Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / llvm / 96edbd34c7b74589956c8a877ed145cd38de7fee / . / docs / LangRef.html
blob: dc6d7f2dae10335ee9b16b44a61c36b299e8527d [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>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]44 </ol>
45 </li>
46 <li><a href="#t_derived">Derived Types</a>
47 <ol>
Chris Lattner251ab812007-12-18 06:18:21 +0000[diff] [blame]48 <li><a href="#t_integer">Integer Type</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]49 <li><a href="#t_array">Array Type</a></li>
50 <li><a href="#t_function">Function Type</a></li>
51 <li><a href="#t_pointer">Pointer Type</a></li>
52 <li><a href="#t_struct">Structure Type</a></li>
53 <li><a href="#t_pstruct">Packed Structure Type</a></li>
54 <li><a href="#t_vector">Vector Type</a></li>
55 <li><a href="#t_opaque">Opaque Type</a></li>
56 </ol>
57 </li>
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]58 <li><a href="#t_uprefs">Type Up-references</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]59 </ol>
60 </li>
61 <li><a href="#constants">Constants</a>
62 <ol>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]63 <li><a href="#simpleconstants">Simple Constants</a></li>
64 <li><a href="#aggregateconstants">Aggregate Constants</a></li>
65 <li><a href="#globalconstants">Global Variable and Function Addresses</a></li>
66 <li><a href="#undefvalues">Undefined Values</a></li>
67 <li><a href="#constantexprs">Constant Expressions</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]68 </ol>
69 </li>
70 <li><a href="#othervalues">Other Values</a>
71 <ol>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]72 <li><a href="#inlineasm">Inline Assembler Expressions</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]73 </ol>
74 </li>
75 <li><a href="#instref">Instruction Reference</a>
76 <ol>
77 <li><a href="#terminators">Terminator Instructions</a>
78 <ol>
79 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
80 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
81 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
82 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
83 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
84 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
85 </ol>
86 </li>
87 <li><a href="#binaryops">Binary Operations</a>
88 <ol>
89 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
90 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
91 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
92 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
93 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
94 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
95 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
96 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
97 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
98 </ol>
99 </li>
100 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
101 <ol>
102 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
103 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
104 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
105 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
106 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
107 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
108 </ol>
109 </li>
110 <li><a href="#vectorops">Vector Operations</a>
111 <ol>
112 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
113 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
114 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
115 </ol>
116 </li>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]117 <li><a href="#aggregateops">Aggregate Operations</a>
118 <ol>
119 <li><a href="#i_extractvalue">'<tt>extractvalue</tt>' Instruction</a></li>
120 <li><a href="#i_insertvalue">'<tt>insertvalue</tt>' Instruction</a></li>
121 </ol>
122 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]123 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
124 <ol>
125 <li><a href="#i_malloc">'<tt>malloc</tt>' Instruction</a></li>
126 <li><a href="#i_free">'<tt>free</tt>' Instruction</a></li>
127 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
128 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
129 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
130 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
131 </ol>
132 </li>
133 <li><a href="#convertops">Conversion Operations</a>
134 <ol>
135 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
136 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
137 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
138 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
139 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
140 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
141 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
142 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
143 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
144 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
145 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
146 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
147 </ol>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]148 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]149 <li><a href="#otherops">Other Operations</a>
150 <ol>
151 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
152 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]153 <li><a href="#i_vicmp">'<tt>vicmp</tt>' Instruction</a></li>
154 <li><a href="#i_vfcmp">'<tt>vfcmp</tt>' Instruction</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]155 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
156 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
157 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
158 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
159 </ol>
160 </li>
161 </ol>
162 </li>
163 <li><a href="#intrinsics">Intrinsic Functions</a>
164 <ol>
165 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
166 <ol>
167 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
168 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
169 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
170 </ol>
171 </li>
172 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
173 <ol>
174 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
175 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
176 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
177 </ol>
178 </li>
179 <li><a href="#int_codegen">Code Generator Intrinsics</a>
180 <ol>
181 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
182 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
183 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
184 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
185 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
186 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
187 <li><a href="#int_readcyclecounter"><tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
188 </ol>
189 </li>
190 <li><a href="#int_libc">Standard C Library Intrinsics</a>
191 <ol>
192 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
193 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
194 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
195 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
196 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]197 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
198 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
199 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]200 </ol>
201 </li>
202 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
203 <ol>
204 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
205 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
206 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
207 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
208 <li><a href="#int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic </a></li>
209 <li><a href="#int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic </a></li>
210 </ol>
211 </li>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]212 <li><a href="#int_overflow">Arithmetic with Overflow Intrinsics</a>
213 <ol>
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]214 <li><a href="#int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt> Intrinsics</a></li>
215 <li><a href="#int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt> Intrinsics</a></li>
216 <li><a href="#int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt> Intrinsics</a></li>
217 <li><a href="#int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt> Intrinsics</a></li>
218 <li><a href="#int_smul_overflow">'<tt>llvm.smul.with.overflow.*</tt> Intrinsics</a></li>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]219 </ol>
220 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]221 <li><a href="#int_debugger">Debugger intrinsics</a></li>
222 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]223 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]224 <ol>
225 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]226 </ol>
227 </li>
Bill Wendling9127adb2008-11-18 22:10:53 +0000[diff] [blame]228 <li><a href="#int_atomics">Atomic intrinsics</a>
229 <ol>
230 <li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
231 <li><a href="#int_atomic_cmp_swap"><tt>llvm.atomic.cmp.swap</tt></a></li>
232 <li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
233 <li><a href="#int_atomic_load_add"><tt>llvm.atomic.load.add</tt></a></li>
234 <li><a href="#int_atomic_load_sub"><tt>llvm.atomic.load.sub</tt></a></li>
235 <li><a href="#int_atomic_load_and"><tt>llvm.atomic.load.and</tt></a></li>
236 <li><a href="#int_atomic_load_nand"><tt>llvm.atomic.load.nand</tt></a></li>
237 <li><a href="#int_atomic_load_or"><tt>llvm.atomic.load.or</tt></a></li>
238 <li><a href="#int_atomic_load_xor"><tt>llvm.atomic.load.xor</tt></a></li>
239 <li><a href="#int_atomic_load_max"><tt>llvm.atomic.load.max</tt></a></li>
240 <li><a href="#int_atomic_load_min"><tt>llvm.atomic.load.min</tt></a></li>
241 <li><a href="#int_atomic_load_umax"><tt>llvm.atomic.load.umax</tt></a></li>
242 <li><a href="#int_atomic_load_umin"><tt>llvm.atomic.load.umin</tt></a></li>
243 </ol>
244 </li>
Reid Spencerb043f672007-07-20 19:59:11 +0000[diff] [blame]245 <li><a href="#int_general">General intrinsics</a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]246 <ol>
Reid Spencerb043f672007-07-20 19:59:11 +0000[diff] [blame]247 <li><a href="#int_var_annotation">
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]248 '<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]249 <li><a href="#int_annotation">
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]250 '<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]251 <li><a href="#int_trap">
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]252 '<tt>llvm.trap</tt>' Intrinsic</a></li>
253 <li><a href="#int_stackprotector">
254 '<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]255 </ol>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]256 </li>
257 </ol>
258 </li>
259</ol>
260
261<div class="doc_author">
262 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
263 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
264</div>
265
266<!-- *********************************************************************** -->
267<div class="doc_section"> <a name="abstract">Abstract </a></div>
268<!-- *********************************************************************** -->
269
270<div class="doc_text">
271<p>This document is a reference manual for the LLVM assembly language.
Bill Wendlinge7846a52008-08-05 22:29:16 +0000[diff] [blame]272LLVM is a Static Single Assignment (SSA) based representation that provides
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]273type safety, low-level operations, flexibility, and the capability of
274representing 'all' high-level languages cleanly. It is the common code
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]275representation used throughout all phases of the LLVM compilation
276strategy.</p>
277</div>
278
279<!-- *********************************************************************** -->
280<div class="doc_section"> <a name="introduction">Introduction</a> </div>
281<!-- *********************************************************************** -->
282
283<div class="doc_text">
284
285<p>The LLVM code representation is designed to be used in three
286different forms: as an in-memory compiler IR, as an on-disk bitcode
287representation (suitable for fast loading by a Just-In-Time compiler),
288and as a human readable assembly language representation. This allows
289LLVM to provide a powerful intermediate representation for efficient
290compiler transformations and analysis, while providing a natural means
291to debug and visualize the transformations. The three different forms
292of LLVM are all equivalent. This document describes the human readable
293representation and notation.</p>
294
295<p>The LLVM representation aims to be light-weight and low-level
296while being expressive, typed, and extensible at the same time. It
297aims to be a "universal IR" of sorts, by being at a low enough level
298that high-level ideas may be cleanly mapped to it (similar to how
299microprocessors are "universal IR's", allowing many source languages to
300be mapped to them). By providing type information, LLVM can be used as
301the target of optimizations: for example, through pointer analysis, it
302can be proven that a C automatic variable is never accessed outside of
303the current function... allowing it to be promoted to a simple SSA
304value instead of a memory location.</p>
305
306</div>
307
308<!-- _______________________________________________________________________ -->
309<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
310
311<div class="doc_text">
312
313<p>It is important to note that this document describes 'well formed'
314LLVM assembly language. There is a difference between what the parser
315accepts and what is considered 'well formed'. For example, the
316following instruction is syntactically okay, but not well formed:</p>
317
318<div class="doc_code">
319<pre>
320%x = <a href="#i_add">add</a> i32 1, %x
321</pre>
322</div>
323
324<p>...because the definition of <tt>%x</tt> does not dominate all of
325its uses. The LLVM infrastructure provides a verification pass that may
326be used to verify that an LLVM module is well formed. This pass is
327automatically run by the parser after parsing input assembly and by
328the optimizer before it outputs bitcode. The violations pointed out
329by the verifier pass indicate bugs in transformation passes or input to
330the parser.</p>
331</div>
332
Chris Lattnera83fdc02007-10-03 17:34:29 +0000[diff] [blame]333<!-- Describe the typesetting conventions here. -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]334
335<!-- *********************************************************************** -->
336<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
337<!-- *********************************************************************** -->
338
339<div class="doc_text">
340
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]341 <p>LLVM identifiers come in two basic types: global and local. Global
342 identifiers (functions, global variables) begin with the @ character. Local
343 identifiers (register names, types) begin with the % character. Additionally,
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]344 there are three different formats for identifiers, for different purposes:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]345
346<ol>
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]347 <li>Named values are represented as a string of characters with their prefix.
348 For example, %foo, @DivisionByZero, %a.really.long.identifier. The actual
349 regular expression used is '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]350 Identifiers which require other characters in their names can be surrounded
Daniel Dunbar73f9a772008-10-14 23:51:43 +0000[diff] [blame]351 with quotes. Special characters may be escaped using "\xx" where xx is the
352 ASCII code for the character in hexadecimal. In this way, any character can
353 be used in a name value, even quotes themselves.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]354
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]355 <li>Unnamed values are represented as an unsigned numeric value with their
356 prefix. For example, %12, @2, %44.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]357
358 <li>Constants, which are described in a <a href="#constants">section about
359 constants</a>, below.</li>
360</ol>
361
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]362<p>LLVM requires that values start with a prefix for two reasons: Compilers
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]363don't need to worry about name clashes with reserved words, and the set of
364reserved words may be expanded in the future without penalty. Additionally,
365unnamed identifiers allow a compiler to quickly come up with a temporary
366variable without having to avoid symbol table conflicts.</p>
367
368<p>Reserved words in LLVM are very similar to reserved words in other
369languages. There are keywords for different opcodes
370('<tt><a href="#i_add">add</a></tt>',
371 '<tt><a href="#i_bitcast">bitcast</a></tt>',
372 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
373href="#t_void">void</a></tt>', '<tt><a href="#t_primitive">i32</a></tt>', etc...),
374and others. These reserved words cannot conflict with variable names, because
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]375none of them start with a prefix character ('%' or '@').</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]376
377<p>Here is an example of LLVM code to multiply the integer variable
378'<tt>%X</tt>' by 8:</p>
379
380<p>The easy way:</p>
381
382<div class="doc_code">
383<pre>
384%result = <a href="#i_mul">mul</a> i32 %X, 8
385</pre>
386</div>
387
388<p>After strength reduction:</p>
389
390<div class="doc_code">
391<pre>
392%result = <a href="#i_shl">shl</a> i32 %X, i8 3
393</pre>
394</div>
395
396<p>And the hard way:</p>
397
398<div class="doc_code">
399<pre>
400<a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
401<a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
402%result = <a href="#i_add">add</a> i32 %1, %1
403</pre>
404</div>
405
406<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
407important lexical features of LLVM:</p>
408
409<ol>
410
411 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
412 line.</li>
413
414 <li>Unnamed temporaries are created when the result of a computation is not
415 assigned to a named value.</li>
416
417 <li>Unnamed temporaries are numbered sequentially</li>
418
419</ol>
420
421<p>...and it also shows a convention that we follow in this document. When
422demonstrating instructions, we will follow an instruction with a comment that
423defines the type and name of value produced. Comments are shown in italic
424text.</p>
425
426</div>
427
428<!-- *********************************************************************** -->
429<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
430<!-- *********************************************************************** -->
431
432<!-- ======================================================================= -->
433<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
434</div>
435
436<div class="doc_text">
437
438<p>LLVM programs are composed of "Module"s, each of which is a
439translation unit of the input programs. Each module consists of
440functions, global variables, and symbol table entries. Modules may be
441combined together with the LLVM linker, which merges function (and
442global variable) definitions, resolves forward declarations, and merges
443symbol table entries. Here is an example of the "hello world" module:</p>
444
445<div class="doc_code">
446<pre><i>; Declare the string constant as a global constant...</i>
447<a href="#identifiers">@.LC0</a> = <a href="#linkage_internal">internal</a> <a
448 href="#globalvars">constant</a> <a href="#t_array">[13 x i8]</a> c"hello world\0A\00" <i>; [13 x i8]*</i>
449
450<i>; External declaration of the puts function</i>
451<a href="#functionstructure">declare</a> i32 @puts(i8 *) <i>; i32(i8 *)* </i>
452
453<i>; Definition of main function</i>
454define i32 @main() { <i>; i32()* </i>
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]455 <i>; Convert [13 x i8]* to i8 *...</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]456 %cast210 = <a
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]457 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]458
459 <i>; Call puts function to write out the string to stdout...</i>
460 <a
461 href="#i_call">call</a> i32 @puts(i8 * %cast210) <i>; i32</i>
462 <a
463 href="#i_ret">ret</a> i32 0<br>}<br>
464</pre>
465</div>
466
467<p>This example is made up of a <a href="#globalvars">global variable</a>
468named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
469function, and a <a href="#functionstructure">function definition</a>
470for "<tt>main</tt>".</p>
471
472<p>In general, a module is made up of a list of global values,
473where both functions and global variables are global values. Global values are
474represented by a pointer to a memory location (in this case, a pointer to an
475array of char, and a pointer to a function), and have one of the following <a
476href="#linkage">linkage types</a>.</p>
477
478</div>
479
480<!-- ======================================================================= -->
481<div class="doc_subsection">
482 <a name="linkage">Linkage Types</a>
483</div>
484
485<div class="doc_text">
486
487<p>
488All Global Variables and Functions have one of the following types of linkage:
489</p>
490
491<dl>
492
Rafael Espindolaa168fc92009-01-15 20:18:42 +0000[diff] [blame]493 <dt><tt><b><a name="linkage_private">private</a></b></tt>: </dt>
494
495 <dd>Global values with private linkage are only directly accessible by
496 objects in the current module. In particular, linking code into a module with
497 an private global value may cause the private to be renamed as necessary to
498 avoid collisions. Because the symbol is private to the module, all
499 references can be updated. This doesn't show up in any symbol table in the
500 object file.
501 </dd>
502
Dale Johannesen96e7e092008-05-23 23:13:41 +0000[diff] [blame]503 <dt><tt><b><a name="linkage_internal">internal</a></b></tt>: </dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]504
Duncan Sandsa75223a2009-01-16 09:29:46 +0000[diff] [blame]505 <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]506 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]507 '<tt>static</tt>' keyword in C.
508 </dd>
509
510 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
511
512 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
513 the same name when linkage occurs. This is typically used to implement
514 inline functions, templates, or other code which must be generated in each
515 translation unit that uses it. Unreferenced <tt>linkonce</tt> globals are
516 allowed to be discarded.
517 </dd>
518
Dale Johannesen96e7e092008-05-23 23:13:41 +0000[diff] [blame]519 <dt><tt><b><a name="linkage_common">common</a></b></tt>: </dt>
520
521 <dd>"<tt>common</tt>" linkage is exactly the same as <tt>linkonce</tt>
522 linkage, except that unreferenced <tt>common</tt> globals may not be
523 discarded. This is used for globals that may be emitted in multiple
524 translation units, but that are not guaranteed to be emitted into every
525 translation unit that uses them. One example of this is tentative
526 definitions in C, such as "<tt>int X;</tt>" at global scope.
527 </dd>
528
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]529 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
530
Dale Johannesen96e7e092008-05-23 23:13:41 +0000[diff] [blame]531 <dd>"<tt>weak</tt>" linkage is the same as <tt>common</tt> linkage, except
532 that some targets may choose to emit different assembly sequences for them
533 for target-dependent reasons. This is used for globals that are declared
534 "weak" in C source code.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]535 </dd>
536
537 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
538
539 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
540 pointer to array type. When two global variables with appending linkage are
541 linked together, the two global arrays are appended together. This is the
542 LLVM, typesafe, equivalent of having the system linker append together
543 "sections" with identical names when .o files are linked.
544 </dd>
545
546 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt>: </dt>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]547 <dd>The semantics of this linkage follow the ELF object file model: the
548 symbol is weak until linked, if not linked, the symbol becomes null instead
549 of being an undefined reference.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]550 </dd>
551
552 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
553
554 <dd>If none of the above identifiers are used, the global is externally
555 visible, meaning that it participates in linkage and can be used to resolve
556 external symbol references.
557 </dd>
558</dl>
559
560 <p>
561 The next two types of linkage are targeted for Microsoft Windows platform
562 only. They are designed to support importing (exporting) symbols from (to)
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]563 DLLs (Dynamic Link Libraries).
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]564 </p>
565
566 <dl>
567 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt>: </dt>
568
569 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
570 or variable via a global pointer to a pointer that is set up by the DLL
571 exporting the symbol. On Microsoft Windows targets, the pointer name is
Dan Gohman5ec99832009-01-12 21:35:55 +0000[diff] [blame]572 formed by combining <code>__imp_</code> and the function or variable name.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]573 </dd>
574
575 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt>: </dt>
576
577 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
578 pointer to a pointer in a DLL, so that it can be referenced with the
579 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
Dan Gohman5ec99832009-01-12 21:35:55 +0000[diff] [blame]580 name is formed by combining <code>__imp_</code> and the function or variable
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]581 name.
582 </dd>
583
584</dl>
585
Dan Gohman4dfac702008-11-24 17:18:39 +0000[diff] [blame]586<p>For example, since the "<tt>.LC0</tt>"
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]587variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
588variable and was linked with this one, one of the two would be renamed,
589preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
590external (i.e., lacking any linkage declarations), they are accessible
591outside of the current module.</p>
592<p>It is illegal for a function <i>declaration</i>
593to have any linkage type other than "externally visible", <tt>dllimport</tt>,
594or <tt>extern_weak</tt>.</p>
595<p>Aliases can have only <tt>external</tt>, <tt>internal</tt> and <tt>weak</tt>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]596linkages.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]597</div>
598
599<!-- ======================================================================= -->
600<div class="doc_subsection">
601 <a name="callingconv">Calling Conventions</a>
602</div>
603
604<div class="doc_text">
605
606<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
607and <a href="#i_invoke">invokes</a> can all have an optional calling convention
608specified for the call. The calling convention of any pair of dynamic
609caller/callee must match, or the behavior of the program is undefined. The
610following calling conventions are supported by LLVM, and more may be added in
611the future:</p>
612
613<dl>
614 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
615
616 <dd>This calling convention (the default if no other calling convention is
617 specified) matches the target C calling conventions. This calling convention
618 supports varargs function calls and tolerates some mismatch in the declared
619 prototype and implemented declaration of the function (as does normal C).
620 </dd>
621
622 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
623
624 <dd>This calling convention attempts to make calls as fast as possible
625 (e.g. by passing things in registers). This calling convention allows the
626 target to use whatever tricks it wants to produce fast code for the target,
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]627 without having to conform to an externally specified ABI (Application Binary
628 Interface). Implementations of this convention should allow arbitrary
Arnold Schwaighofer07444922008-05-14 09:17:12 +0000[diff] [blame]629 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> to be
630 supported. This calling convention does not support varargs and requires the
631 prototype of all callees to exactly match the prototype of the function
632 definition.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]633 </dd>
634
635 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
636
637 <dd>This calling convention attempts to make code in the caller as efficient
638 as possible under the assumption that the call is not commonly executed. As
639 such, these calls often preserve all registers so that the call does not break
640 any live ranges in the caller side. This calling convention does not support
641 varargs and requires the prototype of all callees to exactly match the
642 prototype of the function definition.
643 </dd>
644
645 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
646
647 <dd>Any calling convention may be specified by number, allowing
648 target-specific calling conventions to be used. Target specific calling
649 conventions start at 64.
650 </dd>
651</dl>
652
653<p>More calling conventions can be added/defined on an as-needed basis, to
654support pascal conventions or any other well-known target-independent
655convention.</p>
656
657</div>
658
659<!-- ======================================================================= -->
660<div class="doc_subsection">
661 <a name="visibility">Visibility Styles</a>
662</div>
663
664<div class="doc_text">
665
666<p>
667All Global Variables and Functions have one of the following visibility styles:
668</p>
669
670<dl>
671 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
672
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]673 <dd>On targets that use the ELF object file format, default visibility means
674 that the declaration is visible to other
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]675 modules and, in shared libraries, means that the declared entity may be
676 overridden. On Darwin, default visibility means that the declaration is
677 visible to other modules. Default visibility corresponds to "external
678 linkage" in the language.
679 </dd>
680
681 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
682
683 <dd>Two declarations of an object with hidden visibility refer to the same
684 object if they are in the same shared object. Usually, hidden visibility
685 indicates that the symbol will not be placed into the dynamic symbol table,
686 so no other module (executable or shared library) can reference it
687 directly.
688 </dd>
689
690 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
691
692 <dd>On ELF, protected visibility indicates that the symbol will be placed in
693 the dynamic symbol table, but that references within the defining module will
694 bind to the local symbol. That is, the symbol cannot be overridden by another
695 module.
696 </dd>
697</dl>
698
699</div>
700
701<!-- ======================================================================= -->
702<div class="doc_subsection">
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]703 <a name="namedtypes">Named Types</a>
704</div>
705
706<div class="doc_text">
707
708<p>LLVM IR allows you to specify name aliases for certain types. This can make
709it easier to read the IR and make the IR more condensed (particularly when
710recursive types are involved). An example of a name specification is:
711</p>
712
713<div class="doc_code">
714<pre>
715%mytype = type { %mytype*, i32 }
716</pre>
717</div>
718
719<p>You may give a name to any <a href="#typesystem">type</a> except "<a
720href="t_void">void</a>". Type name aliases may be used anywhere a type is
721expected with the syntax "%mytype".</p>
722
723<p>Note that type names are aliases for the structural type that they indicate,
724and that you can therefore specify multiple names for the same type. This often
725leads to confusing behavior when dumping out a .ll file. Since LLVM IR uses
726structural typing, the name is not part of the type. When printing out LLVM IR,
727the printer will pick <em>one name</em> to render all types of a particular
728shape. This means that if you have code where two different source types end up
729having the same LLVM type, that the dumper will sometimes print the "wrong" or
730unexpected type. This is an important design point and isn't going to
731change.</p>
732
733</div>
734
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]735<!-- ======================================================================= -->
736<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]737 <a name="globalvars">Global Variables</a>
738</div>
739
740<div class="doc_text">
741
742<p>Global variables define regions of memory allocated at compilation time
743instead of run-time. Global variables may optionally be initialized, may have
744an explicit section to be placed in, and may have an optional explicit alignment
745specified. A variable may be defined as "thread_local", which means that it
746will not be shared by threads (each thread will have a separated copy of the
747variable). A variable may be defined as a global "constant," which indicates
748that the contents of the variable will <b>never</b> be modified (enabling better
749optimization, allowing the global data to be placed in the read-only section of
750an executable, etc). Note that variables that need runtime initialization
751cannot be marked "constant" as there is a store to the variable.</p>
752
753<p>
754LLVM explicitly allows <em>declarations</em> of global variables to be marked
755constant, even if the final definition of the global is not. This capability
756can be used to enable slightly better optimization of the program, but requires
757the language definition to guarantee that optimizations based on the
758'constantness' are valid for the translation units that do not include the
759definition.
760</p>
761
762<p>As SSA values, global variables define pointer values that are in
763scope (i.e. they dominate) all basic blocks in the program. Global
764variables always define a pointer to their "content" type because they
765describe a region of memory, and all memory objects in LLVM are
766accessed through pointers.</p>
767
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]768<p>A global variable may be declared to reside in a target-specifc numbered
769address space. For targets that support them, address spaces may affect how
770optimizations are performed and/or what target instructions are used to access
Christopher Lamb20a39e92007-12-12 08:44:39 +0000[diff] [blame]771the variable. The default address space is zero. The address space qualifier
772must precede any other attributes.</p>
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]773
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]774<p>LLVM allows an explicit section to be specified for globals. If the target
775supports it, it will emit globals to the section specified.</p>
776
777<p>An explicit alignment may be specified for a global. If not present, or if
778the alignment is set to zero, the alignment of the global is set by the target
779to whatever it feels convenient. If an explicit alignment is specified, the
780global is forced to have at least that much alignment. All alignments must be
781a power of 2.</p>
782
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]783<p>For example, the following defines a global in a numbered address space with
784an initializer, section, and alignment:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]785
786<div class="doc_code">
787<pre>
Dan Gohman21ef02c2009-01-11 00:40:00 +0000[diff] [blame]788@G = addrspace(5) constant float 1.0, section "foo", align 4
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]789</pre>
790</div>
791
792</div>
793
794
795<!-- ======================================================================= -->
796<div class="doc_subsection">
797 <a name="functionstructure">Functions</a>
798</div>
799
800<div class="doc_text">
801
802<p>LLVM function definitions consist of the "<tt>define</tt>" keyord,
803an optional <a href="#linkage">linkage type</a>, an optional
804<a href="#visibility">visibility style</a>, an optional
805<a href="#callingconv">calling convention</a>, a return type, an optional
806<a href="#paramattrs">parameter attribute</a> for the return type, a function
807name, a (possibly empty) argument list (each with optional
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]808<a href="#paramattrs">parameter attributes</a>), optional
809<a href="#fnattrs">function attributes</a>, an optional section,
810an optional alignment, an optional <a href="#gc">garbage collector name</a>,
Chris Lattneree202542008-10-04 18:10:21 +0000[diff] [blame]811an opening curly brace, a list of basic blocks, and a closing curly brace.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]812
813LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
814optional <a href="#linkage">linkage type</a>, an optional
815<a href="#visibility">visibility style</a>, an optional
816<a href="#callingconv">calling convention</a>, a return type, an optional
817<a href="#paramattrs">parameter attribute</a> for the return type, a function
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]818name, a possibly empty list of arguments, an optional alignment, and an optional
Gordon Henriksend606f5b2007-12-10 03:30:21 +0000[diff] [blame]819<a href="#gc">garbage collector name</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]820
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]821<p>A function definition contains a list of basic blocks, forming the CFG
822(Control Flow Graph) for
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]823the function. Each basic block may optionally start with a label (giving the
824basic block a symbol table entry), contains a list of instructions, and ends
825with a <a href="#terminators">terminator</a> instruction (such as a branch or
826function return).</p>
827
828<p>The first basic block in a function is special in two ways: it is immediately
829executed on entrance to the function, and it is not allowed to have predecessor
830basic blocks (i.e. there can not be any branches to the entry block of a
831function). Because the block can have no predecessors, it also cannot have any
832<a href="#i_phi">PHI nodes</a>.</p>
833
834<p>LLVM allows an explicit section to be specified for functions. If the target
835supports it, it will emit functions to the section specified.</p>
836
837<p>An explicit alignment may be specified for a function. If not present, or if
838the alignment is set to zero, the alignment of the function is set by the target
839to whatever it feels convenient. If an explicit alignment is specified, the
840function is forced to have at least that much alignment. All alignments must be
841a power of 2.</p>
842
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]843 <h5>Syntax:</h5>
844
845<div class="doc_code">
Chris Lattner1e5c5cd02008-10-13 16:55:18 +0000[diff] [blame]846<tt>
847define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
848 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
849 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
850 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
851 [<a href="#gc">gc</a>] { ... }
852</tt>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]853</div>
854
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]855</div>
856
857
858<!-- ======================================================================= -->
859<div class="doc_subsection">
860 <a name="aliasstructure">Aliases</a>
861</div>
862<div class="doc_text">
863 <p>Aliases act as "second name" for the aliasee value (which can be either
Anton Korobeynikov96822812008-03-22 08:36:14 +0000[diff] [blame]864 function, global variable, another alias or bitcast of global value). Aliases
865 may have an optional <a href="#linkage">linkage type</a>, and an
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]866 optional <a href="#visibility">visibility style</a>.</p>
867
868 <h5>Syntax:</h5>
869
870<div class="doc_code">
871<pre>
Duncan Sandsd7bfabf2008-09-12 20:48:21 +0000[diff] [blame]872@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]873</pre>
874</div>
875
876</div>
877
878
879
880<!-- ======================================================================= -->
881<div class="doc_subsection"><a name="paramattrs">Parameter Attributes</a></div>
882<div class="doc_text">
883 <p>The return type and each parameter of a function type may have a set of
884 <i>parameter attributes</i> associated with them. Parameter attributes are
885 used to communicate additional information about the result or parameters of
Duncan Sandsf5588dc2007-11-27 13:23:08 +0000[diff] [blame]886 a function. Parameter attributes are considered to be part of the function,
887 not of the function type, so functions with different parameter attributes
888 can have the same function type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]889
890 <p>Parameter attributes are simple keywords that follow the type specified. If
891 multiple parameter attributes are needed, they are space separated. For
892 example:</p>
893
894<div class="doc_code">
895<pre>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]896declare i32 @printf(i8* noalias , ...)
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]897declare i32 @atoi(i8 zeroext)
898declare signext i8 @returns_signed_char()
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]899</pre>
900</div>
901
Duncan Sandsf5588dc2007-11-27 13:23:08 +0000[diff] [blame]902 <p>Note that any attributes for the function result (<tt>nounwind</tt>,
903 <tt>readonly</tt>) come immediately after the argument list.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]904
905 <p>Currently, only the following parameter attributes are defined:</p>
906 <dl>
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]907 <dt><tt>zeroext</tt></dt>
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]908 <dd>This indicates to the code generator that the parameter or return value
909 should be zero-extended to a 32-bit value by the caller (for a parameter)
910 or the callee (for a return value).</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]911
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]912 <dt><tt>signext</tt></dt>
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]913 <dd>This indicates to the code generator that the parameter or return value
914 should be sign-extended to a 32-bit value by the caller (for a parameter)
915 or the callee (for a return value).</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]916
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]917 <dt><tt>inreg</tt></dt>
Dale Johannesenc08a0e22008-09-25 20:47:45 +0000[diff] [blame]918 <dd>This indicates that this parameter or return value should be treated
919 in a special target-dependent fashion during while emitting code for a
920 function call or return (usually, by putting it in a register as opposed
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]921 to memory, though some targets use it to distinguish between two different
922 kinds of registers). Use of this attribute is target-specific.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]923
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]924 <dt><tt><a name="byval">byval</a></tt></dt>
Chris Lattner04c86182008-01-15 04:34:22 +0000[diff] [blame]925 <dd>This indicates that the pointer parameter should really be passed by
926 value to the function. The attribute implies that a hidden copy of the
927 pointee is made between the caller and the callee, so the callee is unable
Chris Lattner6a9f3c42008-08-05 18:21:08 +0000[diff] [blame]928 to modify the value in the callee. This attribute is only valid on LLVM
Chris Lattner04c86182008-01-15 04:34:22 +0000[diff] [blame]929 pointer arguments. It is generally used to pass structs and arrays by
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]930 value, but is also valid on pointers to scalars. The copy is considered to
931 belong to the caller not the callee (for example,
932 <tt><a href="#readonly">readonly</a></tt> functions should not write to
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]933 <tt>byval</tt> parameters). This is not a valid attribute for return
Chris Lattnere39c5bf2009-02-05 05:42:28 +0000[diff] [blame]934 values. The byval attribute also supports specifying an alignment with the
935 align attribute. This has a target-specific effect on the code generator
936 that usually indicates a desired alignment for the synthesized stack
937 slot.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]938
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]939 <dt><tt>sret</tt></dt>
Duncan Sands616cc032008-02-18 04:19:38 +0000[diff] [blame]940 <dd>This indicates that the pointer parameter specifies the address of a
941 structure that is the return value of the function in the source program.
Chris Lattnerf33b8452008-10-04 18:33:34 +0000[diff] [blame]942 This pointer must be guaranteed by the caller to be valid: loads and stores
943 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]944 be applied to the first parameter. This is not a valid attribute for
945 return values. </dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]946
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]947 <dt><tt>noalias</tt></dt>
Nick Lewyckyff384472008-11-24 03:41:24 +0000[diff] [blame]948 <dd>This indicates that the pointer does not alias any global or any other
949 parameter. The caller is responsible for ensuring that this is the
Nick Lewyckya604a942008-11-24 05:00:44 +0000[diff] [blame]950 case. On a function return value, <tt>noalias</tt> additionally indicates
951 that the pointer does not alias any other pointers visible to the
Nick Lewycky40fb6f22008-12-19 06:39:12 +0000[diff] [blame]952 caller. For further details, please see the discussion of the NoAlias
953 response in
954 <a href="http://llvm.org/docs/AliasAnalysis.html#MustMayNo">alias
955 analysis</a>.</dd>
956
957 <dt><tt>nocapture</tt></dt>
958 <dd>This indicates that the callee does not make any copies of the pointer
959 that outlive the callee itself. This is not a valid attribute for return
960 values.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]961
Duncan Sands4ee46812007-07-27 19:57:41 +0000[diff] [blame]962 <dt><tt>nest</tt></dt>
Duncan Sandsf1a7d4c2008-07-08 09:27:25 +0000[diff] [blame]963 <dd>This indicates that the pointer parameter can be excised using the
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]964 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
965 attribute for return values.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]966 </dl>
967
968</div>
969
970<!-- ======================================================================= -->
971<div class="doc_subsection">
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]972 <a name="gc">Garbage Collector Names</a>
973</div>
974
975<div class="doc_text">
976<p>Each function may specify a garbage collector name, which is simply a
977string.</p>
978
979<div class="doc_code"><pre
980>define void @f() gc "name" { ...</pre></div>
981
982<p>The compiler declares the supported values of <i>name</i>. Specifying a
983collector which will cause the compiler to alter its output in order to support
984the named garbage collection algorithm.</p>
985</div>
986
987<!-- ======================================================================= -->
988<div class="doc_subsection">
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]989 <a name="fnattrs">Function Attributes</a>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]990</div>
991
992<div class="doc_text">
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]993
994<p>Function attributes are set to communicate additional information about
995 a function. Function attributes are considered to be part of the function,
996 not of the function type, so functions with different parameter attributes
997 can have the same function type.</p>
998
999 <p>Function attributes are simple keywords that follow the type specified. If
1000 multiple attributes are needed, they are space separated. For
1001 example:</p>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1002
1003<div class="doc_code">
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1004<pre>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1005define void @f() noinline { ... }
1006define void @f() alwaysinline { ... }
1007define void @f() alwaysinline optsize { ... }
1008define void @f() optsize
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1009</pre>
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1010</div>
1011
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1012<dl>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1013<dt><tt>alwaysinline</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]1014<dd>This attribute indicates that the inliner should attempt to inline this
1015function into callers whenever possible, ignoring any active inlining size
1016threshold for this caller.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1017
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1018<dt><tt>noinline</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]1019<dd>This attribute indicates that the inliner should never inline this function
Chris Lattner377f3ae2008-10-05 17:14:59 +0000[diff] [blame]1020in any situation. This attribute may not be used together with the
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]1021<tt>alwaysinline</tt> attribute.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1022
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1023<dt><tt>optsize</tt></dt>
Devang Patelc29099b2008-09-29 18:34:44 +0000[diff] [blame]1024<dd>This attribute suggests that optimization passes and code generator passes
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]1025make choices that keep the code size of this function low, and otherwise do
1026optimizations specifically to reduce code size.</dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1027
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1028<dt><tt>noreturn</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]1029<dd>This function attribute indicates that the function never returns normally.
1030This produces undefined behavior at runtime if the function ever does
1031dynamically return.</dd>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1032
1033<dt><tt>nounwind</tt></dt>
Chris Lattner82d70f92008-10-04 18:23:17 +0000[diff] [blame]1034<dd>This function attribute indicates that the function never returns with an
1035unwind or exceptional control flow. If the function does unwind, its runtime
1036behavior is undefined.</dd>
1037
1038<dt><tt>readnone</tt></dt>
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]1039<dd>This attribute indicates that the function computes its result (or the
1040exception it throws) based strictly on its arguments, without dereferencing any
1041pointer arguments or otherwise accessing any mutable state (e.g. memory, control
1042registers, etc) visible to caller functions. It does not write through any
1043pointer arguments (including <tt><a href="#byval">byval</a></tt> arguments) and
1044never changes any state visible to callers.</dd>
Devang Patel008cd3e2008-09-26 23:51:19 +0000[diff] [blame]1045
Duncan Sands7d94e5a2008-10-06 08:14:18 +0000[diff] [blame]1046<dt><tt><a name="readonly">readonly</a></tt></dt>
1047<dd>This attribute indicates that the function does not write through any
1048pointer arguments (including <tt><a href="#byval">byval</a></tt> arguments)
1049or otherwise modify any state (e.g. memory, control registers, etc) visible to
1050caller functions. It may dereference pointer arguments and read state that may
1051be set in the caller. A readonly function always returns the same value (or
1052throws the same exception) when called with the same set of arguments and global
1053state.</dd>
Bill Wendlingdfaabba2008-11-13 01:02:51 +0000[diff] [blame]1054
1055<dt><tt><a name="ssp">ssp</a></tt></dt>
Bill Wendling34f70d82008-11-26 19:19:05 +0000[diff] [blame]1056<dd>This attribute indicates that the function should emit a stack smashing
Bill Wendlingdfaabba2008-11-13 01:02:51 +0000[diff] [blame]1057protector. It is in the form of a "canary"&mdash;a random value placed on the
1058stack before the local variables that's checked upon return from the function to
1059see if it has been overwritten. A heuristic is used to determine if a function
Bill Wendling34f70d82008-11-26 19:19:05 +0000[diff] [blame]1060needs stack protectors or not.
Bill Wendlingdfaabba2008-11-13 01:02:51 +0000[diff] [blame]1061
Bill Wendlingbe9ec3f2008-11-26 19:07:40 +0000[diff] [blame]1062<p>If a function that has an <tt>ssp</tt> attribute is inlined into a function
1063that doesn't have an <tt>ssp</tt> attribute, then the resulting function will
1064have an <tt>ssp</tt> attribute.</p></dd>
1065
1066<dt><tt>sspreq</tt></dt>
Bill Wendling34f70d82008-11-26 19:19:05 +0000[diff] [blame]1067<dd>This attribute indicates that the function should <em>always</em> emit a
Bill Wendlingdfaabba2008-11-13 01:02:51 +0000[diff] [blame]1068stack smashing protector. This overrides the <tt><a href="#ssp">ssp</a></tt>
Bill Wendling34f70d82008-11-26 19:19:05 +0000[diff] [blame]1069function attribute.
Bill Wendlingbe9ec3f2008-11-26 19:07:40 +0000[diff] [blame]1070
1071<p>If a function that has an <tt>sspreq</tt> attribute is inlined into a
1072function that doesn't have an <tt>sspreq</tt> attribute or which has
1073an <tt>ssp</tt> attribute, then the resulting function will have
1074an <tt>sspreq</tt> attribute.</p></dd>
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1075</dl>
1076
Devang Pateld468f1c2008-09-04 23:05:13 +0000[diff] [blame]1077</div>
1078
1079<!-- ======================================================================= -->
1080<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1081 <a name="moduleasm">Module-Level Inline Assembly</a>
1082</div>
1083
1084<div class="doc_text">
1085<p>
1086Modules may contain "module-level inline asm" blocks, which corresponds to the
1087GCC "file scope inline asm" blocks. These blocks are internally concatenated by
1088LLVM and treated as a single unit, but may be separated in the .ll file if
1089desired. The syntax is very simple:
1090</p>
1091
1092<div class="doc_code">
1093<pre>
1094module asm "inline asm code goes here"
1095module asm "more can go here"
1096</pre>
1097</div>
1098
1099<p>The strings can contain any character by escaping non-printable characters.
1100 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
1101 for the number.
1102</p>
1103
1104<p>
1105 The inline asm code is simply printed to the machine code .s file when
1106 assembly code is generated.
1107</p>
1108</div>
1109
1110<!-- ======================================================================= -->
1111<div class="doc_subsection">
1112 <a name="datalayout">Data Layout</a>
1113</div>
1114
1115<div class="doc_text">
1116<p>A module may specify a target specific data layout string that specifies how
1117data is to be laid out in memory. The syntax for the data layout is simply:</p>
1118<pre> target datalayout = "<i>layout specification</i>"</pre>
1119<p>The <i>layout specification</i> consists of a list of specifications
1120separated by the minus sign character ('-'). Each specification starts with a
1121letter and may include other information after the letter to define some
1122aspect of the data layout. The specifications accepted are as follows: </p>
1123<dl>
1124 <dt><tt>E</tt></dt>
1125 <dd>Specifies that the target lays out data in big-endian form. That is, the
1126 bits with the most significance have the lowest address location.</dd>
1127 <dt><tt>e</tt></dt>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1128 <dd>Specifies that the target lays out data in little-endian form. That is,
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1129 the bits with the least significance have the lowest address location.</dd>
1130 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1131 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
1132 <i>preferred</i> alignments. All sizes are in bits. Specifying the <i>pref</i>
1133 alignment is optional. If omitted, the preceding <tt>:</tt> should be omitted
1134 too.</dd>
1135 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1136 <dd>This specifies the alignment for an integer type of a given bit
1137 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1138 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1139 <dd>This specifies the alignment for a vector type of a given bit
1140 <i>size</i>.</dd>
1141 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1142 <dd>This specifies the alignment for a floating point type of a given bit
1143 <i>size</i>. The value of <i>size</i> must be either 32 (float) or 64
1144 (double).</dd>
1145 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1146 <dd>This specifies the alignment for an aggregate type of a given bit
1147 <i>size</i>.</dd>
1148</dl>
1149<p>When constructing the data layout for a given target, LLVM starts with a
1150default set of specifications which are then (possibly) overriden by the
1151specifications in the <tt>datalayout</tt> keyword. The default specifications
1152are given in this list:</p>
1153<ul>
1154 <li><tt>E</tt> - big endian</li>
1155 <li><tt>p:32:64:64</tt> - 32-bit pointers with 64-bit alignment</li>
1156 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1157 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1158 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1159 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1160 <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]1161 alignment of 64-bits</li>
1162 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1163 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1164 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1165 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1166 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
1167</ul>
Chris Lattner6a9f3c42008-08-05 18:21:08 +0000[diff] [blame]1168<p>When LLVM is determining the alignment for a given type, it uses the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]1169following rules:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1170<ol>
1171 <li>If the type sought is an exact match for one of the specifications, that
1172 specification is used.</li>
1173 <li>If no match is found, and the type sought is an integer type, then the
1174 smallest integer type that is larger than the bitwidth of the sought type is
1175 used. If none of the specifications are larger than the bitwidth then the the
1176 largest integer type is used. For example, given the default specifications
1177 above, the i7 type will use the alignment of i8 (next largest) while both
1178 i65 and i256 will use the alignment of i64 (largest specified).</li>
1179 <li>If no match is found, and the type sought is a vector type, then the
1180 largest vector type that is smaller than the sought vector type will be used
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]1181 as a fall back. This happens because &lt;128 x double&gt; can be implemented
1182 in terms of 64 &lt;2 x double&gt;, for example.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1183</ol>
1184</div>
1185
1186<!-- *********************************************************************** -->
1187<div class="doc_section"> <a name="typesystem">Type System</a> </div>
1188<!-- *********************************************************************** -->
1189
1190<div class="doc_text">
1191
1192<p>The LLVM type system is one of the most important features of the
1193intermediate representation. Being typed enables a number of
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]1194optimizations to be performed on the intermediate representation directly,
1195without having to do
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1196extra analyses on the side before the transformation. A strong type
1197system makes it easier to read the generated code and enables novel
1198analyses and transformations that are not feasible to perform on normal
1199three address code representations.</p>
1200
1201</div>
1202
1203<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1204<div class="doc_subsection"> <a name="t_classifications">Type
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1205Classifications</a> </div>
1206<div class="doc_text">
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1207<p>The types fall into a few useful
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1208classifications:</p>
1209
1210<table border="1" cellspacing="0" cellpadding="4">
1211 <tbody>
1212 <tr><th>Classification</th><th>Types</th></tr>
1213 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1214 <td><a href="#t_integer">integer</a></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1215 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
1216 </tr>
1217 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1218 <td><a href="#t_floating">floating point</a></td>
1219 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1220 </tr>
1221 <tr>
1222 <td><a name="t_firstclass">first class</a></td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1223 <td><a href="#t_integer">integer</a>,
1224 <a href="#t_floating">floating point</a>,
1225 <a href="#t_pointer">pointer</a>,
Dan Gohmanf6237db2008-06-18 18:42:13 +0000[diff] [blame]1226 <a href="#t_vector">vector</a>,
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]1227 <a href="#t_struct">structure</a>,
1228 <a href="#t_array">array</a>,
Dan Gohmand13951e2008-05-23 22:50:26 +0000[diff] [blame]1229 <a href="#t_label">label</a>.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1230 </td>
1231 </tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1232 <tr>
1233 <td><a href="#t_primitive">primitive</a></td>
1234 <td><a href="#t_label">label</a>,
1235 <a href="#t_void">void</a>,
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1236 <a href="#t_floating">floating point</a>.</td>
1237 </tr>
1238 <tr>
1239 <td><a href="#t_derived">derived</a></td>
1240 <td><a href="#t_integer">integer</a>,
1241 <a href="#t_array">array</a>,
1242 <a href="#t_function">function</a>,
1243 <a href="#t_pointer">pointer</a>,
1244 <a href="#t_struct">structure</a>,
1245 <a href="#t_pstruct">packed structure</a>,
1246 <a href="#t_vector">vector</a>,
1247 <a href="#t_opaque">opaque</a>.
Dan Gohman032ba852008-10-14 16:32:04 +0000[diff] [blame]1248 </td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1249 </tr>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1250 </tbody>
1251</table>
1252
1253<p>The <a href="#t_firstclass">first class</a> types are perhaps the
1254most important. Values of these types are the only ones which can be
1255produced by instructions, passed as arguments, or used as operands to
Dan Gohman4f29e422008-05-23 21:53:15 +0000[diff] [blame]1256instructions.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1257</div>
1258
1259<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1260<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1261
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1262<div class="doc_text">
1263<p>The primitive types are the fundamental building blocks of the LLVM
1264system.</p>
1265
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1266</div>
1267
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1268<!-- _______________________________________________________________________ -->
1269<div class="doc_subsubsection"> <a name="t_floating">Floating Point Types</a> </div>
1270
1271<div class="doc_text">
1272 <table>
1273 <tbody>
1274 <tr><th>Type</th><th>Description</th></tr>
1275 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1276 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1277 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1278 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1279 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1280 </tbody>
1281 </table>
1282</div>
1283
1284<!-- _______________________________________________________________________ -->
1285<div class="doc_subsubsection"> <a name="t_void">Void Type</a> </div>
1286
1287<div class="doc_text">
1288<h5>Overview:</h5>
1289<p>The void type does not represent any value and has no size.</p>
1290
1291<h5>Syntax:</h5>
1292
1293<pre>
1294 void
1295</pre>
1296</div>
1297
1298<!-- _______________________________________________________________________ -->
1299<div class="doc_subsubsection"> <a name="t_label">Label Type</a> </div>
1300
1301<div class="doc_text">
1302<h5>Overview:</h5>
1303<p>The label type represents code labels.</p>
1304
1305<h5>Syntax:</h5>
1306
1307<pre>
1308 label
1309</pre>
1310</div>
1311
1312
1313<!-- ======================================================================= -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1314<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
1315
1316<div class="doc_text">
1317
1318<p>The real power in LLVM comes from the derived types in the system.
1319This is what allows a programmer to represent arrays, functions,
1320pointers, and other useful types. Note that these derived types may be
1321recursive: For example, it is possible to have a two dimensional array.</p>
1322
1323</div>
1324
1325<!-- _______________________________________________________________________ -->
1326<div class="doc_subsubsection"> <a name="t_integer">Integer Type</a> </div>
1327
1328<div class="doc_text">
1329
1330<h5>Overview:</h5>
1331<p>The integer type is a very simple derived type that simply specifies an
1332arbitrary bit width for the integer type desired. Any bit width from 1 bit to
13332^23-1 (about 8 million) can be specified.</p>
1334
1335<h5>Syntax:</h5>
1336
1337<pre>
1338 iN
1339</pre>
1340
1341<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1342value.</p>
1343
1344<h5>Examples:</h5>
1345<table class="layout">
Chris Lattner251ab812007-12-18 06:18:21 +0000[diff] [blame]1346 <tbody>
1347 <tr>
1348 <td><tt>i1</tt></td>
1349 <td>a single-bit integer.</td>
1350 </tr><tr>
1351 <td><tt>i32</tt></td>
1352 <td>a 32-bit integer.</td>
1353 </tr><tr>
1354 <td><tt>i1942652</tt></td>
1355 <td>a really big integer of over 1 million bits.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1356 </tr>
Chris Lattner251ab812007-12-18 06:18:21 +0000[diff] [blame]1357 </tbody>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1358</table>
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1359
1360<p>Note that the code generator does not yet support large integer types
1361to be used as function return types. The specific limit on how large a
1362return type the code generator can currently handle is target-dependent;
1363currently it's often 64 bits for 32-bit targets and 128 bits for 64-bit
1364targets.</p>
1365
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1366</div>
1367
1368<!-- _______________________________________________________________________ -->
1369<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
1370
1371<div class="doc_text">
1372
1373<h5>Overview:</h5>
1374
1375<p>The array type is a very simple derived type that arranges elements
1376sequentially in memory. The array type requires a size (number of
1377elements) and an underlying data type.</p>
1378
1379<h5>Syntax:</h5>
1380
1381<pre>
1382 [&lt;# elements&gt; x &lt;elementtype&gt;]
1383</pre>
1384
1385<p>The number of elements is a constant integer value; elementtype may
1386be any type with a size.</p>
1387
1388<h5>Examples:</h5>
1389<table class="layout">
1390 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1391 <td class="left"><tt>[40 x i32]</tt></td>
1392 <td class="left">Array of 40 32-bit integer values.</td>
1393 </tr>
1394 <tr class="layout">
1395 <td class="left"><tt>[41 x i32]</tt></td>
1396 <td class="left">Array of 41 32-bit integer values.</td>
1397 </tr>
1398 <tr class="layout">
1399 <td class="left"><tt>[4 x i8]</tt></td>
1400 <td class="left">Array of 4 8-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1401 </tr>
1402</table>
1403<p>Here are some examples of multidimensional arrays:</p>
1404<table class="layout">
1405 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1406 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1407 <td class="left">3x4 array of 32-bit integer values.</td>
1408 </tr>
1409 <tr class="layout">
1410 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1411 <td class="left">12x10 array of single precision floating point values.</td>
1412 </tr>
1413 <tr class="layout">
1414 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1415 <td class="left">2x3x4 array of 16-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1416 </tr>
1417</table>
1418
1419<p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
1420length array. Normally, accesses past the end of an array are undefined in
1421LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
1422As a special case, however, zero length arrays are recognized to be variable
1423length. This allows implementation of 'pascal style arrays' with the LLVM
1424type "{ i32, [0 x float]}", for example.</p>
1425
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1426<p>Note that the code generator does not yet support large aggregate types
1427to be used as function return types. The specific limit on how large an
1428aggregate return type the code generator can currently handle is
1429target-dependent, and also dependent on the aggregate element types.</p>
1430
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1431</div>
1432
1433<!-- _______________________________________________________________________ -->
1434<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
1435<div class="doc_text">
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1436
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1437<h5>Overview:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1438
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1439<p>The function type can be thought of as a function signature. It
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1440consists of a return type and a list of formal parameter types. The
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1441return 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]1442If 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]1443class types, and the struct must have at least one element.</p>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1444
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1445<h5>Syntax:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1446
1447<pre>
1448 &lt;returntype list&gt; (&lt;parameter list&gt;)
1449</pre>
1450
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1451<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
1452specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1453which indicates that the function takes a variable number of arguments.
1454Variable argument functions can access their arguments with the <a
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1455 href="#int_varargs">variable argument handling intrinsic</a> functions.
1456'<tt>&lt;returntype list&gt;</tt>' is a comma-separated list of
1457<a href="#t_firstclass">first class</a> type specifiers.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1458
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1459<h5>Examples:</h5>
1460<table class="layout">
1461 <tr class="layout">
1462 <td class="left"><tt>i32 (i32)</tt></td>
1463 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
1464 </td>
1465 </tr><tr class="layout">
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]1466 <td class="left"><tt>float&nbsp;(i16&nbsp;signext,&nbsp;i32&nbsp;*)&nbsp;*
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1467 </tt></td>
1468 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
1469 an <tt>i16</tt> that should be sign extended and a
1470 <a href="#t_pointer">pointer</a> to <tt>i32</tt>, returning
1471 <tt>float</tt>.
1472 </td>
1473 </tr><tr class="layout">
1474 <td class="left"><tt>i32 (i8*, ...)</tt></td>
1475 <td class="left">A vararg function that takes at least one
1476 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1477 which returns an integer. This is the signature for <tt>printf</tt> in
1478 LLVM.
1479 </td>
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1480 </tr><tr class="layout">
1481 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Misha Brukmanafc88b02008-11-27 06:41:20 +0000[diff] [blame]1482 <td class="left">A function taking an <tt>i32</tt>, returning two
1483 <tt>i32</tt> values as an aggregate of type <tt>{ i32, i32 }</tt>
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1484 </td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1485 </tr>
1486</table>
1487
1488</div>
1489<!-- _______________________________________________________________________ -->
1490<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
1491<div class="doc_text">
1492<h5>Overview:</h5>
1493<p>The structure type is used to represent a collection of data members
1494together in memory. The packing of the field types is defined to match
1495the ABI of the underlying processor. The elements of a structure may
1496be any type that has a size.</p>
1497<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1498and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1499field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1500instruction.</p>
1501<h5>Syntax:</h5>
1502<pre> { &lt;type list&gt; }<br></pre>
1503<h5>Examples:</h5>
1504<table class="layout">
1505 <tr class="layout">
1506 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1507 <td class="left">A triple of three <tt>i32</tt> values</td>
1508 </tr><tr class="layout">
1509 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1510 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1511 second element is a <a href="#t_pointer">pointer</a> to a
1512 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1513 an <tt>i32</tt>.</td>
1514 </tr>
1515</table>
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1516
1517<p>Note that the code generator does not yet support large aggregate types
1518to be used as function return types. The specific limit on how large an
1519aggregate return type the code generator can currently handle is
1520target-dependent, and also dependent on the aggregate element types.</p>
1521
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1522</div>
1523
1524<!-- _______________________________________________________________________ -->
1525<div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
1526</div>
1527<div class="doc_text">
1528<h5>Overview:</h5>
1529<p>The packed structure type is used to represent a collection of data members
1530together in memory. There is no padding between fields. Further, the alignment
1531of a packed structure is 1 byte. The elements of a packed structure may
1532be any type that has a size.</p>
1533<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1534and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1535field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1536instruction.</p>
1537<h5>Syntax:</h5>
1538<pre> &lt; { &lt;type list&gt; } &gt; <br></pre>
1539<h5>Examples:</h5>
1540<table class="layout">
1541 <tr class="layout">
1542 <td class="left"><tt>&lt; { i32, i32, i32 } &gt;</tt></td>
1543 <td class="left">A triple of three <tt>i32</tt> values</td>
1544 </tr><tr class="layout">
Bill Wendling74d3eac2008-09-07 10:26:33 +0000[diff] [blame]1545 <td class="left">
1546<tt>&lt;&nbsp;{&nbsp;float,&nbsp;i32&nbsp;(i32)*&nbsp;}&nbsp;&gt;</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1547 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1548 second element is a <a href="#t_pointer">pointer</a> to a
1549 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1550 an <tt>i32</tt>.</td>
1551 </tr>
1552</table>
1553</div>
1554
1555<!-- _______________________________________________________________________ -->
1556<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
1557<div class="doc_text">
1558<h5>Overview:</h5>
1559<p>As in many languages, the pointer type represents a pointer or
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]1560reference to another object, which must live in memory. Pointer types may have
1561an optional address space attribute defining the target-specific numbered
1562address space where the pointed-to object resides. The default address space is
1563zero.</p>
Chris Lattner96edbd32009-02-08 19:53:29 +0000[diff] [blame^]1564
1565<p>Note that LLVM does not permit pointers to void (<tt>void*</tt>) nor does
1566it permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</aa> intead.</p>
1567
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1568<h5>Syntax:</h5>
1569<pre> &lt;type&gt; *<br></pre>
1570<h5>Examples:</h5>
1571<table class="layout">
1572 <tr class="layout">
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]1573 <td class="left"><tt>[4 x i32]*</tt></td>
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1574 <td class="left">A <a href="#t_pointer">pointer</a> to <a
1575 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
1576 </tr>
1577 <tr class="layout">
1578 <td class="left"><tt>i32 (i32 *) *</tt></td>
1579 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1580 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1581 <tt>i32</tt>.</td>
1582 </tr>
1583 <tr class="layout">
1584 <td class="left"><tt>i32 addrspace(5)*</tt></td>
1585 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
1586 that resides in address space #5.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1587 </tr>
1588</table>
1589</div>
1590
1591<!-- _______________________________________________________________________ -->
1592<div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
1593<div class="doc_text">
1594
1595<h5>Overview:</h5>
1596
1597<p>A vector type is a simple derived type that represents a vector
1598of elements. Vector types are used when multiple primitive data
1599are operated in parallel using a single instruction (SIMD).
1600A vector type requires a size (number of
1601elements) and an underlying primitive data type. Vectors must have a power
1602of two length (1, 2, 4, 8, 16 ...). Vector types are
1603considered <a href="#t_firstclass">first class</a>.</p>
1604
1605<h5>Syntax:</h5>
1606
1607<pre>
1608 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1609</pre>
1610
1611<p>The number of elements is a constant integer value; elementtype may
1612be any integer or floating point type.</p>
1613
1614<h5>Examples:</h5>
1615
1616<table class="layout">
1617 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1618 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
1619 <td class="left">Vector of 4 32-bit integer values.</td>
1620 </tr>
1621 <tr class="layout">
1622 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
1623 <td class="left">Vector of 8 32-bit floating-point values.</td>
1624 </tr>
1625 <tr class="layout">
1626 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
1627 <td class="left">Vector of 2 64-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1628 </tr>
1629</table>
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]1630
1631<p>Note that the code generator does not yet support large vector types
1632to be used as function return types. The specific limit on how large a
1633vector return type codegen can currently handle is target-dependent;
1634currently it's often a few times longer than a hardware vector register.</p>
1635
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1636</div>
1637
1638<!-- _______________________________________________________________________ -->
1639<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1640<div class="doc_text">
1641
1642<h5>Overview:</h5>
1643
1644<p>Opaque types are used to represent unknown types in the system. This
Gordon Henriksenda0706e2007-10-14 00:34:53 +0000[diff] [blame]1645corresponds (for example) to the C notion of a forward declared structure type.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1646In LLVM, opaque types can eventually be resolved to any type (not just a
1647structure type).</p>
1648
1649<h5>Syntax:</h5>
1650
1651<pre>
1652 opaque
1653</pre>
1654
1655<h5>Examples:</h5>
1656
1657<table class="layout">
1658 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1659 <td class="left"><tt>opaque</tt></td>
1660 <td class="left">An opaque type.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1661 </tr>
1662</table>
1663</div>
1664
Chris Lattner515195a2009-02-02 07:32:36 +0000[diff] [blame]1665<!-- ======================================================================= -->
1666<div class="doc_subsection">
1667 <a name="t_uprefs">Type Up-references</a>
1668</div>
1669
1670<div class="doc_text">
1671<h5>Overview:</h5>
1672<p>
1673An "up reference" allows you to refer to a lexically enclosing type without
1674requiring it to have a name. For instance, a structure declaration may contain a
1675pointer to any of the types it is lexically a member of. Example of up
1676references (with their equivalent as named type declarations) include:</p>
1677
1678<pre>
1679 { \2 * } %x = type { %t* }
1680 { \2 }* %y = type { %y }*
1681 \1* %z = type %z*
1682</pre>
1683
1684<p>
1685An up reference is needed by the asmprinter for printing out cyclic types when
1686there is no declared name for a type in the cycle. Because the asmprinter does
1687not want to print out an infinite type string, it needs a syntax to handle
1688recursive types that have no names (all names are optional in llvm IR).
1689</p>
1690
1691<h5>Syntax:</h5>
1692<pre>
1693 \&lt;level&gt;
1694</pre>
1695
1696<p>
1697The level is the count of the lexical type that is being referred to.
1698</p>
1699
1700<h5>Examples:</h5>
1701
1702<table class="layout">
1703 <tr class="layout">
1704 <td class="left"><tt>\1*</tt></td>
1705 <td class="left">Self-referential pointer.</td>
1706 </tr>
1707 <tr class="layout">
1708 <td class="left"><tt>{ { \3*, i8 }, i32 }</tt></td>
1709 <td class="left">Recursive structure where the upref refers to the out-most
1710 structure.</td>
1711 </tr>
1712</table>
1713</div>
1714
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1715
1716<!-- *********************************************************************** -->
1717<div class="doc_section"> <a name="constants">Constants</a> </div>
1718<!-- *********************************************************************** -->
1719
1720<div class="doc_text">
1721
1722<p>LLVM has several different basic types of constants. This section describes
1723them all and their syntax.</p>
1724
1725</div>
1726
1727<!-- ======================================================================= -->
1728<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
1729
1730<div class="doc_text">
1731
1732<dl>
1733 <dt><b>Boolean constants</b></dt>
1734
1735 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
1736 constants of the <tt><a href="#t_primitive">i1</a></tt> type.
1737 </dd>
1738
1739 <dt><b>Integer constants</b></dt>
1740
1741 <dd>Standard integers (such as '4') are constants of the <a
1742 href="#t_integer">integer</a> type. Negative numbers may be used with
1743 integer types.
1744 </dd>
1745
1746 <dt><b>Floating point constants</b></dt>
1747
1748 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
1749 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]1750 notation (see below). The assembler requires the exact decimal value of
1751 a floating-point constant. For example, the assembler accepts 1.25 but
1752 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
1753 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1754
1755 <dt><b>Null pointer constants</b></dt>
1756
1757 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
1758 and must be of <a href="#t_pointer">pointer type</a>.</dd>
1759
1760</dl>
1761
1762<p>The one non-intuitive notation for constants is the optional hexadecimal form
1763of floating point constants. For example, the form '<tt>double
17640x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
17654.5e+15</tt>'. The only time hexadecimal floating point constants are required
1766(and the only time that they are generated by the disassembler) is when a
1767floating point constant must be emitted but it cannot be represented as a
1768decimal floating point number. For example, NaN's, infinities, and other
1769special values are represented in their IEEE hexadecimal format so that
1770assembly and disassembly do not cause any bits to change in the constants.</p>
1771
1772</div>
1773
1774<!-- ======================================================================= -->
1775<div class="doc_subsection"><a name="aggregateconstants">Aggregate Constants</a>
1776</div>
1777
1778<div class="doc_text">
1779<p>Aggregate constants arise from aggregation of simple constants
1780and smaller aggregate constants.</p>
1781
1782<dl>
1783 <dt><b>Structure constants</b></dt>
1784
1785 <dd>Structure constants are represented with notation similar to structure
1786 type definitions (a comma separated list of elements, surrounded by braces
Chris Lattner6c8de962007-12-25 20:34:52 +0000[diff] [blame]1787 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
1788 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]1789 must have <a href="#t_struct">structure type</a>, and the number and
1790 types of elements must match those specified by the type.
1791 </dd>
1792
1793 <dt><b>Array constants</b></dt>
1794
1795 <dd>Array constants are represented with notation similar to array type
1796 definitions (a comma separated list of elements, surrounded by square brackets
1797 (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74 ]</tt>". Array
1798 constants must have <a href="#t_array">array type</a>, and the number and
1799 types of elements must match those specified by the type.
1800 </dd>
1801
1802 <dt><b>Vector constants</b></dt>
1803
1804 <dd>Vector constants are represented with notation similar to vector type
1805 definitions (a comma separated list of elements, surrounded by
1806 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32 42,
1807 i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must have <a
1808 href="#t_vector">vector type</a>, and the number and types of elements must
1809 match those specified by the type.
1810 </dd>
1811
1812 <dt><b>Zero initialization</b></dt>
1813
1814 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1815 value to zero of <em>any</em> type, including scalar and aggregate types.
1816 This is often used to avoid having to print large zero initializers (e.g. for
1817 large arrays) and is always exactly equivalent to using explicit zero
1818 initializers.
1819 </dd>
1820</dl>
1821
1822</div>
1823
1824<!-- ======================================================================= -->
1825<div class="doc_subsection">
1826 <a name="globalconstants">Global Variable and Function Addresses</a>
1827</div>
1828
1829<div class="doc_text">
1830
1831<p>The addresses of <a href="#globalvars">global variables</a> and <a
1832href="#functionstructure">functions</a> are always implicitly valid (link-time)
1833constants. These constants are explicitly referenced when the <a
1834href="#identifiers">identifier for the global</a> is used and always have <a
1835href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1836file:</p>
1837
1838<div class="doc_code">
1839<pre>
1840@X = global i32 17
1841@Y = global i32 42
1842@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
1843</pre>
1844</div>
1845
1846</div>
1847
1848<!-- ======================================================================= -->
1849<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
1850<div class="doc_text">
1851 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
1852 no specific value. Undefined values may be of any type and be used anywhere
1853 a constant is permitted.</p>
1854
1855 <p>Undefined values indicate to the compiler that the program is well defined
1856 no matter what value is used, giving the compiler more freedom to optimize.
1857 </p>
1858</div>
1859
1860<!-- ======================================================================= -->
1861<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1862</div>
1863
1864<div class="doc_text">
1865
1866<p>Constant expressions are used to allow expressions involving other constants
1867to be used as constants. Constant expressions may be of any <a
1868href="#t_firstclass">first class</a> type and may involve any LLVM operation
1869that does not have side effects (e.g. load and call are not supported). The
1870following is the syntax for constant expressions:</p>
1871
1872<dl>
1873 <dt><b><tt>trunc ( CST to TYPE )</tt></b></dt>
1874 <dd>Truncate a constant to another type. The bit size of CST must be larger
1875 than the bit size of TYPE. Both types must be integers.</dd>
1876
1877 <dt><b><tt>zext ( CST to TYPE )</tt></b></dt>
1878 <dd>Zero extend a constant to another type. The bit size of CST must be
1879 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1880
1881 <dt><b><tt>sext ( CST to TYPE )</tt></b></dt>
1882 <dd>Sign extend a constant to another type. The bit size of CST must be
1883 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1884
1885 <dt><b><tt>fptrunc ( CST to TYPE )</tt></b></dt>
1886 <dd>Truncate a floating point constant to another floating point type. The
1887 size of CST must be larger than the size of TYPE. Both types must be
1888 floating point.</dd>
1889
1890 <dt><b><tt>fpext ( CST to TYPE )</tt></b></dt>
1891 <dd>Floating point extend a constant to another type. The size of CST must be
1892 smaller or equal to the size of TYPE. Both types must be floating point.</dd>
1893
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]1894 <dt><b><tt>fptoui ( CST to TYPE )</tt></b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1895 <dd>Convert a floating point constant to the corresponding unsigned integer
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1896 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1897 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1898 of the same number of elements. If the value won't fit in the integer type,
1899 the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1900
1901 <dt><b><tt>fptosi ( CST to TYPE )</tt></b></dt>
1902 <dd>Convert a floating point constant to the corresponding signed integer
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1903 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1904 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1905 of the same number of elements. If the value won't fit in the integer type,
1906 the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1907
1908 <dt><b><tt>uitofp ( CST to TYPE )</tt></b></dt>
1909 <dd>Convert an unsigned integer constant to the corresponding floating point
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1910 constant. TYPE must be a scalar or vector floating point type. CST must be of
1911 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1912 of the same number of elements. If the value won't fit in the floating point
1913 type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1914
1915 <dt><b><tt>sitofp ( CST to TYPE )</tt></b></dt>
1916 <dd>Convert a signed integer constant to the corresponding floating point
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1917 constant. TYPE must be a scalar or vector floating point type. CST must be of
1918 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1919 of the same number of elements. If the value won't fit in the floating point
1920 type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1921
1922 <dt><b><tt>ptrtoint ( CST to TYPE )</tt></b></dt>
1923 <dd>Convert a pointer typed constant to the corresponding integer constant
1924 TYPE must be an integer type. CST must be of pointer type. The CST value is
1925 zero extended, truncated, or unchanged to make it fit in TYPE.</dd>
1926
1927 <dt><b><tt>inttoptr ( CST to TYPE )</tt></b></dt>
1928 <dd>Convert a integer constant to a pointer constant. TYPE must be a
1929 pointer type. CST must be of integer type. The CST value is zero extended,
1930 truncated, or unchanged to make it fit in a pointer size. This one is
1931 <i>really</i> dangerous!</dd>
1932
1933 <dt><b><tt>bitcast ( CST to TYPE )</tt></b></dt>
1934 <dd>Convert a constant, CST, to another TYPE. The size of CST and TYPE must be
1935 identical (same number of bits). The conversion is done as if the CST value
1936 was stored to memory and read back as TYPE. In other words, no bits change
1937 with this operator, just the type. This can be used for conversion of
1938 vector types to any other type, as long as they have the same bit width. For
Dan Gohman7305fa02008-09-08 16:45:59 +0000[diff] [blame]1939 pointers it is only valid to cast to another pointer type. It is not valid
1940 to bitcast to or from an aggregate type.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1941 </dd>
1942
1943 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
1944
1945 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
1946 constants. As with the <a href="#i_getelementptr">getelementptr</a>
1947 instruction, the index list may have zero or more indexes, which are required
1948 to make sense for the type of "CSTPTR".</dd>
1949
1950 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
1951
1952 <dd>Perform the <a href="#i_select">select operation</a> on
1953 constants.</dd>
1954
1955 <dt><b><tt>icmp COND ( VAL1, VAL2 )</tt></b></dt>
1956 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
1957
1958 <dt><b><tt>fcmp COND ( VAL1, VAL2 )</tt></b></dt>
1959 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
1960
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]1961 <dt><b><tt>vicmp COND ( VAL1, VAL2 )</tt></b></dt>
1962 <dd>Performs the <a href="#i_vicmp">vicmp operation</a> on constants.</dd>
1963
1964 <dt><b><tt>vfcmp COND ( VAL1, VAL2 )</tt></b></dt>
1965 <dd>Performs the <a href="#i_vfcmp">vfcmp operation</a> on constants.</dd>
1966
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1967 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
1968
1969 <dd>Perform the <a href="#i_extractelement">extractelement
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]1970 operation</a> on constants.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1971
1972 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
1973
1974 <dd>Perform the <a href="#i_insertelement">insertelement
1975 operation</a> on constants.</dd>
1976
1977
1978 <dt><b><tt>shufflevector ( VEC1, VEC2, IDXMASK )</tt></b></dt>
1979
1980 <dd>Perform the <a href="#i_shufflevector">shufflevector
1981 operation</a> on constants.</dd>
1982
1983 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
1984
1985 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
1986 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
1987 binary</a> operations. The constraints on operands are the same as those for
1988 the corresponding instruction (e.g. no bitwise operations on floating point
1989 values are allowed).</dd>
1990</dl>
1991</div>
1992
1993<!-- *********************************************************************** -->
1994<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
1995<!-- *********************************************************************** -->
1996
1997<!-- ======================================================================= -->
1998<div class="doc_subsection">
1999<a name="inlineasm">Inline Assembler Expressions</a>
2000</div>
2001
2002<div class="doc_text">
2003
2004<p>
2005LLVM supports inline assembler expressions (as opposed to <a href="#moduleasm">
2006Module-Level Inline Assembly</a>) through the use of a special value. This
2007value represents the inline assembler as a string (containing the instructions
2008to emit), a list of operand constraints (stored as a string), and a flag that
2009indicates whether or not the inline asm expression has side effects. An example
2010inline assembler expression is:
2011</p>
2012
2013<div class="doc_code">
2014<pre>
2015i32 (i32) asm "bswap $0", "=r,r"
2016</pre>
2017</div>
2018
2019<p>
2020Inline assembler expressions may <b>only</b> be used as the callee operand of
2021a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we have:
2022</p>
2023
2024<div class="doc_code">
2025<pre>
2026%X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
2027</pre>
2028</div>
2029
2030<p>
2031Inline asms with side effects not visible in the constraint list must be marked
2032as having side effects. This is done through the use of the
2033'<tt>sideeffect</tt>' keyword, like so:
2034</p>
2035
2036<div class="doc_code">
2037<pre>
2038call void asm sideeffect "eieio", ""()
2039</pre>
2040</div>
2041
2042<p>TODO: The format of the asm and constraints string still need to be
2043documented here. Constraints on what can be done (e.g. duplication, moving, etc
Chris Lattner1cdd64e2008-10-04 18:36:02 +0000[diff] [blame]2044need to be documented). This is probably best done by reference to another
2045document that covers inline asm from a holistic perspective.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2046</p>
2047
2048</div>
2049
2050<!-- *********************************************************************** -->
2051<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
2052<!-- *********************************************************************** -->
2053
2054<div class="doc_text">
2055
2056<p>The LLVM instruction set consists of several different
2057classifications of instructions: <a href="#terminators">terminator
2058instructions</a>, <a href="#binaryops">binary instructions</a>,
2059<a href="#bitwiseops">bitwise binary instructions</a>, <a
2060 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
2061instructions</a>.</p>
2062
2063</div>
2064
2065<!-- ======================================================================= -->
2066<div class="doc_subsection"> <a name="terminators">Terminator
2067Instructions</a> </div>
2068
2069<div class="doc_text">
2070
2071<p>As mentioned <a href="#functionstructure">previously</a>, every
2072basic block in a program ends with a "Terminator" instruction, which
2073indicates which block should be executed after the current block is
2074finished. These terminator instructions typically yield a '<tt>void</tt>'
2075value: they produce control flow, not values (the one exception being
2076the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
2077<p>There are six different terminator instructions: the '<a
2078 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
2079instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
2080the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
2081 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
2082 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
2083
2084</div>
2085
2086<!-- _______________________________________________________________________ -->
2087<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
2088Instruction</a> </div>
2089<div class="doc_text">
2090<h5>Syntax:</h5>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2091<pre>
2092 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]2093 ret void <i>; Return from void function</i>
2094</pre>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2095
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2096<h5>Overview:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2097
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2098<p>The '<tt>ret</tt>' instruction is used to return control flow (and
2099optionally a value) from a function back to the caller.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2100<p>There are two forms of the '<tt>ret</tt>' instruction: one that
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2101returns a value and then causes control flow, and one that just causes
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2102control flow to occur.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2103
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2104<h5>Arguments:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2105
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2106<p>The '<tt>ret</tt>' instruction optionally accepts a single argument,
2107the return value. The type of the return value must be a
2108'<a href="#t_firstclass">first class</a>' type.</p>
2109
2110<p>A function is not <a href="#wellformed">well formed</a> if
2111it it has a non-void return type and contains a '<tt>ret</tt>'
2112instruction with no return value or a return value with a type that
2113does not match its type, or if it has a void return type and contains
2114a '<tt>ret</tt>' instruction with a return value.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2115
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2116<h5>Semantics:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2117
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2118<p>When the '<tt>ret</tt>' instruction is executed, control flow
2119returns back to the calling function's context. If the caller is a "<a
2120 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
2121the instruction after the call. If the caller was an "<a
2122 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
2123at the beginning of the "normal" destination block. If the instruction
2124returns a value, that value shall set the call or invoke instruction's
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]2125return value.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2126
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2127<h5>Example:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]2128
2129<pre>
2130 ret i32 5 <i>; Return an integer value of 5</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2131 ret void <i>; Return from a void function</i>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]2132 ret { i32, i8 } { i32 4, i8 2 } <i>; Return an aggregate of values 4 and 2</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2133</pre>
Dan Gohman60967192009-01-12 23:12:39 +0000[diff] [blame]2134
djge93155c2009-01-24 15:58:40 +0000[diff] [blame]2135<p>Note that the code generator does not yet fully support large
2136 return values. The specific sizes that are currently supported are
2137 dependent on the target. For integers, on 32-bit targets the limit
2138 is often 64 bits, and on 64-bit targets the limit is often 128 bits.
2139 For aggregate types, the current limits are dependent on the element
2140 types; for example targets are often limited to 2 total integer
2141 elements and 2 total floating-point elements.</p>
Dan Gohman60967192009-01-12 23:12:39 +0000[diff] [blame]2142
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2143</div>
2144<!-- _______________________________________________________________________ -->
2145<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
2146<div class="doc_text">
2147<h5>Syntax:</h5>
2148<pre> br i1 &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;<br> br label &lt;dest&gt; <i>; Unconditional branch</i>
2149</pre>
2150<h5>Overview:</h5>
2151<p>The '<tt>br</tt>' instruction is used to cause control flow to
2152transfer to a different basic block in the current function. There are
2153two forms of this instruction, corresponding to a conditional branch
2154and an unconditional branch.</p>
2155<h5>Arguments:</h5>
2156<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
2157single '<tt>i1</tt>' value and two '<tt>label</tt>' values. The
2158unconditional form of the '<tt>br</tt>' instruction takes a single
2159'<tt>label</tt>' value as a target.</p>
2160<h5>Semantics:</h5>
2161<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
2162argument is evaluated. If the value is <tt>true</tt>, control flows
2163to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
2164control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
2165<h5>Example:</h5>
2166<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
2167 href="#i_ret">ret</a> i32 1<br>IfUnequal:<br> <a href="#i_ret">ret</a> i32 0<br></pre>
2168</div>
2169<!-- _______________________________________________________________________ -->
2170<div class="doc_subsubsection">
2171 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
2172</div>
2173
2174<div class="doc_text">
2175<h5>Syntax:</h5>
2176
2177<pre>
2178 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
2179</pre>
2180
2181<h5>Overview:</h5>
2182
2183<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
2184several different places. It is a generalization of the '<tt>br</tt>'
2185instruction, allowing a branch to occur to one of many possible
2186destinations.</p>
2187
2188
2189<h5>Arguments:</h5>
2190
2191<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
2192comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
2193an array of pairs of comparison value constants and '<tt>label</tt>'s. The
2194table is not allowed to contain duplicate constant entries.</p>
2195
2196<h5>Semantics:</h5>
2197
2198<p>The <tt>switch</tt> instruction specifies a table of values and
2199destinations. When the '<tt>switch</tt>' instruction is executed, this
2200table is searched for the given value. If the value is found, control flow is
2201transfered to the corresponding destination; otherwise, control flow is
2202transfered to the default destination.</p>
2203
2204<h5>Implementation:</h5>
2205
2206<p>Depending on properties of the target machine and the particular
2207<tt>switch</tt> instruction, this instruction may be code generated in different
2208ways. For example, it could be generated as a series of chained conditional
2209branches or with a lookup table.</p>
2210
2211<h5>Example:</h5>
2212
2213<pre>
2214 <i>; Emulate a conditional br instruction</i>
2215 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]2216 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2217
2218 <i>; Emulate an unconditional br instruction</i>
2219 switch i32 0, label %dest [ ]
2220
2221 <i>; Implement a jump table:</i>
Dan Gohman01852382009-01-04 23:44:43 +0000[diff] [blame]2222 switch i32 %val, label %otherwise [ i32 0, label %onzero
2223 i32 1, label %onone
2224 i32 2, label %ontwo ]
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2225</pre>
2226</div>
2227
2228<!-- _______________________________________________________________________ -->
2229<div class="doc_subsubsection">
2230 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
2231</div>
2232
2233<div class="doc_text">
2234
2235<h5>Syntax:</h5>
2236
2237<pre>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]2238 &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]2239 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
2240</pre>
2241
2242<h5>Overview:</h5>
2243
2244<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
2245function, with the possibility of control flow transfer to either the
2246'<tt>normal</tt>' label or the
2247'<tt>exception</tt>' label. If the callee function returns with the
2248"<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
2249"normal" label. If the callee (or any indirect callees) returns with the "<a
2250href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]2251continued at the dynamically nearest "exception" label.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2252
2253<h5>Arguments:</h5>
2254
2255<p>This instruction requires several arguments:</p>
2256
2257<ol>
2258 <li>
2259 The optional "cconv" marker indicates which <a href="#callingconv">calling
2260 convention</a> the call should use. If none is specified, the call defaults
2261 to using C calling conventions.
2262 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]2263
2264 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
2265 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>',
2266 and '<tt>inreg</tt>' attributes are valid here.</li>
2267
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2268 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
2269 function value being invoked. In most cases, this is a direct function
2270 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
2271 an arbitrary pointer to function value.
2272 </li>
2273
2274 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
2275 function to be invoked. </li>
2276
2277 <li>'<tt>function args</tt>': argument list whose types match the function
2278 signature argument types. If the function signature indicates the function
2279 accepts a variable number of arguments, the extra arguments can be
2280 specified. </li>
2281
2282 <li>'<tt>normal label</tt>': the label reached when the called function
2283 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
2284
2285 <li>'<tt>exception label</tt>': the label reached when a callee returns with
2286 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
2287
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]2288 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]2289 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
2290 '<tt>readnone</tt>' attributes are valid here.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2291</ol>
2292
2293<h5>Semantics:</h5>
2294
2295<p>This instruction is designed to operate as a standard '<tt><a
2296href="#i_call">call</a></tt>' instruction in most regards. The primary
2297difference is that it establishes an association with a label, which is used by
2298the runtime library to unwind the stack.</p>
2299
2300<p>This instruction is used in languages with destructors to ensure that proper
2301cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
2302exception. Additionally, this is important for implementation of
2303'<tt>catch</tt>' clauses in high-level languages that support them.</p>
2304
2305<h5>Example:</h5>
2306<pre>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2307 %retval = invoke i32 @Test(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2308 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2309 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2310 unwind label %TestCleanup <i>; {i32}:retval set</i>
2311</pre>
2312</div>
2313
2314
2315<!-- _______________________________________________________________________ -->
2316
2317<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
2318Instruction</a> </div>
2319
2320<div class="doc_text">
2321
2322<h5>Syntax:</h5>
2323<pre>
2324 unwind
2325</pre>
2326
2327<h5>Overview:</h5>
2328
2329<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
2330at the first callee in the dynamic call stack which used an <a
2331href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
2332primarily used to implement exception handling.</p>
2333
2334<h5>Semantics:</h5>
2335
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]2336<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2337immediately halt. The dynamic call stack is then searched for the first <a
2338href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
2339execution continues at the "exceptional" destination block specified by the
2340<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
2341dynamic call chain, undefined behavior results.</p>
2342</div>
2343
2344<!-- _______________________________________________________________________ -->
2345
2346<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
2347Instruction</a> </div>
2348
2349<div class="doc_text">
2350
2351<h5>Syntax:</h5>
2352<pre>
2353 unreachable
2354</pre>
2355
2356<h5>Overview:</h5>
2357
2358<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
2359instruction is used to inform the optimizer that a particular portion of the
2360code is not reachable. This can be used to indicate that the code after a
2361no-return function cannot be reached, and other facts.</p>
2362
2363<h5>Semantics:</h5>
2364
2365<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
2366</div>
2367
2368
2369
2370<!-- ======================================================================= -->
2371<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
2372<div class="doc_text">
2373<p>Binary operators are used to do most of the computation in a
Chris Lattnerab596d92008-04-01 18:47:32 +0000[diff] [blame]2374program. They require two operands of the same type, execute an operation on them, and
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2375produce a single value. The operands might represent
2376multiple 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]2377The result value has the same type as its operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2378<p>There are several different binary operators:</p>
2379</div>
2380<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2381<div class="doc_subsubsection">
2382 <a name="i_add">'<tt>add</tt>' Instruction</a>
2383</div>
2384
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2385<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2386
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2387<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2388
2389<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2390 &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]2391</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2392
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2393<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2394
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2395<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2396
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2397<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2398
2399<p>The two arguments to the '<tt>add</tt>' instruction must be <a
2400 href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>, or
2401 <a href="#t_vector">vector</a> values. Both arguments must have identical
2402 types.</p>
2403
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2404<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2405
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2406<p>The value produced is the integer or floating point sum of the two
2407operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2408
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2409<p>If an integer sum has unsigned overflow, the result returned is the
2410mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2411the result.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2412
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2413<p>Because LLVM integers use a two's complement representation, this
2414instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2415
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2416<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2417
2418<pre>
2419 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2420</pre>
2421</div>
2422<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2423<div class="doc_subsubsection">
2424 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
2425</div>
2426
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2427<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2428
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2429<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2430
2431<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2432 &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]2433</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2434
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2435<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2436
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2437<p>The '<tt>sub</tt>' instruction returns the difference of its two
2438operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2439
2440<p>Note that the '<tt>sub</tt>' instruction is used to represent the
2441'<tt>neg</tt>' instruction present in most other intermediate
2442representations.</p>
2443
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2444<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2445
2446<p>The two arguments to the '<tt>sub</tt>' instruction must be <a
2447 href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>,
2448 or <a href="#t_vector">vector</a> values. Both arguments must have identical
2449 types.</p>
2450
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2451<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2452
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2453<p>The value produced is the integer or floating point difference of
2454the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2455
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2456<p>If an integer difference has unsigned overflow, the result returned is the
2457mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2458the result.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2459
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2460<p>Because LLVM integers use a two's complement representation, this
2461instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2462
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2463<h5>Example:</h5>
2464<pre>
2465 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
2466 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
2467</pre>
2468</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2469
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2470<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2471<div class="doc_subsubsection">
2472 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
2473</div>
2474
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2475<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2476
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2477<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2478<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]2479</pre>
2480<h5>Overview:</h5>
2481<p>The '<tt>mul</tt>' instruction returns the product of its two
2482operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2483
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2484<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2485
2486<p>The two arguments to the '<tt>mul</tt>' instruction must be <a
2487href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>,
2488or <a href="#t_vector">vector</a> values. Both arguments must have identical
2489types.</p>
2490
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2491<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2492
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2493<p>The value produced is the integer or floating point product of the
2494two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2495
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2496<p>If the result of an integer multiplication has unsigned overflow,
2497the result returned is the mathematical result modulo
24982<sup>n</sup>, where n is the bit width of the result.</p>
2499<p>Because LLVM integers use a two's complement representation, and the
2500result is the same width as the operands, this instruction returns the
2501correct result for both signed and unsigned integers. If a full product
2502(e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands
2503should be sign-extended or zero-extended as appropriate to the
2504width of the full product.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2505<h5>Example:</h5>
2506<pre> &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
2507</pre>
2508</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2509
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2510<!-- _______________________________________________________________________ -->
2511<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
2512</a></div>
2513<div class="doc_text">
2514<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2515<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]2516</pre>
2517<h5>Overview:</h5>
2518<p>The '<tt>udiv</tt>' instruction returns the quotient of its two
2519operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2520
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2521<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2522
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2523<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2524<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2525values. Both arguments must have identical types.</p>
2526
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2527<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2528
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2529<p>The value produced is the unsigned integer quotient of the two operands.</p>
2530<p>Note that unsigned integer division and signed integer division are distinct
2531operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
2532<p>Division by zero leads to undefined behavior.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2533<h5>Example:</h5>
2534<pre> &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2535</pre>
2536</div>
2537<!-- _______________________________________________________________________ -->
2538<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
2539</a> </div>
2540<div class="doc_text">
2541<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2542<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2543 &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]2544</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2545
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2546<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2547
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2548<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two
2549operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2550
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2551<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2552
2553<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
2554<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2555values. Both arguments must have identical types.</p>
2556
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2557<h5>Semantics:</h5>
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2558<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]2559<p>Note that signed integer division and unsigned integer division are distinct
2560operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
2561<p>Division by zero leads to undefined behavior. Overflow also leads to
2562undefined behavior; this is a rare case, but can occur, for example,
2563by doing a 32-bit division of -2147483648 by -1.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2564<h5>Example:</h5>
2565<pre> &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2566</pre>
2567</div>
2568<!-- _______________________________________________________________________ -->
2569<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
2570Instruction</a> </div>
2571<div class="doc_text">
2572<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2573<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2574 &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]2575</pre>
2576<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2577
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2578<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two
2579operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2580
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2581<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2582
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2583<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2584<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2585of floating point values. Both arguments must have identical types.</p>
2586
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2587<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2588
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2589<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2590
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2591<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2592
2593<pre>
2594 &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]2595</pre>
2596</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2597
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2598<!-- _______________________________________________________________________ -->
2599<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
2600</div>
2601<div class="doc_text">
2602<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2603<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]2604</pre>
2605<h5>Overview:</h5>
2606<p>The '<tt>urem</tt>' instruction returns the remainder from the
2607unsigned division of its two arguments.</p>
2608<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2609<p>The two arguments to the '<tt>urem</tt>' instruction must be
2610<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2611values. Both arguments must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2612<h5>Semantics:</h5>
2613<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2614This instruction always performs an unsigned division to get the remainder.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2615<p>Note that unsigned integer remainder and signed integer remainder are
2616distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
2617<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2618<h5>Example:</h5>
2619<pre> &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2620</pre>
2621
2622</div>
2623<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2624<div class="doc_subsubsection">
2625 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
2626</div>
2627
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2628<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2629
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2630<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2631
2632<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2633 &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]2634</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2635
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2636<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2637
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2638<p>The '<tt>srem</tt>' instruction returns the remainder from the
Dan Gohman3e3fd8c2007-11-05 23:35:22 +0000[diff] [blame]2639signed division of its two operands. This instruction can also take
2640<a href="#t_vector">vector</a> versions of the values in which case
2641the elements must be integers.</p>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]2642
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2643<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2644
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2645<p>The two arguments to the '<tt>srem</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2646<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2647values. Both arguments must have identical types.</p>
2648
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2649<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2650
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2651<p>This instruction returns the <i>remainder</i> of a division (where the result
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2652has the same sign as the dividend, <tt>op1</tt>), not the <i>modulo</i>
2653operator (where the result has the same sign as the divisor, <tt>op2</tt>) of
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2654a value. For more information about the difference, see <a
2655 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
2656Math Forum</a>. For a table of how this is implemented in various languages,
2657please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
2658Wikipedia: modulo operation</a>.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2659<p>Note that signed integer remainder and unsigned integer remainder are
2660distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
2661<p>Taking the remainder of a division by zero leads to undefined behavior.
2662Overflow also leads to undefined behavior; this is a rare case, but can occur,
2663for example, by taking the remainder of a 32-bit division of -2147483648 by -1.
2664(The remainder doesn't actually overflow, but this rule lets srem be
2665implemented using instructions that return both the result of the division
2666and the remainder.)</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2667<h5>Example:</h5>
2668<pre> &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2669</pre>
2670
2671</div>
2672<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2673<div class="doc_subsubsection">
2674 <a name="i_frem">'<tt>frem</tt>' Instruction</a> </div>
2675
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2676<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2677
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2678<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2679<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]2680</pre>
2681<h5>Overview:</h5>
2682<p>The '<tt>frem</tt>' instruction returns the remainder from the
2683division of its two operands.</p>
2684<h5>Arguments:</h5>
2685<p>The two arguments to the '<tt>frem</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2686<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2687of floating point values. Both arguments must have identical types.</p>
2688
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2689<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2690
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2691<p>This instruction returns the <i>remainder</i> of a division.
2692The remainder has the same sign as the dividend.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2693
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2694<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2695
2696<pre>
2697 &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]2698</pre>
2699</div>
2700
2701<!-- ======================================================================= -->
2702<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
2703Operations</a> </div>
2704<div class="doc_text">
2705<p>Bitwise binary operators are used to do various forms of
2706bit-twiddling in a program. They are generally very efficient
2707instructions and can commonly be strength reduced from other
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2708instructions. They require two operands of the same type, execute an operation on them,
2709and 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]2710</div>
2711
2712<!-- _______________________________________________________________________ -->
2713<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
2714Instruction</a> </div>
2715<div class="doc_text">
2716<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2717<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]2718</pre>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2719
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2720<h5>Overview:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2721
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2722<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
2723the left a specified number of bits.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2724
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2725<h5>Arguments:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2726
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2727<p>Both arguments to the '<tt>shl</tt>' instruction must be the same <a
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]2728 href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2729type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2730
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2731<h5>Semantics:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2732
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2733<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod 2<sup>n</sup>,
2734where 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]2735equal to or larger than the number of bits in <tt>op1</tt>, the result is undefined.
2736If the arguments are vectors, each vector element of <tt>op1</tt> is shifted by the
2737corresponding shift amount in <tt>op2</tt>.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2738
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2739<h5>Example:</h5><pre>
2740 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
2741 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
2742 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2743 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Mon P Wang9901e732008-12-09 05:46:39 +0000[diff] [blame]2744 &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]2745</pre>
2746</div>
2747<!-- _______________________________________________________________________ -->
2748<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
2749Instruction</a> </div>
2750<div class="doc_text">
2751<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2752<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]2753</pre>
2754
2755<h5>Overview:</h5>
2756<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
2757operand shifted to the right a specified number of bits with zero fill.</p>
2758
2759<h5>Arguments:</h5>
2760<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]2761<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]2762type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2763
2764<h5>Semantics:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2765
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2766<p>This instruction always performs a logical shift right operation. The most
2767significant bits of the result will be filled with zero bits after the
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2768shift. If <tt>op2</tt> is (statically or dynamically) equal to or larger than
Mon P Wangb0f51822008-12-10 08:55:09 +0000[diff] [blame]2769the number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
2770vectors, each vector element of <tt>op1</tt> is shifted by the corresponding shift
2771amount in <tt>op2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2772
2773<h5>Example:</h5>
2774<pre>
2775 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
2776 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
2777 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
2778 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2779 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Mon P Wang9901e732008-12-09 05:46:39 +0000[diff] [blame]2780 &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]2781</pre>
2782</div>
2783
2784<!-- _______________________________________________________________________ -->
2785<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
2786Instruction</a> </div>
2787<div class="doc_text">
2788
2789<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2790<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]2791</pre>
2792
2793<h5>Overview:</h5>
2794<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
2795operand shifted to the right a specified number of bits with sign extension.</p>
2796
2797<h5>Arguments:</h5>
2798<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Nate Begemanbb1ce942008-07-29 15:49:41 +0000[diff] [blame]2799<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]2800type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2801
2802<h5>Semantics:</h5>
2803<p>This instruction always performs an arithmetic shift right operation,
2804The most significant bits of the result will be filled with the sign bit
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2805of <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]2806larger than the number of bits in <tt>op1</tt>, the result is undefined. If the
2807arguments are vectors, each vector element of <tt>op1</tt> is shifted by the
2808corresponding shift amount in <tt>op2</tt>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2809
2810<h5>Example:</h5>
2811<pre>
2812 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
2813 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
2814 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
2815 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2816 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Mon P Wang9901e732008-12-09 05:46:39 +0000[diff] [blame]2817 &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]2818</pre>
2819</div>
2820
2821<!-- _______________________________________________________________________ -->
2822<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
2823Instruction</a> </div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2824
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2825<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2826
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2827<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2828
2829<pre>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2830 &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]2831</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2832
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2833<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2834
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2835<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
2836its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2837
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2838<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2839
2840<p>The two arguments to the '<tt>and</tt>' instruction must be
2841<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2842values. Both arguments must have identical types.</p>
2843
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2844<h5>Semantics:</h5>
2845<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
2846<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]2847<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2848<table border="1" cellspacing="0" cellpadding="4">
2849 <tbody>
2850 <tr>
2851 <td>In0</td>
2852 <td>In1</td>
2853 <td>Out</td>
2854 </tr>
2855 <tr>
2856 <td>0</td>
2857 <td>0</td>
2858 <td>0</td>
2859 </tr>
2860 <tr>
2861 <td>0</td>
2862 <td>1</td>
2863 <td>0</td>
2864 </tr>
2865 <tr>
2866 <td>1</td>
2867 <td>0</td>
2868 <td>0</td>
2869 </tr>
2870 <tr>
2871 <td>1</td>
2872 <td>1</td>
2873 <td>1</td>
2874 </tr>
2875 </tbody>
2876</table>
2877</div>
2878<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2879<pre>
2880 &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]2881 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
2882 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
2883</pre>
2884</div>
2885<!-- _______________________________________________________________________ -->
2886<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
2887<div class="doc_text">
2888<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2889<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]2890</pre>
2891<h5>Overview:</h5>
2892<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
2893or of its two operands.</p>
2894<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2895
2896<p>The two arguments to the '<tt>or</tt>' instruction must be
2897<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2898values. Both arguments must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2899<h5>Semantics:</h5>
2900<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
2901<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]2902<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2903<table border="1" cellspacing="0" cellpadding="4">
2904 <tbody>
2905 <tr>
2906 <td>In0</td>
2907 <td>In1</td>
2908 <td>Out</td>
2909 </tr>
2910 <tr>
2911 <td>0</td>
2912 <td>0</td>
2913 <td>0</td>
2914 </tr>
2915 <tr>
2916 <td>0</td>
2917 <td>1</td>
2918 <td>1</td>
2919 </tr>
2920 <tr>
2921 <td>1</td>
2922 <td>0</td>
2923 <td>1</td>
2924 </tr>
2925 <tr>
2926 <td>1</td>
2927 <td>1</td>
2928 <td>1</td>
2929 </tr>
2930 </tbody>
2931</table>
2932</div>
2933<h5>Example:</h5>
2934<pre> &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
2935 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
2936 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
2937</pre>
2938</div>
2939<!-- _______________________________________________________________________ -->
2940<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
2941Instruction</a> </div>
2942<div class="doc_text">
2943<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]2944<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]2945</pre>
2946<h5>Overview:</h5>
2947<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
2948or of its two operands. The <tt>xor</tt> is used to implement the
2949"one's complement" operation, which is the "~" operator in C.</p>
2950<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2951<p>The two arguments to the '<tt>xor</tt>' instruction must be
2952<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2953values. Both arguments must have identical types.</p>
2954
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2955<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2956
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2957<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
2958<p> </p>
Bill Wendling1e19eac2008-09-07 10:29:20 +0000[diff] [blame]2959<div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2960<table border="1" cellspacing="0" cellpadding="4">
2961 <tbody>
2962 <tr>
2963 <td>In0</td>
2964 <td>In1</td>
2965 <td>Out</td>
2966 </tr>
2967 <tr>
2968 <td>0</td>
2969 <td>0</td>
2970 <td>0</td>
2971 </tr>
2972 <tr>
2973 <td>0</td>
2974 <td>1</td>
2975 <td>1</td>
2976 </tr>
2977 <tr>
2978 <td>1</td>
2979 <td>0</td>
2980 <td>1</td>
2981 </tr>
2982 <tr>
2983 <td>1</td>
2984 <td>1</td>
2985 <td>0</td>
2986 </tr>
2987 </tbody>
2988</table>
2989</div>
2990<p> </p>
2991<h5>Example:</h5>
2992<pre> &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
2993 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
2994 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
2995 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
2996</pre>
2997</div>
2998
2999<!-- ======================================================================= -->
3000<div class="doc_subsection">
3001 <a name="vectorops">Vector Operations</a>
3002</div>
3003
3004<div class="doc_text">
3005
3006<p>LLVM supports several instructions to represent vector operations in a
3007target-independent manner. These instructions cover the element-access and
3008vector-specific operations needed to process vectors effectively. While LLVM
3009does directly support these vector operations, many sophisticated algorithms
3010will want to use target-specific intrinsics to take full advantage of a specific
3011target.</p>
3012
3013</div>
3014
3015<!-- _______________________________________________________________________ -->
3016<div class="doc_subsubsection">
3017 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
3018</div>
3019
3020<div class="doc_text">
3021
3022<h5>Syntax:</h5>
3023
3024<pre>
3025 &lt;result&gt; = extractelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, i32 &lt;idx&gt; <i>; yields &lt;ty&gt;</i>
3026</pre>
3027
3028<h5>Overview:</h5>
3029
3030<p>
3031The '<tt>extractelement</tt>' instruction extracts a single scalar
3032element from a vector at a specified index.
3033</p>
3034
3035
3036<h5>Arguments:</h5>
3037
3038<p>
3039The first operand of an '<tt>extractelement</tt>' instruction is a
3040value of <a href="#t_vector">vector</a> type. The second operand is
3041an index indicating the position from which to extract the element.
3042The index may be a variable.</p>
3043
3044<h5>Semantics:</h5>
3045
3046<p>
3047The result is a scalar of the same type as the element type of
3048<tt>val</tt>. Its value is the value at position <tt>idx</tt> of
3049<tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
3050results are undefined.
3051</p>
3052
3053<h5>Example:</h5>
3054
3055<pre>
3056 %result = extractelement &lt;4 x i32&gt; %vec, i32 0 <i>; yields i32</i>
3057</pre>
3058</div>
3059
3060
3061<!-- _______________________________________________________________________ -->
3062<div class="doc_subsubsection">
3063 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
3064</div>
3065
3066<div class="doc_text">
3067
3068<h5>Syntax:</h5>
3069
3070<pre>
Dan Gohmanbcc3c502008-05-12 23:38:42 +0000[diff] [blame]3071 &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]3072</pre>
3073
3074<h5>Overview:</h5>
3075
3076<p>
3077The '<tt>insertelement</tt>' instruction inserts a scalar
3078element into a vector at a specified index.
3079</p>
3080
3081
3082<h5>Arguments:</h5>
3083
3084<p>
3085The first operand of an '<tt>insertelement</tt>' instruction is a
3086value of <a href="#t_vector">vector</a> type. The second operand is a
3087scalar value whose type must equal the element type of the first
3088operand. The third operand is an index indicating the position at
3089which to insert the value. The index may be a variable.</p>
3090
3091<h5>Semantics:</h5>
3092
3093<p>
3094The result is a vector of the same type as <tt>val</tt>. Its
3095element values are those of <tt>val</tt> except at position
3096<tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
3097exceeds the length of <tt>val</tt>, the results are undefined.
3098</p>
3099
3100<h5>Example:</h5>
3101
3102<pre>
3103 %result = insertelement &lt;4 x i32&gt; %vec, i32 1, i32 0 <i>; yields &lt;4 x i32&gt;</i>
3104</pre>
3105</div>
3106
3107<!-- _______________________________________________________________________ -->
3108<div class="doc_subsubsection">
3109 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
3110</div>
3111
3112<div class="doc_text">
3113
3114<h5>Syntax:</h5>
3115
3116<pre>
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3117 &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]3118</pre>
3119
3120<h5>Overview:</h5>
3121
3122<p>
3123The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3124from two input vectors, returning a vector with the same element type as
3125the input and length that is the same as the shuffle mask.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3126</p>
3127
3128<h5>Arguments:</h5>
3129
3130<p>
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3131The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
3132with types that match each other. The third argument is a shuffle mask whose
3133element type is always 'i32'. The result of the instruction is a vector whose
3134length is the same as the shuffle mask and whose element type is the same as
3135the element type of the first two operands.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3136</p>
3137
3138<p>
3139The shuffle mask operand is required to be a constant vector with either
3140constant integer or undef values.
3141</p>
3142
3143<h5>Semantics:</h5>
3144
3145<p>
3146The elements of the two input vectors are numbered from left to right across
3147both of the vectors. The shuffle mask operand specifies, for each element of
Mon P Wangbff5d9c2008-11-10 04:46:22 +0000[diff] [blame]3148the result vector, which element of the two input vectors the result element
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3149gets. The element selector may be undef (meaning "don't care") and the second
3150operand may be undef if performing a shuffle from only one vector.
3151</p>
3152
3153<h5>Example:</h5>
3154
3155<pre>
3156 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
3157 &lt;4 x i32&gt; &lt;i32 0, i32 4, i32 1, i32 5&gt; <i>; yields &lt;4 x i32&gt;</i>
3158 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; undef,
3159 &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]3160 %result = shufflevector &lt;8 x i32&gt; %v1, &lt;8 x i32&gt; undef,
3161 &lt;4 x i32&gt; &lt;i32 0, i32 1, i32 2, i32 3&gt; <i>; yields &lt;4 x i32&gt;</i>
3162 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
3163 &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]3164</pre>
3165</div>
3166
3167
3168<!-- ======================================================================= -->
3169<div class="doc_subsection">
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3170 <a name="aggregateops">Aggregate Operations</a>
3171</div>
3172
3173<div class="doc_text">
3174
3175<p>LLVM supports several instructions for working with aggregate values.
3176</p>
3177
3178</div>
3179
3180<!-- _______________________________________________________________________ -->
3181<div class="doc_subsubsection">
3182 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
3183</div>
3184
3185<div class="doc_text">
3186
3187<h5>Syntax:</h5>
3188
3189<pre>
3190 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
3191</pre>
3192
3193<h5>Overview:</h5>
3194
3195<p>
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3196The '<tt>extractvalue</tt>' instruction extracts the value of a struct field
3197or array element from an aggregate value.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3198</p>
3199
3200
3201<h5>Arguments:</h5>
3202
3203<p>
3204The first operand of an '<tt>extractvalue</tt>' instruction is a
3205value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a>
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3206type. The operands are constant indices to specify which value to extract
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3207in a similar manner as indices in a
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3208'<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
3209</p>
3210
3211<h5>Semantics:</h5>
3212
3213<p>
3214The result is the value at the position in the aggregate specified by
3215the index operands.
3216</p>
3217
3218<h5>Example:</h5>
3219
3220<pre>
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3221 %result = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3222</pre>
3223</div>
3224
3225
3226<!-- _______________________________________________________________________ -->
3227<div class="doc_subsubsection">
3228 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
3229</div>
3230
3231<div class="doc_text">
3232
3233<h5>Syntax:</h5>
3234
3235<pre>
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3236 &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]3237</pre>
3238
3239<h5>Overview:</h5>
3240
3241<p>
3242The '<tt>insertvalue</tt>' instruction inserts a value
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3243into a struct field or array element in an aggregate.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3244</p>
3245
3246
3247<h5>Arguments:</h5>
3248
3249<p>
3250The first operand of an '<tt>insertvalue</tt>' instruction is a
3251value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a> type.
3252The second operand is a first-class value to insert.
Dan Gohman4f29e422008-05-23 21:53:15 +0000[diff] [blame]3253The following operands are constant indices
Dan Gohmane5febe42008-05-31 00:58:22 +0000[diff] [blame]3254indicating the position at which to insert the value in a similar manner as
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3255indices in a
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3256'<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
3257The value to insert must have the same type as the value identified
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3258by the indices.
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3259</p>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3260
3261<h5>Semantics:</h5>
3262
3263<p>
3264The result is an aggregate of the same type as <tt>val</tt>. Its
3265value is that of <tt>val</tt> except that the value at the position
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]3266specified by the indices is that of <tt>elt</tt>.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3267</p>
3268
3269<h5>Example:</h5>
3270
3271<pre>
Dan Gohmanb1aab4e2008-06-23 15:26:37 +0000[diff] [blame]3272 %result = insertvalue {i32, float} %agg, i32 1, 0 <i>; yields {i32, float}</i>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]3273</pre>
3274</div>
3275
3276
3277<!-- ======================================================================= -->
3278<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3279 <a name="memoryops">Memory Access and Addressing Operations</a>
3280</div>
3281
3282<div class="doc_text">
3283
3284<p>A key design point of an SSA-based representation is how it
3285represents memory. In LLVM, no memory locations are in SSA form, which
3286makes things very simple. This section describes how to read, write,
3287allocate, and free memory in LLVM.</p>
3288
3289</div>
3290
3291<!-- _______________________________________________________________________ -->
3292<div class="doc_subsubsection">
3293 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
3294</div>
3295
3296<div class="doc_text">
3297
3298<h5>Syntax:</h5>
3299
3300<pre>
3301 &lt;result&gt; = malloc &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
3302</pre>
3303
3304<h5>Overview:</h5>
3305
3306<p>The '<tt>malloc</tt>' instruction allocates memory from the system
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]3307heap and returns a pointer to it. The object is always allocated in the generic
3308address space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3309
3310<h5>Arguments:</h5>
3311
3312<p>The '<tt>malloc</tt>' instruction allocates
3313<tt>sizeof(&lt;type&gt;)*NumElements</tt>
3314bytes of memory from the operating system and returns a pointer of the
3315appropriate type to the program. If "NumElements" is specified, it is the
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3316number of elements allocated, otherwise "NumElements" is defaulted to be one.
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3317If a constant alignment is specified, the value result of the allocation is guaranteed to
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3318be aligned to at least that boundary. If not specified, or if zero, the target can
3319choose to align the allocation on any convenient boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3320
3321<p>'<tt>type</tt>' must be a sized type.</p>
3322
3323<h5>Semantics:</h5>
3324
3325<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
Nick Lewyckyff384472008-11-24 03:41:24 +0000[diff] [blame]3326a pointer is returned. The result of a zero byte allocation is undefined. The
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]3327result is null if there is insufficient memory available.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3328
3329<h5>Example:</h5>
3330
3331<pre>
Dan Gohmanf54f50a2009-01-04 23:49:44 +0000[diff] [blame]3332 %array = malloc [4 x i8] <i>; yields {[%4 x i8]*}:array</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3333
3334 %size = <a href="#i_add">add</a> i32 2, 2 <i>; yields {i32}:size = i32 4</i>
3335 %array1 = malloc i8, i32 4 <i>; yields {i8*}:array1</i>
3336 %array2 = malloc [12 x i8], i32 %size <i>; yields {[12 x i8]*}:array2</i>
3337 %array3 = malloc i32, i32 4, align 1024 <i>; yields {i32*}:array3</i>
3338 %array4 = malloc i32, align 1024 <i>; yields {i32*}:array4</i>
3339</pre>
Dan Gohman60967192009-01-12 23:12:39 +0000[diff] [blame]3340
3341<p>Note that the code generator does not yet respect the
3342 alignment value.</p>
3343
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3344</div>
3345
3346<!-- _______________________________________________________________________ -->
3347<div class="doc_subsubsection">
3348 <a name="i_free">'<tt>free</tt>' Instruction</a>
3349</div>
3350
3351<div class="doc_text">
3352
3353<h5>Syntax:</h5>
3354
3355<pre>
Dan Gohmanf54f50a2009-01-04 23:49:44 +0000[diff] [blame]3356 free &lt;type&gt; &lt;value&gt; <i>; yields {void}</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3357</pre>
3358
3359<h5>Overview:</h5>
3360
3361<p>The '<tt>free</tt>' instruction returns memory back to the unused
3362memory heap to be reallocated in the future.</p>
3363
3364<h5>Arguments:</h5>
3365
3366<p>'<tt>value</tt>' shall be a pointer value that points to a value
3367that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
3368instruction.</p>
3369
3370<h5>Semantics:</h5>
3371
3372<p>Access to the memory pointed to by the pointer is no longer defined
Chris Lattner329d6532008-04-19 22:41:32 +0000[diff] [blame]3373after this instruction executes. If the pointer is null, the operation
3374is a noop.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3375
3376<h5>Example:</h5>
3377
3378<pre>
Dan Gohmanf54f50a2009-01-04 23:49:44 +0000[diff] [blame]3379 %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]3380 free [4 x i8]* %array
3381</pre>
3382</div>
3383
3384<!-- _______________________________________________________________________ -->
3385<div class="doc_subsubsection">
3386 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
3387</div>
3388
3389<div class="doc_text">
3390
3391<h5>Syntax:</h5>
3392
3393<pre>
3394 &lt;result&gt; = alloca &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
3395</pre>
3396
3397<h5>Overview:</h5>
3398
3399<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
3400currently executing function, to be automatically released when this function
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]3401returns to its caller. The object is always allocated in the generic address
3402space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3403
3404<h5>Arguments:</h5>
3405
3406<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
3407bytes of memory on the runtime stack, returning a pointer of the
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3408appropriate type to the program. If "NumElements" is specified, it is the
3409number of elements allocated, otherwise "NumElements" is defaulted to be one.
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3410If a constant alignment is specified, the value result of the allocation is guaranteed
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3411to be aligned to at least that boundary. If not specified, or if zero, the target
3412can choose to align the allocation on any convenient boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3413
3414<p>'<tt>type</tt>' may be any sized type.</p>
3415
3416<h5>Semantics:</h5>
3417
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]3418<p>Memory is allocated; a pointer is returned. The operation is undefiend if
3419there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3420memory is automatically released when the function returns. The '<tt>alloca</tt>'
3421instruction is commonly used to represent automatic variables that must
3422have an address available. When the function returns (either with the <tt><a
3423 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]3424instructions), the memory is reclaimed. Allocating zero bytes
3425is legal, but the result is undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3426
3427<h5>Example:</h5>
3428
3429<pre>
Dan Gohmanf54f50a2009-01-04 23:49:44 +0000[diff] [blame]3430 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
3431 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
3432 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
3433 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3434</pre>
3435</div>
3436
3437<!-- _______________________________________________________________________ -->
3438<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
3439Instruction</a> </div>
3440<div class="doc_text">
3441<h5>Syntax:</h5>
3442<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>
3443<h5>Overview:</h5>
3444<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
3445<h5>Arguments:</h5>
3446<p>The argument to the '<tt>load</tt>' instruction specifies the memory
3447address from which to load. The pointer must point to a <a
3448 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
3449marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
3450the number or order of execution of this <tt>load</tt> with other
3451volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
3452instructions. </p>
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3453<p>
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3454The optional constant "align" argument specifies the alignment of the operation
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3455(that is, the alignment of the memory address). A value of 0 or an
3456omitted "align" argument means that the operation has the preferential
3457alignment for the target. It is the responsibility of the code emitter
3458to ensure that the alignment information is correct. Overestimating
3459the alignment results in an undefined behavior. Underestimating the
3460alignment may produce less efficient code. An alignment of 1 is always
3461safe.
3462</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3463<h5>Semantics:</h5>
3464<p>The location of memory pointed to is loaded.</p>
3465<h5>Examples:</h5>
3466<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
3467 <a
3468 href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
3469 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
3470</pre>
3471</div>
3472<!-- _______________________________________________________________________ -->
3473<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
3474Instruction</a> </div>
3475<div class="doc_text">
3476<h5>Syntax:</h5>
3477<pre> store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
3478 volatile store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
3479</pre>
3480<h5>Overview:</h5>
3481<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
3482<h5>Arguments:</h5>
3483<p>There are two arguments to the '<tt>store</tt>' instruction: a value
3484to 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]3485operand must be a pointer to the <a href="#t_firstclass">first class</a> type
3486of the '<tt>&lt;value&gt;</tt>'
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3487operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
3488optimizer is not allowed to modify the number or order of execution of
3489this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
3490 href="#i_store">store</a></tt> instructions.</p>
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3491<p>
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3492The optional constant "align" argument specifies the alignment of the operation
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3493(that is, the alignment of the memory address). A value of 0 or an
3494omitted "align" argument means that the operation has the preferential
3495alignment for the target. It is the responsibility of the code emitter
3496to ensure that the alignment information is correct. Overestimating
3497the alignment results in an undefined behavior. Underestimating the
3498alignment may produce less efficient code. An alignment of 1 is always
3499safe.
3500</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3501<h5>Semantics:</h5>
3502<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>'
3503at the location specified by the '<tt>&lt;pointer&gt;</tt>' operand.</p>
3504<h5>Example:</h5>
3505<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling63ffa142007-10-22 05:10:05 +0000[diff] [blame]3506 store i32 3, i32* %ptr <i>; yields {void}</i>
3507 %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]3508</pre>
3509</div>
3510
3511<!-- _______________________________________________________________________ -->
3512<div class="doc_subsubsection">
3513 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
3514</div>
3515
3516<div class="doc_text">
3517<h5>Syntax:</h5>
3518<pre>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3519 &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]3520</pre>
3521
3522<h5>Overview:</h5>
3523
3524<p>
3525The '<tt>getelementptr</tt>' instruction is used to get the address of a
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3526subelement of an aggregate data structure. It performs address calculation only
3527and does not access memory.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3528
3529<h5>Arguments:</h5>
3530
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3531<p>The first argument is always a pointer, and forms the basis of the
3532calculation. The remaining arguments are indices, that indicate which of the
3533elements of the aggregate object are indexed. The interpretation of each index
3534is dependent on the type being indexed into. The first index always indexes the
3535pointer value given as the first argument, the second index indexes a value of
3536the type pointed to (not necessarily the value directly pointed to, since the
3537first index can be non-zero), etc. The first type indexed into must be a pointer
3538value, subsequent types can be arrays, vectors and structs. Note that subsequent
3539types being indexed into can never be pointers, since that would require loading
3540the pointer before continuing calculation.</p>
3541
3542<p>The type of each index argument depends on the type it is indexing into.
3543When indexing into a (packed) structure, only <tt>i32</tt> integer
3544<b>constants</b> are allowed. When indexing into an array, pointer or vector,
3545only integers of 32 or 64 bits are allowed (also non-constants). 32-bit values
3546will be sign extended to 64-bits if required.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3547
3548<p>For example, let's consider a C code fragment and how it gets
3549compiled to LLVM:</p>
3550
3551<div class="doc_code">
3552<pre>
3553struct RT {
3554 char A;
3555 int B[10][20];
3556 char C;
3557};
3558struct ST {
3559 int X;
3560 double Y;
3561 struct RT Z;
3562};
3563
3564int *foo(struct ST *s) {
3565 return &amp;s[1].Z.B[5][13];
3566}
3567</pre>
3568</div>
3569
3570<p>The LLVM code generated by the GCC frontend is:</p>
3571
3572<div class="doc_code">
3573<pre>
Chris Lattner5b6dc6e2009-01-11 20:53:49 +0000[diff] [blame]3574%RT = <a href="#namedtypes">type</a> { i8 , [10 x [20 x i32]], i8 }
3575%ST = <a href="#namedtypes">type</a> { i32, double, %RT }
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3576
3577define i32* %foo(%ST* %s) {
3578entry:
3579 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
3580 ret i32* %reg
3581}
3582</pre>
3583</div>
3584
3585<h5>Semantics:</h5>
3586
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3587<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
3588type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
3589}</tt>' type, a structure. The second index indexes into the third element of
3590the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
3591i8 }</tt>' type, another structure. The third index indexes into the second
3592element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
3593array. The two dimensions of the array are subscripted into, yielding an
3594'<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
3595to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
3596
3597<p>Note that it is perfectly legal to index partially through a
3598structure, returning a pointer to an inner element. Because of this,
3599the LLVM code for the given testcase is equivalent to:</p>
3600
3601<pre>
3602 define i32* %foo(%ST* %s) {
3603 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
3604 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
3605 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
3606 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
3607 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
3608 ret i32* %t5
3609 }
3610</pre>
3611
3612<p>Note that it is undefined to access an array out of bounds: array and
3613pointer indexes must always be within the defined bounds of the array type.
Chris Lattnera7d94ba2008-04-24 05:59:56 +0000[diff] [blame]3614The one exception for this rule is zero length arrays. These arrays are
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3615defined to be accessible as variable length arrays, which requires access
3616beyond the zero'th element.</p>
3617
3618<p>The getelementptr instruction is often confusing. For some more insight
3619into how it works, see <a href="GetElementPtr.html">the getelementptr
3620FAQ</a>.</p>
3621
3622<h5>Example:</h5>
3623
3624<pre>
3625 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijman32a080e2008-10-13 13:44:15 +0000[diff] [blame]3626 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
3627 <i>; yields i8*:vptr</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3628 %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]3629 <i>; yields i8*:eptr</i>
3630 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3631</pre>
3632</div>
3633
3634<!-- ======================================================================= -->
3635<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
3636</div>
3637<div class="doc_text">
3638<p>The instructions in this category are the conversion instructions (casting)
3639which all take a single operand and a type. They perform various bit conversions
3640on the operand.</p>
3641</div>
3642
3643<!-- _______________________________________________________________________ -->
3644<div class="doc_subsubsection">
3645 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
3646</div>
3647<div class="doc_text">
3648
3649<h5>Syntax:</h5>
3650<pre>
3651 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3652</pre>
3653
3654<h5>Overview:</h5>
3655<p>
3656The '<tt>trunc</tt>' instruction truncates its operand to the type <tt>ty2</tt>.
3657</p>
3658
3659<h5>Arguments:</h5>
3660<p>
3661The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
3662be an <a href="#t_integer">integer</a> type, and a type that specifies the size
3663and type of the result, which must be an <a href="#t_integer">integer</a>
3664type. The bit size of <tt>value</tt> must be larger than the bit size of
3665<tt>ty2</tt>. Equal sized types are not allowed.</p>
3666
3667<h5>Semantics:</h5>
3668<p>
3669The '<tt>trunc</tt>' instruction truncates the high order bits in <tt>value</tt>
3670and converts the remaining bits to <tt>ty2</tt>. Since the source size must be
3671larger than the destination size, <tt>trunc</tt> cannot be a <i>no-op cast</i>.
3672It will always truncate bits.</p>
3673
3674<h5>Example:</h5>
3675<pre>
3676 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
3677 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
3678 %Y = trunc i32 122 to i1 <i>; yields i1:false</i>
3679</pre>
3680</div>
3681
3682<!-- _______________________________________________________________________ -->
3683<div class="doc_subsubsection">
3684 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
3685</div>
3686<div class="doc_text">
3687
3688<h5>Syntax:</h5>
3689<pre>
3690 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3691</pre>
3692
3693<h5>Overview:</h5>
3694<p>The '<tt>zext</tt>' instruction zero extends its operand to type
3695<tt>ty2</tt>.</p>
3696
3697
3698<h5>Arguments:</h5>
3699<p>The '<tt>zext</tt>' instruction takes a value to cast, which must be of
3700<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3701also be of <a href="#t_integer">integer</a> type. The bit size of the
3702<tt>value</tt> must be smaller than the bit size of the destination type,
3703<tt>ty2</tt>.</p>
3704
3705<h5>Semantics:</h5>
3706<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
3707bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
3708
3709<p>When zero extending from i1, the result will always be either 0 or 1.</p>
3710
3711<h5>Example:</h5>
3712<pre>
3713 %X = zext i32 257 to i64 <i>; yields i64:257</i>
3714 %Y = zext i1 true to i32 <i>; yields i32:1</i>
3715</pre>
3716</div>
3717
3718<!-- _______________________________________________________________________ -->
3719<div class="doc_subsubsection">
3720 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
3721</div>
3722<div class="doc_text">
3723
3724<h5>Syntax:</h5>
3725<pre>
3726 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3727</pre>
3728
3729<h5>Overview:</h5>
3730<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
3731
3732<h5>Arguments:</h5>
3733<p>
3734The '<tt>sext</tt>' instruction takes a value to cast, which must be of
3735<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3736also be of <a href="#t_integer">integer</a> type. The bit size of the
3737<tt>value</tt> must be smaller than the bit size of the destination type,
3738<tt>ty2</tt>.</p>
3739
3740<h5>Semantics:</h5>
3741<p>
3742The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
3743bit (highest order bit) of the <tt>value</tt> until it reaches the bit size of
3744the type <tt>ty2</tt>.</p>
3745
3746<p>When sign extending from i1, the extension always results in -1 or 0.</p>
3747
3748<h5>Example:</h5>
3749<pre>
3750 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
3751 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
3752</pre>
3753</div>
3754
3755<!-- _______________________________________________________________________ -->
3756<div class="doc_subsubsection">
3757 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
3758</div>
3759
3760<div class="doc_text">
3761
3762<h5>Syntax:</h5>
3763
3764<pre>
3765 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3766</pre>
3767
3768<h5>Overview:</h5>
3769<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
3770<tt>ty2</tt>.</p>
3771
3772
3773<h5>Arguments:</h5>
3774<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
3775 point</a> value to cast and a <a href="#t_floating">floating point</a> type to
3776cast it to. The size of <tt>value</tt> must be larger than the size of
3777<tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
3778<i>no-op cast</i>.</p>
3779
3780<h5>Semantics:</h5>
3781<p> The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
3782<a href="#t_floating">floating point</a> type to a smaller
3783<a href="#t_floating">floating point</a> type. If the value cannot fit within
3784the destination type, <tt>ty2</tt>, then the results are undefined.</p>
3785
3786<h5>Example:</h5>
3787<pre>
3788 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
3789 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
3790</pre>
3791</div>
3792
3793<!-- _______________________________________________________________________ -->
3794<div class="doc_subsubsection">
3795 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
3796</div>
3797<div class="doc_text">
3798
3799<h5>Syntax:</h5>
3800<pre>
3801 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3802</pre>
3803
3804<h5>Overview:</h5>
3805<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
3806floating point value.</p>
3807
3808<h5>Arguments:</h5>
3809<p>The '<tt>fpext</tt>' instruction takes a
3810<a href="#t_floating">floating point</a> <tt>value</tt> to cast,
3811and a <a href="#t_floating">floating point</a> type to cast it to. The source
3812type must be smaller than the destination type.</p>
3813
3814<h5>Semantics:</h5>
3815<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
3816<a href="#t_floating">floating point</a> type to a larger
3817<a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
3818used to make a <i>no-op cast</i> because it always changes bits. Use
3819<tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
3820
3821<h5>Example:</h5>
3822<pre>
3823 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
3824 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
3825</pre>
3826</div>
3827
3828<!-- _______________________________________________________________________ -->
3829<div class="doc_subsubsection">
3830 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
3831</div>
3832<div class="doc_text">
3833
3834<h5>Syntax:</h5>
3835<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3836 &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]3837</pre>
3838
3839<h5>Overview:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3840<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3841unsigned integer equivalent of type <tt>ty2</tt>.
3842</p>
3843
3844<h5>Arguments:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3845<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]3846scalar or vector <a href="#t_floating">floating point</a> value, and a type
3847to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
3848type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
3849vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3850
3851<h5>Semantics:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3852<p> The '<tt>fptoui</tt>' instruction converts its
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3853<a href="#t_floating">floating point</a> operand into the nearest (rounding
3854towards zero) unsigned integer value. If the value cannot fit in <tt>ty2</tt>,
3855the results are undefined.</p>
3856
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3857<h5>Example:</h5>
3858<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3859 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]3860 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3861 %X = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3862</pre>
3863</div>
3864
3865<!-- _______________________________________________________________________ -->
3866<div class="doc_subsubsection">
3867 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
3868</div>
3869<div class="doc_text">
3870
3871<h5>Syntax:</h5>
3872<pre>
3873 &lt;result&gt; = fptosi &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3874</pre>
3875
3876<h5>Overview:</h5>
3877<p>The '<tt>fptosi</tt>' instruction converts
3878<a href="#t_floating">floating point</a> <tt>value</tt> to type <tt>ty2</tt>.
3879</p>
3880
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3881<h5>Arguments:</h5>
3882<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]3883scalar or vector <a href="#t_floating">floating point</a> value, and a type
3884to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
3885type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
3886vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3887
3888<h5>Semantics:</h5>
3889<p>The '<tt>fptosi</tt>' instruction converts its
3890<a href="#t_floating">floating point</a> operand into the nearest (rounding
3891towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
3892the results are undefined.</p>
3893
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3894<h5>Example:</h5>
3895<pre>
3896 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]3897 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3898 %X = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
3899</pre>
3900</div>
3901
3902<!-- _______________________________________________________________________ -->
3903<div class="doc_subsubsection">
3904 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
3905</div>
3906<div class="doc_text">
3907
3908<h5>Syntax:</h5>
3909<pre>
3910 &lt;result&gt; = uitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3911</pre>
3912
3913<h5>Overview:</h5>
3914<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
3915integer and converts that value to the <tt>ty2</tt> type.</p>
3916
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3917<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]3918<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be a
3919scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
3920to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
3921type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
3922floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3923
3924<h5>Semantics:</h5>
3925<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
3926integer quantity and converts it to the corresponding floating point value. If
3927the value cannot fit in the floating point value, the results are undefined.</p>
3928
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3929<h5>Example:</h5>
3930<pre>
3931 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3932 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3933</pre>
3934</div>
3935
3936<!-- _______________________________________________________________________ -->
3937<div class="doc_subsubsection">
3938 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
3939</div>
3940<div class="doc_text">
3941
3942<h5>Syntax:</h5>
3943<pre>
3944 &lt;result&gt; = sitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3945</pre>
3946
3947<h5>Overview:</h5>
3948<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed
3949integer and converts that value to the <tt>ty2</tt> type.</p>
3950
3951<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]3952<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be a
3953scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
3954to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
3955type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
3956floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3957
3958<h5>Semantics:</h5>
3959<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed
3960integer quantity and converts it to the corresponding floating point value. If
3961the value cannot fit in the floating point value, the results are undefined.</p>
3962
3963<h5>Example:</h5>
3964<pre>
3965 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]3966 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3967</pre>
3968</div>
3969
3970<!-- _______________________________________________________________________ -->
3971<div class="doc_subsubsection">
3972 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
3973</div>
3974<div class="doc_text">
3975
3976<h5>Syntax:</h5>
3977<pre>
3978 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3979</pre>
3980
3981<h5>Overview:</h5>
3982<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
3983the integer type <tt>ty2</tt>.</p>
3984
3985<h5>Arguments:</h5>
3986<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
3987must 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]3988<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]3989
3990<h5>Semantics:</h5>
3991<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
3992<tt>ty2</tt> by interpreting the pointer value as an integer and either
3993truncating or zero extending that value to the size of the integer type. If
3994<tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
3995<tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
3996are the same size, then nothing is done (<i>no-op cast</i>) other than a type
3997change.</p>
3998
3999<h5>Example:</h5>
4000<pre>
4001 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
4002 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
4003</pre>
4004</div>
4005
4006<!-- _______________________________________________________________________ -->
4007<div class="doc_subsubsection">
4008 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
4009</div>
4010<div class="doc_text">
4011
4012<h5>Syntax:</h5>
4013<pre>
4014 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4015</pre>
4016
4017<h5>Overview:</h5>
4018<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to
4019a pointer type, <tt>ty2</tt>.</p>
4020
4021<h5>Arguments:</h5>
4022<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
4023value to cast, and a type to cast it to, which must be a
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4024<a href="#t_pointer">pointer</a> type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4025
4026<h5>Semantics:</h5>
4027<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
4028<tt>ty2</tt> by applying either a zero extension or a truncation depending on
4029the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
4030size of a pointer then a truncation is done. If <tt>value</tt> is smaller than
4031the size of a pointer then a zero extension is done. If they are the same size,
4032nothing is done (<i>no-op cast</i>).</p>
4033
4034<h5>Example:</h5>
4035<pre>
4036 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
4037 %X = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
4038 %Y = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
4039</pre>
4040</div>
4041
4042<!-- _______________________________________________________________________ -->
4043<div class="doc_subsubsection">
4044 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
4045</div>
4046<div class="doc_text">
4047
4048<h5>Syntax:</h5>
4049<pre>
4050 &lt;result&gt; = bitcast &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4051</pre>
4052
4053<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4054
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4055<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
4056<tt>ty2</tt> without changing any bits.</p>
4057
4058<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4059
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4060<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be
Dan Gohman7305fa02008-09-08 16:45:59 +0000[diff] [blame]4061a non-aggregate first class value, and a type to cast it to, which must also be
4062a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes of
4063<tt>value</tt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4064and the destination type, <tt>ty2</tt>, must be identical. If the source
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4065type is a pointer, the destination type must also be a pointer. This
4066instruction supports bitwise conversion of vectors to integers and to vectors
4067of other types (as long as they have the same size).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4068
4069<h5>Semantics:</h5>
4070<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
4071<tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
4072this conversion. The conversion is done as if the <tt>value</tt> had been
4073stored to memory and read back as type <tt>ty2</tt>. Pointer types may only be
4074converted to other pointer types with this instruction. To convert pointers to
4075other types, use the <a href="#i_inttoptr">inttoptr</a> or
4076<a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
4077
4078<h5>Example:</h5>
4079<pre>
4080 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
4081 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4082 %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]4083</pre>
4084</div>
4085
4086<!-- ======================================================================= -->
4087<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
4088<div class="doc_text">
4089<p>The instructions in this category are the "miscellaneous"
4090instructions, which defy better classification.</p>
4091</div>
4092
4093<!-- _______________________________________________________________________ -->
4094<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
4095</div>
4096<div class="doc_text">
4097<h5>Syntax:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4098<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]4099</pre>
4100<h5>Overview:</h5>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4101<p>The '<tt>icmp</tt>' instruction returns a boolean value or
4102a vector of boolean values based on comparison
4103of its two integer, integer vector, or pointer operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4104<h5>Arguments:</h5>
4105<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
4106the condition code indicating the kind of comparison to perform. It is not
4107a value, just a keyword. The possible condition code are:
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4108</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4109<ol>
4110 <li><tt>eq</tt>: equal</li>
4111 <li><tt>ne</tt>: not equal </li>
4112 <li><tt>ugt</tt>: unsigned greater than</li>
4113 <li><tt>uge</tt>: unsigned greater or equal</li>
4114 <li><tt>ult</tt>: unsigned less than</li>
4115 <li><tt>ule</tt>: unsigned less or equal</li>
4116 <li><tt>sgt</tt>: signed greater than</li>
4117 <li><tt>sge</tt>: signed greater or equal</li>
4118 <li><tt>slt</tt>: signed less than</li>
4119 <li><tt>sle</tt>: signed less or equal</li>
4120</ol>
4121<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4122<a href="#t_pointer">pointer</a>
4123or integer <a href="#t_vector">vector</a> typed.
4124They must also be identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4125<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4126<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]4127the condition code given as <tt>cond</tt>. The comparison performed always
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4128yields 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]4129</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4130<ol>
4131 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
4132 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
4133 </li>
4134 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4135 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4136 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4137 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4138 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4139 <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]4140 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4141 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4142 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4143 <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]4144 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4145 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4146 <li><tt>sge</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4147 <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]4148 <li><tt>slt</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4149 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4150 <li><tt>sle</tt>: interprets the operands as signed values and yields
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4151 <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]4152</ol>
4153<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
4154values are compared as if they were integers.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4155<p>If the operands are integer vectors, then they are compared
4156element by element. The result is an <tt>i1</tt> vector with
4157the same number of elements as the values being compared.
4158Otherwise, the result is an <tt>i1</tt>.
4159</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4160
4161<h5>Example:</h5>
4162<pre> &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
4163 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
4164 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
4165 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
4166 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
4167 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
4168</pre>
Dan Gohmana53eb382009-01-22 01:39:38 +0000[diff] [blame]4169
4170<p>Note that the code generator does not yet support vector types with
4171 the <tt>icmp</tt> instruction.</p>
4172
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4173</div>
4174
4175<!-- _______________________________________________________________________ -->
4176<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
4177</div>
4178<div class="doc_text">
4179<h5>Syntax:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4180<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]4181</pre>
4182<h5>Overview:</h5>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4183<p>The '<tt>fcmp</tt>' instruction returns a boolean value
4184or vector of boolean values based on comparison
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4185of its operands.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4186<p>
4187If the operands are floating point scalars, then the result
4188type is a boolean (<a href="#t_primitive"><tt>i1</tt></a>).
4189</p>
4190<p>If the operands are floating point vectors, then the result type
4191is a vector of boolean with the same number of elements as the
4192operands being compared.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4193<h5>Arguments:</h5>
4194<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
4195the condition code indicating the kind of comparison to perform. It is not
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4196a value, just a keyword. The possible condition code are:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4197<ol>
4198 <li><tt>false</tt>: no comparison, always returns false</li>
4199 <li><tt>oeq</tt>: ordered and equal</li>
4200 <li><tt>ogt</tt>: ordered and greater than </li>
4201 <li><tt>oge</tt>: ordered and greater than or equal</li>
4202 <li><tt>olt</tt>: ordered and less than </li>
4203 <li><tt>ole</tt>: ordered and less than or equal</li>
4204 <li><tt>one</tt>: ordered and not equal</li>
4205 <li><tt>ord</tt>: ordered (no nans)</li>
4206 <li><tt>ueq</tt>: unordered or equal</li>
4207 <li><tt>ugt</tt>: unordered or greater than </li>
4208 <li><tt>uge</tt>: unordered or greater than or equal</li>
4209 <li><tt>ult</tt>: unordered or less than </li>
4210 <li><tt>ule</tt>: unordered or less than or equal</li>
4211 <li><tt>une</tt>: unordered or not equal</li>
4212 <li><tt>uno</tt>: unordered (either nans)</li>
4213 <li><tt>true</tt>: no comparison, always returns true</li>
4214</ol>
4215<p><i>Ordered</i> means that neither operand is a QNAN while
4216<i>unordered</i> means that either operand may be a QNAN.</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4217<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be
4218either a <a href="#t_floating">floating point</a> type
4219or a <a href="#t_vector">vector</a> of floating point type.
4220They must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4221<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4222<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4223according to the condition code given as <tt>cond</tt>.
4224If the operands are vectors, then the vectors are compared
4225element by element.
4226Each comparison performed
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4227always yields an <a href="#t_primitive">i1</a> result, as follows:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4228<ol>
4229 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
4230 <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]4231 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4232 <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]4233 <tt>op1</tt> is greather than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4234 <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]4235 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4236 <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]4237 <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4238 <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]4239 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4240 <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]4241 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4242 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
4243 <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]4244 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4245 <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]4246 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4247 <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]4248 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4249 <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]4250 <tt>op1</tt> is less than <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4251 <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]4252 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4253 <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]4254 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4255 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
4256 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
4257</ol>
4258
4259<h5>Example:</h5>
4260<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]4261 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
4262 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
4263 &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]4264</pre>
Dan Gohmana53eb382009-01-22 01:39:38 +0000[diff] [blame]4265
4266<p>Note that the code generator does not yet support vector types with
4267 the <tt>fcmp</tt> instruction.</p>
4268
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4269</div>
4270
4271<!-- _______________________________________________________________________ -->
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4272<div class="doc_subsubsection">
4273 <a name="i_vicmp">'<tt>vicmp</tt>' Instruction</a>
4274</div>
4275<div class="doc_text">
4276<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4277<pre> &lt;result&gt; = vicmp &lt;cond&gt; &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4278</pre>
4279<h5>Overview:</h5>
4280<p>The '<tt>vicmp</tt>' instruction returns an integer vector value based on
4281element-wise comparison of its two integer vector operands.</p>
4282<h5>Arguments:</h5>
4283<p>The '<tt>vicmp</tt>' instruction takes three operands. The first operand is
4284the condition code indicating the kind of comparison to perform. It is not
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4285a value, just a keyword. The possible condition code are:</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4286<ol>
4287 <li><tt>eq</tt>: equal</li>
4288 <li><tt>ne</tt>: not equal </li>
4289 <li><tt>ugt</tt>: unsigned greater than</li>
4290 <li><tt>uge</tt>: unsigned greater or equal</li>
4291 <li><tt>ult</tt>: unsigned less than</li>
4292 <li><tt>ule</tt>: unsigned less or equal</li>
4293 <li><tt>sgt</tt>: signed greater than</li>
4294 <li><tt>sge</tt>: signed greater or equal</li>
4295 <li><tt>slt</tt>: signed less than</li>
4296 <li><tt>sle</tt>: signed less or equal</li>
4297</ol>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4298<p>The remaining two arguments must be <a href="#t_vector">vector</a> or
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4299<a href="#t_integer">integer</a> typed. They must also be identical types.</p>
4300<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4301<p>The '<tt>vicmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4302according to the condition code given as <tt>cond</tt>. The comparison yields a
4303<a href="#t_vector">vector</a> of <a href="#t_integer">integer</a> result, of
4304identical type as the values being compared. The most significant bit in each
4305element is 1 if the element-wise comparison evaluates to true, and is 0
4306otherwise. All other bits of the result are undefined. The condition codes
4307are evaluated identically to the <a href="#i_icmp">'<tt>icmp</tt>'
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4308instruction</a>.</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4309
4310<h5>Example:</h5>
4311<pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4312 &lt;result&gt; = vicmp eq &lt;2 x i32&gt; &lt; i32 4, i32 0&gt;, &lt; i32 5, i32 0&gt; <i>; yields: result=&lt;2 x i32&gt; &lt; i32 0, i32 -1 &gt;</i>
4313 &lt;result&gt; = vicmp ult &lt;2 x i8 &gt; &lt; i8 1, i8 2&gt;, &lt; i8 2, i8 2 &gt; <i>; yields: result=&lt;2 x i8&gt; &lt; i8 -1, i8 0 &gt;</i>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4314</pre>
4315</div>
4316
4317<!-- _______________________________________________________________________ -->
4318<div class="doc_subsubsection">
4319 <a name="i_vfcmp">'<tt>vfcmp</tt>' Instruction</a>
4320</div>
4321<div class="doc_text">
4322<h5>Syntax:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4323<pre> &lt;result&gt; = vfcmp &lt;cond&gt; &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt;</pre>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4324<h5>Overview:</h5>
4325<p>The '<tt>vfcmp</tt>' instruction returns an integer vector value based on
4326element-wise comparison of its two floating point vector operands. The output
4327elements have the same width as the input elements.</p>
4328<h5>Arguments:</h5>
4329<p>The '<tt>vfcmp</tt>' instruction takes three operands. The first operand is
4330the condition code indicating the kind of comparison to perform. It is not
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4331a value, just a keyword. The possible condition code are:</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4332<ol>
4333 <li><tt>false</tt>: no comparison, always returns false</li>
4334 <li><tt>oeq</tt>: ordered and equal</li>
4335 <li><tt>ogt</tt>: ordered and greater than </li>
4336 <li><tt>oge</tt>: ordered and greater than or equal</li>
4337 <li><tt>olt</tt>: ordered and less than </li>
4338 <li><tt>ole</tt>: ordered and less than or equal</li>
4339 <li><tt>one</tt>: ordered and not equal</li>
4340 <li><tt>ord</tt>: ordered (no nans)</li>
4341 <li><tt>ueq</tt>: unordered or equal</li>
4342 <li><tt>ugt</tt>: unordered or greater than </li>
4343 <li><tt>uge</tt>: unordered or greater than or equal</li>
4344 <li><tt>ult</tt>: unordered or less than </li>
4345 <li><tt>ule</tt>: unordered or less than or equal</li>
4346 <li><tt>une</tt>: unordered or not equal</li>
4347 <li><tt>uno</tt>: unordered (either nans)</li>
4348 <li><tt>true</tt>: no comparison, always returns true</li>
4349</ol>
4350<p>The remaining two arguments must be <a href="#t_vector">vector</a> of
4351<a href="#t_floating">floating point</a> typed. They must also be identical
4352types.</p>
4353<h5>Semantics:</h5>
Gabor Greifd9068fe2008-08-07 21:46:00 +0000[diff] [blame]4354<p>The '<tt>vfcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4355according to the condition code given as <tt>cond</tt>. The comparison yields a
4356<a href="#t_vector">vector</a> of <a href="#t_integer">integer</a> result, with
4357an identical number of elements as the values being compared, and each element
4358having identical with to the width of the floating point elements. The most
4359significant bit in each element is 1 if the element-wise comparison evaluates to
4360true, and is 0 otherwise. All other bits of the result are undefined. The
4361condition codes are evaluated identically to the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4362<a href="#i_fcmp">'<tt>fcmp</tt>' instruction</a>.</p>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4363
4364<h5>Example:</h5>
4365<pre>
Chris Lattner1e5c5cd02008-10-13 16:55:18 +0000[diff] [blame]4366 <i>; yields: result=&lt;2 x i32&gt; &lt; i32 0, i32 -1 &gt;</i>
4367 &lt;result&gt; = vfcmp oeq &lt;2 x float&gt; &lt; float 4, float 0 &gt;, &lt; float 5, float 0 &gt;
4368
4369 <i>; yields: result=&lt;2 x i64&gt; &lt; i64 -1, i64 0 &gt;</i>
4370 &lt;result&gt; = vfcmp ult &lt;2 x double&gt; &lt; double 1, double 2 &gt;, &lt; double 2, double 2&gt;
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4371</pre>
4372</div>
4373
4374<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4375<div class="doc_subsubsection">
4376 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
4377</div>
4378
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4379<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4380
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4381<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4382
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4383<pre> &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...<br></pre>
4384<h5>Overview:</h5>
4385<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in
4386the SSA graph representing the function.</p>
4387<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4388
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4389<p>The type of the incoming values is specified with the first type
4390field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
4391as arguments, with one pair for each predecessor basic block of the
4392current block. Only values of <a href="#t_firstclass">first class</a>
4393type may be used as the value arguments to the PHI node. Only labels
4394may be used as the label arguments.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4395
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4396<p>There must be no non-phi instructions between the start of a basic
4397block and the PHI instructions: i.e. PHI instructions must be first in
4398a basic block.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4399
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4400<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4401
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4402<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
4403specified by the pair corresponding to the predecessor basic block that executed
4404just prior to the current block.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4405
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4406<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4407<pre>
4408Loop: ; Infinite loop that counts from 0 on up...
4409 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
4410 %nextindvar = add i32 %indvar, 1
4411 br label %Loop
4412</pre>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4413</div>
4414
4415<!-- _______________________________________________________________________ -->
4416<div class="doc_subsubsection">
4417 <a name="i_select">'<tt>select</tt>' Instruction</a>
4418</div>
4419
4420<div class="doc_text">
4421
4422<h5>Syntax:</h5>
4423
4424<pre>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4425 &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>
4426
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4427 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4428</pre>
4429
4430<h5>Overview:</h5>
4431
4432<p>
4433The '<tt>select</tt>' instruction is used to choose one value based on a
4434condition, without branching.
4435</p>
4436
4437
4438<h5>Arguments:</h5>
4439
4440<p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4441The '<tt>select</tt>' instruction requires an 'i1' value or
4442a vector of 'i1' values indicating the
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4443condition, and two values of the same <a href="#t_firstclass">first class</a>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4444type. If the val1/val2 are vectors and
4445the condition is a scalar, then entire vectors are selected, not
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4446individual elements.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4447</p>
4448
4449<h5>Semantics:</h5>
4450
4451<p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4452If 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]4453value argument; otherwise, it returns the second value argument.
4454</p>
Dan Gohmanb60ca3c2008-09-09 01:02:47 +0000[diff] [blame]4455<p>
4456If the condition is a vector of i1, then the value arguments must
4457be vectors of the same size, and the selection is done element
4458by element.
4459</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4460
4461<h5>Example:</h5>
4462
4463<pre>
4464 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
4465</pre>
Dan Gohmana53eb382009-01-22 01:39:38 +0000[diff] [blame]4466
4467<p>Note that the code generator does not yet support conditions
4468 with vector type.</p>
4469
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4470</div>
4471
4472
4473<!-- _______________________________________________________________________ -->
4474<div class="doc_subsubsection">
4475 <a name="i_call">'<tt>call</tt>' Instruction</a>
4476</div>
4477
4478<div class="doc_text">
4479
4480<h5>Syntax:</h5>
4481<pre>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]4482 &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]4483</pre>
4484
4485<h5>Overview:</h5>
4486
4487<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
4488
4489<h5>Arguments:</h5>
4490
4491<p>This instruction requires several arguments:</p>
4492
4493<ol>
4494 <li>
4495 <p>The optional "tail" marker indicates whether the callee function accesses
4496 any allocas or varargs in the caller. If the "tail" marker is present, the
4497 function call is eligible for tail call optimization. Note that calls may
4498 be marked "tail" even if they do not occur before a <a
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4499 href="#i_ret"><tt>ret</tt></a> instruction.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4500 </li>
4501 <li>
4502 <p>The optional "cconv" marker indicates which <a href="#callingconv">calling
4503 convention</a> the call should use. If none is specified, the call defaults
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4504 to using C calling conventions.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4505 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4506
4507 <li>
4508 <p>The optional <a href="#paramattrs">Parameter Attributes</a> list for
4509 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>',
4510 and '<tt>inreg</tt>' attributes are valid here.</p>
4511 </li>
4512
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4513 <li>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]4514 <p>'<tt>ty</tt>': the type of the call instruction itself which is also
4515 the type of the return value. Functions that return no value are marked
4516 <tt><a href="#t_void">void</a></tt>.</p>
4517 </li>
4518 <li>
4519 <p>'<tt>fnty</tt>': shall be the signature of the pointer to function
4520 value being invoked. The argument types must match the types implied by
4521 this signature. This type can be omitted if the function is not varargs
4522 and if the function type does not return a pointer to a function.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4523 </li>
4524 <li>
4525 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
4526 be invoked. In most cases, this is a direct function invocation, but
4527 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
4528 to function value.</p>
4529 </li>
4530 <li>
4531 <p>'<tt>function args</tt>': argument list whose types match the
4532 function signature argument types. All arguments must be of
4533 <a href="#t_firstclass">first class</a> type. If the function signature
4534 indicates the function accepts a variable number of arguments, the extra
4535 arguments can be specified.</p>
4536 </li>
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4537 <li>
Devang Pateld0bfcc72008-10-07 17:48:33 +0000[diff] [blame]4538 <p>The optional <a href="#fnattrs">function attributes</a> list. Only
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4539 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
4540 '<tt>readnone</tt>' attributes are valid here.</p>
4541 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4542</ol>
4543
4544<h5>Semantics:</h5>
4545
4546<p>The '<tt>call</tt>' instruction is used to cause control flow to
4547transfer to a specified function, with its incoming arguments bound to
4548the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
4549instruction in the called function, control flow continues with the
4550instruction after the function call, and the return value of the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4551function is bound to the result argument.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4552
4553<h5>Example:</h5>
4554
4555<pre>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]4556 %retval = call i32 @test(i32 %argc)
Chris Lattner5e893ef2008-03-21 17:24:17 +0000[diff] [blame]4557 call i32 (i8 *, ...)* @printf(i8 * %msg, i32 12, i8 42) <i>; yields i32</i>
4558 %X = tail call i32 @foo() <i>; yields i32</i>
4559 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
4560 call void %foo(i8 97 signext)
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]4561
4562 %struct.A = type { i32, i8 }
Devang Patelac2fc272008-10-06 18:50:38 +0000[diff] [blame]4563 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohman3e700032008-10-04 19:00:07 +0000[diff] [blame]4564 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
4565 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattnerac454b32008-10-08 06:26:11 +0000[diff] [blame]4566 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijman2c4e05a2008-10-07 10:03:45 +0000[diff] [blame]4567 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4568</pre>
4569
4570</div>
4571
4572<!-- _______________________________________________________________________ -->
4573<div class="doc_subsubsection">
4574 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
4575</div>
4576
4577<div class="doc_text">
4578
4579<h5>Syntax:</h5>
4580
4581<pre>
4582 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
4583</pre>
4584
4585<h5>Overview:</h5>
4586
4587<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
4588the "variable argument" area of a function call. It is used to implement the
4589<tt>va_arg</tt> macro in C.</p>
4590
4591<h5>Arguments:</h5>
4592
4593<p>This instruction takes a <tt>va_list*</tt> value and the type of
4594the argument. It returns a value of the specified argument type and
4595increments the <tt>va_list</tt> to point to the next argument. The
4596actual type of <tt>va_list</tt> is target specific.</p>
4597
4598<h5>Semantics:</h5>
4599
4600<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
4601type from the specified <tt>va_list</tt> and causes the
4602<tt>va_list</tt> to point to the next argument. For more information,
4603see the variable argument handling <a href="#int_varargs">Intrinsic
4604Functions</a>.</p>
4605
4606<p>It is legal for this instruction to be called in a function which does not
4607take a variable number of arguments, for example, the <tt>vfprintf</tt>
4608function.</p>
4609
4610<p><tt>va_arg</tt> is an LLVM instruction instead of an <a
4611href="#intrinsics">intrinsic function</a> because it takes a type as an
4612argument.</p>
4613
4614<h5>Example:</h5>
4615
4616<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
4617
Dan Gohman60967192009-01-12 23:12:39 +0000[diff] [blame]4618<p>Note that the code generator does not yet fully support va_arg
4619 on many targets. Also, it does not currently support va_arg with
4620 aggregate types on any target.</p>
4621
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4622</div>
4623
4624<!-- *********************************************************************** -->
4625<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
4626<!-- *********************************************************************** -->
4627
4628<div class="doc_text">
4629
4630<p>LLVM supports the notion of an "intrinsic function". These functions have
4631well known names and semantics and are required to follow certain restrictions.
4632Overall, these intrinsics represent an extension mechanism for the LLVM
4633language that does not require changing all of the transformations in LLVM when
4634adding to the language (or the bitcode reader/writer, the parser, etc...).</p>
4635
4636<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
4637prefix is reserved in LLVM for intrinsic names; thus, function names may not
4638begin with this prefix. Intrinsic functions must always be external functions:
4639you cannot define the body of intrinsic functions. Intrinsic functions may
4640only be used in call or invoke instructions: it is illegal to take the address
4641of an intrinsic function. Additionally, because intrinsic functions are part
4642of the LLVM language, it is required if any are added that they be documented
4643here.</p>
4644
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4645<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents
4646a family of functions that perform the same operation but on different data
4647types. Because LLVM can represent over 8 million different integer types,
4648overloading is used commonly to allow an intrinsic function to operate on any
4649integer type. One or more of the argument types or the result type can be
4650overloaded to accept any integer type. Argument types may also be defined as
4651exactly matching a previous argument's type or the result type. This allows an
4652intrinsic function which accepts multiple arguments, but needs all of them to
4653be of the same type, to only be overloaded with respect to a single argument or
4654the result.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4655
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4656<p>Overloaded intrinsics will have the names of its overloaded argument types
4657encoded into its function name, each preceded by a period. Only those types
4658which are overloaded result in a name suffix. Arguments whose type is matched
4659against another type do not. For example, the <tt>llvm.ctpop</tt> function can
4660take an integer of any width and returns an integer of exactly the same integer
4661width. This leads to a family of functions such as
4662<tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29 %val)</tt>.
4663Only one type, the return type, is overloaded, and only one type suffix is
4664required. Because the argument's type is matched against the return type, it
4665does not require its own name suffix.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4666
4667<p>To learn how to add an intrinsic function, please see the
4668<a href="ExtendingLLVM.html">Extending LLVM Guide</a>.
4669</p>
4670
4671</div>
4672
4673<!-- ======================================================================= -->
4674<div class="doc_subsection">
4675 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
4676</div>
4677
4678<div class="doc_text">
4679
4680<p>Variable argument support is defined in LLVM with the <a
4681 href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
4682intrinsic functions. These functions are related to the similarly
4683named macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
4684
4685<p>All of these functions operate on arguments that use a
4686target-specific value type "<tt>va_list</tt>". The LLVM assembly
4687language reference manual does not define what this type is, so all
4688transformations should be prepared to handle these functions regardless of
4689the type used.</p>
4690
4691<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
4692instruction and the variable argument handling intrinsic functions are
4693used.</p>
4694
4695<div class="doc_code">
4696<pre>
4697define i32 @test(i32 %X, ...) {
4698 ; Initialize variable argument processing
4699 %ap = alloca i8*
4700 %ap2 = bitcast i8** %ap to i8*
4701 call void @llvm.va_start(i8* %ap2)
4702
4703 ; Read a single integer argument
4704 %tmp = va_arg i8** %ap, i32
4705
4706 ; Demonstrate usage of llvm.va_copy and llvm.va_end
4707 %aq = alloca i8*
4708 %aq2 = bitcast i8** %aq to i8*
4709 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
4710 call void @llvm.va_end(i8* %aq2)
4711
4712 ; Stop processing of arguments.
4713 call void @llvm.va_end(i8* %ap2)
4714 ret i32 %tmp
4715}
4716
4717declare void @llvm.va_start(i8*)
4718declare void @llvm.va_copy(i8*, i8*)
4719declare void @llvm.va_end(i8*)
4720</pre>
4721</div>
4722
4723</div>
4724
4725<!-- _______________________________________________________________________ -->
4726<div class="doc_subsubsection">
4727 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
4728</div>
4729
4730
4731<div class="doc_text">
4732<h5>Syntax:</h5>
4733<pre> declare void %llvm.va_start(i8* &lt;arglist&gt;)<br></pre>
4734<h5>Overview:</h5>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4735<p>The '<tt>llvm.va_start</tt>' intrinsic initializes
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4736<tt>*&lt;arglist&gt;</tt> for subsequent use by <tt><a
4737href="#i_va_arg">va_arg</a></tt>.</p>
4738
4739<h5>Arguments:</h5>
4740
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4741<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]4742
4743<h5>Semantics:</h5>
4744
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]4745<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]4746macro available in C. In a target-dependent way, it initializes the
4747<tt>va_list</tt> element to which the argument points, so that the next call to
4748<tt>va_arg</tt> will produce the first variable argument passed to the function.
4749Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
4750last argument of the function as the compiler can figure that out.</p>
4751
4752</div>
4753
4754<!-- _______________________________________________________________________ -->
4755<div class="doc_subsubsection">
4756 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
4757</div>
4758
4759<div class="doc_text">
4760<h5>Syntax:</h5>
4761<pre> declare void @llvm.va_end(i8* &lt;arglist&gt;)<br></pre>
4762<h5>Overview:</h5>
4763
4764<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
4765which has been initialized previously with <tt><a href="#int_va_start">llvm.va_start</a></tt>
4766or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
4767
4768<h5>Arguments:</h5>
4769
4770<p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
4771
4772<h5>Semantics:</h5>
4773
4774<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
4775macro available in C. In a target-dependent way, it destroys the
4776<tt>va_list</tt> element to which the argument points. Calls to <a
4777href="#int_va_start"><tt>llvm.va_start</tt></a> and <a href="#int_va_copy">
4778<tt>llvm.va_copy</tt></a> must be matched exactly with calls to
4779<tt>llvm.va_end</tt>.</p>
4780
4781</div>
4782
4783<!-- _______________________________________________________________________ -->
4784<div class="doc_subsubsection">
4785 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
4786</div>
4787
4788<div class="doc_text">
4789
4790<h5>Syntax:</h5>
4791
4792<pre>
4793 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
4794</pre>
4795
4796<h5>Overview:</h5>
4797
4798<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
4799from the source argument list to the destination argument list.</p>
4800
4801<h5>Arguments:</h5>
4802
4803<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
4804The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
4805
4806
4807<h5>Semantics:</h5>
4808
4809<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
4810macro available in C. In a target-dependent way, it copies the source
4811<tt>va_list</tt> element into the destination <tt>va_list</tt> element. This
4812intrinsic is necessary because the <tt><a href="#int_va_start">
4813llvm.va_start</a></tt> intrinsic may be arbitrarily complex and require, for
4814example, memory allocation.</p>
4815
4816</div>
4817
4818<!-- ======================================================================= -->
4819<div class="doc_subsection">
4820 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
4821</div>
4822
4823<div class="doc_text">
4824
4825<p>
4826LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]4827Collection</a> (GC) requires the implementation and generation of these
4828intrinsics.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4829These intrinsics allow identification of <a href="#int_gcroot">GC roots on the
4830stack</a>, as well as garbage collector implementations that require <a
4831href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a> barriers.
4832Front-ends for type-safe garbage collected languages should generate these
4833intrinsics to make use of the LLVM garbage collectors. For more details, see <a
4834href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
4835</p>
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]4836
4837<p>The garbage collection intrinsics only operate on objects in the generic
4838 address space (address space zero).</p>
4839
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4840</div>
4841
4842<!-- _______________________________________________________________________ -->
4843<div class="doc_subsubsection">
4844 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
4845</div>
4846
4847<div class="doc_text">
4848
4849<h5>Syntax:</h5>
4850
4851<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4852 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4853</pre>
4854
4855<h5>Overview:</h5>
4856
4857<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
4858the code generator, and allows some metadata to be associated with it.</p>
4859
4860<h5>Arguments:</h5>
4861
4862<p>The first argument specifies the address of a stack object that contains the
4863root pointer. The second pointer (which must be either a constant or a global
4864value address) contains the meta-data to be associated with the root.</p>
4865
4866<h5>Semantics:</h5>
4867
Chris Lattnera7d94ba2008-04-24 05:59:56 +0000[diff] [blame]4868<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]4869location. At compile-time, the code generator generates information to allow
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4870the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
4871intrinsic may only be used in a function which <a href="#gc">specifies a GC
4872algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4873
4874</div>
4875
4876
4877<!-- _______________________________________________________________________ -->
4878<div class="doc_subsubsection">
4879 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
4880</div>
4881
4882<div class="doc_text">
4883
4884<h5>Syntax:</h5>
4885
4886<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4887 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4888</pre>
4889
4890<h5>Overview:</h5>
4891
4892<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
4893locations, allowing garbage collector implementations that require read
4894barriers.</p>
4895
4896<h5>Arguments:</h5>
4897
4898<p>The second argument is the address to read from, which should be an address
4899allocated from the garbage collector. The first object is a pointer to the
4900start of the referenced object, if needed by the language runtime (otherwise
4901null).</p>
4902
4903<h5>Semantics:</h5>
4904
4905<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
4906instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4907garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
4908may only be used in a function which <a href="#gc">specifies a GC
4909algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4910
4911</div>
4912
4913
4914<!-- _______________________________________________________________________ -->
4915<div class="doc_subsubsection">
4916 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
4917</div>
4918
4919<div class="doc_text">
4920
4921<h5>Syntax:</h5>
4922
4923<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4924 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4925</pre>
4926
4927<h5>Overview:</h5>
4928
4929<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
4930locations, allowing garbage collector implementations that require write
4931barriers (such as generational or reference counting collectors).</p>
4932
4933<h5>Arguments:</h5>
4934
4935<p>The first argument is the reference to store, the second is the start of the
4936object to store it to, and the third is the address of the field of Obj to
4937store to. If the runtime does not require a pointer to the object, Obj may be
4938null.</p>
4939
4940<h5>Semantics:</h5>
4941
4942<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
4943instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4944garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
4945may only be used in a function which <a href="#gc">specifies a GC
4946algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4947
4948</div>
4949
4950
4951
4952<!-- ======================================================================= -->
4953<div class="doc_subsection">
4954 <a name="int_codegen">Code Generator Intrinsics</a>
4955</div>
4956
4957<div class="doc_text">
4958<p>
4959These intrinsics are provided by LLVM to expose special features that may only
4960be implemented with code generator support.
4961</p>
4962
4963</div>
4964
4965<!-- _______________________________________________________________________ -->
4966<div class="doc_subsubsection">
4967 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
4968</div>
4969
4970<div class="doc_text">
4971
4972<h5>Syntax:</h5>
4973<pre>
4974 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
4975</pre>
4976
4977<h5>Overview:</h5>
4978
4979<p>
4980The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
4981target-specific value indicating the return address of the current function
4982or one of its callers.
4983</p>
4984
4985<h5>Arguments:</h5>
4986
4987<p>
4988The argument to this intrinsic indicates which function to return the address
4989for. Zero indicates the calling function, one indicates its caller, etc. The
4990argument is <b>required</b> to be a constant integer value.
4991</p>
4992
4993<h5>Semantics:</h5>
4994
4995<p>
4996The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
4997the return address of the specified call frame, or zero if it cannot be
4998identified. The value returned by this intrinsic is likely to be incorrect or 0
4999for arguments other than zero, so it should only be used for debugging purposes.
5000</p>
5001
5002<p>
5003Note that calling this intrinsic does not prevent function inlining or other
5004aggressive transformations, so the value returned may not be that of the obvious
5005source-language caller.
5006</p>
5007</div>
5008
5009
5010<!-- _______________________________________________________________________ -->
5011<div class="doc_subsubsection">
5012 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
5013</div>
5014
5015<div class="doc_text">
5016
5017<h5>Syntax:</h5>
5018<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5019 declare i8 *@llvm.frameaddress(i32 &lt;level&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5020</pre>
5021
5022<h5>Overview:</h5>
5023
5024<p>
5025The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
5026target-specific frame pointer value for the specified stack frame.
5027</p>
5028
5029<h5>Arguments:</h5>
5030
5031<p>
5032The argument to this intrinsic indicates which function to return the frame
5033pointer for. Zero indicates the calling function, one indicates its caller,
5034etc. The argument is <b>required</b> to be a constant integer value.
5035</p>
5036
5037<h5>Semantics:</h5>
5038
5039<p>
5040The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
5041the frame address of the specified call frame, or zero if it cannot be
5042identified. The value returned by this intrinsic is likely to be incorrect or 0
5043for arguments other than zero, so it should only be used for debugging purposes.
5044</p>
5045
5046<p>
5047Note that calling this intrinsic does not prevent function inlining or other
5048aggressive transformations, so the value returned may not be that of the obvious
5049source-language caller.
5050</p>
5051</div>
5052
5053<!-- _______________________________________________________________________ -->
5054<div class="doc_subsubsection">
5055 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
5056</div>
5057
5058<div class="doc_text">
5059
5060<h5>Syntax:</h5>
5061<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5062 declare i8 *@llvm.stacksave()
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5063</pre>
5064
5065<h5>Overview:</h5>
5066
5067<p>
5068The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
5069the function stack, for use with <a href="#int_stackrestore">
5070<tt>llvm.stackrestore</tt></a>. This is useful for implementing language
5071features like scoped automatic variable sized arrays in C99.
5072</p>
5073
5074<h5>Semantics:</h5>
5075
5076<p>
5077This intrinsic returns a opaque pointer value that can be passed to <a
5078href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
5079<tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
5080<tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
5081state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
5082practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
5083that were allocated after the <tt>llvm.stacksave</tt> was executed.
5084</p>
5085
5086</div>
5087
5088<!-- _______________________________________________________________________ -->
5089<div class="doc_subsubsection">
5090 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
5091</div>
5092
5093<div class="doc_text">
5094
5095<h5>Syntax:</h5>
5096<pre>
5097 declare void @llvm.stackrestore(i8 * %ptr)
5098</pre>
5099
5100<h5>Overview:</h5>
5101
5102<p>
5103The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
5104the function stack to the state it was in when the corresponding <a
5105href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
5106useful for implementing language features like scoped automatic variable sized
5107arrays in C99.
5108</p>
5109
5110<h5>Semantics:</h5>
5111
5112<p>
5113See the description for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.
5114</p>
5115
5116</div>
5117
5118
5119<!-- _______________________________________________________________________ -->
5120<div class="doc_subsubsection">
5121 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
5122</div>
5123
5124<div class="doc_text">
5125
5126<h5>Syntax:</h5>
5127<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5128 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]5129</pre>
5130
5131<h5>Overview:</h5>
5132
5133
5134<p>
5135The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
5136a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
5137no
5138effect on the behavior of the program but can change its performance
5139characteristics.
5140</p>
5141
5142<h5>Arguments:</h5>
5143
5144<p>
5145<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
5146determining if the fetch should be for a read (0) or write (1), and
5147<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
5148locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
5149<tt>locality</tt> arguments must be constant integers.
5150</p>
5151
5152<h5>Semantics:</h5>
5153
5154<p>
5155This intrinsic does not modify the behavior of the program. In particular,
5156prefetches cannot trap and do not produce a value. On targets that support this
5157intrinsic, the prefetch can provide hints to the processor cache for better
5158performance.
5159</p>
5160
5161</div>
5162
5163<!-- _______________________________________________________________________ -->
5164<div class="doc_subsubsection">
5165 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
5166</div>
5167
5168<div class="doc_text">
5169
5170<h5>Syntax:</h5>
5171<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]5172 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5173</pre>
5174
5175<h5>Overview:</h5>
5176
5177
5178<p>
5179The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]5180(PC) in a region of
5181code to simulators and other tools. The method is target specific, but it is
5182expected that the marker will use exported symbols to transmit the PC of the
5183marker.
5184The marker makes no guarantees that it will remain with any specific instruction
5185after optimizations. It is possible that the presence of a marker will inhibit
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5186optimizations. The intended use is to be inserted after optimizations to allow
5187correlations of simulation runs.
5188</p>
5189
5190<h5>Arguments:</h5>
5191
5192<p>
5193<tt>id</tt> is a numerical id identifying the marker.
5194</p>
5195
5196<h5>Semantics:</h5>
5197
5198<p>
5199This intrinsic does not modify the behavior of the program. Backends that do not
5200support this intrinisic may ignore it.
5201</p>
5202
5203</div>
5204
5205<!-- _______________________________________________________________________ -->
5206<div class="doc_subsubsection">
5207 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
5208</div>
5209
5210<div class="doc_text">
5211
5212<h5>Syntax:</h5>
5213<pre>
5214 declare i64 @llvm.readcyclecounter( )
5215</pre>
5216
5217<h5>Overview:</h5>
5218
5219
5220<p>
5221The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
5222counter register (or similar low latency, high accuracy clocks) on those targets
5223that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
5224As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
5225should only be used for small timings.
5226</p>
5227
5228<h5>Semantics:</h5>
5229
5230<p>
5231When directly supported, reading the cycle counter should not modify any memory.
5232Implementations are allowed to either return a application specific value or a
5233system wide value. On backends without support, this is lowered to a constant 0.
5234</p>
5235
5236</div>
5237
5238<!-- ======================================================================= -->
5239<div class="doc_subsection">
5240 <a name="int_libc">Standard C Library Intrinsics</a>
5241</div>
5242
5243<div class="doc_text">
5244<p>
5245LLVM provides intrinsics for a few important standard C library functions.
5246These intrinsics allow source-language front-ends to pass information about the
5247alignment of the pointer arguments to the code generator, providing opportunity
5248for more efficient code generation.
5249</p>
5250
5251</div>
5252
5253<!-- _______________________________________________________________________ -->
5254<div class="doc_subsubsection">
5255 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
5256</div>
5257
5258<div class="doc_text">
5259
5260<h5>Syntax:</h5>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5261<p>This is an overloaded intrinsic. You can use llvm.memcpy on any integer bit
5262width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5263<pre>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5264 declare void @llvm.memcpy.i8(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5265 i8 &lt;len&gt;, i32 &lt;align&gt;)
5266 declare void @llvm.memcpy.i16(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5267 i16 &lt;len&gt;, i32 &lt;align&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5268 declare void @llvm.memcpy.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5269 i32 &lt;len&gt;, i32 &lt;align&gt;)
5270 declare void @llvm.memcpy.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5271 i64 &lt;len&gt;, i32 &lt;align&gt;)
5272</pre>
5273
5274<h5>Overview:</h5>
5275
5276<p>
5277The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
5278location to the destination location.
5279</p>
5280
5281<p>
5282Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
5283intrinsics do not return a value, and takes an extra alignment argument.
5284</p>
5285
5286<h5>Arguments:</h5>
5287
5288<p>
5289The first argument is a pointer to the destination, the second is a pointer to
5290the source. The third argument is an integer argument
5291specifying the number of bytes to copy, and the fourth argument is the alignment
5292of the source and destination locations.
5293</p>
5294
5295<p>
5296If the call to this intrinisic has an alignment value that is not 0 or 1, then
5297the caller guarantees that both the source and destination pointers are aligned
5298to that boundary.
5299</p>
5300
5301<h5>Semantics:</h5>
5302
5303<p>
5304The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
5305location to the destination location, which are not allowed to overlap. It
5306copies "len" bytes of memory over. If the argument is known to be aligned to
5307some boundary, this can be specified as the fourth argument, otherwise it should
5308be set to 0 or 1.
5309</p>
5310</div>
5311
5312
5313<!-- _______________________________________________________________________ -->
5314<div class="doc_subsubsection">
5315 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
5316</div>
5317
5318<div class="doc_text">
5319
5320<h5>Syntax:</h5>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5321<p>This is an overloaded intrinsic. You can use llvm.memmove on any integer bit
5322width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5323<pre>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5324 declare void @llvm.memmove.i8(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5325 i8 &lt;len&gt;, i32 &lt;align&gt;)
5326 declare void @llvm.memmove.i16(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5327 i16 &lt;len&gt;, i32 &lt;align&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5328 declare void @llvm.memmove.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5329 i32 &lt;len&gt;, i32 &lt;align&gt;)
5330 declare void @llvm.memmove.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5331 i64 &lt;len&gt;, i32 &lt;align&gt;)
5332</pre>
5333
5334<h5>Overview:</h5>
5335
5336<p>
5337The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the source
5338location to the destination location. It is similar to the
Chris Lattnerdba16ea2008-01-06 19:51:52 +0000[diff] [blame]5339'<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to overlap.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5340</p>
5341
5342<p>
5343Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
5344intrinsics do not return a value, and takes an extra alignment argument.
5345</p>
5346
5347<h5>Arguments:</h5>
5348
5349<p>
5350The first argument is a pointer to the destination, the second is a pointer to
5351the source. The third argument is an integer argument
5352specifying the number of bytes to copy, and the fourth argument is the alignment
5353of the source and destination locations.
5354</p>
5355
5356<p>
5357If the call to this intrinisic has an alignment value that is not 0 or 1, then
5358the caller guarantees that the source and destination pointers are aligned to
5359that boundary.
5360</p>
5361
5362<h5>Semantics:</h5>
5363
5364<p>
5365The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the source
5366location to the destination location, which may overlap. It
5367copies "len" bytes of memory over. If the argument is known to be aligned to
5368some boundary, this can be specified as the fourth argument, otherwise it should
5369be set to 0 or 1.
5370</p>
5371</div>
5372
5373
5374<!-- _______________________________________________________________________ -->
5375<div class="doc_subsubsection">
5376 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
5377</div>
5378
5379<div class="doc_text">
5380
5381<h5>Syntax:</h5>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5382<p>This is an overloaded intrinsic. You can use llvm.memset on any integer bit
5383width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5384<pre>
Chris Lattner82c2e432008-11-21 16:42:48 +0000[diff] [blame]5385 declare void @llvm.memset.i8(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5386 i8 &lt;len&gt;, i32 &lt;align&gt;)
5387 declare void @llvm.memset.i16(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5388 i16 &lt;len&gt;, i32 &lt;align&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5389 declare void @llvm.memset.i32(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5390 i32 &lt;len&gt;, i32 &lt;align&gt;)
5391 declare void @llvm.memset.i64(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5392 i64 &lt;len&gt;, i32 &lt;align&gt;)
5393</pre>
5394
5395<h5>Overview:</h5>
5396
5397<p>
5398The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a particular
5399byte value.
5400</p>
5401
5402<p>
5403Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
5404does not return a value, and takes an extra alignment argument.
5405</p>
5406
5407<h5>Arguments:</h5>
5408
5409<p>
5410The first argument is a pointer to the destination to fill, the second is the
5411byte value to fill it with, the third argument is an integer
5412argument specifying the number of bytes to fill, and the fourth argument is the
5413known alignment of destination location.
5414</p>
5415
5416<p>
5417If the call to this intrinisic has an alignment value that is not 0 or 1, then
5418the caller guarantees that the destination pointer is aligned to that boundary.
5419</p>
5420
5421<h5>Semantics:</h5>
5422
5423<p>
5424The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting at
5425the
5426destination location. If the argument is known to be aligned to some boundary,
5427this can be specified as the fourth argument, otherwise it should be set to 0 or
54281.
5429</p>
5430</div>
5431
5432
5433<!-- _______________________________________________________________________ -->
5434<div class="doc_subsubsection">
5435 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
5436</div>
5437
5438<div class="doc_text">
5439
5440<h5>Syntax:</h5>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5441<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]5442floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5443types however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5444<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5445 declare float @llvm.sqrt.f32(float %Val)
5446 declare double @llvm.sqrt.f64(double %Val)
5447 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
5448 declare fp128 @llvm.sqrt.f128(fp128 %Val)
5449 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5450</pre>
5451
5452<h5>Overview:</h5>
5453
5454<p>
5455The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5456returning the same value as the libm '<tt>sqrt</tt>' functions would. Unlike
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5457<tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
Chris Lattnerf914a002008-01-29 07:00:44 +0000[diff] [blame]5458negative numbers other than -0.0 (which allows for better optimization, because
5459there is no need to worry about errno being set). <tt>llvm.sqrt(-0.0)</tt> is
5460defined to return -0.0 like IEEE sqrt.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5461</p>
5462
5463<h5>Arguments:</h5>
5464
5465<p>
5466The argument and return value are floating point numbers of the same type.
5467</p>
5468
5469<h5>Semantics:</h5>
5470
5471<p>
5472This function returns the sqrt of the specified operand if it is a nonnegative
5473floating point number.
5474</p>
5475</div>
5476
5477<!-- _______________________________________________________________________ -->
5478<div class="doc_subsubsection">
5479 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
5480</div>
5481
5482<div class="doc_text">
5483
5484<h5>Syntax:</h5>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5485<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]5486floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5487types however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5488<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5489 declare float @llvm.powi.f32(float %Val, i32 %power)
5490 declare double @llvm.powi.f64(double %Val, i32 %power)
5491 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
5492 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
5493 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5494</pre>
5495
5496<h5>Overview:</h5>
5497
5498<p>
5499The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
5500specified (positive or negative) power. The order of evaluation of
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5501multiplications is not defined. When a vector of floating point type is
5502used, the second argument remains a scalar integer value.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5503</p>
5504
5505<h5>Arguments:</h5>
5506
5507<p>
5508The second argument is an integer power, and the first is a value to raise to
5509that power.
5510</p>
5511
5512<h5>Semantics:</h5>
5513
5514<p>
5515This function returns the first value raised to the second power with an
5516unspecified sequence of rounding operations.</p>
5517</div>
5518
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5519<!-- _______________________________________________________________________ -->
5520<div class="doc_subsubsection">
5521 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
5522</div>
5523
5524<div class="doc_text">
5525
5526<h5>Syntax:</h5>
5527<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
5528floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5529types however.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5530<pre>
5531 declare float @llvm.sin.f32(float %Val)
5532 declare double @llvm.sin.f64(double %Val)
5533 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
5534 declare fp128 @llvm.sin.f128(fp128 %Val)
5535 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
5536</pre>
5537
5538<h5>Overview:</h5>
5539
5540<p>
5541The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.
5542</p>
5543
5544<h5>Arguments:</h5>
5545
5546<p>
5547The argument and return value are floating point numbers of the same type.
5548</p>
5549
5550<h5>Semantics:</h5>
5551
5552<p>
5553This function returns the sine of the specified operand, returning the
5554same values as the libm <tt>sin</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5555conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5556</div>
5557
5558<!-- _______________________________________________________________________ -->
5559<div class="doc_subsubsection">
5560 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
5561</div>
5562
5563<div class="doc_text">
5564
5565<h5>Syntax:</h5>
5566<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
5567floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5568types however.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5569<pre>
5570 declare float @llvm.cos.f32(float %Val)
5571 declare double @llvm.cos.f64(double %Val)
5572 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
5573 declare fp128 @llvm.cos.f128(fp128 %Val)
5574 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
5575</pre>
5576
5577<h5>Overview:</h5>
5578
5579<p>
5580The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.
5581</p>
5582
5583<h5>Arguments:</h5>
5584
5585<p>
5586The argument and return value are floating point numbers of the same type.
5587</p>
5588
5589<h5>Semantics:</h5>
5590
5591<p>
5592This function returns the cosine of the specified operand, returning the
5593same values as the libm <tt>cos</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5594conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5595</div>
5596
5597<!-- _______________________________________________________________________ -->
5598<div class="doc_subsubsection">
5599 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
5600</div>
5601
5602<div class="doc_text">
5603
5604<h5>Syntax:</h5>
5605<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
5606floating point or vector of floating point type. Not all targets support all
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5607types however.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5608<pre>
5609 declare float @llvm.pow.f32(float %Val, float %Power)
5610 declare double @llvm.pow.f64(double %Val, double %Power)
5611 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
5612 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
5613 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
5614</pre>
5615
5616<h5>Overview:</h5>
5617
5618<p>
5619The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
5620specified (positive or negative) power.
5621</p>
5622
5623<h5>Arguments:</h5>
5624
5625<p>
5626The second argument is a floating point power, and the first is a value to
5627raise to that power.
5628</p>
5629
5630<h5>Semantics:</h5>
5631
5632<p>
5633This function returns the first value raised to the second power,
5634returning the
5635same values as the libm <tt>pow</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5636conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5637</div>
5638
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5639
5640<!-- ======================================================================= -->
5641<div class="doc_subsection">
5642 <a name="int_manip">Bit Manipulation Intrinsics</a>
5643</div>
5644
5645<div class="doc_text">
5646<p>
5647LLVM provides intrinsics for a few important bit manipulation operations.
5648These allow efficient code generation for some algorithms.
5649</p>
5650
5651</div>
5652
5653<!-- _______________________________________________________________________ -->
5654<div class="doc_subsubsection">
5655 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
5656</div>
5657
5658<div class="doc_text">
5659
5660<h5>Syntax:</h5>
5661<p>This is an overloaded intrinsic function. You can use bswap on any integer
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5662type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5663<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5664 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
5665 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
5666 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5667</pre>
5668
5669<h5>Overview:</h5>
5670
5671<p>
5672The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
5673values with an even number of bytes (positive multiple of 16 bits). These are
5674useful for performing operations on data that is not in the target's native
5675byte order.
5676</p>
5677
5678<h5>Semantics:</h5>
5679
5680<p>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5681The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5682and low byte of the input i16 swapped. Similarly, the <tt>llvm.bswap.i32</tt>
5683intrinsic returns an i32 value that has the four bytes of the input i32
5684swapped, 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]5685i32 will have its bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i48</tt>,
5686<tt>llvm.bswap.i64</tt> and other intrinsics extend this concept to
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5687additional even-byte lengths (6 bytes, 8 bytes and more, respectively).
5688</p>
5689
5690</div>
5691
5692<!-- _______________________________________________________________________ -->
5693<div class="doc_subsubsection">
5694 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
5695</div>
5696
5697<div class="doc_text">
5698
5699<h5>Syntax:</h5>
5700<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]5701width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5702<pre>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5703 declare i8 @llvm.ctpop.i8(i8 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5704 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5705 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5706 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
5707 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5708</pre>
5709
5710<h5>Overview:</h5>
5711
5712<p>
5713The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
5714value.
5715</p>
5716
5717<h5>Arguments:</h5>
5718
5719<p>
5720The only argument is the value to be counted. The argument may be of any
5721integer type. The return type must match the argument type.
5722</p>
5723
5724<h5>Semantics:</h5>
5725
5726<p>
5727The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
5728</p>
5729</div>
5730
5731<!-- _______________________________________________________________________ -->
5732<div class="doc_subsubsection">
5733 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
5734</div>
5735
5736<div class="doc_text">
5737
5738<h5>Syntax:</h5>
5739<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]5740integer bit width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5741<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5742 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
5743 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5744 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5745 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
5746 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5747</pre>
5748
5749<h5>Overview:</h5>
5750
5751<p>
5752The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
5753leading zeros in a variable.
5754</p>
5755
5756<h5>Arguments:</h5>
5757
5758<p>
5759The only argument is the value to be counted. The argument may be of any
5760integer type. The return type must match the argument type.
5761</p>
5762
5763<h5>Semantics:</h5>
5764
5765<p>
5766The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
5767in a variable. If the src == 0 then the result is the size in bits of the type
5768of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.
5769</p>
5770</div>
5771
5772
5773
5774<!-- _______________________________________________________________________ -->
5775<div class="doc_subsubsection">
5776 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
5777</div>
5778
5779<div class="doc_text">
5780
5781<h5>Syntax:</h5>
5782<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]5783integer bit width. Not all targets support all bit widths however.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5784<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5785 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
5786 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5787 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5788 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
5789 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5790</pre>
5791
5792<h5>Overview:</h5>
5793
5794<p>
5795The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
5796trailing zeros.
5797</p>
5798
5799<h5>Arguments:</h5>
5800
5801<p>
5802The only argument is the value to be counted. The argument may be of any
5803integer type. The return type must match the argument type.
5804</p>
5805
5806<h5>Semantics:</h5>
5807
5808<p>
5809The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
5810in a variable. If the src == 0 then the result is the size in bits of the type
5811of src. For example, <tt>llvm.cttz(2) = 1</tt>.
5812</p>
5813</div>
5814
5815<!-- _______________________________________________________________________ -->
5816<div class="doc_subsubsection">
5817 <a name="int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic</a>
5818</div>
5819
5820<div class="doc_text">
5821
5822<h5>Syntax:</h5>
5823<p>This is an overloaded intrinsic. You can use <tt>llvm.part.select</tt>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5824on any integer bit width.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5825<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5826 declare i17 @llvm.part.select.i17 (i17 %val, i32 %loBit, i32 %hiBit)
5827 declare i29 @llvm.part.select.i29 (i29 %val, i32 %loBit, i32 %hiBit)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5828</pre>
5829
5830<h5>Overview:</h5>
5831<p>The '<tt>llvm.part.select</tt>' family of intrinsic functions selects a
5832range of bits from an integer value and returns them in the same bit width as
5833the original value.</p>
5834
5835<h5>Arguments:</h5>
5836<p>The first argument, <tt>%val</tt> and the result may be integer types of
5837any bit width but they must have the same bit width. The second and third
5838arguments must be <tt>i32</tt> type since they specify only a bit index.</p>
5839
5840<h5>Semantics:</h5>
5841<p>The operation of the '<tt>llvm.part.select</tt>' intrinsic has two modes
5842of operation: forwards and reverse. If <tt>%loBit</tt> is greater than
5843<tt>%hiBits</tt> then the intrinsic operates in reverse mode. Otherwise it
5844operates in forward mode.</p>
5845<p>In forward mode, this intrinsic is the equivalent of shifting <tt>%val</tt>
5846right by <tt>%loBit</tt> bits and then ANDing it with a mask with
5847only the <tt>%hiBit - %loBit</tt> bits set, as follows:</p>
5848<ol>
5849 <li>The <tt>%val</tt> is shifted right (LSHR) by the number of bits specified
5850 by <tt>%loBits</tt>. This normalizes the value to the low order bits.</li>
5851 <li>The <tt>%loBits</tt> value is subtracted from the <tt>%hiBits</tt> value
5852 to determine the number of bits to retain.</li>
5853 <li>A mask of the retained bits is created by shifting a -1 value.</li>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5854 <li>The mask is ANDed with <tt>%val</tt> to produce the result.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5855</ol>
5856<p>In reverse mode, a similar computation is made except that the bits are
5857returned in the reverse order. So, for example, if <tt>X</tt> has the value
5858<tt>i16 0x0ACF (101011001111)</tt> and we apply
5859<tt>part.select(i16 X, 8, 3)</tt> to it, we get back the value
5860<tt>i16 0x0026 (000000100110)</tt>.</p>
5861</div>
5862
5863<div class="doc_subsubsection">
5864 <a name="int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic</a>
5865</div>
5866
5867<div class="doc_text">
5868
5869<h5>Syntax:</h5>
5870<p>This is an overloaded intrinsic. You can use <tt>llvm.part.set</tt>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5871on any integer bit width.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5872<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5873 declare i17 @llvm.part.set.i17.i9 (i17 %val, i9 %repl, i32 %lo, i32 %hi)
5874 declare i29 @llvm.part.set.i29.i9 (i29 %val, i9 %repl, i32 %lo, i32 %hi)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5875</pre>
5876
5877<h5>Overview:</h5>
5878<p>The '<tt>llvm.part.set</tt>' family of intrinsic functions replaces a range
5879of bits in an integer value with another integer value. It returns the integer
5880with the replaced bits.</p>
5881
5882<h5>Arguments:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5883<p>The first argument, <tt>%val</tt>, and the result may be integer types of
5884any bit width, but they must have the same bit width. <tt>%val</tt> is the value
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5885whose bits will be replaced. The second argument, <tt>%repl</tt> may be an
5886integer of any bit width. The third and fourth arguments must be <tt>i32</tt>
5887type since they specify only a bit index.</p>
5888
5889<h5>Semantics:</h5>
5890<p>The operation of the '<tt>llvm.part.set</tt>' intrinsic has two modes
5891of operation: forwards and reverse. If <tt>%lo</tt> is greater than
5892<tt>%hi</tt> then the intrinsic operates in reverse mode. Otherwise it
5893operates in forward mode.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5894
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5895<p>For both modes, the <tt>%repl</tt> value is prepared for use by either
5896truncating it down to the size of the replacement area or zero extending it
5897up to that size.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5898
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5899<p>In forward mode, the bits between <tt>%lo</tt> and <tt>%hi</tt> (inclusive)
5900are replaced with corresponding bits from <tt>%repl</tt>. That is the 0th bit
5901in <tt>%repl</tt> replaces the <tt>%lo</tt>th bit in <tt>%val</tt> and etc. up
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5902to the <tt>%hi</tt>th bit.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5903
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5904<p>In reverse mode, a similar computation is made except that the bits are
5905reversed. That is, the <tt>0</tt>th bit in <tt>%repl</tt> replaces the
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]5906<tt>%hi</tt> bit in <tt>%val</tt> and etc. down to the <tt>%lo</tt>th bit.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5907
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5908<h5>Examples:</h5>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5909
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5910<pre>
5911 llvm.part.set(0xFFFF, 0, 4, 7) -&gt; 0xFF0F
5912 llvm.part.set(0xFFFF, 0, 7, 4) -&gt; 0xFF0F
5913 llvm.part.set(0xFFFF, 1, 7, 4) -&gt; 0xFF8F
5914 llvm.part.set(0xFFFF, F, 8, 3) -&gt; 0xFFE7
5915 llvm.part.set(0xFFFF, 0, 3, 8) -&gt; 0xFE07
5916</pre>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5917
5918</div>
5919
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]5920<!-- ======================================================================= -->
5921<div class="doc_subsection">
5922 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
5923</div>
5924
5925<div class="doc_text">
5926<p>
5927LLVM provides intrinsics for some arithmetic with overflow operations.
5928</p>
5929
5930</div>
5931
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5932<!-- _______________________________________________________________________ -->
5933<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]5934 <a name="int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5935</div>
5936
5937<div class="doc_text">
5938
5939<h5>Syntax:</h5>
5940
5941<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]5942on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5943
5944<pre>
5945 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
5946 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
5947 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
5948</pre>
5949
5950<h5>Overview:</h5>
5951
5952<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
5953a signed addition of the two arguments, and indicate whether an overflow
5954occurred during the signed summation.</p>
5955
5956<h5>Arguments:</h5>
5957
5958<p>The arguments (%a and %b) and the first element of the result structure may
5959be of integer types of any bit width, but they must have the same bit width. The
5960second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
5961and <tt>%b</tt> are the two values that will undergo signed addition.</p>
5962
5963<h5>Semantics:</h5>
5964
5965<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
5966a signed addition of the two variables. They return a structure &mdash; the
5967first element of which is the signed summation, and the second element of which
5968is a bit specifying if the signed summation resulted in an overflow.</p>
5969
5970<h5>Examples:</h5>
5971<pre>
5972 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
5973 %sum = extractvalue {i32, i1} %res, 0
5974 %obit = extractvalue {i32, i1} %res, 1
5975 br i1 %obit, label %overflow, label %normal
5976</pre>
5977
5978</div>
5979
5980<!-- _______________________________________________________________________ -->
5981<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]5982 <a name="int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5983</div>
5984
5985<div class="doc_text">
5986
5987<h5>Syntax:</h5>
5988
5989<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]5990on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]5991
5992<pre>
5993 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
5994 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
5995 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
5996</pre>
5997
5998<h5>Overview:</h5>
5999
6000<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
6001an unsigned addition of the two arguments, and indicate whether a carry occurred
6002during the unsigned summation.</p>
6003
6004<h5>Arguments:</h5>
6005
6006<p>The arguments (%a and %b) and the first element of the result structure may
6007be of integer types of any bit width, but they must have the same bit width. The
6008second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6009and <tt>%b</tt> are the two values that will undergo unsigned addition.</p>
6010
6011<h5>Semantics:</h5>
6012
6013<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
6014an unsigned addition of the two arguments. They return a structure &mdash; the
6015first element of which is the sum, and the second element of which is a bit
6016specifying if the unsigned summation resulted in a carry.</p>
6017
6018<h5>Examples:</h5>
6019<pre>
6020 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6021 %sum = extractvalue {i32, i1} %res, 0
6022 %obit = extractvalue {i32, i1} %res, 1
6023 br i1 %obit, label %carry, label %normal
6024</pre>
6025
6026</div>
6027
6028<!-- _______________________________________________________________________ -->
6029<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6030 <a name="int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6031</div>
6032
6033<div class="doc_text">
6034
6035<h5>Syntax:</h5>
6036
6037<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]6038on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6039
6040<pre>
6041 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
6042 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6043 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
6044</pre>
6045
6046<h5>Overview:</h5>
6047
6048<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
6049a signed subtraction of the two arguments, and indicate whether an overflow
6050occurred during the signed subtraction.</p>
6051
6052<h5>Arguments:</h5>
6053
6054<p>The arguments (%a and %b) and the first element of the result structure may
6055be of integer types of any bit width, but they must have the same bit width. The
6056second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6057and <tt>%b</tt> are the two values that will undergo signed subtraction.</p>
6058
6059<h5>Semantics:</h5>
6060
6061<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
6062a signed subtraction of the two arguments. They return a structure &mdash; the
6063first element of which is the subtraction, and the second element of which is a bit
6064specifying if the signed subtraction resulted in an overflow.</p>
6065
6066<h5>Examples:</h5>
6067<pre>
6068 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6069 %sum = extractvalue {i32, i1} %res, 0
6070 %obit = extractvalue {i32, i1} %res, 1
6071 br i1 %obit, label %overflow, label %normal
6072</pre>
6073
6074</div>
6075
6076<!-- _______________________________________________________________________ -->
6077<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6078 <a name="int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6079</div>
6080
6081<div class="doc_text">
6082
6083<h5>Syntax:</h5>
6084
6085<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]6086on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6087
6088<pre>
6089 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
6090 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6091 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
6092</pre>
6093
6094<h5>Overview:</h5>
6095
6096<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
6097an unsigned subtraction of the two arguments, and indicate whether an overflow
6098occurred during the unsigned subtraction.</p>
6099
6100<h5>Arguments:</h5>
6101
6102<p>The arguments (%a and %b) and the first element of the result structure may
6103be of integer types of any bit width, but they must have the same bit width. The
6104second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6105and <tt>%b</tt> are the two values that will undergo unsigned subtraction.</p>
6106
6107<h5>Semantics:</h5>
6108
6109<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
6110an unsigned subtraction of the two arguments. They return a structure &mdash; the
6111first element of which is the subtraction, and the second element of which is a bit
6112specifying if the unsigned subtraction resulted in an overflow.</p>
6113
6114<h5>Examples:</h5>
6115<pre>
6116 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6117 %sum = extractvalue {i32, i1} %res, 0
6118 %obit = extractvalue {i32, i1} %res, 1
6119 br i1 %obit, label %overflow, label %normal
6120</pre>
6121
6122</div>
6123
6124<!-- _______________________________________________________________________ -->
6125<div class="doc_subsubsection">
Bill Wendling3e1258b2009-02-08 04:04:40 +0000[diff] [blame]6126 <a name="int_smul_overflow">'<tt>llvm.smul.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6127</div>
6128
6129<div class="doc_text">
6130
6131<h5>Syntax:</h5>
6132
6133<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]6134on any integer bit width.</p>
Bill Wendling3f8cebe2009-02-08 01:40:31 +0000[diff] [blame]6135
6136<pre>
6137 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
6138 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6139 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
6140</pre>
6141
6142<h5>Overview:</h5>
6143
6144<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
6145a signed multiplication of the two arguments, and indicate whether an overflow
6146occurred during the signed multiplication.</p>
6147
6148<h5>Arguments:</h5>
6149
6150<p>The arguments (%a and %b) and the first element of the result structure may
6151be of integer types of any bit width, but they must have the same bit width. The
6152second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6153and <tt>%b</tt> are the two values that will undergo signed multiplication.</p>
6154
6155<h5>Semantics:</h5>
6156
6157<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
6158a signed multiplication of the two arguments. They return a structure &mdash;
6159the first element of which is the multiplication, and the second element of
6160which is a bit specifying if the signed multiplication resulted in an
6161overflow.</p>
6162
6163<h5>Examples:</h5>
6164<pre>
6165 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6166 %sum = extractvalue {i32, i1} %res, 0
6167 %obit = extractvalue {i32, i1} %res, 1
6168 br i1 %obit, label %overflow, label %normal
6169</pre>
6170
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6171</div>
6172
6173<!-- ======================================================================= -->
6174<div class="doc_subsection">
6175 <a name="int_debugger">Debugger Intrinsics</a>
6176</div>
6177
6178<div class="doc_text">
6179<p>
6180The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
6181are described in the <a
6182href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
6183Debugging</a> document.
6184</p>
6185</div>
6186
6187
6188<!-- ======================================================================= -->
6189<div class="doc_subsection">
6190 <a name="int_eh">Exception Handling Intrinsics</a>
6191</div>
6192
6193<div class="doc_text">
6194<p> The LLVM exception handling intrinsics (which all start with
6195<tt>llvm.eh.</tt> prefix), are described in the <a
6196href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
6197Handling</a> document. </p>
6198</div>
6199
6200<!-- ======================================================================= -->
6201<div class="doc_subsection">
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6202 <a name="int_trampoline">Trampoline Intrinsic</a>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6203</div>
6204
6205<div class="doc_text">
6206<p>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6207 This intrinsic makes it possible to excise one parameter, marked with
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6208 the <tt>nest</tt> attribute, from a function. The result is a callable
6209 function pointer lacking the nest parameter - the caller does not need
6210 to provide a value for it. Instead, the value to use is stored in
6211 advance in a "trampoline", a block of memory usually allocated
6212 on the stack, which also contains code to splice the nest value into the
6213 argument list. This is used to implement the GCC nested function address
6214 extension.
6215</p>
6216<p>
6217 For example, if the function is
6218 <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]6219 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]6220<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6221 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
6222 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
6223 %p = call i8* @llvm.init.trampoline( i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8*), i8* %nval )
6224 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6225</pre>
Bill Wendlinge262b4c2007-09-22 09:23:55 +0000[diff] [blame]6226 <p>The call <tt>%val = call i32 %fp( i32 %x, i32 %y )</tt> is then equivalent
6227 to <tt>%val = call i32 %f( i8* %nval, i32 %x, i32 %y )</tt>.</p>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6228</div>
6229
6230<!-- _______________________________________________________________________ -->
6231<div class="doc_subsubsection">
6232 <a name="int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a>
6233</div>
6234<div class="doc_text">
6235<h5>Syntax:</h5>
6236<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6237declare 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]6238</pre>
6239<h5>Overview:</h5>
6240<p>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6241 This fills the memory pointed to by <tt>tramp</tt> with code
6242 and returns a function pointer suitable for executing it.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6243</p>
6244<h5>Arguments:</h5>
6245<p>
6246 The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
6247 pointers. The <tt>tramp</tt> argument must point to a sufficiently large
6248 and sufficiently aligned block of memory; this memory is written to by the
Duncan Sands35012212007-08-22 23:39:54 +0000[diff] [blame]6249 intrinsic. Note that the size and the alignment are target-specific - LLVM
6250 currently provides no portable way of determining them, so a front-end that
6251 generates this intrinsic needs to have some target-specific knowledge.
6252 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]6253</p>
6254<h5>Semantics:</h5>
6255<p>
6256 The block of memory pointed to by <tt>tramp</tt> is filled with target
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6257 dependent code, turning it into a function. A pointer to this function is
6258 returned, but needs to be bitcast to an
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6259 <a href="#int_trampoline">appropriate function pointer type</a>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]6260 before being called. The new function's signature is the same as that of
6261 <tt>func</tt> with any arguments marked with the <tt>nest</tt> attribute
6262 removed. At most one such <tt>nest</tt> argument is allowed, and it must be
6263 of pointer type. Calling the new function is equivalent to calling
6264 <tt>func</tt> with the same argument list, but with <tt>nval</tt> used for the
6265 missing <tt>nest</tt> argument. If, after calling
6266 <tt>llvm.init.trampoline</tt>, the memory pointed to by <tt>tramp</tt> is
6267 modified, then the effect of any later call to the returned function pointer is
6268 undefined.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]6269</p>
6270</div>
6271
6272<!-- ======================================================================= -->
6273<div class="doc_subsection">
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6274 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
6275</div>
6276
6277<div class="doc_text">
6278<p>
6279 These intrinsic functions expand the "universal IR" of LLVM to represent
6280 hardware constructs for atomic operations and memory synchronization. This
6281 provides an interface to the hardware, not an interface to the programmer. It
Chris Lattner96451482008-08-05 18:29:16 +0000[diff] [blame]6282 is aimed at a low enough level to allow any programming models or APIs
6283 (Application Programming Interfaces) which
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6284 need atomic behaviors to map cleanly onto it. It is also modeled primarily on
6285 hardware behavior. Just as hardware provides a "universal IR" for source
6286 languages, it also provides a starting point for developing a "universal"
6287 atomic operation and synchronization IR.
6288</p>
6289<p>
6290 These do <em>not</em> form an API such as high-level threading libraries,
6291 software transaction memory systems, atomic primitives, and intrinsic
6292 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
6293 application libraries. The hardware interface provided by LLVM should allow
6294 a clean implementation of all of these APIs and parallel programming models.
6295 No one model or paradigm should be selected above others unless the hardware
6296 itself ubiquitously does so.
6297
6298</p>
6299</div>
6300
6301<!-- _______________________________________________________________________ -->
6302<div class="doc_subsubsection">
6303 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
6304</div>
6305<div class="doc_text">
6306<h5>Syntax:</h5>
6307<pre>
6308declare void @llvm.memory.barrier( i1 &lt;ll&gt;, i1 &lt;ls&gt;, i1 &lt;sl&gt;, i1 &lt;ss&gt;,
6309i1 &lt;device&gt; )
6310
6311</pre>
6312<h5>Overview:</h5>
6313<p>
6314 The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
6315 specific pairs of memory access types.
6316</p>
6317<h5>Arguments:</h5>
6318<p>
6319 The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
6320 The first four arguments enables a specific barrier as listed below. The fith
6321 argument specifies that the barrier applies to io or device or uncached memory.
6322
6323</p>
6324 <ul>
6325 <li><tt>ll</tt>: load-load barrier</li>
6326 <li><tt>ls</tt>: load-store barrier</li>
6327 <li><tt>sl</tt>: store-load barrier</li>
6328 <li><tt>ss</tt>: store-store barrier</li>
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]6329 <li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6330 </ul>
6331<h5>Semantics:</h5>
6332<p>
6333 This intrinsic causes the system to enforce some ordering constraints upon
6334 the loads and stores of the program. This barrier does not indicate
6335 <em>when</em> any events will occur, it only enforces an <em>order</em> in
6336 which they occur. For any of the specified pairs of load and store operations
6337 (f.ex. load-load, or store-load), all of the first operations preceding the
6338 barrier will complete before any of the second operations succeeding the
6339 barrier begin. Specifically the semantics for each pairing is as follows:
6340</p>
6341 <ul>
6342 <li><tt>ll</tt>: All loads before the barrier must complete before any load
6343 after the barrier begins.</li>
6344
6345 <li><tt>ls</tt>: All loads before the barrier must complete before any
6346 store after the barrier begins.</li>
6347 <li><tt>ss</tt>: All stores before the barrier must complete before any
6348 store after the barrier begins.</li>
6349 <li><tt>sl</tt>: All stores before the barrier must complete before any
6350 load after the barrier begins.</li>
6351 </ul>
6352<p>
6353 These semantics are applied with a logical "and" behavior when more than one
6354 is enabled in a single memory barrier intrinsic.
6355</p>
6356<p>
6357 Backends may implement stronger barriers than those requested when they do not
6358 support as fine grained a barrier as requested. Some architectures do not
6359 need all types of barriers and on such architectures, these become noops.
6360</p>
6361<h5>Example:</h5>
6362<pre>
6363%ptr = malloc i32
6364 store i32 4, %ptr
6365
6366%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
6367 call void @llvm.memory.barrier( i1 false, i1 true, i1 false, i1 false )
6368 <i>; guarantee the above finishes</i>
6369 store i32 8, %ptr <i>; before this begins</i>
6370</pre>
6371</div>
6372
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6373<!-- _______________________________________________________________________ -->
6374<div class="doc_subsubsection">
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6375 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6376</div>
6377<div class="doc_text">
6378<h5>Syntax:</h5>
6379<p>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6380 This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
6381 any integer bit width and for different address spaces. Not all targets
6382 support all bit widths however.</p>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6383
6384<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6385declare i8 @llvm.atomic.cmp.swap.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt; )
6386declare i16 @llvm.atomic.cmp.swap.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt; )
6387declare i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt; )
6388declare 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]6389
6390</pre>
6391<h5>Overview:</h5>
6392<p>
6393 This loads a value in memory and compares it to a given value. If they are
6394 equal, it stores a new value into the memory.
6395</p>
6396<h5>Arguments:</h5>
6397<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6398 The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result as
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6399 well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
6400 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
6401 this integer type. While any bit width integer may be used, targets may only
6402 lower representations they support in hardware.
6403
6404</p>
6405<h5>Semantics:</h5>
6406<p>
6407 This entire intrinsic must be executed atomically. It first loads the value
6408 in memory pointed to by <tt>ptr</tt> and compares it with the value
6409 <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the memory. The
6410 loaded value is yielded in all cases. This provides the equivalent of an
6411 atomic compare-and-swap operation within the SSA framework.
6412</p>
6413<h5>Examples:</h5>
6414
6415<pre>
6416%ptr = malloc i32
6417 store i32 4, %ptr
6418
6419%val1 = add i32 4, 4
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6420%result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 4, %val1 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6421 <i>; yields {i32}:result1 = 4</i>
6422%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
6423%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
6424
6425%val2 = add i32 1, 1
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6426%result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 5, %val2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6427 <i>; yields {i32}:result2 = 8</i>
6428%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
6429
6430%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
6431</pre>
6432</div>
6433
6434<!-- _______________________________________________________________________ -->
6435<div class="doc_subsubsection">
6436 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
6437</div>
6438<div class="doc_text">
6439<h5>Syntax:</h5>
6440
6441<p>
6442 This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
6443 integer bit width. Not all targets support all bit widths however.</p>
6444<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6445declare i8 @llvm.atomic.swap.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;val&gt; )
6446declare i16 @llvm.atomic.swap.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;val&gt; )
6447declare i32 @llvm.atomic.swap.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;val&gt; )
6448declare i64 @llvm.atomic.swap.i64.p0i64( i64* &lt;ptr&gt;, i64 &lt;val&gt; )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6449
6450</pre>
6451<h5>Overview:</h5>
6452<p>
6453 This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
6454 the value from memory. It then stores the value in <tt>val</tt> in the memory
6455 at <tt>ptr</tt>.
6456</p>
6457<h5>Arguments:</h5>
6458
6459<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6460 The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both the
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6461 <tt>val</tt> argument and the result must be integers of the same bit width.
6462 The first argument, <tt>ptr</tt>, must be a pointer to a value of this
6463 integer type. The targets may only lower integer representations they
6464 support.
6465</p>
6466<h5>Semantics:</h5>
6467<p>
6468 This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
6469 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
6470 equivalent of an atomic swap operation within the SSA framework.
6471
6472</p>
6473<h5>Examples:</h5>
6474<pre>
6475%ptr = malloc i32
6476 store i32 4, %ptr
6477
6478%val1 = add i32 4, 4
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6479%result1 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val1 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6480 <i>; yields {i32}:result1 = 4</i>
6481%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
6482%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
6483
6484%val2 = add i32 1, 1
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6485%result2 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6486 <i>; yields {i32}:result2 = 8</i>
6487
6488%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
6489%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
6490</pre>
6491</div>
6492
6493<!-- _______________________________________________________________________ -->
6494<div class="doc_subsubsection">
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6495 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6496
6497</div>
6498<div class="doc_text">
6499<h5>Syntax:</h5>
6500<p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6501 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]6502 integer bit width. Not all targets support all bit widths however.</p>
6503<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6504declare i8 @llvm.atomic.load.add.i8..p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6505declare i16 @llvm.atomic.load.add.i16..p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6506declare i32 @llvm.atomic.load.add.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6507declare 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]6508
6509</pre>
6510<h5>Overview:</h5>
6511<p>
6512 This intrinsic adds <tt>delta</tt> to the value stored in memory at
6513 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
6514</p>
6515<h5>Arguments:</h5>
6516<p>
6517
6518 The intrinsic takes two arguments, the first a pointer to an integer value
6519 and the second an integer value. The result is also an integer value. These
6520 integer types can have any bit width, but they must all have the same bit
6521 width. The targets may only lower integer representations they support.
6522</p>
6523<h5>Semantics:</h5>
6524<p>
6525 This intrinsic does a series of operations atomically. It first loads the
6526 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
6527 to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
6528</p>
6529
6530<h5>Examples:</h5>
6531<pre>
6532%ptr = malloc i32
6533 store i32 4, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6534%result1 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 4 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6535 <i>; yields {i32}:result1 = 4</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6536%result2 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 2 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6537 <i>; yields {i32}:result2 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6538%result3 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 5 )
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6539 <i>; yields {i32}:result3 = 10</i>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6540%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]6541</pre>
6542</div>
6543
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6544<!-- _______________________________________________________________________ -->
6545<div class="doc_subsubsection">
6546 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
6547
6548</div>
6549<div class="doc_text">
6550<h5>Syntax:</h5>
6551<p>
6552 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]6553 any integer bit width and for different address spaces. Not all targets
6554 support all bit widths however.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6555<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6556declare i8 @llvm.atomic.load.sub.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6557declare i16 @llvm.atomic.load.sub.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6558declare i32 @llvm.atomic.load.sub.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6559declare 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]6560
6561</pre>
6562<h5>Overview:</h5>
6563<p>
6564 This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
6565 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
6566</p>
6567<h5>Arguments:</h5>
6568<p>
6569
6570 The intrinsic takes two arguments, the first a pointer to an integer value
6571 and the second an integer value. The result is also an integer value. These
6572 integer types can have any bit width, but they must all have the same bit
6573 width. The targets may only lower integer representations they support.
6574</p>
6575<h5>Semantics:</h5>
6576<p>
6577 This intrinsic does a series of operations atomically. It first loads the
6578 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
6579 result to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
6580</p>
6581
6582<h5>Examples:</h5>
6583<pre>
6584%ptr = malloc i32
6585 store i32 8, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6586%result1 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 4 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6587 <i>; yields {i32}:result1 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6588%result2 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 2 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6589 <i>; yields {i32}:result2 = 4</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6590%result3 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 5 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6591 <i>; yields {i32}:result3 = 2</i>
6592%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
6593</pre>
6594</div>
6595
6596<!-- _______________________________________________________________________ -->
6597<div class="doc_subsubsection">
6598 <a name="int_atomic_load_and">'<tt>llvm.atomic.load.and.*</tt>' Intrinsic</a><br>
6599 <a name="int_atomic_load_nand">'<tt>llvm.atomic.load.nand.*</tt>' Intrinsic</a><br>
6600 <a name="int_atomic_load_or">'<tt>llvm.atomic.load.or.*</tt>' Intrinsic</a><br>
6601 <a name="int_atomic_load_xor">'<tt>llvm.atomic.load.xor.*</tt>' Intrinsic</a><br>
6602
6603</div>
6604<div class="doc_text">
6605<h5>Syntax:</h5>
6606<p>
6607 These are overloaded intrinsics. You can use <tt>llvm.atomic.load_and</tt>,
6608 <tt>llvm.atomic.load_nand</tt>, <tt>llvm.atomic.load_or</tt>, and
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6609 <tt>llvm.atomic.load_xor</tt> on any integer bit width and for different
6610 address spaces. Not all targets support all bit widths however.</p>
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6611<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6612declare i8 @llvm.atomic.load.and.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6613declare i16 @llvm.atomic.load.and.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6614declare i32 @llvm.atomic.load.and.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6615declare 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]6616
6617</pre>
6618
6619<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6620declare i8 @llvm.atomic.load.or.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6621declare i16 @llvm.atomic.load.or.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6622declare i32 @llvm.atomic.load.or.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6623declare 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]6624
6625</pre>
6626
6627<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6628declare i8 @llvm.atomic.load.nand.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6629declare i16 @llvm.atomic.load.nand.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6630declare i32 @llvm.atomic.load.nand.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6631declare 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]6632
6633</pre>
6634
6635<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6636declare i8 @llvm.atomic.load.xor.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6637declare i16 @llvm.atomic.load.xor.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6638declare i32 @llvm.atomic.load.xor.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6639declare 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]6640
6641</pre>
6642<h5>Overview:</h5>
6643<p>
6644 These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
6645 the value stored in memory at <tt>ptr</tt>. It yields the original value
6646 at <tt>ptr</tt>.
6647</p>
6648<h5>Arguments:</h5>
6649<p>
6650
6651 These intrinsics take two arguments, the first a pointer to an integer value
6652 and the second an integer value. The result is also an integer value. These
6653 integer types can have any bit width, but they must all have the same bit
6654 width. The targets may only lower integer representations they support.
6655</p>
6656<h5>Semantics:</h5>
6657<p>
6658 These intrinsics does a series of operations atomically. They first load the
6659 value stored at <tt>ptr</tt>. They then do the bitwise operation
6660 <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the original
6661 value stored at <tt>ptr</tt>.
6662</p>
6663
6664<h5>Examples:</h5>
6665<pre>
6666%ptr = malloc i32
6667 store i32 0x0F0F, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6668%result0 = call i32 @llvm.atomic.load.nand.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6669 <i>; yields {i32}:result0 = 0x0F0F</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6670%result1 = call i32 @llvm.atomic.load.and.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6671 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6672%result2 = call i32 @llvm.atomic.load.or.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6673 <i>; yields {i32}:result2 = 0xF0</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6674%result3 = call i32 @llvm.atomic.load.xor.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6675 <i>; yields {i32}:result3 = FF</i>
6676%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
6677</pre>
6678</div>
6679
6680
6681<!-- _______________________________________________________________________ -->
6682<div class="doc_subsubsection">
6683 <a name="int_atomic_load_max">'<tt>llvm.atomic.load.max.*</tt>' Intrinsic</a><br>
6684 <a name="int_atomic_load_min">'<tt>llvm.atomic.load.min.*</tt>' Intrinsic</a><br>
6685 <a name="int_atomic_load_umax">'<tt>llvm.atomic.load.umax.*</tt>' Intrinsic</a><br>
6686 <a name="int_atomic_load_umin">'<tt>llvm.atomic.load.umin.*</tt>' Intrinsic</a><br>
6687
6688</div>
6689<div class="doc_text">
6690<h5>Syntax:</h5>
6691<p>
6692 These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
6693 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6694 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
6695 address spaces. Not all targets
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6696 support all bit widths however.</p>
6697<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6698declare i8 @llvm.atomic.load.max.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6699declare i16 @llvm.atomic.load.max.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6700declare i32 @llvm.atomic.load.max.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6701declare 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]6702
6703</pre>
6704
6705<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6706declare i8 @llvm.atomic.load.min.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6707declare i16 @llvm.atomic.load.min.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6708declare i32 @llvm.atomic.load.min.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6709declare 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]6710
6711</pre>
6712
6713<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6714declare i8 @llvm.atomic.load.umax.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6715declare i16 @llvm.atomic.load.umax.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6716declare i32 @llvm.atomic.load.umax.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6717declare 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]6718
6719</pre>
6720
6721<pre>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6722declare i8 @llvm.atomic.load.umin.i8..p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6723declare i16 @llvm.atomic.load.umin.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6724declare i32 @llvm.atomic.load.umin.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6725declare 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]6726
6727</pre>
6728<h5>Overview:</h5>
6729<p>
6730 These intrinsics takes the signed or unsigned minimum or maximum of
6731 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
6732 original value at <tt>ptr</tt>.
6733</p>
6734<h5>Arguments:</h5>
6735<p>
6736
6737 These intrinsics take two arguments, the first a pointer to an integer value
6738 and the second an integer value. The result is also an integer value. These
6739 integer types can have any bit width, but they must all have the same bit
6740 width. The targets may only lower integer representations they support.
6741</p>
6742<h5>Semantics:</h5>
6743<p>
6744 These intrinsics does a series of operations atomically. They first load the
6745 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or max
6746 <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They yield
6747 the original value stored at <tt>ptr</tt>.
6748</p>
6749
6750<h5>Examples:</h5>
6751<pre>
6752%ptr = malloc i32
6753 store i32 7, %ptr
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6754%result0 = call i32 @llvm.atomic.load.min.i32.p0i32( i32* %ptr, i32 -2 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6755 <i>; yields {i32}:result0 = 7</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6756%result1 = call i32 @llvm.atomic.load.max.i32.p0i32( i32* %ptr, i32 8 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6757 <i>; yields {i32}:result1 = -2</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6758%result2 = call i32 @llvm.atomic.load.umin.i32.p0i32( i32* %ptr, i32 10 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6759 <i>; yields {i32}:result2 = 8</i>
Mon P Wangce3ac892008-07-30 04:36:53 +0000[diff] [blame]6760%result3 = call i32 @llvm.atomic.load.umax.i32.p0i32( i32* %ptr, i32 30 )
Mon P Wang6bde9ec2008-06-25 08:15:39 +0000[diff] [blame]6761 <i>; yields {i32}:result3 = 8</i>
6762%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
6763</pre>
6764</div>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]6765
6766<!-- ======================================================================= -->
6767<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6768 <a name="int_general">General Intrinsics</a>
6769</div>
6770
6771<div class="doc_text">
6772<p> This class of intrinsics is designed to be generic and has
6773no specific purpose. </p>
6774</div>
6775
6776<!-- _______________________________________________________________________ -->
6777<div class="doc_subsubsection">
6778 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
6779</div>
6780
6781<div class="doc_text">
6782
6783<h5>Syntax:</h5>
6784<pre>
6785 declare void @llvm.var.annotation(i8* &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6786</pre>
6787
6788<h5>Overview:</h5>
6789
6790<p>
6791The '<tt>llvm.var.annotation</tt>' intrinsic
6792</p>
6793
6794<h5>Arguments:</h5>
6795
6796<p>
6797The first argument is a pointer to a value, the second is a pointer to a
6798global string, the third is a pointer to a global string which is the source
6799file name, and the last argument is the line number.
6800</p>
6801
6802<h5>Semantics:</h5>
6803
6804<p>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]6805This intrinsic allows annotation of local variables with arbitrary strings.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6806This can be useful for special purpose optimizations that want to look for these
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]6807annotations. These have no other defined use, they are ignored by code
6808generation and optimization.
6809</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6810</div>
6811
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6812<!-- _______________________________________________________________________ -->
6813<div class="doc_subsubsection">
Tanya Lattnerc9869b12007-09-21 23:57:59 +0000[diff] [blame]6814 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6815</div>
6816
6817<div class="doc_text">
6818
6819<h5>Syntax:</h5>
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]6820<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
6821any integer bit width.
6822</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6823<pre>
Tanya Lattner09161fe2007-09-22 00:03:01 +0000[diff] [blame]6824 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6825 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6826 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6827 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6828 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]6829</pre>
6830
6831<h5>Overview:</h5>
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]6832
6833<p>
6834The '<tt>llvm.annotation</tt>' intrinsic.
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6835</p>
6836
6837<h5>Arguments:</h5>
6838
6839<p>
6840The first argument is an integer value (result of some expression),
6841the second is a pointer to a global string, the third is a pointer to a global
6842string which is the source file name, and the last argument is the line number.
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]6843It returns the value of the first argument.
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6844</p>
6845
6846<h5>Semantics:</h5>
6847
6848<p>
6849This intrinsic allows annotations to be put on arbitrary expressions
6850with arbitrary strings. This can be useful for special purpose optimizations
6851that want to look for these annotations. These have no other defined use, they
6852are ignored by code generation and optimization.
Dan Gohman2672f3e2008-10-14 16:51:45 +0000[diff] [blame]6853</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6854</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6855
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]6856<!-- _______________________________________________________________________ -->
6857<div class="doc_subsubsection">
6858 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
6859</div>
6860
6861<div class="doc_text">
6862
6863<h5>Syntax:</h5>
6864<pre>
6865 declare void @llvm.trap()
6866</pre>
6867
6868<h5>Overview:</h5>
6869
6870<p>
6871The '<tt>llvm.trap</tt>' intrinsic
6872</p>
6873
6874<h5>Arguments:</h5>
6875
6876<p>
6877None
6878</p>
6879
6880<h5>Semantics:</h5>
6881
6882<p>
6883This intrinsics is lowered to the target dependent trap instruction. If the
6884target does not have a trap instruction, this intrinsic will be lowered to the
6885call of the abort() function.
6886</p>
6887</div>
6888
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]6889<!-- _______________________________________________________________________ -->
6890<div class="doc_subsubsection">
Misha Brukman5dd7f4d2008-11-22 23:55:29 +0000[diff] [blame]6891 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
Bill Wendlinge4164592008-11-19 05:56:17 +0000[diff] [blame]6892</div>
6893<div class="doc_text">
6894<h5>Syntax:</h5>
6895<pre>
6896declare void @llvm.stackprotector( i8* &lt;guard&gt;, i8** &lt;slot&gt; )
6897
6898</pre>
6899<h5>Overview:</h5>
6900<p>
6901 The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and stores
6902 it onto the stack at <tt>slot</tt>. The stack slot is adjusted to ensure that
6903 it is placed on the stack before local variables.
6904</p>
6905<h5>Arguments:</h5>
6906<p>
6907 The <tt>llvm.stackprotector</tt> intrinsic requires two pointer arguments. The
6908 first argument is the value loaded from the stack guard
6909 <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt> that
6910 has enough space to hold the value of the guard.
6911</p>
6912<h5>Semantics:</h5>
6913<p>
6914 This intrinsic causes the prologue/epilogue inserter to force the position of
6915 the <tt>AllocaInst</tt> stack slot to be before local variables on the
6916 stack. This is to ensure that if a local variable on the stack is overwritten,
6917 it will destroy the value of the guard. When the function exits, the guard on
6918 the stack is checked against the original guard. If they're different, then
6919 the program aborts by calling the <tt>__stack_chk_fail()</tt> function.
6920</p>
6921</div>
6922
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6923<!-- *********************************************************************** -->
6924<hr>
6925<address>
6926 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman947321d2008-12-11 17:34:48 +0000[diff] [blame]6927 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6928 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman947321d2008-12-11 17:34:48 +0000[diff] [blame]6929 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]6930
6931 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
6932 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
6933 Last modified: $Date$
6934</address>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]6935
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6936</body>
6937</html>