents of the variable will <b>never</b> be modified (enabling better
774optimization, allowing the global data to be placed in the read-only section of
775an executable, etc). Note that variables that need runtime initialization
776cannot be marked "constant" as there is a store to the variable.</p>
777
778<p>
779LLVM explicitly allows <em>declarations</em> of global variables to be marked
780constant, even if the final definition of the global is not. This capability
781can be used to enable slightly better optimization of the program, but requires
782the language definition to guarantee that optimizations based on the
783'constantness' are valid for the translation units that do not include the
784definition.
785</p>
786
787<p>As SSA values, global variables define pointer values that are in
788scope (i.e. they dominate) all basic blocks in the program. Global
789variables always define a pointer to their "content" type because they
790describe a region of memory, and all memory objects in LLVM are
791accessed through pointers.</p>
792
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]793<p>A global variable may be declared to reside in a target-specifc numbered
794address space. For targets that support them, address spaces may affect how
795optimizations are performed and/or what target instructions are used to access
Christopher Lamb20a39e92007-12-12 08:44:39 +0000[diff] [blame]796the variable. The default address space is zero. The address space qualifier
797must precede any other attributes.</p>
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]798
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]799<p>LLVM allows an explicit section to be specified for globals. If the target
800supports it, it will emit globals to the section specified.</p>
801
802<p>An explicit alignment may be specified for a global. If not present, or if
803the alignment is set to zero, the alignment of the global is set by the target
804to whatever it feels convenient. If an explicit alignment is specified, the
805global is forced to have at least that much alignment. All alignments must be
806a power of 2.</p>
807
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]808<p>For example, the following defines a global in a numbered address space with
809an initializer, section, and alignment:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]810
811<div class="doc_code">
812<pre>
Dan Gohman21ef02c2009-01-11 00:40:00 +0000[diff] [blame]813@G = addrspace(5) constant float 1.0, section "foo", align 4
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]814</pre>
815</div>
816
817</div>
818
819
820<!-- ======================================================================= -->
821<div class="doc_subsection">
822 <a name="functionstructure">Functions</a>
823</div>
824
825<div class="doc_text">
826
827<p>LLVM function definitions consist of the "<tt>define</tt>" keyord,
828an optional <a href="#linkage">linkage type</a>, an optional
829<a href="#visibility">visibility style</a>, an optional
830<a href="#callingconv">calling convention</a>, a return type, an optional
831<a href="#paramattrs">parameter attribute</a> for the return type, a function
832name, a (possibly empty) argument list (each with optional
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]833<a href="#paramattrs">parameter attributes</a>), optional
834<a href="#fnattrs">function attributes</a>, an optional section,
835an optional alignment, an optional <a href="#gc">garbage collector name</a>,
Chris Lattneree202542008-10-04 18:10:21 +0000[diff] [blame]836an opening curly brace, a list of basic blocks, and a closing curly brace.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]837
838LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
839optional <a href="#linkage">linkage type</a>, an optional
840<a href="#visibility">visibility style</a>, an optional
841<a href="#callingconv">calling convention</a>, a return type, an optional
842<a href="#paramattrs">parameter attribute</a> for the return type, a function
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]843name, a possibly empty list of arguments, an optional alignment, and an optional
Gordon Henriksend606f5b2007-12-10 03:30:21 +0000[diff] [blame]844<a href="#gc">garbage collector name</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]845
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]846<p>A function definition contains a list of basic blocks, forming the CFG
847(Control Flow Graph) for
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]848the function. Each basic block may optionally start with a label (giving the
849basic block a symbol table entry), contains a list of instructions, and ends
850with a <a href="#terminators">terminator</a> instruction (such as a branch or
851function return).</p>
852
853<p>The first basic block in a function is special in two ways: it is immediately
854executed on entrance to the function, and it is not allowed to have predecessor
855basic blocks (i.e. there can not be any branches to the entry block of a
856function). Because the block can have no predecessors, it also cannot have any
857<a href="#i_phi">PHI nodes</a>.</p>
858
859<p>LLVM allows an explicit section to be specified for functions. If the target
860supports it, it will emit functions to the section specified.</p>
861
862<p>An explicit alignment may be specified for a function. If not present, or if
863the alignment is set to zero, the alignment of the function is set by the target
864to whatever it feels convenient. If an explicit alignment is specified, the
865function is forced to have at least that much alignment. All alignments must be
866a power of 2.</p>
867
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]868 <h5>Syntax:</h5>
869
870<div class="doc_code">
Chris Lattner1e5c5cd02008-10-13 16:55:18 +0000[diff] [blame]871<tt>
872define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
873 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
874 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
875 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
876 [<a href="#gc">gc</a>] { ... }
877</tt>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]878</div>
879
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]880</div>
881
882
883<!-- ======================================================================= -->
884<div class="doc_subsection">
885 <a name="aliasstructure">Aliases</a>
886</div>
887<div class="doc_text">
888 <p>Aliases act as "second name" for the aliasee value (which can be either
Anton Korobeynikov96822812008-03-22 08:36:14 +0000[diff] [blame]889 function, global variable, another alias or bitcast of global value). Aliases
890 may have an optional <a href="#linkage">linkage type</a>, and an
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]891 optional <a href="#visibility">visibility style</a>.</p>
892
893 <h5>Syntax:</h5>
894
895<div class="doc_code">
896<pre>
Duncan Sandsd7bfabf2008-09-12 20:48:21 +0000[diff] [blame]897@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]898</pre>
899</div>
900
901</div>
902
903
904
905<!-- ======================================================================= -->
906<div class="doc_subsection"><a name="paramattrs">Parameter Attributes</a></div>
907<div class="doc_text">
908 <p>The return type and each parameter of a function type may have a set of
909 <i>parameter attributes</i> associated with them. Parameter attributes are
910 used to communicate additional information about the result or parameters of
Duncan Sandsf5588dc2007-11-27 13:23:08 +0000[diff] [blame]911 a function. Parameter attributes are considered to be part of the function,
912 not of the function type, so functions with different parameter attributes
913 can have the same function type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]914
915 <p>Parameter attributes are simple keywords that follow the type specified. If
916 multiple parameter attributes are needed, they are space separated. For
917 example:</p>
918
919<div class="doc_code">
920<pre>
Nick Lewycky3022a742009-02-15 23:06:14 +0000[diff] [blame]921declare i32 @printf(i8* noalias nocapture, ...)
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]922declare i32 @atoi(i8 zeroext)
923declare signext i8 @returns_signed_char()
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]924</pre>
925</div>
926
Duncan Sandsf5588dc2007-11-27 13:23:08 +0000[diff] [blame]927 <p>Note that any attributes for the function result (<tt>nounwind</tt>,
928 <tt>readonly</tt>) come immediately after the argument list.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]929
930 <p>Currently, only the following parameter attributes are defined:</p>
931 <dl>
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]932 <dt><tt>zeroext</tt></dt>
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]933 <dd>This indicates to the code generator that the parameter or return value
934 should be zero-extended to a 32-bit value by the caller (for a parameter)
935 or the callee (for a return value).</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]936
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]937 <dt><tt>signext</tt></dt>
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]938 <dd>This indicates to the code generator that the parameter or return value
939 should be sign-extended to a 32-bit value by the caller (for a parameter)
940 or the callee (for a return value).</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]941
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]942 <dt><tt>inreg</tt></dt>
Dale Johannesenc08a0e22008-09-25 20:47:45 +0000[diff] [blame]943 <dd>This indicates that this parameter or return value should be treated
944 in a special target-dependent fashion during while emitting code for a
945 function call or return (usually, by putting it in a register as opposed
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]946 to memory, though some targets use it to distinguish between two different
947 kinds of registers). Use of this attribute is target-specific.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]948
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]949 <dt><tt><a name="byval">byval</a></tt></dt>
Chris Lattner04c86182008-01-15 04:34:22 +0000[diff] [blame]950 <dd>This indicates that the pointer parameter should really be passed by
951 value to the function. The attribute implies that a hidden copy of the
952 pointee is made between the caller and the callee, so the callee is unable
Chris Lattner6a9f3c42008-08-05 18:21:08 +0000[diff] [blame]953 to modify the value in the callee. This attribute is only valid on LLVM
Chris Lattner04c86182008-01-15 04:34:22 +0000[diff] [blame]954 pointer arguments. It is generally used to pass structs and arrays by
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]955 value, but is also valid on pointers to scalars. The copy is considered to
956 belong to the caller not the callee (for example,
957 <tt><a href="#readonly">readonly</a></tt> functions should not write to
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]958 <tt>byval</tt> parameters). This is not a valid attribute for return
Chris Lattnere39c5bf2009-02-05 05:42:28 +0000[diff] [blame]959 values. The byval attribute also supports specifying an alignment with the
960 align attribute. This has a target-specific effect on the code generator
961 that usually indicates a desired alignment for the synthesized stack
962 slot.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]963
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]964 <dt><tt>sret</tt></dt>
Duncan Sands616cc032008-02-18 04:19:38 +0000[diff] [blame]965 <dd>This indicates that the pointer parameter specifies the address of a
966 structure that is the return value of the function in the source program.
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]967 This pointer must be guaranteed by the caller to be valid: loads and stores
968 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]969 be applied to the first parameter. This is not a valid attribute for
970 return values. </dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]971
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]972 <dt><tt>noalias</tt></dt>
Nick Lewyckyff384472008-11-24 03:41:24 +0000[diff] [blame]973 <dd>This indicates that the pointer does not alias any global or any other
974 parameter. The caller is responsible for ensuring that this is the
Nick Lewyckya604a942008-11-24 05:00:44 +0000[diff] [blame]975 case. On a function return value, <tt>noalias</tt> additionally indicates
976 that the pointer does not alias any other pointers visible to the
Nick Lewycky40fb6f22008-12-19 06:39:12 +0000[diff] [blame]977 caller. For further details, please see the discussion of the NoAlias
978 response in
979 <a href="http://llvm.org/docs/AliasAnalysis.html#MustMayNo">alias
980 analysis</a>.</dd>
981
982 <dt><tt>nocapture</tt></dt>
983 <dd>This indicates that the callee does not make any copies of the pointer
984 that outlive the callee itself. This is not a valid attribute for return
985 values.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]986
Duncan Sands4ee46812007-07-27 19:57:41 +0000[diff] [blame]987 <dt><tt>nest</tt></dt>
Duncan Sandsf1a7d4c2008-07-08 09:27:25 +0000[diff] [blame]988 <dd>This indicates that the pointer parameter can be excised using the
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]989 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
990 attribute for return values.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]991 </dl>
992
993</div>
994
995<!-- ======================================================================= -->
996<div class="doc_subsection">
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]997 <a name="gc">Garbage Collector Names</a>
998</div>
999
1000<div class="doc_text">
1001<p>Each function may specify a garbage collector name, which is simply a
1002string.</p>
1003
1004<div class="doc_code"><pre
1005>define void @f() gc "name" { ...</pre></div>
1006
1007<p>The compiler declares the supported values of <i>name</i>. Specifying a
1008collector which will cause the compiler to alter its output in order to support
1009the named garbage collection algorithm.</p>
1010</div>
1011
1012<!-- ======================================================================= -->
1013<div class="doc_subsection">
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1014 <a name="fnattrs">Function Attributes</a>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1015</div>
1016
1017<div class="doc_text">
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1018
1019<p>Function attributes are set to communicate additional information about
1020 a function. Function attributes are considered to be part of the function,
1021 not of the function type, so functions with different parameter attributes
1022 can have the same function type.</p>
1023
1024 <p>Function attributes are simple keywords that follow the type specified. If
1025 multiple attributes are needed, they are space separated. For
1026 example:</p>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1027
1028<div class="doc_code">
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1029<pre>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1030define void @f() noinline { ... }
1031define void @f() alwaysinline { ... }
1032define void @f() alwaysinline optsize { ... }
1033define void @f() optsize
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1034</pre>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1035</div>
1036
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1037<dl>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1038<dt><tt>alwaysinline</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]1039<dd>This attribute indicates that the inliner should attempt to inline this
1040function into callers whenever possible, ignoring any active inlining size
1041threshold for this caller.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1042
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1043<dt><tt>noinline</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]1044<dd>This attribute indicates that the inliner should never inline this function
Chris Lattner377f3ae2008-10-05 17:14:59 +0000[diff] [blame]1045in any situation. This attribute may not be used together with the
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]1046<tt>alwaysinline</tt> attribute.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1047
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1048<dt><tt>optsize</tt></dt>
Devang Patelc29099b2008-09-29 18:34:44 +0000[diff] [blame]1049<dd>This attribute suggests that optimization passes and code generator passes
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]1050make choices that keep the code size of this function low, and otherwise do
1051optimizations specifically to reduce code size.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1052
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1053<dt><tt>noreturn</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]1054<dd>This function attribute indicates that the function never returns normally.
1055This produces undefined behavior at runtime if the function ever does
1056dynamically return.</dd>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1057
1058<dt><tt>nounwind</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]1059<dd>This function attribute indicates that the function never returns with an
1060unwind or exceptional control flow. If the function does unwind, its runtime
1061behavior is undefined.</dd>
1062
1063<dt><tt>readnone</tt></dt>
Duncan Sands2f500832009-05-06 06:49:50 +0000[diff] [blame]1064<dd>This attribute indicates that the function computes its result (or decides to
1065unwind an exception) based strictly on its arguments, without dereferencing any
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]1066pointer arguments or otherwise accessing any mutable state (e.g. memory, control
1067registers, etc) visible to caller functions. It does not write through any
1068pointer arguments (including <tt><a href="#byval">byval</a></tt> arguments) and
Duncan Sands2f500832009-05-06 06:49:50 +0000[diff] [blame]1069never changes any state visible to callers. This means that it cannot unwind
1070exceptions by calling the <tt>C++</tt> exception throwing methods, but could
1071use the <tt>unwind</tt> instruction.</dd>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1072
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]1073<dt><tt><a name="readonly">readonly</a></tt></dt>
1074<dd>This attribute indicates that the function does not write through any
1075pointer arguments (including <tt><a href="#byval">byval</a></tt> arguments)
1076or otherwise modify any state (e.g. memory, control registers, etc) visible to
1077caller functions. It may dereference pointer arguments and read state that may
Duncan Sands2f500832009-05-06 06:49:50 +0000[diff] [blame]1078be set in the caller. A readonly function always returns the same value (or
1079unwinds an exception identically) when called with the same set of arguments
1080and global state. It cannot unwind an exception by calling the <tt>C++</tt>
1081exception throwing methods, but may use the <tt>unwind</tt> instruction.</dd>
Bill Wendlingdfaabba2008-11-13 01:02:51 +0000[diff] [blame]1082
1083<dt><tt><a name="ssp">ssp</a></tt></dt>
Bill Wendling34f70d82008-11-26 19:19:05 +0000[diff] [blame]1084<dd>This attribute indicates that the function should emit a stack smashing
Bill Wendlingdfaabba2008-11-13 01:02:51 +0000[diff] [blame]1085protector. It is in the form of a "canary"&mdash;a random value placed on the
1086stack before the local variables that's checked upon return from the function to
1087see if it has been overwritten. A heuristic is used to determine if a function
Bill Wendling34f70d82008-11-26 19:19:05 +0000[diff] [blame]1088needs stack protectors or not.
Bill Wendlingdfaabba2008-11-13 01:02:51 +0000[diff] [blame]1089
Devang Patela2f9f412009-06-12 19:45:19 +0000[diff] [blame]1090<br><br>If a function that has an <tt>ssp</tt> attribute is inlined into a function
Bill Wendlingbe9ec3f2008-11-26 19:07:40 +0000[diff] [blame]1091that doesn't have an <tt>ssp</tt> attribute, then the resulting function will
Devang Patela2f9f412009-06-12 19:45:19 +0000[diff] [blame]1092have an <tt>ssp</tt> attribute.</dd>
Bill Wendlingbe9ec3f2008-11-26 19:07:40 +0000[diff] [blame]1093
1094<dt><tt>sspreq</tt></dt>
Bill Wendling34f70d82008-11-26 19:19:05 +0000[diff] [blame]1095<dd>This attribute indicates that the function should <em>always</em> emit a
Bill Wendlingdfaabba2008-11-13 01:02:51 +0000[diff] [blame]1096stack smashing protector. This overrides the <tt><a href="#ssp">ssp</a></tt>
Bill Wendling34f70d82008-11-26 19:19:05 +0000[diff] [blame]1097function attribute.
Bill Wendlingbe9ec3f2008-11-26 19:07:40 +0000[diff] [blame]1098
Devang Patela2f9f412009-06-12 19:45:19 +0000[diff] [blame]1099If a function that has an <tt>sspreq</tt> attribute is inlined into a
Bill Wendlingbe9ec3f2008-11-26 19:07:40 +0000[diff] [blame]1100function that doesn't have an <tt>sspreq</tt> attribute or which has
1101an <tt>ssp</tt> attribute, then the resulting function will have
Devang Patela2f9f412009-06-12 19:45:19 +0000[diff] [blame]1102an <tt>sspreq</tt> attribute.</dd>
1103
1104<dt><tt>noredzone</tt></dt>
Dan Gohman06c9b732009-06-15 17:37:09 +0000[diff] [blame]1105<dd>This attribute indicates that the code generator should not use a
Dan Gohmanf958d5c2009-06-15 21:18:01 +0000[diff] [blame]1106red zone, even if the target-specific ABI normally permits it.
Dan Gohman06c9b732009-06-15 17:37:09 +0000[diff] [blame]1107</dd>
Devang Patela2f9f412009-06-12 19:45:19 +0000[diff] [blame]1108
1109<dt><tt>noimplicitfloat</tt></dt>
1110<dd>This attributes disables implicit floating point instructions.</dd>
1111
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1112</dl>
1113
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1114</div>
1115
1116<!-- ======================================================================= -->
1117<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1118 <a name="moduleasm">Module-Level Inline Assembly</a>
1119</div>
1120
1121<div class="doc_text">
1122<p>
1123Modules may contain "module-level inline asm" blocks, which corresponds to the
1124GCC "file scope inline asm" blocks. These blocks are internally concatenated by
1125LLVM and treated as a single unit, but may be separated in the .ll file if
1126desired. The syntax is very simple:
1127</p>
1128
1129<div class="doc_code">
1130<pre>
1131module asm "inline asm code goes here"
1132module asm "more can go here"
1133</pre>
1134</div>
1135
1136<p>The strings can contain any character by escaping non-printable characters.
1137 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
1138 for the number.
1139</p>
1140
1141<p>
1142 The inline asm code is simply printed to the machine code .s file when
1143 assembly code is generated.
1144</p>
1145</div>
1146
1147<!-- ======================================================================= -->
1148<div class="doc_subsection">
1149 <a name="datalayout">Data Layout</a>
1150</div>
1151
1152<div class="doc_text">
1153<p>A module may specify a target specific data layout string that specifies how
1154data is to be laid out in memory. The syntax for the data layout is simply:</p>
1155<pre> target datalayout = "<i>layout specification</i>"</pre>
1156<p>The <i>layout specification</i> consists of a list of specifications
1157separated by the minus sign character ('-'). Each specification starts with a
1158letter and may include other information after the letter to define some
1159aspect of the data layout. The specifications accepted are as follows: </p>
1160<dl>
1161 <dt><tt>E</tt></dt>
1162 <dd>Specifies that the target lays out data in big-endian form. That is, the
1163 bits with the most significance have the lowest address location.</dd>
1164 <dt><tt>e</tt></dt>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1165 <dd>Specifies that the target lays out data in little-endian form. That is,
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1166 the bits with the least significance have the lowest address location.</dd>
1167 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1168 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
1169 <i>preferred</i> alignments. All sizes are in bits. Specifying the <i>pref</i>
1170 alignment is optional. If omitted, the preceding <tt>:</tt> should be omitted
1171 too.</dd>
1172 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1173 <dd>This specifies the alignment for an integer type of a given bit
1174 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1175 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1176 <dd>This specifies the alignment for a vector type of a given bit
1177 <i>size</i>.</dd>
1178 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1179 <dd>This specifies the alignment for a floating point type of a given bit
1180 <i>size</i>. The value of <i>size</i> must be either 32 (float) or 64
1181 (double).</dd>
1182 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1183 <dd>This specifies the alignment for an aggregate type of a given bit
1184 <i>size</i>.</dd>
Daniel Dunbard88a97b2009-06-08 22:17:53 +0000[diff] [blame]1185 <dt><tt>s<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1186 <dd>This specifies the alignment for a stack object of a given bit
1187 <i>size</i>.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1188</dl>
1189<p>When constructing the data layout for a given target, LLVM starts with a
1190default set of specifications which are then (possibly) overriden by the
1191specifications in the <tt>datalayout</tt> keyword. The default specifications
1192are given in this list:</p>
1193<ul>
1194 <li><tt>E</tt> - big endian</li>
1195 <li><tt>p:32:64:64</tt> - 32-bit pointers with 64-bit alignment</li>
1196 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1197 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1198 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1199 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1200 <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]1201 alignment of 64-bits</li>
1202 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1203 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1204 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1205 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1206 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
Daniel Dunbard88a97b2009-06-08 22:17:53 +0000[diff] [blame]1207 <li><tt>s0:64:64</tt> - stack objects are 64-bit aligned</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1208</ul>
Chris Lattner6a9f3c42008-08-05 18:21:08 +0000[diff] [blame]1209<p>When LLVM is determining the alignment for a given type, it uses the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]1210following rules:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1211<ol>
1212 <li>If the type sought is an exact match for one of the specifications, that
1213 specification is used.</li>
1214 <li>If no match is found, and the type sought is an integer type, then the
1215 smallest integer type that is larger than the bitwidth of the sought type is
1216 used. If none of the specifications are larger than the bitwidth then the the
1217 largest integer type is used. For example, given the default specifications
1218 above, the i7 type will use the alignment of i8 (next largest) while both
1219 i65 and i256 will use the alignment of i64 (largest specified).</li>
1220 <li>If no match is found, and the type sought is a vector type, then the
1221 largest vector type that is smaller than the sought vector type will be used
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]1222 as a fall back. This happens because &lt;128 x double&gt; can be implemented
1223 in terms of 64 &lt;2 x double&gt;, for example.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1224</ol>
1225</div>
1226
1227<!-- *********************************************************************** -->
1228<div class="doc_section"> <a name="typesystem">Type System</a> </div>
1229<!-- *********************************************************************** -->
1230
1231<div class="doc_text">
1232
1233<p>The LLVM type system is one of the most important features of the
1234intermediate representation. Being typed enables a number of
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1235optimizations to be performed on the intermediate representation directly,
1236without having to do
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1237extra analyses on the side before the transformation. A strong type
1238system makes it easier to read the generated code and enables novel
1239analyses and transformations that are not feasible to perform on normal
1240three address code representations.</p>
1241
1242</div>
1243
1244<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1245<div class="doc_subsection"> <a name="t_classifications">Type
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1246Classifications</a> </div>
1247<div class="doc_text">
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1248<p>The types fall into a few useful
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1249classifications:</p>
1250
1251<table border="1" cellspacing="0" cellpadding="4">
1252 <tbody>
1253 <tr><th>Classification</th><th>Types</th></tr>
1254 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1255 <td><a href="#t_integer">integer</a></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1256 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
1257 </tr>
1258 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1259 <td><a href="#t_floating">floating point</a></td>
1260 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1261 </tr>
1262 <tr>
1263 <td><a name="t_firstclass">first class</a></td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1264 <td><a href="#t_integer">integer</a>,
1265 <a href="#t_floating">floating point</a>,
1266 <a href="#t_pointer">pointer</a>,
Dan Gohmanf6237db2008-06-18 18:42:13 +0000[diff] [blame]1267 <a href="#t_vector">vector</a>,
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]1268 <a href="#t_struct">structure</a>,
1269 <a href="#t_array">array</a>,
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]1270 <a href="#t_label">label</a>,
1271 <a href="#t_metadata">metadata</a>.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1272 </td>
1273 </tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1274 <tr>
1275 <td><a href="#t_primitive">primitive</a></td>
1276 <td><a href="#t_label">label</a>,
1277 <a href="#t_void">void</a>,
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]1278 <a href="#t_floating">floating point</a>,
1279 <a href="#t_metadata">metadata</a>.</td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1280 </tr>
1281 <tr>
1282 <td><a href="#t_derived">derived</a></td>
1283 <td><a href="#t_integer">integer</a>,
1284 <a href="#t_array">array</a>,
1285 <a href="#t_function">function</a>,
1286 <a href="#t_pointer">pointer</a>,
1287 <a href="#t_struct">structure</a>,
1288 <a href="#t_pstruct">packed structure</a>,
1289 <a href="#t_vector">vector</a>,
1290 <a href="#t_opaque">opaque</a>.
Dan Gohman032ba852008-10-14 16:32:04 +0000[diff] [blame]1291 </td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1292 </tr>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1293 </tbody>
1294</table>
1295
1296<p>The <a href="#t_firstclass">first class</a> types are perhaps the
1297most important. Values of these types are the only ones which can be
1298produced by instructions, passed as arguments, or used as operands to
Dan Gohman4f29e422008-05-23 21:53:15 +0000[diff] [blame]1299instructions.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1300</div>
1301
1302<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1303<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1304
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1305<div class="doc_text">
1306<p>The primitive types are the fundamental building blocks of the LLVM
1307system.</p>
1308
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1309</div>
1310
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1311<!-- _______________________________________________________________________ -->
1312<div class="doc_subsubsection"> <a name="t_floating">Floating Point Types</a> </div>
1313
1314<div class="doc_text">
1315 <table>
1316 <tbody>
1317 <tr><th>Type</th><th>Description</th></tr>
1318 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1319 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1320 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1321 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1322 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1323 </tbody>
1324 </table>
1325</div>
1326
1327<!-- _______________________________________________________________________ -->
1328<div class="doc_subsubsection"> <a name="t_void">Void Type</a> </div>
1329
1330<div class="doc_text">
1331<h5>Overview:</h5>
1332<p>The void type does not represent any value and has no size.</p>
1333
1334<h5>Syntax:</h5>
1335
1336<pre>
1337 void
1338</pre>
1339</div>
1340
1341<!-- _______________________________________________________________________ -->
1342<div class="doc_subsubsection"> <a name="t_label">Label Type</a> </div>
1343
1344<div class="doc_text">
1345<h5>Overview:</h5>
1346<p>The label type represents code labels.</p>
1347
1348<h5>Syntax:</h5>
1349
1350<pre>
1351 label
1352</pre>
1353</div>
1354
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]1355<!-- _______________________________________________________________________ -->
1356<div class="doc_subsubsection"> <a name="t_metadata">Metadata Type</a> </div>
1357
1358<div class="doc_text">
1359<h5>Overview:</h5>
1360<p>The metadata type represents embedded metadata. The only derived type that
1361may contain metadata is <tt>metadata*</tt> or a function type that returns or
1362takes metadata typed parameters, but not pointer to metadata types.</p>
1363
1364<h5>Syntax:</h5>
1365
1366<pre>
1367 metadata
1368</pre>
1369</div>
1370
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1371
1372<!-- ======================================================================= -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1373<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
1374
1375<div class="doc_text">
1376
1377<p>The real power in LLVM comes from the derived types in the system.
1378This is what allows a programmer to represent arrays, functions,
1379pointers, and other useful types. Note that these derived types may be
1380recursive: For example, it is possible to have a two dimensional array.</p>
1381
1382</div>
1383
1384<!-- _______________________________________________________________________ -->
1385<div class="doc_subsubsection"> <a name="t_integer">Integer Type</a> </div>
1386
1387<div class="doc_text">
1388
1389<h5>Overview:</h5>
1390<p>The integer type is a very simple derived type that simply specifies an
1391arbitrary bit width for the integer type desired. Any bit width from 1 bit to
13922^23-1 (about 8 million) can be specified.</p>
1393
1394<h5>Syntax:</h5>
1395
1396<pre>
1397 iN
1398</pre>
1399
1400<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1401value.</p>
1402
1403<h5>Examples:</h5>
1404<table class="layout">
Nick Lewycky39382d62009-05-24 02:46:06 +0000[diff] [blame]1405 <tr class="layout">
1406 <td class="left"><tt>i1</tt></td>
1407 <td class="left">a single-bit integer.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1408 </tr>
Nick Lewycky39382d62009-05-24 02:46:06 +0000[diff] [blame]1409 <tr class="layout">
1410 <td class="left"><tt>i32</tt></td>
1411 <td class="left">a 32-bit integer.</td>
1412 </tr>
1413 <tr class="layout">
1414 <td class="left"><tt>i1942652</tt></td>
1415 <td class="left">a really big integer of over 1 million bits.</td>
1416 </tr>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1417</table>
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1418
1419<p>Note that the code generator does not yet support large integer types
1420to be used as function return types. The specific limit on how large a
1421return type the code generator can currently handle is target-dependent;
1422currently it's often 64 bits for 32-bit targets and 128 bits for 64-bit
1423targets.</p>
1424
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1425</div>
1426
1427<!-- _______________________________________________________________________ -->
1428<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
1429
1430<div class="doc_text">
1431
1432<h5>Overview:</h5>
1433
1434<p>The array type is a very simple derived type that arranges elements
1435sequentially in memory. The array type requires a size (number of
1436elements) and an underlying data type.</p>
1437
1438<h5>Syntax:</h5>
1439
1440<pre>
1441 [&lt;# elements&gt; x &lt;elementtype&gt;]
1442</pre>
1443
1444<p>The number of elements is a constant integer value; elementtype may
1445be any type with a size.</p>
1446
1447<h5>Examples:</h5>
1448<table class="layout">
1449 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1450 <td class="left"><tt>[40 x i32]</tt></td>
1451 <td class="left">Array of 40 32-bit integer values.</td>
1452 </tr>
1453 <tr class="layout">
1454 <td class="left"><tt>[41 x i32]</tt></td>
1455 <td class="left">Array of 41 32-bit integer values.</td>
1456 </tr>
1457 <tr class="layout">
1458 <td class="left"><tt>[4 x i8]</tt></td>
1459 <td class="left">Array of 4 8-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1460 </tr>
1461</table>
1462<p>Here are some examples of multidimensional arrays:</p>
1463<table class="layout">
1464 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1465 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1466 <td class="left">3x4 array of 32-bit integer values.</td>
1467 </tr>
1468 <tr class="layout">
1469 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1470 <td class="left">12x10 array of single precision floating point values.</td>
1471 </tr>
1472 <tr class="layout">
1473 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1474 <td class="left">2x3x4 array of 16-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1475 </tr>
1476</table>
1477
1478<p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
1479length array. Normally, accesses past the end of an array are undefined in
1480LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
1481As a special case, however, zero length arrays are recognized to be variable
1482length. This allows implementation of 'pascal style arrays' with the LLVM
1483type "{ i32, [0 x float]}", for example.</p>
1484
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1485<p>Note that the code generator does not yet support large aggregate types
1486to be used as function return types. The specific limit on how large an
1487aggregate return type the code generator can currently handle is
1488target-dependent, and also dependent on the aggregate element types.</p>
1489
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1490</div>
1491
1492<!-- _______________________________________________________________________ -->
1493<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
1494<div class="doc_text">
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1495
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1496<h5>Overview:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1497
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1498<p>The function type can be thought of as a function signature. It
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1499consists of a return type and a list of formal parameter types. The
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1500return 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]1501If 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]1502class types, and the struct must have at least one element.</p>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1503
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1504<h5>Syntax:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1505
1506<pre>
1507 &lt;returntype list&gt; (&lt;parameter list&gt;)
1508</pre>
1509
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1510<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
1511specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1512which indicates that the function takes a variable number of arguments.
1513Variable argument functions can access their arguments with the <a
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1514 href="#int_varargs">variable argument handling intrinsic</a> functions.
1515'<tt>&lt;returntype list&gt;</tt>' is a comma-separated list of
1516<a href="#t_firstclass">first class</a> type specifiers.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1517
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1518<h5>Examples:</h5>
1519<table class="layout">
1520 <tr class="layout">
1521 <td class="left"><tt>i32 (i32)</tt></td>
1522 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
1523 </td>
1524 </tr><tr class="layout">
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]1525 <td class="left"><tt>float&nbsp;(i16&nbsp;signext,&nbsp;i32&nbsp;*)&nbsp;*
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1526 </tt></td>
1527 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
1528 an <tt>i16</tt> that should be sign extended and a
1529 <a href="#t_pointer">pointer</a> to <tt>i32</tt>, returning
1530 <tt>float</tt>.
1531 </td>
1532 </tr><tr class="layout">
1533 <td class="left"><tt>i32 (i8*, ...)</tt></td>
1534 <td class="left">A vararg function that takes at least one
1535 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1536 which returns an integer. This is the signature for <tt>printf</tt> in
1537 LLVM.
1538 </td>
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1539 </tr><tr class="layout">
1540 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Misha Brukmanafc88b02008-11-27 06:41:20 +0000[diff] [blame]1541 <td class="left">A function taking an <tt>i32</tt>, returning two
1542 <tt>i32</tt> values as an aggregate of type <tt>{ i32, i32 }</tt>
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1543 </td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1544 </tr>
1545</table>
1546
1547</div>
1548<!-- _______________________________________________________________________ -->
1549<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
1550<div class="doc_text">
1551<h5>Overview:</h5>
1552<p>The structure type is used to represent a collection of data members
1553together in memory. The packing of the field types is defined to match
1554the ABI of the underlying processor. The elements of a structure may
1555be any type that has a size.</p>
1556<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1557and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1558field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1559instruction.</p>
1560<h5>Syntax:</h5>
1561<pre> { &lt;type list&gt; }<br></pre>
1562<h5>Examples:</h5>
1563<table class="layout">
1564 <tr class="layout">
1565 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1566 <td class="left">A triple of three <tt>i32</tt> values</td>
1567 </tr><tr class="layout">
1568 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1569 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1570 second element is a <a href="#t_pointer">pointer</a> to a
1571 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1572 an <tt>i32</tt>.</td>
1573 </tr>
1574</table>
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1575
1576<p>Note that the code generator does not yet support large aggregate types
1577to be used as function return types. The specific limit on how large an
1578aggregate return type the code generator can currently handle is
1579target-dependent, and also dependent on the aggregate element types.</p>
1580
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1581</div>
1582
1583<!-- _______________________________________________________________________ -->
1584<div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
1585</div>
1586<div class="doc_text">
1587<h5>Overview:</h5>
1588<p>The packed structure type is used to represent a collection of data members
1589together in memory. There is no padding between fields. Further, the alignment
1590of a packed structure is 1 byte. The elements of a packed structure may
1591be any type that has a size.</p>
1592<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1593and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1594field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1595instruction.</p>
1596<h5>Syntax:</h5>
1597<pre> &lt; { &lt;type list&gt; } &gt; <br></pre>
1598<h5>Examples:</h5>
1599<table class="layout">
1600 <tr class="layout">
1601 <td class="left"><tt>&lt; { i32, i32, i32 } &gt;</tt></td>
1602 <td class="left">A triple of three <tt>i32</tt> values</td>
1603 </tr><tr class="layout">
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1604 <td class="left">
1605<tt>&lt;&nbsp;{&nbsp;float,&nbsp;i32&nbsp;(i32)*&nbsp;}&nbsp;&gt;</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1606 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1607 second element is a <a href="#t_pointer">pointer</a> to a
1608 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1609 an <tt>i32</tt>.</td>
1610 </tr>
1611</table>
1612</div>
1613
1614<!-- _______________________________________________________________________ -->
1615<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
1616<div class="doc_text">
1617<h5>Overview:</h5>
1618<p>As in many languages, the pointer type represents a pointer or
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]1619reference to another object, which must live in memory. Pointer types may have
1620an optional address space attribute defining the target-specific numbered
1621address space where the pointed-to object resides. The default address space is
1622zero.</p>
Chris Lattner96edbd32009-02-08 19:53:29 +0000[diff] [blame]1623
1624<p>Note that LLVM does not permit pointers to void (<tt>void*</tt>) nor does
Chris Lattner0d879e02009-02-08 22:21:28 +0000[diff] [blame]1625it permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</tt> instead.</p>
Chris Lattner96edbd32009-02-08 19:53:29 +0000[diff] [blame]1626
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1627<h5>Syntax:</h5>
1628<pre> &lt;type&gt; *<br></pre>
1629<h5>Examples:</h5>
1630<table class="layout">
1631 <tr class="layout">
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]1632 <td class="left"><tt>[4 x i32]*</tt></td>
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1633 <td class="left">A <a href="#t_pointer">pointer</a> to <a
1634 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
1635 </tr>
1636 <tr class="layout">
1637 <td class="left"><tt>i32 (i32 *) *</tt></td>
1638 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1639 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1640 <tt>i32</tt>.</td>
1641 </tr>
1642 <tr class="layout">
1643 <td class="left"><tt>i32 addrspace(5)*</tt></td>
1644 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
1645 that resides in address space #5.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1646 </tr>
1647</table>
1648</div>
1649
1650<!-- _______________________________________________________________________ -->
1651<div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
1652<div class="doc_text">
1653
1654<h5>Overview:</h5>
1655
1656<p>A vector type is a simple derived type that represents a vector
1657of elements. Vector types are used when multiple primitive data
1658are operated in parallel using a single instruction (SIMD).
1659A vector type requires a size (number of
1660elements) and an underlying primitive data type. Vectors must have a power
1661of two length (1, 2, 4, 8, 16 ...). Vector types are
1662considered <a href="#t_firstclass">first class</a>.</p>
1663
1664<h5>Syntax:</h5>
1665
1666<pre>
1667 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1668</pre>
1669
1670<p>The number of elements is a constant integer value; elementtype may
1671be any integer or floating point type.</p>
1672
1673<h5>Examples:</h5>
1674
1675<table class="layout">
1676 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1677 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
1678 <td class="left">Vector of 4 32-bit integer values.</td>
1679 </tr>
1680 <tr class="layout">
1681 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
1682 <td class="left">Vector of 8 32-bit floating-point values.</td>
1683 </tr>
1684 <tr class="layout">
1685 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
1686 <td class="left">Vector of 2 64-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1687 </tr>
1688</table>
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1689
1690<p>Note that the code generator does not yet support large vector types
1691to be used as function return types. The specific limit on how large a
1692vector return type codegen can currently handle is target-dependent;
1693currently it's often a few times longer than a hardware vector register.</p>
1694
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1695</div>
1696
1697<!-- _______________________________________________________________________ -->
1698<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1699<div class="doc_text">
1700
1701<h5>Overview:</h5>
1702
1703<p>Opaque types are used to represent unknown types in the system. This
Gordon Henriksenda0706e2007-10-14 00:34:53 +0000[diff] [blame]1704corresponds (for example) to the C notion of a forward declared structure type.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1705In LLVM, opaque types can eventually be resolved to any type (not just a
1706structure type).</p>
1707
1708<h5>Syntax:</h5>
1709
1710<pre>
1711 opaque
1712</pre>
1713
1714<h5>Examples:</h5>
1715
1716<table class="layout">
1717 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1718 <td class="left"><tt>opaque</tt></td>
1719 <td class="left">An opaque type.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1720 </tr>
1721</table>
1722</div>
1723
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]1724<!-- ======================================================================= -->
1725<div class="doc_subsection">
1726 <a name="t_uprefs">Type Up-references</a>
1727</div>
1728
1729<div class="doc_text">
1730<h5>Overview:</h5>
1731<p>
1732An "up reference" allows you to refer to a lexically enclosing type without
1733requiring it to have a name. For instance, a structure declaration may contain a
1734pointer to any of the types it is lexically a member of. Example of up
1735references (with their equivalent as named type declarations) include:</p>
1736
1737<pre>
Chris Lattner5ad632d2009-02-09 10:00:56 +0000[diff] [blame]1738 { \2 * } %x = type { %x* }
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]1739 { \2 }* %y = type { %y }*
1740 \1* %z = type %z*
1741</pre>
1742
1743<p>
1744An up reference is needed by the asmprinter for printing out cyclic types when
1745there is no declared name for a type in the cycle. Because the asmprinter does
1746not want to print out an infinite type string, it needs a syntax to handle
1747recursive types that have no names (all names are optional in llvm IR).
1748</p>
1749
1750<h5>Syntax:</h5>
1751<pre>
1752 \&lt;level&gt;
1753</pre>
1754
1755<p>
1756The level is the count of the lexical type that is being referred to.
1757</p>
1758
1759<h5>Examples:</h5>
1760
1761<table class="layout">
1762 <tr class="layout">
1763 <td class="left"><tt>\1*</tt></td>
1764 <td class="left">Self-referential pointer.</td>
1765 </tr>
1766 <tr class="layout">
1767 <td class="left"><tt>{ { \3*, i8 }, i32 }</tt></td>
1768 <td class="left">Recursive structure where the upref refers to the out-most
1769 structure.</td>
1770 </tr>
1771</table>
1772</div>
1773
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1774
1775<!-- *********************************************************************** -->
1776<div class="doc_section"> <a name="constants">Constants</a> </div>
1777<!-- *********************************************************************** -->
1778
1779<div class="doc_text">
1780
1781<p>LLVM has several different basic types of constants. This section describes
1782them all and their syntax.</p>
1783
1784</div>
1785
1786<!-- ======================================================================= -->
1787<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
1788
1789<div class="doc_text">
1790
1791<dl>
1792 <dt><b>Boolean constants</b></dt>
1793
1794 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
1795 constants of the <tt><a href="#t_primitive">i1</a></tt> type.
1796 </dd>
1797
1798 <dt><b>Integer constants</b></dt>
1799
1800 <dd>Standard integers (such as '4') are constants of the <a
1801 href="#t_integer">integer</a> type. Negative numbers may be used with
1802 integer types.
1803 </dd>
1804
1805 <dt><b>Floating point constants</b></dt>
1806
1807 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
1808 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]1809 notation (see below). The assembler requires the exact decimal value of
1810 a floating-point constant. For example, the assembler accepts 1.25 but
1811 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
1812 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1813
1814 <dt><b>Null pointer constants</b></dt>
1815
1816 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
1817 and must be of <a href="#t_pointer">pointer type</a>.</dd>
1818
1819</dl>
1820
Dale Johannesenf82a52f2009-02-11 22:14:51 +0000[diff] [blame]1821<p>The one non-intuitive notation for constants is the hexadecimal form
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1822of floating point constants. For example, the form '<tt>double
18230x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
18244.5e+15</tt>'. The only time hexadecimal floating point constants are required
1825(and the only time that they are generated by the disassembler) is when a
1826floating point constant must be emitted but it cannot be represented as a
Dale Johannesenf82a52f2009-02-11 22:14:51 +0000[diff] [blame]1827decimal floating point number in a reasonable number of digits. For example,
1828NaN's, infinities, and other
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1829special values are represented in their IEEE hexadecimal format so that
1830assembly and disassembly do not cause any bits to change in the constants.</p>
Dale Johannesenf82a52f2009-02-11 22:14:51 +0000[diff] [blame]1831<p>When using the hexadecimal form, constants of types float and double are
1832represented using the 16-digit form shown above (which matches the IEEE754
1833representation for double); float values must, however, be exactly representable
1834as IEE754 single precision.
1835Hexadecimal format is always used for long
1836double, and there are three forms of long double. The 80-bit
1837format used by x86 is represented as <tt>0xK</tt>
1838followed by 20 hexadecimal digits.
1839The 128-bit format used by PowerPC (two adjacent doubles) is represented
1840by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit
1841format is represented
1842by <tt>0xL</tt> followed by 32 hexadecimal digits; no currently supported
1843target uses this format. Long doubles will only work if they match
1844the long double format on your target. All hexadecimal formats are big-endian
1845(sign bit at the left).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1846</div>
1847
1848<!-- ======================================================================= -->
Chris Lattner97063852009-02-28 18:32:25 +0000[diff] [blame]1849<div class="doc_subsection">
1850<a name="aggregateconstants"> <!-- old anchor -->
1851<a name="complexconstants">Complex Constants</a></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1852</div>
1853
1854<div class="doc_text">
Chris Lattner97063852009-02-28 18:32:25 +0000[diff] [blame]1855<p>Complex constants are a (potentially recursive) combination of simple
1856constants and smaller complex constants.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1857
1858<dl>
1859 <dt><b>Structure constants</b></dt>
1860
1861 <dd>Structure constants are represented with notation similar to structure
1862 type definitions (a comma separated list of elements, surrounded by braces
Chris Lattner6c8de962007-12-25 20:34:52 +0000[diff] [blame]1863 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
1864 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]1865 must have <a href="#t_struct">structure type</a>, and the number and
1866 types of elements must match those specified by the type.
1867 </dd>
1868
1869 <dt><b>Array constants</b></dt>
1870
1871 <dd>Array constants are represented with notation similar to array type
1872 definitions (a comma separated list of elements, surrounded by square brackets
1873 (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74 ]</tt>". Array
1874 constants must have <a href="#t_array">array type</a>, and the number and
1875 types of elements must match those specified by the type.
1876 </dd>
1877
1878 <dt><b>Vector constants</b></dt>
1879
1880 <dd>Vector constants are represented with notation similar to vector type
1881 definitions (a comma separated list of elements, surrounded by
1882 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32 42,
1883 i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must have <a
1884 href="#t_vector">vector type</a>, and the number and types of elements must
1885 match those specified by the type.
1886 </dd>
1887
1888 <dt><b>Zero initialization</b></dt>
1889
1890 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1891 value to zero of <em>any</em> type, including scalar and aggregate types.
1892 This is often used to avoid having to print large zero initializers (e.g. for
1893 large arrays) and is always exactly equivalent to using explicit zero
1894 initializers.
1895 </dd>
Nick Lewycky4dcf8102009-04-04 07:22:01 +0000[diff] [blame]1896
1897 <dt><b>Metadata node</b></dt>
1898
Nick Lewyckyf122c7e2009-05-30 16:08:30 +0000[diff] [blame]1899 <dd>A metadata node is a structure-like constant with
1900 <a href="#t_metadata">metadata type</a>. For example:
1901 "<tt>metadata !{ i32 0, metadata !"test" }</tt>". Unlike other constants
1902 that are meant to be interpreted as part of the instruction stream, metadata
1903 is a place to attach additional information such as debug info.
Nick Lewycky4dcf8102009-04-04 07:22:01 +0000[diff] [blame]1904 </dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1905</dl>
1906
1907</div>
1908
1909<!-- ======================================================================= -->
1910<div class="doc_subsection">
1911 <a name="globalconstants">Global Variable and Function Addresses</a>
1912</div>
1913
1914<div class="doc_text">
1915
1916<p>The addresses of <a href="#globalvars">global variables</a> and <a
1917href="#functionstructure">functions</a> are always implicitly valid (link-time)
1918constants. These constants are explicitly referenced when the <a
1919href="#identifiers">identifier for the global</a> is used and always have <a
1920href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1921file:</p>
1922
1923<div class="doc_code">
1924<pre>
1925@X = global i32 17
1926@Y = global i32 42
1927@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
1928</pre>
1929</div>
1930
1931</div>
1932
1933<!-- ======================================================================= -->
1934<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
1935<div class="doc_text">
1936 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
1937 no specific value. Undefined values may be of any type and be used anywhere
1938 a constant is permitted.</p>
1939
1940 <p>Undefined values indicate to the compiler that the program is well defined
1941 no matter what value is used, giving the compiler more freedom to optimize.
1942 </p>
1943</div>
1944
1945<!-- ======================================================================= -->
1946<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1947</div>
1948
1949<div class="doc_text">
1950
1951<p>Constant expressions are used to allow expressions involving other constants
1952to be used as constants. Constant expressions may be of any <a
1953href="#t_firstclass">first class</a> type and may involve any LLVM operation
1954that does not have side effects (e.g. load and call are not supported). The
1955following is the syntax for constant expressions:</p>
1956
1957<dl>
1958 <dt><b><tt>trunc ( CST to TYPE )</tt></b></dt>
1959 <dd>Truncate a constant to another type. The bit size of CST must be larger
1960 than the bit size of TYPE. Both types must be integers.</dd>
1961
1962 <dt><b><tt>zext ( CST to TYPE )</tt></b></dt>
1963 <dd>Zero extend a constant to another type. The bit size of CST must be
1964 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1965
1966 <dt><b><tt>sext ( CST to TYPE )</tt></b></dt>
1967 <dd>Sign extend a constant to another type. The bit size of CST must be
1968 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1969
1970 <dt><b><tt>fptrunc ( CST to TYPE )</tt></b></dt>
1971 <dd>Truncate a floating point constant to another floating point type. The
1972 size of CST must be larger than the size of TYPE. Both types must be
1973 floating point.</dd>
1974
1975 <dt><b><tt>fpext ( CST to TYPE )</tt></b></dt>
1976 <dd>Floating point extend a constant to another type. The size of CST must be
1977 smaller or equal to the size of TYPE. Both types must be floating point.</dd>
1978
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]1979 <dt><b><tt>fptoui ( CST to TYPE )</tt></b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1980 <dd>Convert a floating point constant to the corresponding unsigned integer
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1981 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1982 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1983 of the same number of elements. If the value won't fit in the integer type,
1984 the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1985
1986 <dt><b><tt>fptosi ( CST to TYPE )</tt></b></dt>
1987 <dd>Convert a floating point constant to the corresponding signed integer
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1988 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1989 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1990 of the same number of elements. If the value won't fit in the integer type,
1991 the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1992
1993 <dt><b><tt>uitofp ( CST to TYPE )</tt></b></dt>
1994 <dd>Convert an unsigned integer constant to the corresponding floating point
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1995 constant. TYPE must be a scalar or vector floating point type. CST must be of
1996 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1997 of the same number of elements. If the value won't fit in the floating point
1998 type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1999
2000 <dt><b><tt>sitofp ( CST to TYPE )</tt></b></dt>
2001 <dd>Convert a signed integer constant to the corresponding floating point
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]2002 constant. TYPE must be a scalar or vector floating point type. CST must be of
2003 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
2004 of the same number of elements. If the value won't fit in the floating point
2005 type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2006
2007 <dt><b><tt>ptrtoint ( CST to TYPE )</tt></b></dt>
2008 <dd>Convert a pointer typed constant to the corresponding integer constant
2009 TYPE must be an integer type. CST must be of pointer type. The CST value is
2010 zero extended, truncated, or unchanged to make it fit in TYPE.</dd>
2011
2012 <dt><b><tt>inttoptr ( CST to TYPE )</tt></b></dt>
2013 <dd>Convert a integer constant to a pointer constant. TYPE must be a
2014 pointer type. CST must be of integer type. The CST value is zero extended,
2015 truncated, or unchanged to make it fit in a pointer size. This one is
2016 <i>really</i> dangerous!</dd>
2017
2018 <dt><b><tt>bitcast ( CST to TYPE )</tt></b></dt>
Chris Lattner557bc5d2009-02-28 18:27:03 +0000[diff] [blame]2019 <dd>Convert a constant, CST, to another TYPE. The constraints of the operands
2020 are the same as those for the <a href="#i_bitcast">bitcast
2021 instruction</a>.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2022
2023 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
2024
2025 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
2026 constants. As with the <a href="#i_getelementptr">getelementptr</a>
2027 instruction, the index list may have zero or more indexes, which are required
2028 to make sense for the type of "CSTPTR".</dd>
2029
2030 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
2031
2032 <dd>Perform the <a href="#i_select">select operation</a> on
2033 constants.</dd>
2034
2035 <dt><b><tt>icmp COND ( VAL1, VAL2 )</tt></b></dt>
2036 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
2037
2038 <dt><b><tt>fcmp COND ( VAL1, VAL2 )</tt></b></dt>
2039 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
2040
2041 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
2042
2043 <dd>Perform the <a href="#i_extractelement">extractelement
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]2044 operation</a> on constants.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2045
2046 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
2047
2048 <dd>Perform the <a href="#i_insertelement">insertelement
2049 operation</a> on constants.</dd>
2050
2051
2052 <dt><b><tt>shufflevector ( VEC1, VEC2, IDXMASK )</tt></b></dt>
2053
2054 <dd>Perform the <a href="#i_shufflevector">shufflevector
2055 operation</a> on constants.</dd>
2056
2057 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
2058
2059 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
2060 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
2061 binary</a> operations. The constraints on operands are the same as those for
2062 the corresponding instruction (e.g. no bitwise operations on floating point
2063 values are allowed).</dd>
2064</dl>
2065</div>
2066
Nick Lewycky4dcf8102009-04-04 07:22:01 +0000[diff] [blame]2067<!-- ======================================================================= -->
2068<div class="doc_subsection"><a name="metadata">Embedded Metadata</a>
2069</div>
2070
2071<div class="doc_text">
2072
2073<p>Embedded metadata provides a way to attach arbitrary data to the
2074instruction stream without affecting the behaviour of the program. There are
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]2075two metadata primitives, strings and nodes. All metadata has the
2076<tt>metadata</tt> type and is identified in syntax by a preceding exclamation
2077point ('<tt>!</tt>').
Nick Lewycky4dcf8102009-04-04 07:22:01 +0000[diff] [blame]2078</p>
2079
2080<p>A metadata string is a string surrounded by double quotes. It can contain
2081any character by escaping non-printable characters with "\xx" where "xx" is
2082the two digit hex code. For example: "<tt>!"test\00"</tt>".
2083</p>
2084
2085<p>Metadata nodes are represented with notation similar to structure constants
2086(a comma separated list of elements, surrounded by braces and preceeded by an
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]2087exclamation point). For example: "<tt>!{ metadata !"test\00", i32 10}</tt>".
Nick Lewycky4dcf8102009-04-04 07:22:01 +0000[diff] [blame]2088</p>
2089
Nick Lewycky117f4382009-05-10 20:57:05 +0000[diff] [blame]2090<p>A metadata node will attempt to track changes to the values it holds. In
2091the event that a value is deleted, it will be replaced with a typeless
Nick Lewycky29aaef82009-05-30 05:06:04 +0000[diff] [blame]2092"<tt>null</tt>", such as "<tt>metadata !{null, i32 10}</tt>".</p>
Nick Lewycky117f4382009-05-10 20:57:05 +0000[diff] [blame]2093
Nick Lewycky4dcf8102009-04-04 07:22:01 +0000[diff] [blame]2094<p>Optimizations may rely on metadata to provide additional information about
2095the program that isn't available in the instructions, or that isn't easily
2096computable. Similarly, the code generator may expect a certain metadata format
2097to be used to express debugging information.</p>
2098</div>
2099
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2100<!-- *********************************************************************** -->
2101<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
2102<!-- *********************************************************************** -->
2103
2104<!-- ======================================================================= -->
2105<div class="doc_subsection">
2106<a name="inlineasm">Inline Assembler Expressions</a>
2107</div>
2108
2109<div class="doc_text">
2110
2111<p>
2112LLVM supports inline assembler expressions (as opposed to <a href="#moduleasm">
2113Module-Level Inline Assembly</a>) through the use of a special value. This
2114value represents the inline assembler as a string (containing the instructions
2115to emit), a list of operand constraints (stored as a string), and a flag that
2116indicates whether or not the inline asm expression has side effects. An example
2117inline assembler expression is:
2118</p>
2119
2120<div class="doc_code">
2121<pre>
2122i32 (i32) asm "bswap $0", "=r,r"
2123</pre>
2124</div>
2125
2126<p>
2127Inline assembler expressions may <b>only</b> be used as the callee operand of
2128a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we have:
2129</p>
2130
2131<div class="doc_code">
2132<pre>
2133%X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
2134</pre>
2135</div>
2136
2137<p>
2138Inline asms with side effects not visible in the constraint list must be marked
2139as having side effects. This is done through the use of the
2140'<tt>sideeffect</tt>' keyword, like so:
2141</p>
2142
2143<div class="doc_code">
2144<pre>
2145call void asm sideeffect "eieio", ""()
2146</pre>
2147</div>
2148
2149<p>TODO: The format of the asm and constraints string still need to be
2150documented here. Constraints on what can be done (e.g. duplication, moving, etc
Chris Lattner1cdd64e2008-10-04 18:36:02 +0000[diff] [blame]2151need to be documented). This is probably best done by reference to another
2152document that covers inline asm from a holistic perspective.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2153</p>
2154
2155</div>
2156
2157<!-- *********************************************************************** -->
2158<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
2159<!-- *********************************************************************** -->
2160
2161<div class="doc_text">
2162
2163<p>The LLVM instruction set consists of several different
2164classifications of instructions: <a href="#terminators">terminator
2165instructions</a>, <a href="#binaryops">binary instructions</a>,
2166<a href="#bitwiseops">bitwise binary instructions</a>, <a
2167 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
2168instructions</a>.</p>
2169
2170</div>
2171
2172<!-- ======================================================================= -->
2173<div class="doc_subsection"> <a name="terminators">Terminator
2174Instructions</a> </div>
2175
2176<div class="doc_text">
2177
2178<p>As mentioned <a href="#functionstructure">previously</a>, every
2179basic block in a program ends with a "Terminator" instruction, which
2180indicates which block should be executed after the current block is
2181finished. These terminator instructions typically yield a '<tt>void</tt>'
2182value: they produce control flow, not values (the one exception being
2183the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
2184<p>There are six different terminator instructions: the '<a
2185 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
2186instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
2187the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
2188 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
2189 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
2190
2191</div>
2192
2193<!-- _______________________________________________________________________ -->
2194<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
2195Instruction</a> </div>
2196<div class="doc_text">
2197<h5>Syntax:</h5>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2198<pre>
2199 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]2200 ret void <i>; Return from void function</i>
2201</pre>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2202
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2203<h5>Overview:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2204
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2205<p>The '<tt>ret</tt>' instruction is used to return control flow (and
2206optionally a value) from a function back to the caller.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2207<p>There are two forms of the '<tt>ret</tt>' instruction: one that
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2208returns a value and then causes control flow, and one that just causes
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2209control flow to occur.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2210
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2211<h5>Arguments:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2212
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2213<p>The '<tt>ret</tt>' instruction optionally accepts a single argument,
2214the return value. The type of the return value must be a
2215'<a href="#t_firstclass">first class</a>' type.</p>
2216
2217<p>A function is not <a href="#wellformed">well formed</a> if
2218it it has a non-void return type and contains a '<tt>ret</tt>'
2219instruction with no return value or a return value with a type that
2220does not match its type, or if it has a void return type and contains
2221a '<tt>ret</tt>' instruction with a return value.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2222
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2223<h5>Semantics:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2224
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2225<p>When the '<tt>ret</tt>' instruction is executed, control flow
2226returns back to the calling function's context. If the caller is a "<a
2227 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
2228the instruction after the call. If the caller was an "<a
2229 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
2230at the beginning of the "normal" destination block. If the instruction
2231returns a value, that value shall set the call or invoke instruction's
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]2232return value.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2233
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2234<h5>Example:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2235
2236<pre>
2237 ret i32 5 <i>; Return an integer value of 5</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2238 ret void <i>; Return from a void function</i>
Bill Wendlingd163e2d2009-02-28 22:12:54 +0000[diff] [blame]2239 ret { i32, i8 } { i32 4, i8 2 } <i>; Return a struct of values 4 and 2</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2240</pre>
Dan Gohman60967192009-01-12 23:12:39 +0000[diff] [blame]2241
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]2242<p>Note that the code generator does not yet fully support large
2243 return values. The specific sizes that are currently supported are
2244 dependent on the target. For integers, on 32-bit targets the limit
2245 is often 64 bits, and on 64-bit targets the limit is often 128 bits.
2246 For aggregate types, the current limits are dependent on the element
2247 types; for example targets are often limited to 2 total integer
2248 elements and 2 total floating-point elements.</p>
Dan Gohman60967192009-01-12 23:12:39 +0000[diff] [blame]2249
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2250</div>
2251<!-- _______________________________________________________________________ -->
2252<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
2253<div class="doc_text">
2254<h5>Syntax:</h5>
2255<pre> br i1 &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;<br> br label &lt;dest&gt; <i>; Unconditional branch</i>
2256</pre>
2257<h5>Overview:</h5>
2258<p>The '<tt>br</tt>' instruction is used to cause control flow to
2259transfer to a different basic block in the current function. There are
2260two forms of this instruction, corresponding to a conditional branch
2261and an unconditional branch.</p>
2262<h5>Arguments:</h5>
2263<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
2264single '<tt>i1</tt>' value and two '<tt>label</tt>' values. The
2265unconditional form of the '<tt>br</tt>' instruction takes a single
2266'<tt>label</tt>' value as a target.</p>
2267<h5>Semantics:</h5>
2268<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
2269argument is evaluated. If the value is <tt>true</tt>, control flows
2270to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
2271control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
2272<h5>Example:</h5>
Chris Lattner95127832009-05-09 18:11:50 +0000[diff] [blame]2273<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
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2274 href="#i_ret">ret</a> i32 1<br>IfUnequal:<br> <a href="#i_ret">ret</a> i32 0<br></pre>
2275</div>
2276<!-- _______________________________________________________________________ -->
2277<div class="doc_subsubsection">
2278 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
2279</div>
2280
2281<div class="doc_text">
2282<h5>Syntax:</h5>
2283
2284<pre>
2285 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
2286</pre>
2287
2288<h5>Overview:</h5>
2289
2290<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
2291several different places. It is a generalization of the '<tt>br</tt>'
2292instruction, allowing a branch to occur to one of many possible
2293destinations.</p>
2294
2295
2296<h5>Arguments:</h5>
2297
2298<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
2299comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
2300an array of pairs of comparison value constants and '<tt>label</tt>'s. The
2301table is not allowed to contain duplicate constant entries.</p>
2302
2303<h5>Semantics:</h5>
2304
2305<p>The <tt>switch</tt> instruction specifies a table of values and
2306destinations. When the '<tt>switch</tt>' instruction is executed, this
2307table is searched for the given value. If the value is found, control flow is
2308transfered to the corresponding destination; otherwise, control flow is
2309transfered to the default destination.</p>
2310
2311<h5>Implementation:</h5>
2312
2313<p>Depending on properties of the target machine and the particular
2314<tt>switch</tt> instruction, this instruction may be code generated in different
2315ways. For example, it could be generated as a series of chained conditional
2316branches or with a lookup table.</p>
2317
2318<h5>Example:</h5>
2319
2320<pre>
2321 <i>; Emulate a conditional br instruction</i>
2322 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]2323 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2324
2325 <i>; Emulate an unconditional br instruction</i>
2326 switch i32 0, label %dest [ ]
2327
2328 <i>; Implement a jump table:</i>
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]2329 switch i32 %val, label %otherwise [ i32 0, label %onzero
2330 i32 1, label %onone
2331 i32 2, label %ontwo ]
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2332</pre>
2333</div>
2334
2335<!-- _______________________________________________________________________ -->
2336<div class="doc_subsubsection">
2337 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
2338</div>
2339
2340<div class="doc_text">
2341
2342<h5>Syntax:</h5>
2343
2344<pre>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]2345 &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]2346 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
2347</pre>
2348
2349<h5>Overview:</h5>
2350
2351<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
2352function, with the possibility of control flow transfer to either the
2353'<tt>normal</tt>' label or the
2354'<tt>exception</tt>' label. If the callee function returns with the
2355"<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
2356"normal" label. If the callee (or any indirect callees) returns with the "<a
2357href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]2358continued at the dynamically nearest "exception" label.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2359
2360<h5>Arguments:</h5>
2361
2362<p>This instruction requires several arguments:</p>
2363
2364<ol>
2365 <li>
2366 The optional "cconv" marker indicates which <a href="#callingconv">calling
2367 convention</a> the call should use. If none is specified, the call defaults
2368 to using C calling conventions.
2369 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]2370
2371 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
2372 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>',
2373 and '<tt>inreg</tt>' attributes are valid here.</li>
2374
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2375 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
2376 function value being invoked. In most cases, this is a direct function
2377 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
2378 an arbitrary pointer to function value.
2379 </li>
2380
2381 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
2382 function to be invoked. </li>
2383
2384 <li>'<tt>function args</tt>': argument list whose types match the function
2385 signature argument types. If the function signature indicates the function
2386 accepts a variable number of arguments, the extra arguments can be
2387 specified. </li>
2388
2389 <li>'<tt>normal label</tt>': the label reached when the called function
2390 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
2391
2392 <li>'<tt>exception label</tt>': the label reached when a callee returns with
2393 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
2394
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]2395 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]2396 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
2397 '<tt>readnone</tt>' attributes are valid here.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2398</ol>
2399
2400<h5>Semantics:</h5>
2401
2402<p>This instruction is designed to operate as a standard '<tt><a
2403href="#i_call">call</a></tt>' instruction in most regards. The primary
2404difference is that it establishes an association with a label, which is used by
2405the runtime library to unwind the stack.</p>
2406
2407<p>This instruction is used in languages with destructors to ensure that proper
2408cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
2409exception. Additionally, this is important for implementation of
2410'<tt>catch</tt>' clauses in high-level languages that support them.</p>
2411
Jay Foad8e2fd2c2009-06-03 10:20:10 +0000[diff] [blame]2412<p>For the purposes of the SSA form, the definition of the value
2413returned by the '<tt>invoke</tt>' instruction is deemed to occur on
2414the edge from the current block to the "normal" label. If the callee
2415unwinds then no return value is available.</p>
Dan Gohman140ba5d2009-05-22 21:47:08 +0000[diff] [blame]2416
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2417<h5>Example:</h5>
2418<pre>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2419 %retval = invoke i32 @Test(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2420 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2421 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2422 unwind label %TestCleanup <i>; {i32}:retval set</i>
2423</pre>
2424</div>
2425
2426
2427<!-- _______________________________________________________________________ -->
2428
2429<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
2430Instruction</a> </div>
2431
2432<div class="doc_text">
2433
2434<h5>Syntax:</h5>
2435<pre>
2436 unwind
2437</pre>
2438
2439<h5>Overview:</h5>
2440
2441<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
2442at the first callee in the dynamic call stack which used an <a
2443href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
2444primarily used to implement exception handling.</p>
2445
2446<h5>Semantics:</h5>
2447
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]2448<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2449immediately halt. The dynamic call stack is then searched for the first <a
2450href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
2451execution continues at the "exceptional" destination block specified by the
2452<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
2453dynamic call chain, undefined behavior results.</p>
2454</div>
2455
2456<!-- _______________________________________________________________________ -->
2457
2458<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
2459Instruction</a> </div>
2460
2461<div class="doc_text">
2462
2463<h5>Syntax:</h5>
2464<pre>
2465 unreachable
2466</pre>
2467
2468<h5>Overview:</h5>
2469
2470<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
2471instruction is used to inform the optimizer that a particular portion of the
2472code is not reachable. This can be used to indicate that the code after a
2473no-return function cannot be reached, and other facts.</p>
2474
2475<h5>Semantics:</h5>
2476
2477<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
2478</div>
2479
2480
2481
2482<!-- ======================================================================= -->
2483<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
2484<div class="doc_text">
2485<p>Binary operators are used to do most of the computation in a
Chris Lattnerab596d92008-04-01 18:47:32 +0000[diff] [blame]2486program. They require two operands of the same type, execute an operation on them, and
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2487produce a single value. The operands might represent
2488multiple 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]2489The result value has the same type as its operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2490<p>There are several different binary operators:</p>
2491</div>
2492<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2493<div class="doc_subsubsection">
2494 <a name="i_add">'<tt>add</tt>' Instruction</a>
2495</div>
2496
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2497<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2498
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2499<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2500
2501<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2502 &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]2503</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2504
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2505<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2506
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2507<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2508
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2509<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2510
2511<p>The two arguments to the '<tt>add</tt>' instruction must be <a
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2512 href="#t_integer">integer</a> or
2513 <a href="#t_vector">vector</a> of integer values. Both arguments must
2514 have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2515
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2516<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2517
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2518<p>The value produced is the integer sum of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2519
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2520<p>If the sum has unsigned overflow, the result returned is the
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2521mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2522the result.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2523
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2524<p>Because LLVM integers use a two's complement representation, this
2525instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2526
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2527<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2528
2529<pre>
2530 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2531</pre>
2532</div>
2533<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2534<div class="doc_subsubsection">
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2535 <a name="i_fadd">'<tt>fadd</tt>' Instruction</a>
2536</div>
2537
2538<div class="doc_text">
2539
2540<h5>Syntax:</h5>
2541
2542<pre>
2543 &lt;result&gt; = fadd &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
2544</pre>
2545
2546<h5>Overview:</h5>
2547
2548<p>The '<tt>fadd</tt>' instruction returns the sum of its two operands.</p>
2549
2550<h5>Arguments:</h5>
2551
2552<p>The two arguments to the '<tt>fadd</tt>' instruction must be
2553<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
2554floating point values. Both arguments must have identical types.</p>
2555
2556<h5>Semantics:</h5>
2557
2558<p>The value produced is the floating point sum of the two operands.</p>
2559
2560<h5>Example:</h5>
2561
2562<pre>
2563 &lt;result&gt; = fadd float 4.0, %var <i>; yields {float}:result = 4.0 + %var</i>
2564</pre>
2565</div>
2566<!-- _______________________________________________________________________ -->
2567<div class="doc_subsubsection">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2568 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
2569</div>
2570
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2571<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2572
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2573<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2574
2575<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2576 &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]2577</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2578
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2579<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2580
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2581<p>The '<tt>sub</tt>' instruction returns the difference of its two
2582operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2583
2584<p>Note that the '<tt>sub</tt>' instruction is used to represent the
2585'<tt>neg</tt>' instruction present in most other intermediate
2586representations.</p>
2587
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2588<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2589
2590<p>The two arguments to the '<tt>sub</tt>' instruction must be <a
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2591 href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
2592 integer values. Both arguments must have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2593
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2594<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2595
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2596<p>The value produced is the integer difference of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2597
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2598<p>If the difference has unsigned overflow, the result returned is the
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2599mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2600the result.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2601
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2602<p>Because LLVM integers use a two's complement representation, this
2603instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2604
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2605<h5>Example:</h5>
2606<pre>
2607 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
2608 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
2609</pre>
2610</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2611
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2612<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2613<div class="doc_subsubsection">
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2614 <a name="i_fsub">'<tt>fsub</tt>' Instruction</a>
2615</div>
2616
2617<div class="doc_text">
2618
2619<h5>Syntax:</h5>
2620
2621<pre>
2622 &lt;result&gt; = fsub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
2623</pre>
2624
2625<h5>Overview:</h5>
2626
2627<p>The '<tt>fsub</tt>' instruction returns the difference of its two
2628operands.</p>
2629
2630<p>Note that the '<tt>fsub</tt>' instruction is used to represent the
2631'<tt>fneg</tt>' instruction present in most other intermediate
2632representations.</p>
2633
2634<h5>Arguments:</h5>
2635
2636<p>The two arguments to the '<tt>fsub</tt>' instruction must be <a
2637 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2638 of floating point values. Both arguments must have identical types.</p>
2639
2640<h5>Semantics:</h5>
2641
2642<p>The value produced is the floating point difference of the two operands.</p>
2643
2644<h5>Example:</h5>
2645<pre>
2646 &lt;result&gt; = fsub float 4.0, %var <i>; yields {float}:result = 4.0 - %var</i>
2647 &lt;result&gt; = fsub float -0.0, %val <i>; yields {float}:result = -%var</i>
2648</pre>
2649</div>
2650
2651<!-- _______________________________________________________________________ -->
2652<div class="doc_subsubsection">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2653 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
2654</div>
2655
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2656<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2657
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2658<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2659<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]2660</pre>
2661<h5>Overview:</h5>
2662<p>The '<tt>mul</tt>' instruction returns the product of its two
2663operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2664
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2665<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2666
2667<p>The two arguments to the '<tt>mul</tt>' instruction must be <a
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2668href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2669values. Both arguments must have identical types.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2670
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2671<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2672
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2673<p>The value produced is the integer product of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2674
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2675<p>If the result of the multiplication has unsigned overflow,
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2676the result returned is the mathematical result modulo
26772<sup>n</sup>, where n is the bit width of the result.</p>
2678<p>Because LLVM integers use a two's complement representation, and the
2679result is the same width as the operands, this instruction returns the
2680correct result for both signed and unsigned integers. If a full product
2681(e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands
2682should be sign-extended or zero-extended as appropriate to the
2683width of the full product.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2684<h5>Example:</h5>
2685<pre> &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
2686</pre>
2687</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2688
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2689<!-- _______________________________________________________________________ -->
Dan Gohman7ce405e2009-06-04 22:49:04 +0000[diff] [blame]2690<div class="doc_subsubsection">
2691 <a name="i_fmul">'<tt>fmul</tt>' Instruction</a>
2692</div>
2693
2694<div class="doc_text">
2695
2696<h5>Syntax:</h5>
2697<pre> &lt;result&gt; = fmul &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
2698</pre>
2699<h5>Overview:</h5>
2700<p>The '<tt>fmul</tt>' instruction returns the product of its two
2701operands.</p>
2702
2703<h5>Arguments:</h5>
2704
2705<p>The two arguments to the '<tt>fmul</tt>' instruction must be
2706<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2707of floating point values. Both arguments must have identical types.</p>
2708
2709<h5>Semantics:</h5>
2710
2711<p>The value produced is the floating point product of the two operands.</p>
2712
2713<h5>Example:</h5>
2714<pre> &lt;result&gt; = fmul float 4.0, %var <i>; yields {float}:result = 4.0 * %var</i>
2715</pre>
2716</div>
2717
2718<!-- _______________________________________________________________________ -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2719<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
2720</a></div>
2721<div class="doc_text">
2722<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2723<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]2724</pre>
2725<h5>Overview:</h5>
2726<p>The '<tt>udiv</tt>' instruction returns the quotient of its two
2727operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2728
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2729<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2730
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2731<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2732<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2733values. Both arguments must have identical types.</p>
2734
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2735<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2736
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2737<p>The value produced is the unsigned integer quotient of the two operands.</p>
2738<p>Note that unsigned integer division and signed integer division are distinct
2739operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
2740<p>Division by zero leads to undefined behavior.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2741<h5>Example:</h5>
2742<pre> &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2743</pre>
2744</div>
2745<!-- _______________________________________________________________________ -->
2746<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
2747</a> </div>
2748<div class="doc_text">
2749<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2750<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2751 &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]2752</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2753
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2754<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2755
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2756<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two
2757operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2758
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2759<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2760
2761<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
2762<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2763values. Both arguments must have identical types.</p>
2764
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2765<h5>Semantics:</h5>
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2766<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]2767<p>Note that signed integer division and unsigned integer division are distinct
2768operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
2769<p>Division by zero leads to undefined behavior. Overflow also leads to
2770undefined behavior; this is a rare case, but can occur, for example,
2771by doing a 32-bit division of -2147483648 by -1.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2772<h5>Example:</h5>
2773<pre> &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2774</pre>
2775</div>
2776<!-- _______________________________________________________________________ -->
2777<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
2778Instruction</a> </div>
2779<div class="doc_text">
2780<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2781<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2782 &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]2783</pre>
2784<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2785
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2786<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two
2787operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2788
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2789<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2790
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2791<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2792<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2793of floating point values. Both arguments must have identical types.</p>
2794
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2795<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2796
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2797<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2798
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2799<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2800
2801<pre>
2802 &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]2803</pre>
2804</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2805
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2806<!-- _______________________________________________________________________ -->
2807<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
2808</div>
2809<div class="doc_text">
2810<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2811<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]2812</pre>
2813<h5>Overview:</h5>
2814<p>The '<tt>urem</tt>' instruction returns the remainder from the
2815unsigned division of its two arguments.</p>
2816<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2817<p>The two arguments to the '<tt>urem</tt>' instruction must be
2818<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2819values. Both arguments must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2820<h5>Semantics:</h5>
2821<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2822This instruction always performs an unsigned division to get the remainder.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2823<p>Note that unsigned integer remainder and signed integer remainder are
2824distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
2825<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2826<h5>Example:</h5>
2827<pre> &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2828</pre>
2829
2830</div>
2831<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2832<div class="doc_subsubsection">
2833 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
2834</div>
2835
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2836<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2837
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2838<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2839
2840<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2841 &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]2842</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2843
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2844<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2845
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2846<p>The '<tt>srem</tt>' instruction returns the remainder from the
Dan Gohman3e3fd8c2007-11-05 23:35:22 +0000[diff] [blame]2847signed division of its two operands. This instruction can also take
2848<a href="#t_vector">vector</a> versions of the values in which case
2849the elements must be integers.</p>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]2850
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2851<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2852
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2853<p>The two arguments to the '<tt>srem</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2854<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2855values. Both arguments must have identical types.</p>
2856
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2857<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2858
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2859<p>This instruction returns the <i>remainder</i> of a division (where the result
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2860has the same sign as the dividend, <tt>op1</tt>), not the <i>modulo</i>
2861operator (where the result has the same sign as the divisor, <tt>op2</tt>) of
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2862a value. For more information about the difference, see <a
2863 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
2864Math Forum</a>. For a table of how this is implemented in various languages,
2865please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
2866Wikipedia: modulo operation</a>.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2867<p>Note that signed integer remainder and unsigned integer remainder are
2868distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
2869<p>Taking the remainder of a division by zero leads to undefined behavior.
2870Overflow also leads to undefined behavior; this is a rare case, but can occur,
2871for example, by taking the remainder of a 32-bit division of -2147483648 by -1.
2872(The remainder doesn't actually overflow, but this rule lets srem be
2873implemented using instructions that return both the result of the division
2874and the remainder.)</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2875<h5>Example:</h5>
2876<pre> &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2877</pre>
2878
2879</div>
2880<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2881<div class="doc_subsubsection">
2882 <a name="i_frem">'<tt>frem</tt>' Instruction</a> </div>
2883
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2884<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2885
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2886<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2887<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]2888</pre>
2889<h5>Overview:</h5>
2890<p>The '<tt>frem</tt>' instruction returns the remainder from the
2891division of its two operands.</p>
2892<h5>Arguments:</h5>
2893<p>The two arguments to the '<tt>frem</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2894<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2895of floating point values. Both arguments must have identical types.</p>
2896
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2897<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2898
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2899<p>This instruction returns the <i>remainder</i> of a division.
2900The remainder has the same sign as the dividend.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2901
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2902<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2903
2904<pre>
2905 &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]2906</pre>
2907</div>
2908
2909<!-- ======================================================================= -->
2910<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
2911Operations</a> </div>
2912<div class="doc_text">
2913<p>Bitwise binary operators are used to do various forms of
2914bit-twiddling in a program. They are generally very efficient
2915instructions and can commonly be strength reduced from other
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2916instructions. They require two operands of the same type, execute an operation on them,
2917and 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]2918</div>
2919
2920<!-- _______________________________________________________________________ -->
2921<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
2922Instruction</a> </div>
2923<div class="doc_text">
2924<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2925<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]2926</pre>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2927
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2928<h5>Overview:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2929
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2930<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
2931the left a specified number of bits.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2932
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2933<h5>Arguments:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2934
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2935<p>Both arguments to the '<tt>shl</tt>' instruction must be the same <a
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]2936 href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2937type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2938
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2939<h5>Semantics:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2940
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2941<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod 2<sup>n</sup>,
2942where n is the width of the result. If <tt>op2</tt> is (statically or dynamically) negative or
Mon P Wangb0f51822008-12-10 08:55:09 +0000[diff] [blame]2943equal to or larger than the number of bits in <tt>op1</tt>, the result is undefined.
2944If the arguments are vectors, each vector element of <tt>op1</tt> is shifted by the
2945corresponding shift amount in <tt>op2</tt>.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2946
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2947<h5>Example:</h5><pre>
2948 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
2949 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
2950 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2951 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Mon P Wang9901e732008-12-09 05:46:39 +0000[diff] [blame]2952 &lt;result&gt; = shl &lt;2 x i32&gt; &lt; i32 1, i32 1&gt;, &lt; i32 1, i32 2&gt; <i>; yields: result=&lt;2 x i32&gt; &lt; i32 2, i32 4&gt;</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2953</pre>
2954</div>
2955<!-- _______________________________________________________________________ -->
2956<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
2957Instruction</a> </div>
2958<div class="doc_text">
2959<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2960<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]2961</pre>
2962
2963<h5>Overview:</h5>
2964<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
2965operand shifted to the right a specified number of bits with zero fill.</p>
2966
2967<h5>Arguments:</h5>
2968<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]2969<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]2970type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2971
2972<h5>Semantics:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2973
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2974<p>This instruction always performs a logical shift right operation. The most
2975significant bits of the result will be filled with zero bits after the
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2976shift. If <tt>op2</tt> is (statically or dynamically) equal to or larger than
Mon P Wangb0f51822008-12-10 08:55:09 +0000[diff] [blame]2977the number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
2978vectors, each vector element of <tt>op1</tt> is shifted by the corresponding shift
2979amount in <tt>op2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2980
2981<h5>Example:</h5>
2982<pre>
2983 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
2984 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
2985 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
2986 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2987 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Mon P Wang9901e732008-12-09 05:46:39 +0000[diff] [blame]2988 &lt;result&gt; = lshr &lt;2 x i32&gt; &lt; i32 -2, i32 4&gt;, &lt; i32 1, i32 2&gt; <i>; yields: result=&lt;2 x i32&gt; &lt; i32 0x7FFFFFFF, i32 1&gt;</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2989</pre>
2990</div>
2991
2992<!-- _______________________________________________________________________ -->
2993<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
2994Instruction</a> </div>
2995<div class="doc_text">
2996
2997<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2998<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]2999</pre>
3000
3001<h5>Overview:</h5>
3002<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
3003operand shifted to the right a specified number of bits with sign extension.</p>
3004
3005<h5>Arguments:</h5>
3006<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]3007<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]3008type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3009
3010<h5>Semantics:</h5>
3011<p>This instruction always performs an arithmetic shift right operation,
3012The most significant bits of the result will be filled with the sign bit
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3013of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
Mon P Wangb0f51822008-12-10 08:55:09 +0000[diff] [blame]3014larger than the number of bits in <tt>op1</tt>, the result is undefined. If the
3015arguments are vectors, each vector element of <tt>op1</tt> is shifted by the
3016corresponding shift amount in <tt>op2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3017
3018<h5>Example:</h5>
3019<pre>
3020 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
3021 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
3022 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
3023 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]3024 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Mon P Wang9901e732008-12-09 05:46:39 +0000[diff] [blame]3025 &lt;result&gt; = ashr &lt;2 x i32&gt; &lt; i32 -2, i32 4&gt;, &lt; i32 1, i32 3&gt; <i>; yields: result=&lt;2 x i32&gt; &lt; i32 -1, i32 0&gt;</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3026</pre>
3027</div>
3028
3029<!-- _______________________________________________________________________ -->
3030<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
3031Instruction</a> </div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3032
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3033<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3034
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3035<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3036
3037<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3038 &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]3039</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3040
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3041<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3042
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3043<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
3044its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3045
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3046<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3047
3048<p>The two arguments to the '<tt>and</tt>' instruction must be
3049<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3050values. Both arguments must have identical types.</p>
3051
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3052<h5>Semantics:</h5>
3053<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
3054<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]3055<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3056<table border="1" cellspacing="0" cellpadding="4">
3057 <tbody>
3058 <tr>
3059 <td>In0</td>
3060 <td>In1</td>
3061 <td>Out</td>
3062 </tr>
3063 <tr>
3064 <td>0</td>
3065 <td>0</td>
3066 <td>0</td>
3067 </tr>
3068 <tr>
3069 <td>0</td>
3070 <td>1</td>
3071 <td>0</td>
3072 </tr>
3073 <tr>
3074 <td>1</td>
3075 <td>0</td>
3076 <td>0</td>
3077 </tr>
3078 <tr>
3079 <td>1</td>
3080 <td>1</td>
3081 <td>1</td>
3082 </tr>
3083 </tbody>
3084</table>
3085</div>
3086<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3087<pre>
3088 &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]3089 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
3090 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
3091</pre>
3092</div>
3093<!-- _______________________________________________________________________ -->
3094<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
3095<div class="doc_text">
3096<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3097<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]3098</pre>
3099<h5>Overview:</h5>
3100<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
3101or of its two operands.</p>
3102<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3103
3104<p>The two arguments to the '<tt>or</tt>' instruction must be
3105<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3106values. Both arguments must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3107<h5>Semantics:</h5>
3108<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
3109<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]3110<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3111<table border="1" cellspacing="0" cellpadding="4">
3112 <tbody>
3113 <tr>
3114 <td>In0</td>
3115 <td>In1</td>
3116 <td>Out</td>
3117 </tr>
3118 <tr>
3119 <td>0</td>
3120 <td>0</td>
3121 <td>0</td>
3122 </tr>
3123 <tr>
3124 <td>0</td>
3125 <td>1</td>
3126 <td>1</td>
3127 </tr>
3128 <tr>
3129 <td>1</td>
3130 <td>0</td>
3131 <td>1</td>
3132 </tr>
3133 <tr>
3134 <td>1</td>
3135 <td>1</td>
3136 <td>1</td>
3137 </tr>
3138 </tbody>
3139</table>
3140</div>
3141<h5>Example:</h5>
3142<pre> &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
3143 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
3144 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
3145</pre>
3146</div>
3147<!-- _______________________________________________________________________ -->
3148<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
3149Instruction</a> </div>
3150<div class="doc_text">
3151<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]3152<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]3153</pre>
3154<h5>Overview:</h5>
3155<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
3156or of its two operands. The <tt>xor</tt> is used to implement the
3157"one's complement" operation, which is the "~" operator in C.</p>
3158<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3159<p>The two arguments to the '<tt>xor</tt>' instruction must be
3160<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3161values. Both arguments must have identical types.</p>
3162
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3163<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3164
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3165<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
3166<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]3167<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3168<table border="1" cellspacing="0" cellpadding="4">
3169 <tbody>
3170 <tr>
3171 <td>In0</td>
3172 <td>In1</td>
3173 <td>Out</td>
3174 </tr>
3175 <tr>
3176 <td>0</td>
3177 <td>0</td>
3178 <td>0</td>
3179 </tr>
3180 <tr>
3181 <td>0</td>
3182 <td>1</td>
3183 <td>1</td>
3184 </tr>
3185 <tr>
3186 <td>1</td>
3187 <td>0</td>
3188 <td>1</td>
3189 </tr>
3190 <tr>
3191 <td>1</td>
3192 <td>1</td>
3193 <td>0</td>
3194 </tr>
3195 </tbody>
3196</table>
3197</div>
3198<p> </p>
3199<h5>Example:</h5>
3200<pre> &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
3201 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
3202 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
3203 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
3204</pre>
3205</div>
3206
3207<!-- ======================================================================= -->
3208<div class="doc_subsection">
3209 <a name="vectorops">Vector Operations</a>
3210</div>
3211
3212<div class="doc_text">
3213
3214<p>LLVM supports several instructions to represent vector operations in a
3215target-independent manner. These instructions cover the element-access and
3216vector-specific operations needed to process vectors effectively. While LLVM
3217does directly support these vector operations, many sophisticated algorithms
3218will want to use target-specific intrinsics to take full advantage of a specific
3219target.</p>
3220
3221</div>
3222
3223<!-- _______________________________________________________________________ -->
3224<div class="doc_subsubsection">
3225 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
3226</div>
3227
3228<div class="doc_text">
3229
3230<h5>Syntax:</h5>
3231
3232<pre>
3233 &lt;result&gt; = extractelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, i32 &lt;idx&gt; <i>; yields &lt;ty&gt;</i>
3234</pre>
3235
3236<h5>Overview:</h5>
3237
3238<p>
3239The '<tt>extractelement</tt>' instruction extracts a single scalar
3240element from a vector at a specified index.
3241</p>
3242
3243
3244<h5>Arguments:</h5>
3245
3246<p>
3247The first operand of an '<tt>extractelement</tt>' instruction is a
3248value of <a href="#t_vector">vector</a> type. The second operand is
3249an index indicating the position from which to extract the element.
3250The index may be a variable.</p>
3251
3252<h5>Semantics:</h5>
3253
3254<p>
3255The result is a scalar of the same type as the element type of
3256<tt>val</tt>. Its value is the value at position <tt>idx</tt> of
3257<tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
3258results are undefined.
3259</p>
3260
3261<h5>Example:</h5>
3262
3263<pre>
3264 %result = extractelement &lt;4 x i32&gt; %vec, i32 0 <i>; yields i32</i>
3265</pre>
3266</div>
3267
3268
3269<!-- _______________________________________________________________________ -->
3270<div class="doc_subsubsection">
3271 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
3272</div>
3273
3274<div class="doc_text">
3275
3276<h5>Syntax:</h5>
3277
3278<pre>
Dan Gohmanbcc3c502008-05-12 23:38:42 +0000[diff] [blame]3279 &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]3280</pre>
3281
3282<h5>Overview:</h5>
3283
3284<p>
3285The '<tt>insertelement</tt>' instruction inserts a scalar
3286element into a vector at a specified index.
3287</p>
3288
3289
3290<h5>Arguments:</h5>
3291
3292<p>
3293The first operand of an '<tt>insertelement</tt>' instruction is a
3294value of <a href="#t_vector">vector</a> type. The second operand is a
3295scalar value whose type must equal the element type of the first
3296operand. The third operand is an index indicating the position at
3297which to insert the value. The index may be a variable.</p>
3298
3299<h5>Semantics:</h5>
3300
3301<p>
3302The result is a vector of the same type as <tt>val</tt>. Its
3303element values are those of <tt>val</tt> except at position
3304<tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
3305exceeds the length of <tt>val</tt>, the results are undefined.
3306</p>
3307
3308<h5>Example:</h5>
3309
3310<pre>
3311 %result = insertelement &lt;4 x i32&gt; %vec, i32 1, i32 0 <i>; yields &lt;4 x i32&gt;</i>
3312</pre>
3313</div>
3314
3315<!-- _______________________________________________________________________ -->
3316<div class="doc_subsubsection">
3317 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
3318</div>
3319
3320<div class="doc_text">
3321
3322<h5>Syntax:</h5>
3323
3324<pre>
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3325 &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]3326</pre>
3327
3328<h5>Overview:</h5>
3329
3330<p>
3331The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3332from two input vectors, returning a vector with the same element type as
3333the input and length that is the same as the shuffle mask.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3334</p>
3335
3336<h5>Arguments:</h5>
3337
3338<p>
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3339The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
3340with types that match each other. The third argument is a shuffle mask whose
3341element type is always 'i32'. The result of the instruction is a vector whose
3342length is the same as the shuffle mask and whose element type is the same as
3343the element type of the first two operands.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3344</p>
3345
3346<p>
3347The shuffle mask operand is required to be a constant vector with either
3348constant integer or undef values.
3349</p>
3350
3351<h5>Semantics:</h5>
3352
3353<p>
3354The elements of the two input vectors are numbered from left to right across
3355both of the vectors. The shuffle mask operand specifies, for each element of
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3356the result vector, which element of the two input vectors the result element
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3357gets. The element selector may be undef (meaning "don't care") and the second
3358operand may be undef if performing a shuffle from only one vector.
3359</p>
3360
3361<h5>Example:</h5>
3362
3363<pre>
3364 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
3365 &lt;4 x i32&gt; &lt;i32 0, i32 4, i32 1, i32 5&gt; <i>; yields &lt;4 x i32&gt;</i>
3366 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; undef,
3367 &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]3368 %result = shufflevector &lt;8 x i32&gt; %v1, &lt;8 x i32&gt; undef,
3369 &lt;4 x i32&gt; &lt;i32 0, i32 1, i32 2, i32 3&gt; <i>; yields &lt;4 x i32&gt;</i>
3370 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
3371 &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]3372</pre>
3373</div>
3374
3375
3376<!-- ======================================================================= -->
3377<div class="doc_subsection">
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3378 <a name="aggregateops">Aggregate Operations</a>
3379</div>
3380
3381<div class="doc_text">
3382
3383<p>LLVM supports several instructions for working with aggregate values.
3384</p>
3385
3386</div>
3387
3388<!-- _______________________________________________________________________ -->
3389<div class="doc_subsubsection">
3390 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
3391</div>
3392
3393<div class="doc_text">
3394
3395<h5>Syntax:</h5>
3396
3397<pre>
3398 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
3399</pre>
3400
3401<h5>Overview:</h5>
3402
3403<p>
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3404The '<tt>extractvalue</tt>' instruction extracts the value of a struct field
3405or array element from an aggregate value.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3406</p>
3407
3408
3409<h5>Arguments:</h5>
3410
3411<p>
3412The first operand of an '<tt>extractvalue</tt>' instruction is a
3413value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a>
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3414type. The operands are constant indices to specify which value to extract
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3415in a similar manner as indices in a
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3416'<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
3417</p>
3418
3419<h5>Semantics:</h5>
3420
3421<p>
3422The result is the value at the position in the aggregate specified by
3423the index operands.
3424</p>
3425
3426<h5>Example:</h5>
3427
3428<pre>
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3429 %result = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3430</pre>
3431</div>
3432
3433
3434<!-- _______________________________________________________________________ -->
3435<div class="doc_subsubsection">
3436 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
3437</div>
3438
3439<div class="doc_text">
3440
3441<h5>Syntax:</h5>
3442
3443<pre>
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3444 &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]3445</pre>
3446
3447<h5>Overview:</h5>
3448
3449<p>
3450The '<tt>insertvalue</tt>' instruction inserts a value
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3451into a struct field or array element in an aggregate.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3452</p>
3453
3454
3455<h5>Arguments:</h5>
3456
3457<p>
3458The first operand of an '<tt>insertvalue</tt>' instruction is a
3459value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a> type.
3460The second operand is a first-class value to insert.
Dan Gohman4f29e422008-05-23 21:53:15 +0000[diff] [blame]3461The following operands are constant indices
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3462indicating the position at which to insert the value in a similar manner as
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3463indices in a
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3464'<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
3465The value to insert must have the same type as the value identified
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3466by the indices.
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3467</p>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3468
3469<h5>Semantics:</h5>
3470
3471<p>
3472The result is an aggregate of the same type as <tt>val</tt>. Its
3473value is that of <tt>val</tt> except that the value at the position
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3474specified by the indices is that of <tt>elt</tt>.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3475</p>
3476
3477<h5>Example:</h5>
3478
3479<pre>
Dan Gohmanb1aab4e2008-06-23 15:26:37 +0000[diff] [blame]3480 %result = insertvalue {i32, float} %agg, i32 1, 0 <i>; yields {i32, float}</i>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3481</pre>
3482</div>
3483
3484
3485<!-- ======================================================================= -->
3486<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3487 <a name="memoryops">Memory Access and Addressing Operations</a>
3488</div>
3489
3490<div class="doc_text">
3491
3492<p>A key design point of an SSA-based representation is how it
3493represents memory. In LLVM, no memory locations are in SSA form, which
3494makes things very simple. This section describes how to read, write,
3495allocate, and free memory in LLVM.</p>
3496
3497</div>
3498
3499<!-- _______________________________________________________________________ -->
3500<div class="doc_subsubsection">
3501 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
3502</div>
3503
3504<div class="doc_text">
3505
3506<h5>Syntax:</h5>
3507
3508<pre>
3509 &lt;result&gt; = malloc &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
3510</pre>
3511
3512<h5>Overview:</h5>
3513
3514<p>The '<tt>malloc</tt>' instruction allocates memory from the system
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]3515heap and returns a pointer to it. The object is always allocated in the generic
3516address space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3517
3518<h5>Arguments:</h5>
3519
3520<p>The '<tt>malloc</tt>' instruction allocates
3521<tt>sizeof(&lt;type&gt;)*NumElements</tt>
3522bytes of memory from the operating system and returns a pointer of the
3523appropriate type to the program. If "NumElements" is specified, it is the
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3524number of elements allocated, otherwise "NumElements" is defaulted to be one.
Duncan Sandsb38ce6f2009-06-20 13:26:06 +0000[diff] [blame]3525If a constant alignment is specified, the value result of the allocation is
3526guaranteed to be aligned to at least that boundary. If not specified, or if
3527zero, the target can choose to align the allocation on any convenient boundary
3528compatible with the type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3529
3530<p>'<tt>type</tt>' must be a sized type.</p>
3531
3532<h5>Semantics:</h5>
3533
3534<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
Nick Lewyckyff384472008-11-24 03:41:24 +0000[diff] [blame]3535a pointer is returned. The result of a zero byte allocation is undefined. The
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]3536result is null if there is insufficient memory available.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3537
3538<h5>Example:</h5>
3539
3540<pre>
Dan Gohmanf54f50a2009-01-04 23:49:44 +0000[diff] [blame]3541 %array = malloc [4 x i8] <i>; yields {[%4 x i8]*}:array</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3542
3543 %size = <a href="#i_add">add</a> i32 2, 2 <i>; yields {i32}:size = i32 4</i>
3544 %array1 = malloc i8, i32 4 <i>; yields {i8*}:array1</i>
3545 %array2 = malloc [12 x i8], i32 %size <i>; yields {[12 x i8]*}:array2</i>
3546 %array3 = malloc i32, i32 4, align 1024 <i>; yields {i32*}:array3</i>
3547 %array4 = malloc i32, align 1024 <i>; yields {i32*}:array4</i>
3548</pre>
Dan Gohman60967192009-01-12 23:12:39 +0000[diff] [blame]3549
3550<p>Note that the code generator does not yet respect the
3551 alignment value.</p>
3552
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3553</div>
3554
3555<!-- _______________________________________________________________________ -->
3556<div class="doc_subsubsection">
3557 <a name="i_free">'<tt>free</tt>' Instruction</a>
3558</div>
3559
3560<div class="doc_text">
3561
3562<h5>Syntax:</h5>
3563
3564<pre>
Dan Gohmanf54f50a2009-01-04 23:49:44 +0000[diff] [blame]3565 free &lt;type&gt; &lt;value&gt; <i>; yields {void}</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3566</pre>
3567
3568<h5>Overview:</h5>
3569
3570<p>The '<tt>free</tt>' instruction returns memory back to the unused
3571memory heap to be reallocated in the future.</p>
3572
3573<h5>Arguments:</h5>
3574
3575<p>'<tt>value</tt>' shall be a pointer value that points to a value
3576that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
3577instruction.</p>
3578
3579<h5>Semantics:</h5>
3580
3581<p>Access to the memory pointed to by the pointer is no longer defined
Chris Lattner329d6532008-04-19 22:41:32 +0000[diff] [blame]3582after this instruction executes. If the pointer is null, the operation
3583is a noop.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3584
3585<h5>Example:</h5>
3586
3587<pre>
Dan Gohmanf54f50a2009-01-04 23:49:44 +0000[diff] [blame]3588 %array = <a href="#i_malloc">malloc</a> [4 x i8] <i>; yields {[4 x i8]*}:array</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3589 free [4 x i8]* %array
3590</pre>
3591</div>
3592
3593<!-- _______________________________________________________________________ -->
3594<div class="doc_subsubsection">
3595 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
3596</div>
3597
3598<div class="doc_text">
3599
3600<h5>Syntax:</h5>
3601
3602<pre>
3603 &lt;result&gt; = alloca &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
3604</pre>
3605
3606<h5>Overview:</h5>
3607
3608<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
3609currently executing function, to be automatically released when this function
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]3610returns to its caller. The object is always allocated in the generic address
3611space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3612
3613<h5>Arguments:</h5>
3614
3615<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
3616bytes of memory on the runtime stack, returning a pointer of the
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3617appropriate type to the program. If "NumElements" is specified, it is the
3618number of elements allocated, otherwise "NumElements" is defaulted to be one.
Duncan Sandsb38ce6f2009-06-20 13:26:06 +0000[diff] [blame]3619If a constant alignment is specified, the value result of the allocation is
3620guaranteed to be aligned to at least that boundary. If not specified, or if
3621zero, the target can choose to align the allocation on any convenient boundary
3622compatible with the type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3623
3624<p>'<tt>type</tt>' may be any sized type.</p>
3625
3626<h5>Semantics:</h5>
3627
Bill Wendling2a454572009-05-08 20:49:29 +0000[diff] [blame]3628<p>Memory is allocated; a pointer is returned. The operation is undefined if
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]3629there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3630memory is automatically released when the function returns. The '<tt>alloca</tt>'
3631instruction is commonly used to represent automatic variables that must
3632have an address available. When the function returns (either with the <tt><a
3633 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]3634instructions), the memory is reclaimed. Allocating zero bytes
3635is legal, but the result is undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3636
3637<h5>Example:</h5>
3638
3639<pre>
Dan Gohmanf54f50a2009-01-04 23:49:44 +0000[diff] [blame]3640 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
3641 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
3642 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
3643 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3644</pre>
3645</div>
3646
3647<!-- _______________________________________________________________________ -->
3648<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
3649Instruction</a> </div>
3650<div class="doc_text">
3651<h5>Syntax:</h5>
3652<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>
3653<h5>Overview:</h5>
3654<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
3655<h5>Arguments:</h5>
3656<p>The argument to the '<tt>load</tt>' instruction specifies the memory
3657address from which to load. The pointer must point to a <a
3658 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
3659marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
3660the number or order of execution of this <tt>load</tt> with other
3661volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
3662instructions. </p>
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3663<p>
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3664The optional constant "align" argument specifies the alignment of the operation
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3665(that is, the alignment of the memory address). A value of 0 or an
3666omitted "align" argument means that the operation has the preferential
3667alignment for the target. It is the responsibility of the code emitter
3668to ensure that the alignment information is correct. Overestimating
3669the alignment results in an undefined behavior. Underestimating the
3670alignment may produce less efficient code. An alignment of 1 is always
3671safe.
3672</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3673<h5>Semantics:</h5>
Duncan Sandsfe87fff2009-03-22 11:33:16 +0000[diff] [blame]3674<p>The location of memory pointed to is loaded. If the value being loaded
3675is of scalar type then the number of bytes read does not exceed the minimum
3676number of bytes needed to hold all bits of the type. For example, loading an
3677<tt>i24</tt> reads at most three bytes. When loading a value of a type like
3678<tt>i20</tt> with a size that is not an integral number of bytes, the result
3679is undefined if the value was not originally written using a store of the
3680same type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3681<h5>Examples:</h5>
3682<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
3683 <a
3684 href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
3685 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
3686</pre>
3687</div>
3688<!-- _______________________________________________________________________ -->
3689<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
3690Instruction</a> </div>
3691<div class="doc_text">
3692<h5>Syntax:</h5>
3693<pre> store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
3694 volatile store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
3695</pre>
3696<h5>Overview:</h5>
3697<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
3698<h5>Arguments:</h5>
3699<p>There are two arguments to the '<tt>store</tt>' instruction: a value
3700to 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]3701operand must be a pointer to the <a href="#t_firstclass">first class</a> type
3702of the '<tt>&lt;value&gt;</tt>'
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3703operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
3704optimizer is not allowed to modify the number or order of execution of
3705this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
3706 href="#i_store">store</a></tt> instructions.</p>
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3707<p>
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3708The optional constant "align" argument specifies the alignment of the operation
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3709(that is, the alignment of the memory address). A value of 0 or an
3710omitted "align" argument means that the operation has the preferential
3711alignment for the target. It is the responsibility of the code emitter
3712to ensure that the alignment information is correct. Overestimating
3713the alignment results in an undefined behavior. Underestimating the
3714alignment may produce less efficient code. An alignment of 1 is always
3715safe.
3716</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3717<h5>Semantics:</h5>
3718<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>'
Duncan Sandsfe87fff2009-03-22 11:33:16 +0000[diff] [blame]3719at the location specified by the '<tt>&lt;pointer&gt;</tt>' operand.
3720If '<tt>&lt;value&gt;</tt>' is of scalar type then the number of bytes
3721written does not exceed the minimum number of bytes needed to hold all
3722bits of the type. For example, storing an <tt>i24</tt> writes at most
3723three bytes. When writing a value of a type like <tt>i20</tt> with a
3724size that is not an integral number of bytes, it is unspecified what
3725happens to the extra bits that do not belong to the type, but they will
3726typically be overwritten.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3727<h5>Example:</h5>
3728<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling63ffa142007-10-22 05:10:05 +0000[diff] [blame]3729 store i32 3, i32* %ptr <i>; yields {void}</i>
3730 %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]3731</pre>
3732</div>
3733
3734<!-- _______________________________________________________________________ -->
3735<div class="doc_subsubsection">
3736 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
3737</div>
3738
3739<div class="doc_text">
3740<h5>Syntax:</h5>
3741<pre>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3742 &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]3743</pre>
3744
3745<h5>Overview:</h5>
3746
3747<p>
3748The '<tt>getelementptr</tt>' instruction is used to get the address of a
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3749subelement of an aggregate data structure. It performs address calculation only
3750and does not access memory.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3751
3752<h5>Arguments:</h5>
3753
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3754<p>The first argument is always a pointer, and forms the basis of the
3755calculation. The remaining arguments are indices, that indicate which of the
3756elements of the aggregate object are indexed. The interpretation of each index
3757is dependent on the type being indexed into. The first index always indexes the
3758pointer value given as the first argument, the second index indexes a value of
3759the type pointed to (not necessarily the value directly pointed to, since the
3760first index can be non-zero), etc. The first type indexed into must be a pointer
3761value, subsequent types can be arrays, vectors and structs. Note that subsequent
3762types being indexed into can never be pointers, since that would require loading
3763the pointer before continuing calculation.</p>
3764
3765<p>The type of each index argument depends on the type it is indexing into.
3766When indexing into a (packed) structure, only <tt>i32</tt> integer
3767<b>constants</b> are allowed. When indexing into an array, pointer or vector,
Sanjiv Guptaa2e72d92009-04-27 03:21:00 +0000[diff] [blame]3768integers of any width are allowed (also non-constants).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3769
3770<p>For example, let's consider a C code fragment and how it gets
3771compiled to LLVM:</p>
3772
3773<div class="doc_code">
3774<pre>
3775struct RT {
3776 char A;
3777 int B[10][20];
3778 char C;
3779};
3780struct ST {
3781 int X;
3782 double Y;
3783 struct RT Z;
3784};
3785
3786int *foo(struct ST *s) {
3787 return &amp;s[1].Z.B[5][13];
3788}
3789</pre>
3790</div>
3791
3792<p>The LLVM code generated by the GCC frontend is:</p>
3793
3794<div class="doc_code">
3795<pre>
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]3796%RT = <a href="#namedtypes">type</a> { i8 , [10 x [20 x i32]], i8 }
3797%ST = <a href="#namedtypes">type</a> { i32, double, %RT }
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3798
3799define i32* %foo(%ST* %s) {
3800entry:
3801 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
3802 ret i32* %reg
3803}
3804</pre>
3805</div>
3806
3807<h5>Semantics:</h5>
3808
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3809<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
3810type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
3811}</tt>' type, a structure. The second index indexes into the third element of
3812the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
3813i8 }</tt>' type, another structure. The third index indexes into the second
3814element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
3815array. The two dimensions of the array are subscripted into, yielding an
3816'<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
3817to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
3818
3819<p>Note that it is perfectly legal to index partially through a
3820structure, returning a pointer to an inner element. Because of this,
3821the LLVM code for the given testcase is equivalent to:</p>
3822
3823<pre>
3824 define i32* %foo(%ST* %s) {
3825 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
3826 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
3827 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
3828 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
3829 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
3830 ret i32* %t5
3831 }
3832</pre>
3833
Chris Lattner50609942009-03-09 20:55:18 +0000[diff] [blame]3834<p>Note that it is undefined to access an array out of bounds: array
3835and pointer indexes must always be within the defined bounds of the
3836array type when accessed with an instruction that dereferences the
3837pointer (e.g. a load or store instruction). The one exception for
3838this rule is zero length arrays. These arrays are defined to be
3839accessible as variable length arrays, which requires access beyond the
3840zero'th element.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3841
3842<p>The getelementptr instruction is often confusing. For some more insight
3843into how it works, see <a href="GetElementPtr.html">the getelementptr
3844FAQ</a>.</p>
3845
3846<h5>Example:</h5>
3847
3848<pre>
3849 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3850 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
3851 <i>; yields i8*:vptr</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3852 %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]3853 <i>; yields i8*:eptr</i>
3854 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Sanjiv Gupta4f9a0dc2009-04-25 07:27:44 +0000[diff] [blame]3855 <i>; yields i32*:iptr</i>
Sanjiv Gupta1e46c582009-04-24 16:38:13 +0000[diff] [blame]3856 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3857</pre>
3858</div>
3859
3860<!-- ======================================================================= -->
3861<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
3862</div>
3863<div class="doc_text">
3864<p>The instructions in this category are the conversion instructions (casting)
3865which all take a single operand and a type. They perform various bit conversions
3866on the operand.</p>
3867</div>
3868
3869<!-- _______________________________________________________________________ -->
3870<div class="doc_subsubsection">
3871 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
3872</div>
3873<div class="doc_text">
3874
3875<h5>Syntax:</h5>
3876<pre>
3877 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3878</pre>
3879
3880<h5>Overview:</h5>
3881<p>
3882The '<tt>trunc</tt>' instruction truncates its operand to the type <tt>ty2</tt>.
3883</p>
3884
3885<h5>Arguments:</h5>
3886<p>
3887The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
3888be an <a href="#t_integer">integer</a> type, and a type that specifies the size
3889and type of the result, which must be an <a href="#t_integer">integer</a>
3890type. The bit size of <tt>value</tt> must be larger than the bit size of
3891<tt>ty2</tt>. Equal sized types are not allowed.</p>
3892
3893<h5>Semantics:</h5>
3894<p>
3895The '<tt>trunc</tt>' instruction truncates the high order bits in <tt>value</tt>
3896and converts the remaining bits to <tt>ty2</tt>. Since the source size must be
3897larger than the destination size, <tt>trunc</tt> cannot be a <i>no-op cast</i>.
3898It will always truncate bits.</p>
3899
3900<h5>Example:</h5>
3901<pre>
3902 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
3903 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
3904 %Y = trunc i32 122 to i1 <i>; yields i1:false</i>
3905</pre>
3906</div>
3907
3908<!-- _______________________________________________________________________ -->
3909<div class="doc_subsubsection">
3910 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
3911</div>
3912<div class="doc_text">
3913
3914<h5>Syntax:</h5>
3915<pre>
3916 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3917</pre>
3918
3919<h5>Overview:</h5>
3920<p>The '<tt>zext</tt>' instruction zero extends its operand to type
3921<tt>ty2</tt>.</p>
3922
3923
3924<h5>Arguments:</h5>
3925<p>The '<tt>zext</tt>' instruction takes a value to cast, which must be of
3926<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3927also be of <a href="#t_integer">integer</a> type. The bit size of the
3928<tt>value</tt> must be smaller than the bit size of the destination type,
3929<tt>ty2</tt>.</p>
3930
3931<h5>Semantics:</h5>
3932<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
3933bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
3934
3935<p>When zero extending from i1, the result will always be either 0 or 1.</p>
3936
3937<h5>Example:</h5>
3938<pre>
3939 %X = zext i32 257 to i64 <i>; yields i64:257</i>
3940 %Y = zext i1 true to i32 <i>; yields i32:1</i>
3941</pre>
3942</div>
3943
3944<!-- _______________________________________________________________________ -->
3945<div class="doc_subsubsection">
3946 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
3947</div>
3948<div class="doc_text">
3949
3950<h5>Syntax:</h5>
3951<pre>
3952 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3953</pre>
3954
3955<h5>Overview:</h5>
3956<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
3957
3958<h5>Arguments:</h5>
3959<p>
3960The '<tt>sext</tt>' instruction takes a value to cast, which must be of
3961<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3962also be of <a href="#t_integer">integer</a> type. The bit size of the
3963<tt>value</tt> must be smaller than the bit size of the destination type,
3964<tt>ty2</tt>.</p>
3965
3966<h5>Semantics:</h5>
3967<p>
3968The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
3969bit (highest order bit) of the <tt>value</tt> until it reaches the bit size of
3970the type <tt>ty2</tt>.</p>
3971
3972<p>When sign extending from i1, the extension always results in -1 or 0.</p>
3973
3974<h5>Example:</h5>
3975<pre>
3976 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
3977 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
3978</pre>
3979</div>
3980
3981<!-- _______________________________________________________________________ -->
3982<div class="doc_subsubsection">
3983 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
3984</div>
3985
3986<div class="doc_text">
3987
3988<h5>Syntax:</h5>
3989
3990<pre>
3991 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3992</pre>
3993
3994<h5>Overview:</h5>
3995<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
3996<tt>ty2</tt>.</p>
3997
3998
3999<h5>Arguments:</h5>
4000<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
4001 point</a> value to cast and a <a href="#t_floating">floating point</a> type to
4002cast it to. The size of <tt>value</tt> must be larger than the size of
4003<tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
4004<i>no-op cast</i>.</p>
4005
4006<h5>Semantics:</h5>
4007<p> The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
4008<a href="#t_floating">floating point</a> type to a smaller
4009<a href="#t_floating">floating point</a> type. If the value cannot fit within
4010the destination type, <tt>ty2</tt>, then the results are undefined.</p>
4011
4012<h5>Example:</h5>
4013<pre>
4014 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
4015 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
4016</pre>
4017</div>
4018
4019<!-- _______________________________________________________________________ -->
4020<div class="doc_subsubsection">
4021 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
4022</div>
4023<div class="doc_text">
4024
4025<h5>Syntax:</h5>
4026<pre>
4027 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4028</pre>
4029
4030<h5>Overview:</h5>
4031<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
4032floating point value.</p>
4033
4034<h5>Arguments:</h5>
4035<p>The '<tt>fpext</tt>' instruction takes a
4036<a href="#t_floating">floating point</a> <tt>value</tt> to cast,
4037and a <a href="#t_floating">floating point</a> type to cast it to. The source
4038type must be smaller than the destination type.</p>
4039
4040<h5>Semantics:</h5>
4041<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
4042<a href="#t_floating">floating point</a> type to a larger
4043<a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
4044used to make a <i>no-op cast</i> because it always changes bits. Use
4045<tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
4046
4047<h5>Example:</h5>
4048<pre>
4049 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
4050 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
4051</pre>
4052</div>
4053
4054<!-- _______________________________________________________________________ -->
4055<div class="doc_subsubsection">
4056 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
4057</div>
4058<div class="doc_text">
4059
4060<h5>Syntax:</h5>
4061<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]4062 &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]4063</pre>
4064
4065<h5>Overview:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]4066<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4067unsigned integer equivalent of type <tt>ty2</tt>.
4068</p>
4069
4070<h5>Arguments:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]4071<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]4072scalar or vector <a href="#t_floating">floating point</a> value, and a type
4073to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4074type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4075vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4076
4077<h5>Semantics:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]4078<p> The '<tt>fptoui</tt>' instruction converts its
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4079<a href="#t_floating">floating point</a> operand into the nearest (rounding
4080towards zero) unsigned integer value. If the value cannot fit in <tt>ty2</tt>,
4081the results are undefined.</p>
4082
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4083<h5>Example:</h5>
4084<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]4085 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]4086 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]4087 %X = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4088</pre>
4089</div>
4090
4091<!-- _______________________________________________________________________ -->
4092<div class="doc_subsubsection">
4093 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
4094</div>
4095<div class="doc_text">
4096
4097<h5>Syntax:</h5>
4098<pre>
4099 &lt;result&gt; = fptosi &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4100</pre>
4101
4102<h5>Overview:</h5>
4103<p>The '<tt>fptosi</tt>' instruction converts
4104<a href="#t_floating">floating point</a> <tt>value</tt> to type <tt>ty2</tt>.
4105</p>
4106
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4107<h5>Arguments:</h5>
4108<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]4109scalar or vector <a href="#t_floating">floating point</a> value, and a type
4110to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4111type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4112vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4113
4114<h5>Semantics:</h5>
4115<p>The '<tt>fptosi</tt>' instruction converts its
4116<a href="#t_floating">floating point</a> operand into the nearest (rounding
4117towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
4118the results are undefined.</p>
4119
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4120<h5>Example:</h5>
4121<pre>
4122 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]4123 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4124 %X = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
4125</pre>
4126</div>
4127
4128<!-- _______________________________________________________________________ -->
4129<div class="doc_subsubsection">
4130 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
4131</div>
4132<div class="doc_text">
4133
4134<h5>Syntax:</h5>
4135<pre>
4136 &lt;result&gt; = uitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4137</pre>
4138
4139<h5>Overview:</h5>
4140<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
4141integer and converts that value to the <tt>ty2</tt> type.</p>
4142
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4143<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]4144<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be a
4145scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
4146to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4147type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4148floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4149
4150<h5>Semantics:</h5>
4151<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
4152integer quantity and converts it to the corresponding floating point value. If
4153the value cannot fit in the floating point value, the results are undefined.</p>
4154
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4155<h5>Example:</h5>
4156<pre>
4157 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4158 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4159</pre>
4160</div>
4161
4162<!-- _______________________________________________________________________ -->
4163<div class="doc_subsubsection">
4164 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
4165</div>
4166<div class="doc_text">
4167
4168<h5>Syntax:</h5>
4169<pre>
4170 &lt;result&gt; = sitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4171</pre>
4172
4173<h5>Overview:</h5>
4174<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed
4175integer and converts that value to the <tt>ty2</tt> type.</p>
4176
4177<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]4178<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be a
4179scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
4180to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4181type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4182floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4183
4184<h5>Semantics:</h5>
4185<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed
4186integer quantity and converts it to the corresponding floating point value. If
4187the value cannot fit in the floating point value, the results are undefined.</p>
4188
4189<h5>Example:</h5>
4190<pre>
4191 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4192 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4193</pre>
4194</div>
4195
4196<!-- _______________________________________________________________________ -->
4197<div class="doc_subsubsection">
4198 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
4199</div>
4200<div class="doc_text">
4201
4202<h5>Syntax:</h5>
4203<pre>
4204 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4205</pre>
4206
4207<h5>Overview:</h5>
4208<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
4209the integer type <tt>ty2</tt>.</p>
4210
4211<h5>Arguments:</h5>
4212<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
4213must 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]4214<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]4215
4216<h5>Semantics:</h5>
4217<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
4218<tt>ty2</tt> by interpreting the pointer value as an integer and either
4219truncating or zero extending that value to the size of the integer type. If
4220<tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
4221<tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
4222are the same size, then nothing is done (<i>no-op cast</i>) other than a type
4223change.</p>
4224
4225<h5>Example:</h5>
4226<pre>
4227 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
4228 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
4229</pre>
4230</div>
4231
4232<!-- _______________________________________________________________________ -->
4233<div class="doc_subsubsection">
4234 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
4235</div>
4236<div class="doc_text">
4237
4238<h5>Syntax:</h5>
4239<pre>
4240 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4241</pre>
4242
4243<h5>Overview:</h5>
4244<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to
4245a pointer type, <tt>ty2</tt>.</p>
4246
4247<h5>Arguments:</h5>
4248<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
4249value to cast, and a type to cast it to, which must be a
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4250<a href="#t_pointer">pointer</a> type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4251
4252<h5>Semantics:</h5>
4253<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
4254<tt>ty2</tt> by applying either a zero extension or a truncation depending on
4255the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
4256size of a pointer then a truncation is done. If <tt>value</tt> is smaller than
4257the size of a pointer then a zero extension is done. If they are the same size,
4258nothing is done (<i>no-op cast</i>).</p>
4259
4260<h5>Example:</h5>
4261<pre>
4262 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
4263 %X = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
4264 %Y = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
4265</pre>
4266</div>
4267
4268<!-- _______________________________________________________________________ -->
4269<div class="doc_subsubsection">
4270 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
4271</div>
4272<div class="doc_text">
4273
4274<h5>Syntax:</h5>
4275<pre>
4276 &lt;result&gt; = bitcast &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4277</pre>
4278
4279<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4280
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4281<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
4282<tt>ty2</tt> without changing any bits.</p>
4283
4284<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4285
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4286<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be
Dan Gohman7305fa02008-09-08 16:45:59 +0000[diff] [blame]4287a non-aggregate first class value, and a type to cast it to, which must also be
4288a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes of
4289<tt>value</tt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4290and the destination type, <tt>ty2</tt>, must be identical. If the source
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4291type is a pointer, the destination type must also be a pointer. This
4292instruction supports bitwise conversion of vectors to integers and to vectors
4293of other types (as long as they have the same size).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4294
4295<h5>Semantics:</h5>
4296<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
4297<tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
4298this conversion. The conversion is done as if the <tt>value</tt> had been
4299stored to memory and read back as type <tt>ty2</tt>. Pointer types may only be
4300converted to other pointer types with this instruction. To convert pointers to
4301other types, use the <a href="#i_inttoptr">inttoptr</a> or
4302<a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
4303
4304<h5>Example:</h5>
4305<pre>
4306 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
4307 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4308 %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]4309</pre>
4310</div>
4311
4312<!-- ======================================================================= -->
4313<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
4314<div class="doc_text">
4315<p>The instructions in this category are the "miscellaneous"
4316instructions, which defy better classification.</p>
4317</div>
4318
4319<!-- _______________________________________________________________________ -->
4320<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
4321</div>
4322<div class="doc_text">
4323<h5>Syntax:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4324<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]4325</pre>
4326<h5>Overview:</h5>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4327<p>The '<tt>icmp</tt>' instruction returns a boolean value or
4328a vector of boolean values based on comparison
4329of its two integer, integer vector, or pointer operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4330<h5>Arguments:</h5>
4331<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
4332the condition code indicating the kind of comparison to perform. It is not
4333a value, just a keyword. The possible condition code are:
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4334</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4335<ol>
4336 <li><tt>eq</tt>: equal</li>
4337 <li><tt>ne</tt>: not equal </li>
4338 <li><tt>ugt</tt>: unsigned greater than</li>
4339 <li><tt>uge</tt>: unsigned greater or equal</li>
4340 <li><tt>ult</tt>: unsigned less than</li>
4341 <li><tt>ule</tt>: unsigned less or equal</li>
4342 <li><tt>sgt</tt>: signed greater than</li>
4343 <li><tt>sge</tt>: signed greater or equal</li>
4344 <li><tt>slt</tt>: signed less than</li>
4345 <li><tt>sle</tt>: signed less or equal</li>
4346</ol>
4347<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4348<a href="#t_pointer">pointer</a>
4349or integer <a href="#t_vector">vector</a> typed.
4350They must also be identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4351<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4352<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]4353the condition code given as <tt>cond</tt>. The comparison performed always
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4354yields 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]4355</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4356<ol>
4357 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
4358 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
4359 </li>
4360 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4361 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4362 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4363 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4364 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4365 <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]4366 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4367 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4368 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4369 <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]4370 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4371 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4372 <li><tt>sge</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4373 <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]4374 <li><tt>slt</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4375 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4376 <li><tt>sle</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4377 <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]4378</ol>
4379<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
4380values are compared as if they were integers.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4381<p>If the operands are integer vectors, then they are compared
4382element by element. The result is an <tt>i1</tt> vector with
4383the same number of elements as the values being compared.
4384Otherwise, the result is an <tt>i1</tt>.
4385</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4386
4387<h5>Example:</h5>
4388<pre> &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
4389 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
4390 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
4391 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
4392 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
4393 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
4394</pre>
Dan Gohmana53eb382009-01-22 01:39:38 +0000[diff] [blame]4395
4396<p>Note that the code generator does not yet support vector types with
4397 the <tt>icmp</tt> instruction.</p>
4398
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4399</div>
4400
4401<!-- _______________________________________________________________________ -->
4402<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
4403</div>
4404<div class="doc_text">
4405<h5>Syntax:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4406<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]4407</pre>
4408<h5>Overview:</h5>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4409<p>The '<tt>fcmp</tt>' instruction returns a boolean value
4410or vector of boolean values based on comparison
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4411of its operands.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4412<p>
4413If the operands are floating point scalars, then the result
4414type is a boolean (<a href="#t_primitive"><tt>i1</tt></a>).
4415</p>
4416<p>If the operands are floating point vectors, then the result type
4417is a vector of boolean with the same number of elements as the
4418operands being compared.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4419<h5>Arguments:</h5>
4420<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
4421the condition code indicating the kind of comparison to perform. It is not
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4422a value, just a keyword. The possible condition code are:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4423<ol>
4424 <li><tt>false</tt>: no comparison, always returns false</li>
4425 <li><tt>oeq</tt>: ordered and equal</li>
4426 <li><tt>ogt</tt>: ordered and greater than </li>
4427 <li><tt>oge</tt>: ordered and greater than or equal</li>
4428 <li><tt>olt</tt>: ordered and less than </li>
4429 <li><tt>ole</tt>: ordered and less than or equal</li>
4430 <li><tt>one</tt>: ordered and not equal</li>
4431 <li><tt>ord</tt>: ordered (no nans)</li>
4432 <li><tt>ueq</tt>: unordered or equal</li>
4433 <li><tt>ugt</tt>: unordered or greater than </li>
4434 <li><tt>uge</tt>: unordered or greater than or equal</li>
4435 <li><tt>ult</tt>: unordered or less than </li>
4436 <li><tt>ule</tt>: unordered or less than or equal</li>
4437 <li><tt>une</tt>: unordered or not equal</li>
4438 <li><tt>uno</tt>: unordered (either nans)</li>
4439 <li><tt>true</tt>: no comparison, always returns true</li>
4440</ol>
4441<p><i>Ordered</i> means that neither operand is a QNAN while
4442<i>unordered</i> means that either operand may be a QNAN.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4443<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be
4444either a <a href="#t_floating">floating point</a> type
4445or a <a href="#t_vector">vector</a> of floating point type.
4446They must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4447<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4448<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4449according to the condition code given as <tt>cond</tt>.
4450If the operands are vectors, then the vectors are compared
4451element by element.
4452Each comparison performed
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4453always yields an <a href="#t_primitive">i1</a> result, as follows:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4454<ol>
4455 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
4456 <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]4457 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4458 <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]4459 <tt>op1</tt> is greather than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4460 <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]4461 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4462 <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]4463 <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4464 <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]4465 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4466 <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]4467 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4468 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
4469 <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]4470 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4471 <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]4472 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4473 <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]4474 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4475 <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]4476 <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4477 <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]4478 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4479 <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]4480 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4481 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
4482 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
4483</ol>
4484
4485<h5>Example:</h5>
4486<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]4487 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
4488 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
4489 &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]4490</pre>
Dan Gohmana53eb382009-01-22 01:39:38 +0000[diff] [blame]4491
4492<p>Note that the code generator does not yet support vector types with
4493 the <tt>fcmp</tt> instruction.</p>
4494
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4495</div>
4496
4497<!-- _______________________________________________________________________ -->
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4498<div class="doc_subsubsection">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4499 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
4500</div>
4501
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4502<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4503
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4504<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4505
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4506<pre> &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...<br></pre>
4507<h5>Overview:</h5>
4508<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in
4509the SSA graph representing the function.</p>
4510<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4511
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4512<p>The type of the incoming values is specified with the first type
4513field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
4514as arguments, with one pair for each predecessor basic block of the
4515current block. Only values of <a href="#t_firstclass">first class</a>
4516type may be used as the value arguments to the PHI node. Only labels
4517may be used as the label arguments.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4518
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4519<p>There must be no non-phi instructions between the start of a basic
4520block and the PHI instructions: i.e. PHI instructions must be first in
4521a basic block.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4522
Jay Foad8e2fd2c2009-06-03 10:20:10 +0000[diff] [blame]4523<p>For the purposes of the SSA form, the use of each incoming value is
4524deemed to occur on the edge from the corresponding predecessor block
4525to the current block (but after any definition of an '<tt>invoke</tt>'
4526instruction's return value on the same edge).</p>
4527
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4528<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4529
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4530<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
4531specified by the pair corresponding to the predecessor basic block that executed
4532just prior to the current block.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4533
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4534<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4535<pre>
4536Loop: ; Infinite loop that counts from 0 on up...
4537 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
4538 %nextindvar = add i32 %indvar, 1
4539 br label %Loop
4540</pre>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4541</div>
4542
4543<!-- _______________________________________________________________________ -->
4544<div class="doc_subsubsection">
4545 <a name="i_select">'<tt>select</tt>' Instruction</a>
4546</div>
4547
4548<div class="doc_text">
4549
4550<h5>Syntax:</h5>
4551
4552<pre>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4553 &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>
4554
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4555 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4556</pre>
4557
4558<h5>Overview:</h5>
4559
4560<p>
4561The '<tt>select</tt>' instruction is used to choose one value based on a
4562condition, without branching.
4563</p>
4564
4565
4566<h5>Arguments:</h5>
4567
4568<p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4569The '<tt>select</tt>' instruction requires an 'i1' value or
4570a vector of 'i1' values indicating the
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4571condition, and two values of the same <a href="#t_firstclass">first class</a>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4572type. If the val1/val2 are vectors and
4573the condition is a scalar, then entire vectors are selected, not
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4574individual elements.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4575</p>
4576
4577<h5>Semantics:</h5>
4578
4579<p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4580If 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]4581value argument; otherwise, it returns the second value argument.
4582</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4583<p>
4584If the condition is a vector of i1, then the value arguments must
4585be vectors of the same size, and the selection is done element
4586by element.
4587</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4588
4589<h5>Example:</h5>
4590
4591<pre>
4592 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
4593</pre>
Dan Gohmana53eb382009-01-22 01:39:38 +0000[diff] [blame]4594
4595<p>Note that the code generator does not yet support conditions
4596 with vector type.</p>
4597
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4598</div>
4599
4600
4601<!-- _______________________________________________________________________ -->
4602<div class="doc_subsubsection">
4603 <a name="i_call">'<tt>call</tt>' Instruction</a>
4604</div>
4605
4606<div class="doc_text">
4607
4608<h5>Syntax:</h5>
4609<pre>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]4610 &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]4611</pre>
4612
4613<h5>Overview:</h5>
4614
4615<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
4616
4617<h5>Arguments:</h5>
4618
4619<p>This instruction requires several arguments:</p>
4620
4621<ol>
4622 <li>
4623 <p>The optional "tail" marker indicates whether the callee function accesses
4624 any allocas or varargs in the caller. If the "tail" marker is present, the
4625 function call is eligible for tail call optimization. Note that calls may
4626 be marked "tail" even if they do not occur before a <a
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4627 href="#i_ret"><tt>ret</tt></a> instruction.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4628 </li>
4629 <li>
4630 <p>The optional "cconv" marker indicates which <a href="#callingconv">calling
4631 convention</a> the call should use. If none is specified, the call defaults
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4632 to using C calling conventions.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4633 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4634
4635 <li>
4636 <p>The optional <a href="#paramattrs">Parameter Attributes</a> list for
4637 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>',
4638 and '<tt>inreg</tt>' attributes are valid here.</p>
4639 </li>
4640
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4641 <li>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]4642 <p>'<tt>ty</tt>': the type of the call instruction itself which is also
4643 the type of the return value. Functions that return no value are marked
4644 <tt><a href="#t_void">void</a></tt>.</p>
4645 </li>
4646 <li>
4647 <p>'<tt>fnty</tt>': shall be the signature of the pointer to function
4648 value being invoked. The argument types must match the types implied by
4649 this signature. This type can be omitted if the function is not varargs
4650 and if the function type does not return a pointer to a function.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4651 </li>
4652 <li>
4653 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
4654 be invoked. In most cases, this is a direct function invocation, but
4655 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
4656 to function value.</p>
4657 </li>
4658 <li>
4659 <p>'<tt>function args</tt>': argument list whose types match the
4660 function signature argument types. All arguments must be of
4661 <a href="#t_firstclass">first class</a> type. If the function signature
4662 indicates the function accepts a variable number of arguments, the extra
4663 arguments can be specified.</p>
4664 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4665 <li>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]4666 <p>The optional <a href="#fnattrs">function attributes</a> list. Only
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4667 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
4668 '<tt>readnone</tt>' attributes are valid here.</p>
4669 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4670</ol>
4671
4672<h5>Semantics:</h5>
4673
4674<p>The '<tt>call</tt>' instruction is used to cause control flow to
4675transfer to a specified function, with its incoming arguments bound to
4676the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
4677instruction in the called function, control flow continues with the
4678instruction after the function call, and the return value of the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4679function is bound to the result argument.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4680
4681<h5>Example:</h5>
4682
4683<pre>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]4684 %retval = call i32 @test(i32 %argc)
Chris Lattner5e893ef2008-03-21 17:24:17 +0000[diff] [blame]4685 call i32 (i8 *, ...)* @printf(i8 * %msg, i32 12, i8 42) <i>; yields i32</i>
4686 %X = tail call i32 @foo() <i>; yields i32</i>
4687 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
4688 call void %foo(i8 97 signext)
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]4689
4690 %struct.A = type { i32, i8 }
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4691 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]4692 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
4693 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattnerac454b32008-10-08 06:26:11 +0000[diff] [blame]4694 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijman2c4e05a2008-10-07 10:03:45 +0000[diff] [blame]4695 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4696</pre>
4697
4698</div>
4699
4700<!-- _______________________________________________________________________ -->
4701<div class="doc_subsubsection">
4702 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
4703</div>
4704
4705<div class="doc_text">
4706
4707<h5>Syntax:</h5>
4708
4709<pre>
4710 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
4711</pre>
4712
4713<h5>Overview:</h5>
4714
4715<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
4716the "variable argument" area of a function call. It is used to implement the
4717<tt>va_arg</tt> macro in C.</p>
4718
4719<h5>Arguments:</h5>
4720
4721<p>This instruction takes a <tt>va_list*</tt> value and the type of
4722the argument. It returns a value of the specified argument type and
4723increments the <tt>va_list</tt> to point to the next argument. The
4724actual type of <tt>va_list</tt> is target specific.</p>
4725
4726<h5>Semantics:</h5>
4727
4728<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
4729type from the specified <tt>va_list</tt> and causes the
4730<tt>va_list</tt> to point to the next argument. For more information,
4731see the variable argument handling <a href="#int_varargs">Intrinsic
4732Functions</a>.</p>
4733
4734<p>It is legal for this instruction to be called in a function which does not
4735take a variable number of arguments, for example, the <tt>vfprintf</tt>
4736function.</p>
4737
4738<p><tt>va_arg</tt> is an LLVM instruction instead of an <a
4739href="#intrinsics">intrinsic function</a> because it takes a type as an
4740argument.</p>
4741
4742<h5>Example:</h5>
4743
4744<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
4745
Dan Gohman60967192009-01-12 23:12:39 +0000[diff] [blame]4746<p>Note that the code generator does not yet fully support va_arg
4747 on many targets. Also, it does not currently support va_arg with
4748 aggregate types on any target.</p>
4749
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4750</div>
4751
4752<!-- *********************************************************************** -->
4753<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
4754<!-- *********************************************************************** -->
4755
4756<div class="doc_text">
4757
4758<p>LLVM supports the notion of an "intrinsic function". These functions have
4759well known names and semantics and are required to follow certain restrictions.
4760Overall, these intrinsics represent an extension mechanism for the LLVM
4761language that does not require changing all of the transformations in LLVM when
4762adding to the language (or the bitcode reader/writer, the parser, etc...).</p>
4763
4764<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
4765prefix is reserved in LLVM for intrinsic names; thus, function names may not
4766begin with this prefix. Intrinsic functions must always be external functions:
4767you cannot define the body of intrinsic functions. Intrinsic functions may
4768only be used in call or invoke instructions: it is illegal to take the address
4769of an intrinsic function. Additionally, because intrinsic functions are part
4770of the LLVM language, it is required if any are added that they be documented
4771here.</p>
4772
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4773<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents
4774a family of functions that perform the same operation but on different data
4775types. Because LLVM can represent over 8 million different integer types,
4776overloading is used commonly to allow an intrinsic function to operate on any
4777integer type. One or more of the argument types or the result type can be
4778overloaded to accept any integer type. Argument types may also be defined as
4779exactly matching a previous argument's type or the result type. This allows an
4780intrinsic function which accepts multiple arguments, but needs all of them to
4781be of the same type, to only be overloaded with respect to a single argument or
4782the result.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4783
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4784<p>Overloaded intrinsics will have the names of its overloaded argument types
4785encoded into its function name, each preceded by a period. Only those types
4786which are overloaded result in a name suffix. Arguments whose type is matched
4787against another type do not. For example, the <tt>llvm.ctpop</tt> function can
4788take an integer of any width and returns an integer of exactly the same integer
4789width. This leads to a family of functions such as
4790<tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29 %val)</tt>.
4791Only one type, the return type, is overloaded, and only one type suffix is
4792required. Because the argument's type is matched against the return type, it
4793does not require its own name suffix.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4794
4795<p>To learn how to add an intrinsic function, please see the
4796<a href="ExtendingLLVM.html">Extending LLVM Guide</a>.
4797</p>
4798
4799</div>
4800
4801<!-- ======================================================================= -->
4802<div class="doc_subsection">
4803 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
4804</div>
4805
4806<div class="doc_text">
4807
4808<p>Variable argument support is defined in LLVM with the <a
4809 href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
4810intrinsic functions. These functions are related to the similarly
4811named macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
4812
4813<p>All of these functions operate on arguments that use a
4814target-specific value type "<tt>va_list</tt>". The LLVM assembly
4815language reference manual does not define what this type is, so all
4816transformations should be prepared to handle these functions regardless of
4817the type used.</p>
4818
4819<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
4820instruction and the variable argument handling intrinsic functions are
4821used.</p>
4822
4823<div class="doc_code">
4824<pre>
4825define i32 @test(i32 %X, ...) {
4826 ; Initialize variable argument processing
4827 %ap = alloca i8*
4828 %ap2 = bitcast i8** %ap to i8*
4829 call void @llvm.va_start(i8* %ap2)
4830
4831 ; Read a single integer argument
4832 %tmp = va_arg i8** %ap, i32
4833
4834 ; Demonstrate usage of llvm.va_copy and llvm.va_end
4835 %aq = alloca i8*
4836 %aq2 = bitcast i8** %aq to i8*
4837 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
4838 call void @llvm.va_end(i8* %aq2)
4839
4840 ; Stop processing of arguments.
4841 call void @llvm.va_end(i8* %ap2)
4842 ret i32 %tmp
4843}
4844
4845declare void @llvm.va_start(i8*)
4846declare void @llvm.va_copy(i8*, i8*)
4847declare void @llvm.va_end(i8*)
4848</pre>
4849</div>
4850
4851</div>
4852
4853<!-- _______________________________________________________________________ -->
4854<div class="doc_subsubsection">
4855 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
4856</div>
4857
4858
4859<div class="doc_text">
4860<h5>Syntax:</h5>
4861<pre> declare void %llvm.va_start(i8* &lt;arglist&gt;)<br></pre>
4862<h5>Overview:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4863<p>The '<tt>llvm.va_start</tt>' intrinsic initializes
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4864<tt>*&lt;arglist&gt;</tt> for subsequent use by <tt><a
4865href="#i_va_arg">va_arg</a></tt>.</p>
4866
4867<h5>Arguments:</h5>
4868
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4869<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]4870
4871<h5>Semantics:</h5>
4872
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4873<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]4874macro available in C. In a target-dependent way, it initializes the
4875<tt>va_list</tt> element to which the argument points, so that the next call to
4876<tt>va_arg</tt> will produce the first variable argument passed to the function.
4877Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
4878last argument of the function as the compiler can figure that out.</p>
4879
4880</div>
4881
4882<!-- _______________________________________________________________________ -->
4883<div class="doc_subsubsection">
4884 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
4885</div>
4886
4887<div class="doc_text">
4888<h5>Syntax:</h5>
4889<pre> declare void @llvm.va_end(i8* &lt;arglist&gt;)<br></pre>
4890<h5>Overview:</h5>
4891
4892<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
4893which has been initialized previously with <tt><a href="#int_va_start">llvm.va_start</a></tt>
4894or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
4895
4896<h5>Arguments:</h5>
4897
4898<p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
4899
4900<h5>Semantics:</h5>
4901
4902<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
4903macro available in C. In a target-dependent way, it destroys the
4904<tt>va_list</tt> element to which the argument points. Calls to <a
4905href="#int_va_start"><tt>llvm.va_start</tt></a> and <a href="#int_va_copy">
4906<tt>llvm.va_copy</tt></a> must be matched exactly with calls to
4907<tt>llvm.va_end</tt>.</p>
4908
4909</div>
4910
4911<!-- _______________________________________________________________________ -->
4912<div class="doc_subsubsection">
4913 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
4914</div>
4915
4916<div class="doc_text">
4917
4918<h5>Syntax:</h5>
4919
4920<pre>
4921 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
4922</pre>
4923
4924<h5>Overview:</h5>
4925
4926<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
4927from the source argument list to the destination argument list.</p>
4928
4929<h5>Arguments:</h5>
4930
4931<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
4932The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
4933
4934
4935<h5>Semantics:</h5>
4936
4937<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
4938macro available in C. In a target-dependent way, it copies the source
4939<tt>va_list</tt> element into the destination <tt>va_list</tt> element. This
4940intrinsic is necessary because the <tt><a href="#int_va_start">
4941llvm.va_start</a></tt> intrinsic may be arbitrarily complex and require, for
4942example, memory allocation.</p>
4943
4944</div>
4945
4946<!-- ======================================================================= -->
4947<div class="doc_subsection">
4948 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
4949</div>
4950
4951<div class="doc_text">
4952
4953<p>
4954LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]4955Collection</a> (GC) requires the implementation and generation of these
4956intrinsics.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4957These intrinsics allow identification of <a href="#int_gcroot">GC roots on the
4958stack</a>, as well as garbage collector implementations that require <a
4959href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a> barriers.
4960Front-ends for type-safe garbage collected languages should generate these
4961intrinsics to make use of the LLVM garbage collectors. For more details, see <a
4962href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
4963</p>
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]4964
4965<p>The garbage collection intrinsics only operate on objects in the generic
4966 address space (address space zero).</p>
4967
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4968</div>
4969
4970<!-- _______________________________________________________________________ -->
4971<div class="doc_subsubsection">
4972 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
4973</div>
4974
4975<div class="doc_text">
4976
4977<h5>Syntax:</h5>
4978
4979<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4980 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4981</pre>
4982
4983<h5>Overview:</h5>
4984
4985<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
4986the code generator, and allows some metadata to be associated with it.</p>
4987
4988<h5>Arguments:</h5>
4989
4990<p>The first argument specifies the address of a stack object that contains the
4991root pointer. The second pointer (which must be either a constant or a global
4992value address) contains the meta-data to be associated with the root.</p>
4993
4994<h5>Semantics:</h5>
4995
Chris Lattnera7d94ba2008-04-24 05:59:56 +0000[diff] [blame]4996<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]4997location. At compile-time, the code generator generates information to allow
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4998the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
4999intrinsic may only be used in a function which <a href="#gc">specifies a GC
5000algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5001
5002</div>
5003
5004
5005<!-- _______________________________________________________________________ -->
5006<div class="doc_subsubsection">
5007 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
5008</div>
5009
5010<div class="doc_text">
5011
5012<h5>Syntax:</h5>
5013
5014<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5015 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5016</pre>
5017
5018<h5>Overview:</h5>
5019
5020<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
5021locations, allowing garbage collector implementations that require read
5022barriers.</p>
5023
5024<h5>Arguments:</h5>
5025
5026<p>The second argument is the address to read from, which should be an address
5027allocated from the garbage collector. The first object is a pointer to the
5028start of the referenced object, if needed by the language runtime (otherwise
5029null).</p>
5030
5031<h5>Semantics:</h5>
5032
5033<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
5034instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]5035garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
5036may only be used in a function which <a href="#gc">specifies a GC
5037algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5038
5039</div>
5040
5041
5042<!-- _______________________________________________________________________ -->
5043<div class="doc_subsubsection">
5044 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
5045</div>
5046
5047<div class="doc_text">
5048
5049<h5>Syntax:</h5>
5050
5051<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5052 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5053</pre>
5054
5055<h5>Overview:</h5>
5056
5057<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
5058locations, allowing garbage collector implementations that require write
5059barriers (such as generational or reference counting collectors).</p>
5060
5061<h5>Arguments:</h5>
5062
5063<p>The first argument is the reference to store, the second is the start of the
5064object to store it to, and the third is the address of the field of Obj to
5065store to. If the runtime does not require a pointer to the object, Obj may be
5066null.</p>
5067
5068<h5>Semantics:</h5>
5069
5070<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
5071instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]5072garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
5073may only be used in a function which <a href="#gc">specifies a GC
5074algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5075
5076</div>
5077
5078
5079
5080<!-- ======================================================================= -->
5081<div class="doc_subsection">
5082 <a name="int_codegen">Code Generator Intrinsics</a>
5083</div>
5084
5085<div class="doc_text">
5086<p>
5087These intrinsics are provided by LLVM to expose special features that may only
5088be implemented with code generator support.
5089</p>
5090
5091</div>
5092
5093<!-- _______________________________________________________________________ -->
5094<div class="doc_subsubsection">
5095 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
5096</div>
5097
5098<div class="doc_text">
5099
5100<h5>Syntax:</h5>
5101<pre>
5102 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
5103</pre>
5104
5105<h5>Overview:</h5>
5106
5107<p>
5108The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
5109target-specific value indicating the return address of the current function
5110or one of its callers.
5111</p>
5112
5113<h5>Arguments:</h5>
5114
5115<p>
5116The argument to this intrinsic indicates which function to return the address
5117for. Zero indicates the calling function, one indicates its caller, etc. The
5118argument is <b>required</b> to be a constant integer value.
5119</p>
5120
5121<h5>Semantics:</h5>
5122
5123<p>
5124The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
5125the return address of the specified call frame, or zero if it cannot be
5126identified. The value returned by this intrinsic is likely to be incorrect or 0
5127for arguments other than zero, so it should only be used for debugging purposes.
5128</p>
5129
5130<p>
5131Note that calling this intrinsic does not prevent function inlining or other
5132aggressive transformations, so the value returned may not be that of the obvious
5133source-language caller.
5134</p>
5135</div>
5136
5137
5138<!-- _______________________________________________________________________ -->
5139<div class="doc_subsubsection">
5140 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
5141</div>
5142
5143<div class="doc_text">
5144
5145<h5>Syntax:</h5>
5146<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5147 declare i8 *@llvm.frameaddress(i32 &lt;level&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5148</pre>
5149
5150<h5>Overview:</h5>
5151
5152<p>
5153The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
5154target-specific frame pointer value for the specified stack frame.
5155</p>
5156
5157<h5>Arguments:</h5>
5158
5159<p>
5160The argument to this intrinsic indicates which function to return the frame
5161pointer for. Zero indicates the calling function, one indicates its caller,
5162etc. The argument is <b>required</b> to be a constant integer value.
5163</p>
5164
5165<h5>Semantics:</h5>
5166
5167<p>
5168The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
5169the frame address of the specified call frame, or zero if it cannot be
5170identified. The value returned by this intrinsic is likely to be incorrect or 0
5171for arguments other than zero, so it should only be used for debugging purposes.
5172</p>
5173
5174<p>
5175Note that calling this intrinsic does not prevent function inlining or other
5176aggressive transformations, so the value returned may not be that of the obvious
5177source-language caller.
5178</p>
5179</div>
5180
5181<!-- _______________________________________________________________________ -->
5182<div class="doc_subsubsection">
5183 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
5184</div>
5185
5186<div class="doc_text">
5187
5188<h5>Syntax:</h5>
5189<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5190 declare i8 *@llvm.stacksave()
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5191</pre>
5192
5193<h5>Overview:</h5>
5194
5195<p>
5196The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
5197the function stack, for use with <a href="#int_stackrestore">
5198<tt>llvm.stackrestore</tt></a>. This is useful for implementing language
5199features like scoped automatic variable sized arrays in C99.
5200</p>
5201
5202<h5>Semantics:</h5>
5203
5204<p>
5205This intrinsic returns a opaque pointer value that can be passed to <a
5206href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
5207<tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
5208<tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
5209state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
5210practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
5211that were allocated after the <tt>llvm.stacksave</tt> was executed.
5212</p>
5213
5214</div>
5215
5216<!-- _______________________________________________________________________ -->
5217<div class="doc_subsubsection">
5218 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
5219</div>
5220
5221<div class="doc_text">
5222
5223<h5>Syntax:</h5>
5224<pre>
5225 declare void @llvm.stackrestore(i8 * %ptr)
5226</pre>
5227
5228<h5>Overview:</h5>
5229
5230<p>
5231The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
5232the function stack to the state it was in when the corresponding <a
5233href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
5234useful for implementing language features like scoped automatic variable sized
5235arrays in C99.
5236</p>
5237
5238<h5>Semantics:</h5>
5239
5240<p>
5241See the description for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.
5242</p>
5243
5244</div>
5245
5246
5247<!-- _______________________________________________________________________ -->
5248<div class="doc_subsubsection">
5249 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
5250</div>
5251
5252<div class="doc_text">
5253
5254<h5>Syntax:</h5>
5255<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5256 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]5257</pre>
5258
5259<h5>Overview:</h5>
5260
5261
5262<p>
5263The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
5264a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
5265no
5266effect on the behavior of the program but can change its performance
5267characteristics.
5268</p>
5269
5270<h5>Arguments:</h5>
5271
5272<p>
5273<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
5274determining if the fetch should be for a read (0) or write (1), and
5275<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
5276locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
5277<tt>locality</tt> arguments must be constant integers.
5278</p>
5279
5280<h5>Semantics:</h5>
5281
5282<p>
5283This intrinsic does not modify the behavior of the program. In particular,
5284prefetches cannot trap and do not produce a value. On targets that support this
5285intrinsic, the prefetch can provide hints to the processor cache for better
5286performance.
5287</p>
5288
5289</div>
5290
5291<!-- _______________________________________________________________________ -->
5292<div class="doc_subsubsection">
5293 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
5294</div>
5295
5296<div class="doc_text">
5297
5298<h5>Syntax:</h5>
5299<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5300 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5301</pre>
5302
5303<h5>Overview:</h5>
5304
5305
5306<p>
5307The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]5308(PC) in a region of
5309code to simulators and other tools. The method is target specific, but it is
5310expected that the marker will use exported symbols to transmit the PC of the
5311marker.
5312The marker makes no guarantees that it will remain with any specific instruction
5313after optimizations. It is possible that the presence of a marker will inhibit
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5314optimizations. The intended use is to be inserted after optimizations to allow
5315correlations of simulation runs.
5316</p>
5317
5318<h5>Arguments:</h5>
5319
5320<p>
5321<tt>id</tt> is a numerical id identifying the marker.
5322</p>
5323
5324<h5>Semantics:</h5>
5325
5326<p>
5327This intrinsic does not modify the behavior of the program. Backends that do not
5328support this intrinisic may ignore it.
5329</p>
5330
5331</div>
5332
5333<!-- _______________________________________________________________________ -->
5334<div class="doc_subsubsection">
5335 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
5336</div>
5337
5338<div class="doc_text">
5339
5340<h5>Syntax:</h5>
5341<pre>
5342 declare i64 @llvm.readcyclecounter( )
5343</pre>
5344
5345<h5>Overview:</h5>
5346
5347
5348<p>
5349The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
5350counter register (or similar low latency, high accuracy clocks) on those targets
5351that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
5352As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
5353should only be used for small timings.
5354</p>
5355
5356<h5>Semantics:</h5>
5357
5358<p>
5359When directly supported, reading the cycle counter should not modify any memory.
5360Implementations are allowed to either return a application specific value or a
5361system wide value. On backends without support, this is lowered to a constant 0.
5362</p>
5363
5364</div>
5365
5366<!-- ======================================================================= -->
5367<div class="doc_subsection">
5368 <a name="int_libc">Standard C Library Intrinsics</a>
5369</div>
5370
5371<div class="doc_text">
5372<p>
5373LLVM provides intrinsics for a few important standard C library functions.
5374These intrinsics allow source-language front-ends to pass information about the
5375alignment of the pointer arguments to the code generator, providing opportunity
5376for more efficient code generation.
5377</p>
5378
5379</div>
5380
5381<!-- _______________________________________________________________________ -->
5382<div class="doc_subsubsection">
5383 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
5384</div>
5385
5386<div class="doc_text">
5387
5388<h5>Syntax:</h5>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5389<p>This is an overloaded intrinsic. You can use llvm.memcpy on any integer bit
5390width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5391<pre>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5392 declare void @llvm.memcpy.i8(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5393 i8 &lt;len&gt;, i32 &lt;align&gt;)
5394 declare void @llvm.memcpy.i16(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5395 i16 &lt;len&gt;, i32 &lt;align&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5396 declare void @llvm.memcpy.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5397 i32 &lt;len&gt;, i32 &lt;align&gt;)
5398 declare void @llvm.memcpy.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5399 i64 &lt;len&gt;, i32 &lt;align&gt;)
5400</pre>
5401
5402<h5>Overview:</h5>
5403
5404<p>
5405The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
5406location to the destination location.
5407</p>
5408
5409<p>
5410Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
5411intrinsics do not return a value, and takes an extra alignment argument.
5412</p>
5413
5414<h5>Arguments:</h5>
5415
5416<p>
5417The first argument is a pointer to the destination, the second is a pointer to
5418the source. The third argument is an integer argument
5419specifying the number of bytes to copy, and the fourth argument is the alignment
5420of the source and destination locations.
5421</p>
5422
5423<p>
5424If the call to this intrinisic has an alignment value that is not 0 or 1, then
5425the caller guarantees that both the source and destination pointers are aligned
5426to that boundary.
5427</p>
5428
5429<h5>Semantics:</h5>
5430
5431<p>
5432The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
5433location to the destination location, which are not allowed to overlap. It
5434copies "len" bytes of memory over. If the argument is known to be aligned to
5435some boundary, this can be specified as the fourth argument, otherwise it should
5436be set to 0 or 1.
5437</p>
5438</div>
5439
5440
5441<!-- _______________________________________________________________________ -->
5442<div class="doc_subsubsection">
5443 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
5444</div>
5445
5446<div class="doc_text">
5447
5448<h5>Syntax:</h5>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5449<p>This is an overloaded intrinsic. You can use llvm.memmove on any integer bit
5450width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5451<pre>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5452 declare void @llvm.memmove.i8(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5453 i8 &lt;len&gt;, i32 &lt;align&gt;)
5454 declare void @llvm.memmove.i16(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5455 i16 &lt;len&gt;, i32 &lt;align&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5456 declare void @llvm.memmove.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5457 i32 &lt;len&gt;, i32 &lt;align&gt;)
5458 declare void @llvm.memmove.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5459 i64 &lt;len&gt;, i32 &lt;align&gt;)
5460</pre>
5461
5462<h5>Overview:</h5>
5463
5464<p>
5465The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the source
5466location to the destination location. It is similar to the
Chris Lattnerdba16ea2008-01-06 19:51:52 +0000[diff] [blame]5467'<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to overlap.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5468</p>
5469
5470<p>
5471Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
5472intrinsics do not return a value, and takes an extra alignment argument.
5473</p>
5474
5475<h5>Arguments:</h5>
5476
5477<p>
5478The first argument is a pointer to the destination, the second is a pointer to
5479the source. The third argument is an integer argument
5480specifying the number of bytes to copy, and the fourth argument is the alignment
5481of the source and destination locations.
5482</p>
5483
5484<p>
5485If the call to this intrinisic has an alignment value that is not 0 or 1, then
5486the caller guarantees that the source and destination pointers are aligned to
5487that boundary.
5488</p>
5489
5490<h5>Semantics:</h5>
5491
5492<p>
5493The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the source
5494location to the destination location, which may overlap. It
5495copies "len" bytes of memory over. If the argument is known to be aligned to
5496some boundary, this can be specified as the fourth argument, otherwise it should
5497be set to 0 or 1.
5498</p>
5499</div>
5500
5501
5502<!-- _______________________________________________________________________ -->
5503<div class="doc_subsubsection">
5504 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
5505</div>
5506
5507<div class="doc_text">
5508
5509<h5>Syntax:</h5>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5510<p>This is an overloaded intrinsic. You can use llvm.memset on any integer bit
5511width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5512<pre>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5513 declare void @llvm.memset.i8(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5514 i8 &lt;len&gt;, i32 &lt;align&gt;)
5515 declare void @llvm.memset.i16(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5516 i16 &lt;len&gt;, i32 &lt;align&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5517 declare void @llvm.memset.i32(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5518 i32 &lt;len&gt;, i32 &lt;align&gt;)
5519 declare void @llvm.memset.i64(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5520 i64 &lt;len&gt;, i32 &lt;align&gt;)
5521</pre>
5522
5523<h5>Overview:</h5>
5524
5525<p>
5526The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a particular
5527byte value.
5528</p>
5529
5530<p>
5531Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
5532does not return a value, and takes an extra alignment argument.
5533</p>
5534
5535<h5>Arguments:</h5>
5536
5537<p>
5538The first argument is a pointer to the destination to fill, the second is the
5539byte value to fill it with, the third argument is an integer
5540argument specifying the number of bytes to fill, and the fourth argument is the
5541known alignment of destination location.
5542</p>
5543
5544<p>
5545If the call to this intrinisic has an alignment value that is not 0 or 1, then
5546the caller guarantees that the destination pointer is aligned to that boundary.
5547</p>
5548
5549<h5>Semantics:</h5>
5550
5551<p>
5552The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting at
5553the
5554destination location. If the argument is known to be aligned to some boundary,
5555this can be specified as the fourth argument, otherwise it should be set to 0 or
55561.
5557</p>
5558</div>
5559
5560
5561<!-- _______________________________________________________________________ -->
5562<div class="doc_subsubsection">
5563 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
5564</div>
5565
5566<div class="doc_text">
5567
5568<h5>Syntax:</h5>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5569<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]5570floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5571types however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5572<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5573 declare float @llvm.sqrt.f32(float %Val)
5574 declare double @llvm.sqrt.f64(double %Val)
5575 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
5576 declare fp128 @llvm.sqrt.f128(fp128 %Val)
5577 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5578</pre>
5579
5580<h5>Overview:</h5>
5581
5582<p>
5583The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5584returning the same value as the libm '<tt>sqrt</tt>' functions would. Unlike
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5585<tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
Chris Lattnerf914a002008-01-29 07:00:44 +0000[diff] [blame]5586negative numbers other than -0.0 (which allows for better optimization, because
5587there is no need to worry about errno being set). <tt>llvm.sqrt(-0.0)</tt> is
5588defined to return -0.0 like IEEE sqrt.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5589</p>
5590
5591<h5>Arguments:</h5>
5592
5593<p>
5594The argument and return value are floating point numbers of the same type.
5595</p>
5596
5597<h5>Semantics:</h5>
5598
5599<p>
5600This function returns the sqrt of the specified operand if it is a nonnegative
5601floating point number.
5602</p>
5603</div>
5604
5605<!-- _______________________________________________________________________ -->
5606<div class="doc_subsubsection">
5607 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
5608</div>
5609
5610<div class="doc_text">
5611
5612<h5>Syntax:</h5>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5613<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]5614floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5615types however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5616<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5617 declare float @llvm.powi.f32(float %Val, i32 %power)
5618 declare double @llvm.powi.f64(double %Val, i32 %power)
5619 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
5620 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
5621 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5622</pre>
5623
5624<h5>Overview:</h5>
5625
5626<p>
5627The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
5628specified (positive or negative) power. The order of evaluation of
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5629multiplications is not defined. When a vector of floating point type is
5630used, the second argument remains a scalar integer value.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5631</p>
5632
5633<h5>Arguments:</h5>
5634
5635<p>
5636The second argument is an integer power, and the first is a value to raise to
5637that power.
5638</p>
5639
5640<h5>Semantics:</h5>
5641
5642<p>
5643This function returns the first value raised to the second power with an
5644unspecified sequence of rounding operations.</p>
5645</div>
5646
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5647<!-- _______________________________________________________________________ -->
5648<div class="doc_subsubsection">
5649 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
5650</div>
5651
5652<div class="doc_text">
5653
5654<h5>Syntax:</h5>
5655<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
5656floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5657types however.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5658<pre>
5659 declare float @llvm.sin.f32(float %Val)
5660 declare double @llvm.sin.f64(double %Val)
5661 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
5662 declare fp128 @llvm.sin.f128(fp128 %Val)
5663 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
5664</pre>
5665
5666<h5>Overview:</h5>
5667
5668<p>
5669The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.
5670</p>
5671
5672<h5>Arguments:</h5>
5673
5674<p>
5675The argument and return value are floating point numbers of the same type.
5676</p>
5677
5678<h5>Semantics:</h5>
5679
5680<p>
5681This function returns the sine of the specified operand, returning the
5682same values as the libm <tt>sin</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5683conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5684</div>
5685
5686<!-- _______________________________________________________________________ -->
5687<div class="doc_subsubsection">
5688 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
5689</div>
5690
5691<div class="doc_text">
5692
5693<h5>Syntax:</h5>
5694<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
5695floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5696types however.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5697<pre>
5698 declare float @llvm.cos.f32(float %Val)
5699 declare double @llvm.cos.f64(double %Val)
5700 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
5701 declare fp128 @llvm.cos.f128(fp128 %Val)
5702 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
5703</pre>
5704
5705<h5>Overview:</h5>
5706
5707<p>
5708The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.
5709</p>
5710
5711<h5>Arguments:</h5>
5712
5713<p>
5714The argument and return value are floating point numbers of the same type.
5715</p>
5716
5717<h5>Semantics:</h5>
5718
5719<p>
5720This function returns the cosine of the specified operand, returning the
5721same values as the libm <tt>cos</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5722conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5723</div>
5724
5725<!-- _______________________________________________________________________ -->
5726<div class="doc_subsubsection">
5727 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
5728</div>
5729
5730<div class="doc_text">
5731
5732<h5>Syntax:</h5>
5733<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
5734floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5735types however.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5736<pre>
5737 declare float @llvm.pow.f32(float %Val, float %Power)
5738 declare double @llvm.pow.f64(double %Val, double %Power)
5739 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
5740 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
5741 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
5742</pre>
5743
5744<h5>Overview:</h5>
5745
5746<p>
5747The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
5748specified (positive or negative) power.
5749</p>
5750
5751<h5>Arguments:</h5>
5752
5753<p>
5754The second argument is a floating point power, and the first is a value to
5755raise to that power.
5756</p>
5757
5758<h5>Semantics:</h5>
5759
5760<p>
5761This function returns the first value raised to the second power,
5762returning the
5763same values as the libm <tt>pow</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5764conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5765</div>
5766
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5767
5768<!-- ======================================================================= -->
5769<div class="doc_subsection">
5770 <a name="int_manip">Bit Manipulation Intrinsics</a>
5771</div>
5772
5773<div class="doc_text">
5774<p>
5775LLVM provides intrinsics for a few important bit manipulation operations.
5776These allow efficient code generation for some algorithms.
5777</p>
5778
5779</div>
5780
5781<!-- _______________________________________________________________________ -->
5782<div class="doc_subsubsection">
5783 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
5784</div>
5785
5786<div class="doc_text">
5787
5788<h5>Syntax:</h5>
5789<p>This is an overloaded intrinsic function. You can use bswap on any integer
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5790type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5791<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5792 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
5793 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
5794 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5795</pre>
5796
5797<h5>Overview:</h5>
5798
5799<p>
5800The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
5801values with an even number of bytes (positive multiple of 16 bits). These are
5802useful for performing operations on data that is not in the target's native
5803byte order.
5804</p>
5805
5806<h5>Semantics:</h5>
5807
5808<p>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5809The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5810and low byte of the input i16 swapped. Similarly, the <tt>llvm.bswap.i32</tt>
5811intrinsic returns an i32 value that has the four bytes of the input i32
5812swapped, 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]5813i32 will have its bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i48</tt>,
5814<tt>llvm.bswap.i64</tt> and other intrinsics extend this concept to
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5815additional even-byte lengths (6 bytes, 8 bytes and more, respectively).
5816</p>
5817
5818</div>
5819
5820<!-- _______________________________________________________________________ -->
5821<div class="doc_subsubsection">
5822 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
5823</div>
5824
5825<div class="doc_text">
5826
5827<h5>Syntax:</h5>
5828<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]5829width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5830<pre>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5831 declare i8 @llvm.ctpop.i8(i8 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5832 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5833 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5834 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
5835 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5836</pre>
5837
5838<h5>Overview:</h5>
5839
5840<p>
5841The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
5842value.
5843</p>
5844
5845<h5>Arguments:</h5>
5846
5847<p>
5848The only argument is the value to be counted. The argument may be of any
5849integer type. The return type must match the argument type.
5850</p>
5851
5852<h5>Semantics:</h5>
5853
5854<p>
5855The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
5856</p>
5857</div>
5858
5859<!-- _______________________________________________________________________ -->
5860<div class="doc_subsubsection">
5861 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
5862</div>
5863
5864<div class="doc_text">
5865
5866<h5>Syntax:</h5>
5867<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]5868integer bit width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5869<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5870 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
5871 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5872 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5873 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
5874 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5875</pre>
5876
5877<h5>Overview:</h5>
5878
5879<p>
5880The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
5881leading zeros in a variable.
5882</p>
5883
5884<h5>Arguments:</h5>
5885
5886<p>
5887The only argument is the value to be counted. The argument may be of any
5888integer type. The return type must match the argument type.
5889</p>
5890
5891<h5>Semantics:</h5>
5892
5893<p>
5894The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
5895in a variable. If the src == 0 then the result is the size in bits of the type
5896of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.
5897</p>
5898</div>
5899
5900
5901
5902<!-- _______________________________________________________________________ -->
5903<div class="doc_subsubsection">
5904 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
5905</div>
5906
5907<div class="doc_text">
5908
5909<h5>Syntax:</h5>
5910<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5911integer bit width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5912<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5913 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
5914 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5915 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5916 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
5917 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5918</pre>
5919
5920<h5>Overview:</h5>
5921
5922<p>
5923The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
5924trailing zeros.
5925</p>
5926
5927<h5>Arguments:</h5>
5928
5929<p>
5930The only argument is the value to be counted. The argument may be of any
5931integer type. The return type must match the argument type.
5932</p>
5933
5934<h5>Semantics:</h5>
5935
5936<p>
5937The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
5938in a variable. If the src == 0 then the result is the size in bits of the type
5939of src. For example, <tt>llvm.cttz(2) = 1</tt>.
5940</p>
5941</div>
5942
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5943
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]5944<!-- ======================================================================= -->
5945<div class="doc_subsection">
5946 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
5947</div>
5948
5949<div class="doc_text">
5950<p>
5951LLVM provides intrinsics for some arithmetic with overflow operations.
5952</p>
5953
5954</div>
5955
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5956<!-- _______________________________________________________________________ -->
5957<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]5958 <a name="int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5959</div>
5960
5961<div class="doc_text">
5962
5963<h5>Syntax:</h5>
5964
5965<p>This is an overloaded intrinsic. You can use <tt>llvm.sadd.with.overflow</tt>
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]5966on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5967
5968<pre>
5969 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
5970 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
5971 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
5972</pre>
5973
5974<h5>Overview:</h5>
5975
5976<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
5977a signed addition of the two arguments, and indicate whether an overflow
5978occurred during the signed summation.</p>
5979
5980<h5>Arguments:</h5>
5981
5982<p>The arguments (%a and %b) and the first element of the result structure may
5983be of integer types of any bit width, but they must have the same bit width. The
5984second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
5985and <tt>%b</tt> are the two values that will undergo signed addition.</p>
5986
5987<h5>Semantics:</h5>
5988
5989<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
5990a signed addition of the two variables. They return a structure &mdash; the
5991first element of which is the signed summation, and the second element of which
5992is a bit specifying if the signed summation resulted in an overflow.</p>
5993
5994<h5>Examples:</h5>
5995<pre>
5996 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
5997 %sum = extractvalue {i32, i1} %res, 0
5998 %obit = extractvalue {i32, i1} %res, 1
5999 br i1 %obit, label %overflow, label %normal
6000</pre>
6001
6002</div>
6003
6004<!-- _______________________________________________________________________ -->
6005<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6006 <a name="int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6007</div>
6008
6009<div class="doc_text">
6010
6011<h5>Syntax:</h5>
6012
6013<p>This is an overloaded intrinsic. You can use <tt>llvm.uadd.with.overflow</tt>
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6014on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6015
6016<pre>
6017 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
6018 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6019 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
6020</pre>
6021
6022<h5>Overview:</h5>
6023
6024<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
6025an unsigned addition of the two arguments, and indicate whether a carry occurred
6026during the unsigned summation.</p>
6027
6028<h5>Arguments:</h5>
6029
6030<p>The arguments (%a and %b) and the first element of the result structure may
6031be of integer types of any bit width, but they must have the same bit width. The
6032second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6033and <tt>%b</tt> are the two values that will undergo unsigned addition.</p>
6034
6035<h5>Semantics:</h5>
6036
6037<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
6038an unsigned addition of the two arguments. They return a structure &mdash; the
6039first element of which is the sum, and the second element of which is a bit
6040specifying if the unsigned summation resulted in a carry.</p>
6041
6042<h5>Examples:</h5>
6043<pre>
6044 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6045 %sum = extractvalue {i32, i1} %res, 0
6046 %obit = extractvalue {i32, i1} %res, 1
6047 br i1 %obit, label %carry, label %normal
6048</pre>
6049
6050</div>
6051
6052<!-- _______________________________________________________________________ -->
6053<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6054 <a name="int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6055</div>
6056
6057<div class="doc_text">
6058
6059<h5>Syntax:</h5>
6060
6061<p>This is an overloaded intrinsic. You can use <tt>llvm.ssub.with.overflow</tt>
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6062on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6063
6064<pre>
6065 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
6066 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6067 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
6068</pre>
6069
6070<h5>Overview:</h5>
6071
6072<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
6073a signed subtraction of the two arguments, and indicate whether an overflow
6074occurred during the signed subtraction.</p>
6075
6076<h5>Arguments:</h5>
6077
6078<p>The arguments (%a and %b) and the first element of the result structure may
6079be of integer types of any bit width, but they must have the same bit width. The
6080second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6081and <tt>%b</tt> are the two values that will undergo signed subtraction.</p>
6082
6083<h5>Semantics:</h5>
6084
6085<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
6086a signed subtraction of the two arguments. They return a structure &mdash; the
6087first element of which is the subtraction, and the second element of which is a bit
6088specifying if the signed subtraction resulted in an overflow.</p>
6089
6090<h5>Examples:</h5>
6091<pre>
6092 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6093 %sum = extractvalue {i32, i1} %res, 0
6094 %obit = extractvalue {i32, i1} %res, 1
6095 br i1 %obit, label %overflow, label %normal
6096</pre>
6097
6098</div>
6099
6100<!-- _______________________________________________________________________ -->
6101<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6102 <a name="int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6103</div>
6104
6105<div class="doc_text">
6106
6107<h5>Syntax:</h5>
6108
6109<p>This is an overloaded intrinsic. You can use <tt>llvm.usub.with.overflow</tt>
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6110on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6111
6112<pre>
6113 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
6114 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6115 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
6116</pre>
6117
6118<h5>Overview:</h5>
6119
6120<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
6121an unsigned subtraction of the two arguments, and indicate whether an overflow
6122occurred during the unsigned subtraction.</p>
6123
6124<h5>Arguments:</h5>
6125
6126<p>The arguments (%a and %b) and the first element of the result structure may
6127be of integer types of any bit width, but they must have the same bit width. The
6128second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6129and <tt>%b</tt> are the two values that will undergo unsigned subtraction.</p>
6130
6131<h5>Semantics:</h5>
6132
6133<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
6134an unsigned subtraction of the two arguments. They return a structure &mdash; the
6135first element of which is the subtraction, and the second element of which is a bit
6136specifying if the unsigned subtraction resulted in an overflow.</p>
6137
6138<h5>Examples:</h5>
6139<pre>
6140 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6141 %sum = extractvalue {i32, i1} %res, 0
6142 %obit = extractvalue {i32, i1} %res, 1
6143 br i1 %obit, label %overflow, label %normal
6144</pre>
6145
6146</div>
6147
6148<!-- _______________________________________________________________________ -->
6149<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6150 <a name="int_smul_overflow">'<tt>llvm.smul.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6151</div>
6152
6153<div class="doc_text">
6154
6155<h5>Syntax:</h5>
6156
6157<p>This is an overloaded intrinsic. You can use <tt>llvm.smul.with.overflow</tt>
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6158on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6159
6160<pre>
6161 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
6162 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6163 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
6164</pre>
6165
6166<h5>Overview:</h5>
6167
6168<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
6169a signed multiplication of the two arguments, and indicate whether an overflow
6170occurred during the signed multiplication.</p>
6171
6172<h5>Arguments:</h5>
6173
6174<p>The arguments (%a and %b) and the first element of the result structure may
6175be of integer types of any bit width, but they must have the same bit width. The
6176second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6177and <tt>%b</tt> are the two values that will undergo signed multiplication.</p>
6178
6179<h5>Semantics:</h5>
6180
6181<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
6182a signed multiplication of the two arguments. They return a structure &mdash;
6183the first element of which is the multiplication, and the second element of
6184which is a bit specifying if the signed multiplication resulted in an
6185overflow.</p>
6186
6187<h5>Examples:</h5>
6188<pre>
6189 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6190 %sum = extractvalue {i32, i1} %res, 0
6191 %obit = extractvalue {i32, i1} %res, 1
6192 br i1 %obit, label %overflow, label %normal
6193</pre>
6194
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6195</div>
6196
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]6197<!-- _______________________________________________________________________ -->
6198<div class="doc_subsubsection">
6199 <a name="int_umul_overflow">'<tt>llvm.umul.with.overflow.*</tt>' Intrinsics</a>
6200</div>
6201
6202<div class="doc_text">
6203
6204<h5>Syntax:</h5>
6205
6206<p>This is an overloaded intrinsic. You can use <tt>llvm.umul.with.overflow</tt>
6207on any integer bit width.</p>
6208
6209<pre>
6210 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
6211 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6212 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
6213</pre>
6214
6215<h5>Overview:</h5>
6216
Bill Wendlingbda98b62009-02-08 23:00:09 +0000[diff] [blame]6217<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
6218a unsigned multiplication of the two arguments, and indicate whether an overflow
6219occurred during the unsigned multiplication.</p>
6220
6221<h5>Arguments:</h5>
6222
6223<p>The arguments (%a and %b) and the first element of the result structure may
6224be of integer types of any bit width, but they must have the same bit width. The
6225second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6226and <tt>%b</tt> are the two values that will undergo unsigned
6227multiplication.</p>
6228
6229<h5>Semantics:</h5>
6230
6231<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
6232an unsigned multiplication of the two arguments. They return a structure &mdash;
6233the first element of which is the multiplication, and the second element of
6234which is a bit specifying if the unsigned multiplication resulted in an
6235overflow.</p>
6236
6237<h5>Examples:</h5>
6238<pre>
6239 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6240 %sum = extractvalue {i32, i1} %res, 0
6241 %obit = extractvalue {i32, i1} %res, 1
6242 br i1 %obit, label %overflow, label %normal
6243</pre>
6244
6245</div>
6246
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6247<!-- ======================================================================= -->
6248<div class="doc_subsection">
6249 <a name="int_debugger">Debugger Intrinsics</a>
6250</div>
6251
6252<div class="doc_text">
6253<p>
6254The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
6255are described in the <a
6256href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
6257Debugging</a> document.
6258</p>
6259</div>
6260
6261
6262<!-- ======================================================================= -->
6263<div class="doc_subsection">
6264 <a name="int_eh">Exception Handling Intrinsics</a>
6265</div>
6266
6267<div class="doc_text">
6268<p> The LLVM exception handling intrinsics (which all start with
6269<tt>llvm.eh.</tt> prefix), are described in the <a
6270href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
6271Handling</a> document. </p>
6272</div>
6273
6274<!-- ======================================================================= -->
6275<div class="doc_subsection">
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6276 <a name="int_trampoline">Trampoline Intrinsic</a>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6277</div>
6278
6279<div class="doc_text">
6280<p>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6281 This intrinsic makes it possible to excise one parameter, marked with
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6282 the <tt>nest</tt> attribute, from a function. The result is a callable
6283 function pointer lacking the nest parameter - the caller does not need
6284 to provide a value for it. Instead, the value to use is stored in
6285 advance in a "trampoline", a block of memory usually allocated
6286 on the stack, which also contains code to splice the nest value into the
6287 argument list. This is used to implement the GCC nested function address
6288 extension.
6289</p>
6290<p>
6291 For example, if the function is
6292 <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]6293 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]6294<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6295 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
6296 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
6297 %p = call i8* @llvm.init.trampoline( i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8*), i8* %nval )
6298 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6299</pre>
Bill Wendlinge262b4c2007-09-22 09:23:55 +0000[diff] [blame]6300 <p>The call <tt>%val = call i32 %fp( i32 %x, i32 %y )</tt> is then equivalent
6301 to <tt>%val = call i32 %f( i8* %nval, i32 %x, i32 %y )</tt>.</p>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6302</div>
6303
6304<!-- _______________________________________________________________________ -->
6305<div class="doc_subsubsection">
6306 <a name="int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a>
6307</div>
6308<div class="doc_text">
6309<h5>Syntax:</h5>
6310<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6311declare 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]6312</pre>
6313<h5>Overview:</h5>
6314<p>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6315 This fills the memory pointed to by <tt>tramp</tt> with code
6316 and returns a function pointer suitable for executing it.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6317</p>
6318<h5>Arguments:</h5>
6319<p>
6320 The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
6321 pointers. The <tt>tramp</tt> argument must point to a sufficiently large
6322 and sufficiently aligned block of memory; this memory is written to by the
Duncan Sands35012212007-08-22 23:39:54 +0000[diff] [blame]6323 intrinsic. Note that the size and the alignment are target-specific - LLVM
6324 currently provides no portable way of determining them, so a front-end that
6325 generates this intrinsic needs to have some target-specific knowledge.
6326 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]6327</p>
6328<h5>Semantics:</h5>
6329<p>
6330 The block of memory pointed to by <tt>tramp</tt> is filled with target
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6331 dependent code, turning it into a function. A pointer to this function is
6332 returned, but needs to be bitcast to an
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6333 <a href="#int_trampoline">appropriate function pointer type</a>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6334 before being called. The new function's signature is the same as that of
6335 <tt>func</tt> with any arguments marked with the <tt>nest</tt> attribute
6336 removed. At most one such <tt>nest</tt> argument is allowed, and it must be
6337 of pointer type. Calling the new function is equivalent to calling
6338 <tt>func</tt> with the same argument list, but with <tt>nval</tt> used for the
6339 missing <tt>nest</tt> argument. If, after calling
6340 <tt>llvm.init.trampoline</tt>, the memory pointed to by <tt>tramp</tt> is
6341 modified, then the effect of any later call to the returned function pointer is
6342 undefined.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6343</p>
6344</div>
6345
6346<!-- ======================================================================= -->
6347<div class="doc_subsection">
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6348 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
6349</div>
6350
6351<div class="doc_text">
6352<p>
6353 These intrinsic functions expand the "universal IR" of LLVM to represent
6354 hardware constructs for atomic operations and memory synchronization. This
6355 provides an interface to the hardware, not an interface to the programmer. It
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]6356 is aimed at a low enough level to allow any programming models or APIs
6357 (Application Programming Interfaces) which
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6358 need atomic behaviors to map cleanly onto it. It is also modeled primarily on
6359 hardware behavior. Just as hardware provides a "universal IR" for source
6360 languages, it also provides a starting point for developing a "universal"
6361 atomic operation and synchronization IR.
6362</p>
6363<p>
6364 These do <em>not</em> form an API such as high-level threading libraries,
6365 software transaction memory systems, atomic primitives, and intrinsic
6366 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
6367 application libraries. The hardware interface provided by LLVM should allow
6368 a clean implementation of all of these APIs and parallel programming models.
6369 No one model or paradigm should be selected above others unless the hardware
6370 itself ubiquitously does so.
6371
6372</p>
6373</div>
6374
6375<!-- _______________________________________________________________________ -->
6376<div class="doc_subsubsection">
6377 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
6378</div>
6379<div class="doc_text">
6380<h5>Syntax:</h5>
6381<pre>
6382declare void @llvm.memory.barrier( i1 &lt;ll&gt;, i1 &lt;ls&gt;, i1 &lt;sl&gt;, i1 &lt;ss&gt;,
6383i1 &lt;device&gt; )
6384
6385</pre>
6386<h5>Overview:</h5>
6387<p>
6388 The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
6389 specific pairs of memory access types.
6390</p>
6391<h5>Arguments:</h5>
6392<p>
6393 The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
6394 The first four arguments enables a specific barrier as listed below. The fith
6395 argument specifies that the barrier applies to io or device or uncached memory.
6396
6397</p>
6398 <ul>
6399 <li><tt>ll</tt>: load-load barrier</li>
6400 <li><tt>ls</tt>: load-store barrier</li>
6401 <li><tt>sl</tt>: store-load barrier</li>
6402 <li><tt>ss</tt>: store-store barrier</li>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]6403 <li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6404 </ul>
6405<h5>Semantics:</h5>
6406<p>
6407 This intrinsic causes the system to enforce some ordering constraints upon
6408 the loads and stores of the program. This barrier does not indicate
6409 <em>when</em> any events will occur, it only enforces an <em>order</em> in
6410 which they occur. For any of the specified pairs of load and store operations
6411 (f.ex. load-load, or store-load), all of the first operations preceding the
6412 barrier will complete before any of the second operations succeeding the
6413 barrier begin. Specifically the semantics for each pairing is as follows:
6414</p>
6415 <ul>
6416 <li><tt>ll</tt>: All loads before the barrier must complete before any load
6417 after the barrier begins.</li>
6418
6419 <li><tt>ls</tt>: All loads before the barrier must complete before any
6420 store after the barrier begins.</li>
6421 <li><tt>ss</tt>: All stores before the barrier must complete before any
6422 store after the barrier begins.</li>
6423 <li><tt>sl</tt>: All stores before the barrier must complete before any
6424 load after the barrier begins.</li>
6425 </ul>
6426<p>
6427 These semantics are applied with a logical "and" behavior when more than one
6428 is enabled in a single memory barrier intrinsic.
6429</p>
6430<p>
6431 Backends may implement stronger barriers than those requested when they do not
6432 support as fine grained a barrier as requested. Some architectures do not
6433 need all types of barriers and on such architectures, these become noops.
6434</p>
6435<h5>Example:</h5>
6436<pre>
6437%ptr = malloc i32
6438 store i32 4, %ptr
6439
6440%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
6441 call void @llvm.memory.barrier( i1 false, i1 true, i1 false, i1 false )
6442 <i>; guarantee the above finishes</i>
6443 store i32 8, %ptr <i>; before this begins</i>
6444</pre>
6445</div>
6446
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6447<!-- _______________________________________________________________________ -->
6448<div class="doc_subsubsection">
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6449 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6450</div>
6451<div class="doc_text">
6452<h5>Syntax:</h5>
6453<p>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6454 This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
6455 any integer bit width and for different address spaces. Not all targets
6456 support all bit widths however.</p>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6457
6458<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6459declare i8 @llvm.atomic.cmp.swap.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt; )
6460declare i16 @llvm.atomic.cmp.swap.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt; )
6461declare i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt; )
6462declare 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]6463
6464</pre>
6465<h5>Overview:</h5>
6466<p>
6467 This loads a value in memory and compares it to a given value. If they are
6468 equal, it stores a new value into the memory.
6469</p>
6470<h5>Arguments:</h5>
6471<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6472 The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result as
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6473 well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
6474 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
6475 this integer type. While any bit width integer may be used, targets may only
6476 lower representations they support in hardware.
6477
6478</p>
6479<h5>Semantics:</h5>
6480<p>
6481 This entire intrinsic must be executed atomically. It first loads the value
6482 in memory pointed to by <tt>ptr</tt> and compares it with the value
6483 <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the memory. The
6484 loaded value is yielded in all cases. This provides the equivalent of an
6485 atomic compare-and-swap operation within the SSA framework.
6486</p>
6487<h5>Examples:</h5>
6488
6489<pre>
6490%ptr = malloc i32
6491 store i32 4, %ptr
6492
6493%val1 = add i32 4, 4
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6494%result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 4, %val1 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6495 <i>; yields {i32}:result1 = 4</i>
6496%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
6497%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
6498
6499%val2 = add i32 1, 1
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6500%result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 5, %val2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6501 <i>; yields {i32}:result2 = 8</i>
6502%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
6503
6504%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
6505</pre>
6506</div>
6507
6508<!-- _______________________________________________________________________ -->
6509<div class="doc_subsubsection">
6510 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
6511</div>
6512<div class="doc_text">
6513<h5>Syntax:</h5>
6514
6515<p>
6516 This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
6517 integer bit width. Not all targets support all bit widths however.</p>
6518<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6519declare i8 @llvm.atomic.swap.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;val&gt; )
6520declare i16 @llvm.atomic.swap.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;val&gt; )
6521declare i32 @llvm.atomic.swap.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;val&gt; )
6522declare i64 @llvm.atomic.swap.i64.p0i64( i64* &lt;ptr&gt;, i64 &lt;val&gt; )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6523
6524</pre>
6525<h5>Overview:</h5>
6526<p>
6527 This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
6528 the value from memory. It then stores the value in <tt>val</tt> in the memory
6529 at <tt>ptr</tt>.
6530</p>
6531<h5>Arguments:</h5>
6532
6533<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6534 The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both the
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6535 <tt>val</tt> argument and the result must be integers of the same bit width.
6536 The first argument, <tt>ptr</tt>, must be a pointer to a value of this
6537 integer type. The targets may only lower integer representations they
6538 support.
6539</p>
6540<h5>Semantics:</h5>
6541<p>
6542 This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
6543 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
6544 equivalent of an atomic swap operation within the SSA framework.
6545
6546</p>
6547<h5>Examples:</h5>
6548<pre>
6549%ptr = malloc i32
6550 store i32 4, %ptr
6551
6552%val1 = add i32 4, 4
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6553%result1 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val1 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6554 <i>; yields {i32}:result1 = 4</i>
6555%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
6556%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
6557
6558%val2 = add i32 1, 1
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6559%result2 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6560 <i>; yields {i32}:result2 = 8</i>
6561
6562%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
6563%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
6564</pre>
6565</div>
6566
6567<!-- _______________________________________________________________________ -->
6568<div class="doc_subsubsection">
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6569 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6570
6571</div>
6572<div class="doc_text">
6573<h5>Syntax:</h5>
6574<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6575 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]6576 integer bit width. Not all targets support all bit widths however.</p>
6577<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6578declare i8 @llvm.atomic.load.add.i8..p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6579declare i16 @llvm.atomic.load.add.i16..p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6580declare i32 @llvm.atomic.load.add.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6581declare 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]6582
6583</pre>
6584<h5>Overview:</h5>
6585<p>
6586 This intrinsic adds <tt>delta</tt> to the value stored in memory at
6587 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
6588</p>
6589<h5>Arguments:</h5>
6590<p>
6591
6592 The intrinsic takes two arguments, the first a pointer to an integer value
6593 and the second an integer value. The result is also an integer value. These
6594 integer types can have any bit width, but they must all have the same bit
6595 width. The targets may only lower integer representations they support.
6596</p>
6597<h5>Semantics:</h5>
6598<p>
6599 This intrinsic does a series of operations atomically. It first loads the
6600 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
6601 to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
6602</p>
6603
6604<h5>Examples:</h5>
6605<pre>
6606%ptr = malloc i32
6607 store i32 4, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6608%result1 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 4 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6609 <i>; yields {i32}:result1 = 4</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6610%result2 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6611 <i>; yields {i32}:result2 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6612%result3 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 5 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6613 <i>; yields {i32}:result3 = 10</i>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6614%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6615</pre>
6616</div>
6617
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6618<!-- _______________________________________________________________________ -->
6619<div class="doc_subsubsection">
6620 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
6621
6622</div>
6623<div class="doc_text">
6624<h5>Syntax:</h5>
6625<p>
6626 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]6627 any integer bit width and for different address spaces. Not all targets
6628 support all bit widths however.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6629<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6630declare i8 @llvm.atomic.load.sub.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6631declare i16 @llvm.atomic.load.sub.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6632declare i32 @llvm.atomic.load.sub.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6633declare 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]6634
6635</pre>
6636<h5>Overview:</h5>
6637<p>
6638 This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
6639 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
6640</p>
6641<h5>Arguments:</h5>
6642<p>
6643
6644 The intrinsic takes two arguments, the first a pointer to an integer value
6645 and the second an integer value. The result is also an integer value. These
6646 integer types can have any bit width, but they must all have the same bit
6647 width. The targets may only lower integer representations they support.
6648</p>
6649<h5>Semantics:</h5>
6650<p>
6651 This intrinsic does a series of operations atomically. It first loads the
6652 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
6653 result to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
6654</p>
6655
6656<h5>Examples:</h5>
6657<pre>
6658%ptr = malloc i32
6659 store i32 8, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6660%result1 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 4 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6661 <i>; yields {i32}:result1 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6662%result2 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 2 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6663 <i>; yields {i32}:result2 = 4</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6664%result3 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 5 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6665 <i>; yields {i32}:result3 = 2</i>
6666%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
6667</pre>
6668</div>
6669
6670<!-- _______________________________________________________________________ -->
6671<div class="doc_subsubsection">
6672 <a name="int_atomic_load_and">'<tt>llvm.atomic.load.and.*</tt>' Intrinsic</a><br>
6673 <a name="int_atomic_load_nand">'<tt>llvm.atomic.load.nand.*</tt>' Intrinsic</a><br>
6674 <a name="int_atomic_load_or">'<tt>llvm.atomic.load.or.*</tt>' Intrinsic</a><br>
6675 <a name="int_atomic_load_xor">'<tt>llvm.atomic.load.xor.*</tt>' Intrinsic</a><br>
6676
6677</div>
6678<div class="doc_text">
6679<h5>Syntax:</h5>
6680<p>
6681 These are overloaded intrinsics. You can use <tt>llvm.atomic.load_and</tt>,
6682 <tt>llvm.atomic.load_nand</tt>, <tt>llvm.atomic.load_or</tt>, and
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6683 <tt>llvm.atomic.load_xor</tt> on any integer bit width and for different
6684 address spaces. Not all targets support all bit widths however.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6685<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6686declare i8 @llvm.atomic.load.and.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6687declare i16 @llvm.atomic.load.and.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6688declare i32 @llvm.atomic.load.and.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6689declare 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]6690
6691</pre>
6692
6693<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6694declare i8 @llvm.atomic.load.or.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6695declare i16 @llvm.atomic.load.or.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6696declare i32 @llvm.atomic.load.or.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6697declare 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]6698
6699</pre>
6700
6701<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6702declare i8 @llvm.atomic.load.nand.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6703declare i16 @llvm.atomic.load.nand.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6704declare i32 @llvm.atomic.load.nand.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6705declare 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]6706
6707</pre>
6708
6709<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6710declare i8 @llvm.atomic.load.xor.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6711declare i16 @llvm.atomic.load.xor.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6712declare i32 @llvm.atomic.load.xor.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6713declare 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]6714
6715</pre>
6716<h5>Overview:</h5>
6717<p>
6718 These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
6719 the value stored in memory at <tt>ptr</tt>. It yields the original value
6720 at <tt>ptr</tt>.
6721</p>
6722<h5>Arguments:</h5>
6723<p>
6724
6725 These intrinsics take two arguments, the first a pointer to an integer value
6726 and the second an integer value. The result is also an integer value. These
6727 integer types can have any bit width, but they must all have the same bit
6728 width. The targets may only lower integer representations they support.
6729</p>
6730<h5>Semantics:</h5>
6731<p>
6732 These intrinsics does a series of operations atomically. They first load the
6733 value stored at <tt>ptr</tt>. They then do the bitwise operation
6734 <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the original
6735 value stored at <tt>ptr</tt>.
6736</p>
6737
6738<h5>Examples:</h5>
6739<pre>
6740%ptr = malloc i32
6741 store i32 0x0F0F, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6742%result0 = call i32 @llvm.atomic.load.nand.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6743 <i>; yields {i32}:result0 = 0x0F0F</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6744%result1 = call i32 @llvm.atomic.load.and.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6745 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6746%result2 = call i32 @llvm.atomic.load.or.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6747 <i>; yields {i32}:result2 = 0xF0</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6748%result3 = call i32 @llvm.atomic.load.xor.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6749 <i>; yields {i32}:result3 = FF</i>
6750%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
6751</pre>
6752</div>
6753
6754
6755<!-- _______________________________________________________________________ -->
6756<div class="doc_subsubsection">
6757 <a name="int_atomic_load_max">'<tt>llvm.atomic.load.max.*</tt>' Intrinsic</a><br>
6758 <a name="int_atomic_load_min">'<tt>llvm.atomic.load.min.*</tt>' Intrinsic</a><br>
6759 <a name="int_atomic_load_umax">'<tt>llvm.atomic.load.umax.*</tt>' Intrinsic</a><br>
6760 <a name="int_atomic_load_umin">'<tt>llvm.atomic.load.umin.*</tt>' Intrinsic</a><br>
6761
6762</div>
6763<div class="doc_text">
6764<h5>Syntax:</h5>
6765<p>
6766 These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
6767 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6768 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
6769 address spaces. Not all targets
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6770 support all bit widths however.</p>
6771<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6772declare i8 @llvm.atomic.load.max.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6773declare i16 @llvm.atomic.load.max.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6774declare i32 @llvm.atomic.load.max.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6775declare 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]6776
6777</pre>
6778
6779<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6780declare i8 @llvm.atomic.load.min.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6781declare i16 @llvm.atomic.load.min.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6782declare i32 @llvm.atomic.load.min.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6783declare 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]6784
6785</pre>
6786
6787<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6788declare i8 @llvm.atomic.load.umax.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6789declare i16 @llvm.atomic.load.umax.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6790declare i32 @llvm.atomic.load.umax.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6791declare 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]6792
6793</pre>
6794
6795<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6796declare i8 @llvm.atomic.load.umin.i8..p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6797declare i16 @llvm.atomic.load.umin.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6798declare i32 @llvm.atomic.load.umin.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6799declare 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]6800
6801</pre>
6802<h5>Overview:</h5>
6803<p>
6804 These intrinsics takes the signed or unsigned minimum or maximum of
6805 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
6806 original value at <tt>ptr</tt>.
6807</p>
6808<h5>Arguments:</h5>
6809<p>
6810
6811 These intrinsics take two arguments, the first a pointer to an integer value
6812 and the second an integer value. The result is also an integer value. These
6813 integer types can have any bit width, but they must all have the same bit
6814 width. The targets may only lower integer representations they support.
6815</p>
6816<h5>Semantics:</h5>
6817<p>
6818 These intrinsics does a series of operations atomically. They first load the
6819 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or max
6820 <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They yield
6821 the original value stored at <tt>ptr</tt>.
6822</p>
6823
6824<h5>Examples:</h5>
6825<pre>
6826%ptr = malloc i32
6827 store i32 7, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6828%result0 = call i32 @llvm.atomic.load.min.i32.p0i32( i32* %ptr, i32 -2 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6829 <i>; yields {i32}:result0 = 7</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6830%result1 = call i32 @llvm.atomic.load.max.i32.p0i32( i32* %ptr, i32 8 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6831 <i>; yields {i32}:result1 = -2</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6832%result2 = call i32 @llvm.atomic.load.umin.i32.p0i32( i32* %ptr, i32 10 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6833 <i>; yields {i32}:result2 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6834%result3 = call i32 @llvm.atomic.load.umax.i32.p0i32( i32* %ptr, i32 30 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6835 <i>; yields {i32}:result3 = 8</i>
6836%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
6837</pre>
6838</div>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6839
6840<!-- ======================================================================= -->
6841<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6842 <a name="int_general">General Intrinsics</a>
6843</div>
6844
6845<div class="doc_text">
6846<p> This class of intrinsics is designed to be generic and has
6847no specific purpose. </p>
6848</div>
6849
6850<!-- _______________________________________________________________________ -->
6851<div class="doc_subsubsection">
6852 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
6853</div>
6854
6855<div class="doc_text">
6856
6857<h5>Syntax:</h5>
6858<pre>
6859 declare void @llvm.var.annotation(i8* &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6860</pre>
6861
6862<h5>Overview:</h5>
6863
6864<p>
6865The '<tt>llvm.var.annotation</tt>' intrinsic
6866</p>
6867
6868<h5>Arguments:</h5>
6869
6870<p>
6871The first argument is a pointer to a value, the second is a pointer to a
6872global string, the third is a pointer to a global string which is the source
6873file name, and the last argument is the line number.
6874</p>
6875
6876<h5>Semantics:</h5>
6877
6878<p>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]6879This intrinsic allows annotation of local variables with arbitrary strings.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6880This can be useful for special purpose optimizations that want to look for these
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]6881annotations. These have no other defined use, they are ignored by code
6882generation and optimization.
6883</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6884</div>
6885
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6886<!-- _______________________________________________________________________ -->
6887<div class="doc_subsubsection">
Tanya Lattnerc9869b12007-09-21 23:57:59 +0000[diff] [blame]6888 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6889</div>
6890
6891<div class="doc_text">
6892
6893<h5>Syntax:</h5>
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]6894<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
6895any integer bit width.
6896</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6897<pre>
Tanya Lattner09161fe2007-09-22 00:03:01 +0000[diff] [blame]6898 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6899 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6900 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6901 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6902 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]6903</pre>
6904
6905<h5>Overview:</h5>
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]6906
6907<p>
6908The '<tt>llvm.annotation</tt>' intrinsic.
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6909</p>
6910
6911<h5>Arguments:</h5>
6912
6913<p>
6914The first argument is an integer value (result of some expression),
6915the second is a pointer to a global string, the third is a pointer to a global
6916string which is the source file name, and the last argument is the line number.
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]6917It returns the value of the first argument.
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6918</p>
6919
6920<h5>Semantics:</h5>
6921
6922<p>
6923This intrinsic allows annotations to be put on arbitrary expressions
6924with arbitrary strings. This can be useful for special purpose optimizations
6925that want to look for these annotations. These have no other defined use, they
6926are ignored by code generation and optimization.
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]6927</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6928</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6929
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]6930<!-- _______________________________________________________________________ -->
6931<div class="doc_subsubsection">
6932 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
6933</div>
6934
6935<div class="doc_text">
6936
6937<h5>Syntax:</h5>
6938<pre>
6939 declare void @llvm.trap()
6940</pre>
6941
6942<h5>Overview:</h5>
6943
6944<p>
6945The '<tt>llvm.trap</tt>' intrinsic
6946</p>
6947
6948<h5>Arguments:</h5>
6949
6950<p>
6951None
6952</p>
6953
6954<h5>Semantics:</h5>
6955
6956<p>
6957This intrinsics is lowered to the target dependent trap instruction. If the
6958target does not have a trap instruction, this intrinsic will be lowered to the
6959call of the abort() function.
6960</p>
6961</div>
6962
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]6963<!-- _______________________________________________________________________ -->
6964<div class="doc_subsubsection">
Misha Brukman5dd7f4d2008-11-22 23:55:29 +0000[diff] [blame]6965 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]6966</div>
6967<div class="doc_text">
6968<h5>Syntax:</h5>
6969<pre>
6970declare void @llvm.stackprotector( i8* &lt;guard&gt;, i8** &lt;slot&gt; )
6971
6972</pre>
6973<h5>Overview:</h5>
6974<p>
6975 The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and stores
6976 it onto the stack at <tt>slot</tt>. The stack slot is adjusted to ensure that
6977 it is placed on the stack before local variables.
6978</p>
6979<h5>Arguments:</h5>
6980<p>
6981 The <tt>llvm.stackprotector</tt> intrinsic requires two pointer arguments. The
6982 first argument is the value loaded from the stack guard
6983 <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt> that
6984 has enough space to hold the value of the guard.
6985</p>
6986<h5>Semantics:</h5>
6987<p>
6988 This intrinsic causes the prologue/epilogue inserter to force the position of
6989 the <tt>AllocaInst</tt> stack slot to be before local variables on the
6990 stack. This is to ensure that if a local variable on the stack is overwritten,
6991 it will destroy the value of the guard. When the function exits, the guard on
6992 the stack is checked against the original guard. If they're different, then
6993 the program aborts by calling the <tt>__stack_chk_fail()</tt> function.
6994</p>
6995</div>
6996
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6997<!-- *********************************************************************** -->
6998<hr>
6999<address>
7000 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman947321d2008-12-11 17:34:48 +0000[diff] [blame]7001 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7002 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman947321d2008-12-11 17:34:48 +0000[diff] [blame]7003 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7004
7005 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
7006 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
7007 Last modified: $Date$
7008</address>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]7009
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7010</body>
7011</html>