Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / llvm / 4f29e421f0e05ec1d0b9f4a04d61bbcd4c10b960 / . / docs / LangRef.html
blob: 0fcc26313ef04906ce98180e386d192c808eb99a [file] [log] [blame]
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
3<html>
4<head>
5 <title>LLVM Assembly Language Reference Manual</title>
6 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
7 <meta name="author" content="Chris Lattner">
8 <meta name="description"
9 content="LLVM Assembly Language Reference Manual.">
10 <link rel="stylesheet" href="llvm.css" type="text/css">
11</head>
12
13<body>
14
15<div class="doc_title"> LLVM Language Reference Manual </div>
16<ol>
17 <li><a href="#abstract">Abstract</a></li>
18 <li><a href="#introduction">Introduction</a></li>
19 <li><a href="#identifiers">Identifiers</a></li>
20 <li><a href="#highlevel">High Level Structure</a>
21 <ol>
22 <li><a href="#modulestructure">Module Structure</a></li>
23 <li><a href="#linkage">Linkage Types</a></li>
24 <li><a href="#callingconv">Calling Conventions</a></li>
25 <li><a href="#globalvars">Global Variables</a></li>
26 <li><a href="#functionstructure">Functions</a></li>
27 <li><a href="#aliasstructure">Aliases</a>
28 <li><a href="#paramattrs">Parameter Attributes</a></li>
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]29 <li><a href="#gc">Garbage Collector Names</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]30 <li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
31 <li><a href="#datalayout">Data Layout</a></li>
32 </ol>
33 </li>
34 <li><a href="#typesystem">Type System</a>
35 <ol>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]36 <li><a href="#t_classifications">Type Classifications</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]37 <li><a href="#t_primitive">Primitive Types</a>
38 <ol>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]39 <li><a href="#t_floating">Floating Point Types</a></li>
40 <li><a href="#t_void">Void Type</a></li>
41 <li><a href="#t_label">Label Type</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]42 </ol>
43 </li>
44 <li><a href="#t_derived">Derived Types</a>
45 <ol>
Chris Lattner251ab812007-12-18 06:18:21 +0000[diff] [blame]46 <li><a href="#t_integer">Integer Type</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]47 <li><a href="#t_array">Array Type</a></li>
48 <li><a href="#t_function">Function Type</a></li>
49 <li><a href="#t_pointer">Pointer Type</a></li>
50 <li><a href="#t_struct">Structure Type</a></li>
51 <li><a href="#t_pstruct">Packed Structure Type</a></li>
52 <li><a href="#t_vector">Vector Type</a></li>
53 <li><a href="#t_opaque">Opaque Type</a></li>
54 </ol>
55 </li>
56 </ol>
57 </li>
58 <li><a href="#constants">Constants</a>
59 <ol>
60 <li><a href="#simpleconstants">Simple Constants</a>
61 <li><a href="#aggregateconstants">Aggregate Constants</a>
62 <li><a href="#globalconstants">Global Variable and Function Addresses</a>
63 <li><a href="#undefvalues">Undefined Values</a>
64 <li><a href="#constantexprs">Constant Expressions</a>
65 </ol>
66 </li>
67 <li><a href="#othervalues">Other Values</a>
68 <ol>
69 <li><a href="#inlineasm">Inline Assembler Expressions</a>
70 </ol>
71 </li>
72 <li><a href="#instref">Instruction Reference</a>
73 <ol>
74 <li><a href="#terminators">Terminator Instructions</a>
75 <ol>
76 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
77 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
78 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
79 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
80 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
81 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
82 </ol>
83 </li>
84 <li><a href="#binaryops">Binary Operations</a>
85 <ol>
86 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
87 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
88 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
89 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
90 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
91 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
92 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
93 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
94 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
95 </ol>
96 </li>
97 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
98 <ol>
99 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
100 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
101 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
102 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
103 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
104 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
105 </ol>
106 </li>
107 <li><a href="#vectorops">Vector Operations</a>
108 <ol>
109 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
110 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
111 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
112 </ol>
113 </li>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]114 <li><a href="#aggregateops">Aggregate Operations</a>
115 <ol>
116 <li><a href="#i_extractvalue">'<tt>extractvalue</tt>' Instruction</a></li>
117 <li><a href="#i_insertvalue">'<tt>insertvalue</tt>' Instruction</a></li>
118 </ol>
119 </li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]120 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
121 <ol>
122 <li><a href="#i_malloc">'<tt>malloc</tt>' Instruction</a></li>
123 <li><a href="#i_free">'<tt>free</tt>' Instruction</a></li>
124 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
125 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
126 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
127 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
128 </ol>
129 </li>
130 <li><a href="#convertops">Conversion Operations</a>
131 <ol>
132 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
133 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
134 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
135 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
136 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
137 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
138 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
139 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
140 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
141 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
142 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
143 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
144 </ol>
145 <li><a href="#otherops">Other Operations</a>
146 <ol>
147 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
148 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]149 <li><a href="#i_vicmp">'<tt>vicmp</tt>' Instruction</a></li>
150 <li><a href="#i_vfcmp">'<tt>vfcmp</tt>' Instruction</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]151 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
152 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
153 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
154 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]155 <li><a href="#i_getresult">'<tt>getresult</tt>' Instruction</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]156 </ol>
157 </li>
158 </ol>
159 </li>
160 <li><a href="#intrinsics">Intrinsic Functions</a>
161 <ol>
162 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
163 <ol>
164 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
165 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
166 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
167 </ol>
168 </li>
169 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
170 <ol>
171 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
172 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
173 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
174 </ol>
175 </li>
176 <li><a href="#int_codegen">Code Generator Intrinsics</a>
177 <ol>
178 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
179 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
180 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
181 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
182 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
183 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
184 <li><a href="#int_readcyclecounter"><tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
185 </ol>
186 </li>
187 <li><a href="#int_libc">Standard C Library Intrinsics</a>
188 <ol>
189 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
190 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
191 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
192 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
193 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]194 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
195 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
196 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]197 </ol>
198 </li>
199 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
200 <ol>
201 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
202 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
203 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
204 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
205 <li><a href="#int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic </a></li>
206 <li><a href="#int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic </a></li>
207 </ol>
208 </li>
209 <li><a href="#int_debugger">Debugger intrinsics</a></li>
210 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]211 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]212 <ol>
213 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]214 </ol>
215 </li>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]216 <li><a href="#int_atomics">Atomic intrinsics</a>
217 <ol>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]218 <li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
219 <li><a href="#int_atomic_lcs"><tt>llvm.atomic.lcs</tt></a></li>
220 <li><a href="#int_atomic_las"><tt>llvm.atomic.las</tt></a></li>
221 <li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]222 </ol>
223 </li>
Reid Spencerb043f672007-07-20 19:59:11 +0000[diff] [blame]224 <li><a href="#int_general">General intrinsics</a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]225 <ol>
Reid Spencerb043f672007-07-20 19:59:11 +0000[diff] [blame]226 <li><a href="#int_var_annotation">
Tanya Lattner51369f32007-09-22 00:01:26 +0000[diff] [blame]227 <tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]228 <li><a href="#int_annotation">
Tanya Lattner51369f32007-09-22 00:01:26 +0000[diff] [blame]229 <tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]230 <li><a href="#int_trap">
231 <tt>llvm.trap</tt>' Intrinsic</a></li>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]232 </ol>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]233 </li>
234 </ol>
235 </li>
236</ol>
237
238<div class="doc_author">
239 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
240 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
241</div>
242
243<!-- *********************************************************************** -->
244<div class="doc_section"> <a name="abstract">Abstract </a></div>
245<!-- *********************************************************************** -->
246
247<div class="doc_text">
248<p>This document is a reference manual for the LLVM assembly language.
249LLVM is an SSA based representation that provides type safety,
250low-level operations, flexibility, and the capability of representing
251'all' high-level languages cleanly. It is the common code
252representation used throughout all phases of the LLVM compilation
253strategy.</p>
254</div>
255
256<!-- *********************************************************************** -->
257<div class="doc_section"> <a name="introduction">Introduction</a> </div>
258<!-- *********************************************************************** -->
259
260<div class="doc_text">
261
262<p>The LLVM code representation is designed to be used in three
263different forms: as an in-memory compiler IR, as an on-disk bitcode
264representation (suitable for fast loading by a Just-In-Time compiler),
265and as a human readable assembly language representation. This allows
266LLVM to provide a powerful intermediate representation for efficient
267compiler transformations and analysis, while providing a natural means
268to debug and visualize the transformations. The three different forms
269of LLVM are all equivalent. This document describes the human readable
270representation and notation.</p>
271
272<p>The LLVM representation aims to be light-weight and low-level
273while being expressive, typed, and extensible at the same time. It
274aims to be a "universal IR" of sorts, by being at a low enough level
275that high-level ideas may be cleanly mapped to it (similar to how
276microprocessors are "universal IR's", allowing many source languages to
277be mapped to them). By providing type information, LLVM can be used as
278the target of optimizations: for example, through pointer analysis, it
279can be proven that a C automatic variable is never accessed outside of
280the current function... allowing it to be promoted to a simple SSA
281value instead of a memory location.</p>
282
283</div>
284
285<!-- _______________________________________________________________________ -->
286<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
287
288<div class="doc_text">
289
290<p>It is important to note that this document describes 'well formed'
291LLVM assembly language. There is a difference between what the parser
292accepts and what is considered 'well formed'. For example, the
293following instruction is syntactically okay, but not well formed:</p>
294
295<div class="doc_code">
296<pre>
297%x = <a href="#i_add">add</a> i32 1, %x
298</pre>
299</div>
300
301<p>...because the definition of <tt>%x</tt> does not dominate all of
302its uses. The LLVM infrastructure provides a verification pass that may
303be used to verify that an LLVM module is well formed. This pass is
304automatically run by the parser after parsing input assembly and by
305the optimizer before it outputs bitcode. The violations pointed out
306by the verifier pass indicate bugs in transformation passes or input to
307the parser.</p>
308</div>
309
Chris Lattnera83fdc02007-10-03 17:34:29 +0000[diff] [blame]310<!-- Describe the typesetting conventions here. -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]311
312<!-- *********************************************************************** -->
313<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
314<!-- *********************************************************************** -->
315
316<div class="doc_text">
317
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]318 <p>LLVM identifiers come in two basic types: global and local. Global
319 identifiers (functions, global variables) begin with the @ character. Local
320 identifiers (register names, types) begin with the % character. Additionally,
321 there are three different formats for identifiers, for different purposes:
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]322
323<ol>
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]324 <li>Named values are represented as a string of characters with their prefix.
325 For example, %foo, @DivisionByZero, %a.really.long.identifier. The actual
326 regular expression used is '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]327 Identifiers which require other characters in their names can be surrounded
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]328 with quotes. In this way, anything except a <tt>&quot;</tt> character can
329 be used in a named value.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]330
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]331 <li>Unnamed values are represented as an unsigned numeric value with their
332 prefix. For example, %12, @2, %44.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]333
334 <li>Constants, which are described in a <a href="#constants">section about
335 constants</a>, below.</li>
336</ol>
337
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]338<p>LLVM requires that values start with a prefix for two reasons: Compilers
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]339don't need to worry about name clashes with reserved words, and the set of
340reserved words may be expanded in the future without penalty. Additionally,
341unnamed identifiers allow a compiler to quickly come up with a temporary
342variable without having to avoid symbol table conflicts.</p>
343
344<p>Reserved words in LLVM are very similar to reserved words in other
345languages. There are keywords for different opcodes
346('<tt><a href="#i_add">add</a></tt>',
347 '<tt><a href="#i_bitcast">bitcast</a></tt>',
348 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
349href="#t_void">void</a></tt>', '<tt><a href="#t_primitive">i32</a></tt>', etc...),
350and others. These reserved words cannot conflict with variable names, because
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]351none of them start with a prefix character ('%' or '@').</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]352
353<p>Here is an example of LLVM code to multiply the integer variable
354'<tt>%X</tt>' by 8:</p>
355
356<p>The easy way:</p>
357
358<div class="doc_code">
359<pre>
360%result = <a href="#i_mul">mul</a> i32 %X, 8
361</pre>
362</div>
363
364<p>After strength reduction:</p>
365
366<div class="doc_code">
367<pre>
368%result = <a href="#i_shl">shl</a> i32 %X, i8 3
369</pre>
370</div>
371
372<p>And the hard way:</p>
373
374<div class="doc_code">
375<pre>
376<a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
377<a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
378%result = <a href="#i_add">add</a> i32 %1, %1
379</pre>
380</div>
381
382<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
383important lexical features of LLVM:</p>
384
385<ol>
386
387 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
388 line.</li>
389
390 <li>Unnamed temporaries are created when the result of a computation is not
391 assigned to a named value.</li>
392
393 <li>Unnamed temporaries are numbered sequentially</li>
394
395</ol>
396
397<p>...and it also shows a convention that we follow in this document. When
398demonstrating instructions, we will follow an instruction with a comment that
399defines the type and name of value produced. Comments are shown in italic
400text.</p>
401
402</div>
403
404<!-- *********************************************************************** -->
405<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
406<!-- *********************************************************************** -->
407
408<!-- ======================================================================= -->
409<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
410</div>
411
412<div class="doc_text">
413
414<p>LLVM programs are composed of "Module"s, each of which is a
415translation unit of the input programs. Each module consists of
416functions, global variables, and symbol table entries. Modules may be
417combined together with the LLVM linker, which merges function (and
418global variable) definitions, resolves forward declarations, and merges
419symbol table entries. Here is an example of the "hello world" module:</p>
420
421<div class="doc_code">
422<pre><i>; Declare the string constant as a global constant...</i>
423<a href="#identifiers">@.LC0</a> = <a href="#linkage_internal">internal</a> <a
424 href="#globalvars">constant</a> <a href="#t_array">[13 x i8]</a> c"hello world\0A\00" <i>; [13 x i8]*</i>
425
426<i>; External declaration of the puts function</i>
427<a href="#functionstructure">declare</a> i32 @puts(i8 *) <i>; i32(i8 *)* </i>
428
429<i>; Definition of main function</i>
430define i32 @main() { <i>; i32()* </i>
431 <i>; Convert [13x i8 ]* to i8 *...</i>
432 %cast210 = <a
433 href="#i_getelementptr">getelementptr</a> [13 x i8 ]* @.LC0, i64 0, i64 0 <i>; i8 *</i>
434
435 <i>; Call puts function to write out the string to stdout...</i>
436 <a
437 href="#i_call">call</a> i32 @puts(i8 * %cast210) <i>; i32</i>
438 <a
439 href="#i_ret">ret</a> i32 0<br>}<br>
440</pre>
441</div>
442
443<p>This example is made up of a <a href="#globalvars">global variable</a>
444named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
445function, and a <a href="#functionstructure">function definition</a>
446for "<tt>main</tt>".</p>
447
448<p>In general, a module is made up of a list of global values,
449where both functions and global variables are global values. Global values are
450represented by a pointer to a memory location (in this case, a pointer to an
451array of char, and a pointer to a function), and have one of the following <a
452href="#linkage">linkage types</a>.</p>
453
454</div>
455
456<!-- ======================================================================= -->
457<div class="doc_subsection">
458 <a name="linkage">Linkage Types</a>
459</div>
460
461<div class="doc_text">
462
463<p>
464All Global Variables and Functions have one of the following types of linkage:
465</p>
466
467<dl>
468
469 <dt><tt><b><a name="linkage_internal">internal</a></b></tt> </dt>
470
471 <dd>Global values with internal linkage are only directly accessible by
472 objects in the current module. In particular, linking code into a module with
473 an internal global value may cause the internal to be renamed as necessary to
474 avoid collisions. Because the symbol is internal to the module, all
475 references can be updated. This corresponds to the notion of the
476 '<tt>static</tt>' keyword in C.
477 </dd>
478
479 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
480
481 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
482 the same name when linkage occurs. This is typically used to implement
483 inline functions, templates, or other code which must be generated in each
484 translation unit that uses it. Unreferenced <tt>linkonce</tt> globals are
485 allowed to be discarded.
486 </dd>
487
488 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
489
490 <dd>"<tt>weak</tt>" linkage is exactly the same as <tt>linkonce</tt> linkage,
491 except that unreferenced <tt>weak</tt> globals may not be discarded. This is
492 used for globals that may be emitted in multiple translation units, but that
493 are not guaranteed to be emitted into every translation unit that uses them.
494 One example of this are common globals in C, such as "<tt>int X;</tt>" at
495 global scope.
496 </dd>
497
498 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
499
500 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
501 pointer to array type. When two global variables with appending linkage are
502 linked together, the two global arrays are appended together. This is the
503 LLVM, typesafe, equivalent of having the system linker append together
504 "sections" with identical names when .o files are linked.
505 </dd>
506
507 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt>: </dt>
508 <dd>The semantics of this linkage follow the ELF model: the symbol is weak
509 until linked, if not linked, the symbol becomes null instead of being an
510 undefined reference.
511 </dd>
512
513 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
514
515 <dd>If none of the above identifiers are used, the global is externally
516 visible, meaning that it participates in linkage and can be used to resolve
517 external symbol references.
518 </dd>
519</dl>
520
521 <p>
522 The next two types of linkage are targeted for Microsoft Windows platform
523 only. They are designed to support importing (exporting) symbols from (to)
524 DLLs.
525 </p>
526
527 <dl>
528 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt>: </dt>
529
530 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
531 or variable via a global pointer to a pointer that is set up by the DLL
532 exporting the symbol. On Microsoft Windows targets, the pointer name is
533 formed by combining <code>_imp__</code> and the function or variable name.
534 </dd>
535
536 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt>: </dt>
537
538 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
539 pointer to a pointer in a DLL, so that it can be referenced with the
540 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
541 name is formed by combining <code>_imp__</code> and the function or variable
542 name.
543 </dd>
544
545</dl>
546
547<p><a name="linkage_external"></a>For example, since the "<tt>.LC0</tt>"
548variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
549variable and was linked with this one, one of the two would be renamed,
550preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
551external (i.e., lacking any linkage declarations), they are accessible
552outside of the current module.</p>
553<p>It is illegal for a function <i>declaration</i>
554to have any linkage type other than "externally visible", <tt>dllimport</tt>,
555or <tt>extern_weak</tt>.</p>
556<p>Aliases can have only <tt>external</tt>, <tt>internal</tt> and <tt>weak</tt>
557linkages.
558</div>
559
560<!-- ======================================================================= -->
561<div class="doc_subsection">
562 <a name="callingconv">Calling Conventions</a>
563</div>
564
565<div class="doc_text">
566
567<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
568and <a href="#i_invoke">invokes</a> can all have an optional calling convention
569specified for the call. The calling convention of any pair of dynamic
570caller/callee must match, or the behavior of the program is undefined. The
571following calling conventions are supported by LLVM, and more may be added in
572the future:</p>
573
574<dl>
575 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
576
577 <dd>This calling convention (the default if no other calling convention is
578 specified) matches the target C calling conventions. This calling convention
579 supports varargs function calls and tolerates some mismatch in the declared
580 prototype and implemented declaration of the function (as does normal C).
581 </dd>
582
583 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
584
585 <dd>This calling convention attempts to make calls as fast as possible
586 (e.g. by passing things in registers). This calling convention allows the
587 target to use whatever tricks it wants to produce fast code for the target,
588 without having to conform to an externally specified ABI. Implementations of
Arnold Schwaighofer07444922008-05-14 09:17:12 +0000[diff] [blame]589 this convention should allow arbitrary
590 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> to be
591 supported. This calling convention does not support varargs and requires the
592 prototype of all callees to exactly match the prototype of the function
593 definition.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]594 </dd>
595
596 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
597
598 <dd>This calling convention attempts to make code in the caller as efficient
599 as possible under the assumption that the call is not commonly executed. As
600 such, these calls often preserve all registers so that the call does not break
601 any live ranges in the caller side. This calling convention does not support
602 varargs and requires the prototype of all callees to exactly match the
603 prototype of the function definition.
604 </dd>
605
606 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
607
608 <dd>Any calling convention may be specified by number, allowing
609 target-specific calling conventions to be used. Target specific calling
610 conventions start at 64.
611 </dd>
612</dl>
613
614<p>More calling conventions can be added/defined on an as-needed basis, to
615support pascal conventions or any other well-known target-independent
616convention.</p>
617
618</div>
619
620<!-- ======================================================================= -->
621<div class="doc_subsection">
622 <a name="visibility">Visibility Styles</a>
623</div>
624
625<div class="doc_text">
626
627<p>
628All Global Variables and Functions have one of the following visibility styles:
629</p>
630
631<dl>
632 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
633
634 <dd>On ELF, default visibility means that the declaration is visible to other
635 modules and, in shared libraries, means that the declared entity may be
636 overridden. On Darwin, default visibility means that the declaration is
637 visible to other modules. Default visibility corresponds to "external
638 linkage" in the language.
639 </dd>
640
641 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
642
643 <dd>Two declarations of an object with hidden visibility refer to the same
644 object if they are in the same shared object. Usually, hidden visibility
645 indicates that the symbol will not be placed into the dynamic symbol table,
646 so no other module (executable or shared library) can reference it
647 directly.
648 </dd>
649
650 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
651
652 <dd>On ELF, protected visibility indicates that the symbol will be placed in
653 the dynamic symbol table, but that references within the defining module will
654 bind to the local symbol. That is, the symbol cannot be overridden by another
655 module.
656 </dd>
657</dl>
658
659</div>
660
661<!-- ======================================================================= -->
662<div class="doc_subsection">
663 <a name="globalvars">Global Variables</a>
664</div>
665
666<div class="doc_text">
667
668<p>Global variables define regions of memory allocated at compilation time
669instead of run-time. Global variables may optionally be initialized, may have
670an explicit section to be placed in, and may have an optional explicit alignment
671specified. A variable may be defined as "thread_local", which means that it
672will not be shared by threads (each thread will have a separated copy of the
673variable). A variable may be defined as a global "constant," which indicates
674that the contents of the variable will <b>never</b> be modified (enabling better
675optimization, allowing the global data to be placed in the read-only section of
676an executable, etc). Note that variables that need runtime initialization
677cannot be marked "constant" as there is a store to the variable.</p>
678
679<p>
680LLVM explicitly allows <em>declarations</em> of global variables to be marked
681constant, even if the final definition of the global is not. This capability
682can be used to enable slightly better optimization of the program, but requires
683the language definition to guarantee that optimizations based on the
684'constantness' are valid for the translation units that do not include the
685definition.
686</p>
687
688<p>As SSA values, global variables define pointer values that are in
689scope (i.e. they dominate) all basic blocks in the program. Global
690variables always define a pointer to their "content" type because they
691describe a region of memory, and all memory objects in LLVM are
692accessed through pointers.</p>
693
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]694<p>A global variable may be declared to reside in a target-specifc numbered
695address space. For targets that support them, address spaces may affect how
696optimizations are performed and/or what target instructions are used to access
Christopher Lamb20a39e92007-12-12 08:44:39 +0000[diff] [blame]697the variable. The default address space is zero. The address space qualifier
698must precede any other attributes.</p>
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]699
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]700<p>LLVM allows an explicit section to be specified for globals. If the target
701supports it, it will emit globals to the section specified.</p>
702
703<p>An explicit alignment may be specified for a global. If not present, or if
704the alignment is set to zero, the alignment of the global is set by the target
705to whatever it feels convenient. If an explicit alignment is specified, the
706global is forced to have at least that much alignment. All alignments must be
707a power of 2.</p>
708
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]709<p>For example, the following defines a global in a numbered address space with
710an initializer, section, and alignment:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]711
712<div class="doc_code">
713<pre>
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]714@G = constant float 1.0 addrspace(5), section "foo", align 4
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]715</pre>
716</div>
717
718</div>
719
720
721<!-- ======================================================================= -->
722<div class="doc_subsection">
723 <a name="functionstructure">Functions</a>
724</div>
725
726<div class="doc_text">
727
728<p>LLVM function definitions consist of the "<tt>define</tt>" keyord,
729an optional <a href="#linkage">linkage type</a>, an optional
730<a href="#visibility">visibility style</a>, an optional
731<a href="#callingconv">calling convention</a>, a return type, an optional
732<a href="#paramattrs">parameter attribute</a> for the return type, a function
733name, a (possibly empty) argument list (each with optional
734<a href="#paramattrs">parameter attributes</a>), an optional section, an
Gordon Henriksend606f5b2007-12-10 03:30:21 +0000[diff] [blame]735optional alignment, an optional <a href="#gc">garbage collector name</a>, an
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]736opening curly brace, a list of basic blocks, and a closing curly brace.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]737
738LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
739optional <a href="#linkage">linkage type</a>, an optional
740<a href="#visibility">visibility style</a>, an optional
741<a href="#callingconv">calling convention</a>, a return type, an optional
742<a href="#paramattrs">parameter attribute</a> for the return type, a function
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]743name, a possibly empty list of arguments, an optional alignment, and an optional
Gordon Henriksend606f5b2007-12-10 03:30:21 +0000[diff] [blame]744<a href="#gc">garbage collector name</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]745
746<p>A function definition contains a list of basic blocks, forming the CFG for
747the function. Each basic block may optionally start with a label (giving the
748basic block a symbol table entry), contains a list of instructions, and ends
749with a <a href="#terminators">terminator</a> instruction (such as a branch or
750function return).</p>
751
752<p>The first basic block in a function is special in two ways: it is immediately
753executed on entrance to the function, and it is not allowed to have predecessor
754basic blocks (i.e. there can not be any branches to the entry block of a
755function). Because the block can have no predecessors, it also cannot have any
756<a href="#i_phi">PHI nodes</a>.</p>
757
758<p>LLVM allows an explicit section to be specified for functions. If the target
759supports it, it will emit functions to the section specified.</p>
760
761<p>An explicit alignment may be specified for a function. If not present, or if
762the alignment is set to zero, the alignment of the function is set by the target
763to whatever it feels convenient. If an explicit alignment is specified, the
764function is forced to have at least that much alignment. All alignments must be
765a power of 2.</p>
766
767</div>
768
769
770<!-- ======================================================================= -->
771<div class="doc_subsection">
772 <a name="aliasstructure">Aliases</a>
773</div>
774<div class="doc_text">
775 <p>Aliases act as "second name" for the aliasee value (which can be either
Anton Korobeynikov96822812008-03-22 08:36:14 +0000[diff] [blame]776 function, global variable, another alias or bitcast of global value). Aliases
777 may have an optional <a href="#linkage">linkage type</a>, and an
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]778 optional <a href="#visibility">visibility style</a>.</p>
779
780 <h5>Syntax:</h5>
781
782<div class="doc_code">
783<pre>
784@&lt;Name&gt; = [Linkage] [Visibility] alias &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
785</pre>
786</div>
787
788</div>
789
790
791
792<!-- ======================================================================= -->
793<div class="doc_subsection"><a name="paramattrs">Parameter Attributes</a></div>
794<div class="doc_text">
795 <p>The return type and each parameter of a function type may have a set of
796 <i>parameter attributes</i> associated with them. Parameter attributes are
797 used to communicate additional information about the result or parameters of
Duncan Sandsf5588dc2007-11-27 13:23:08 +0000[diff] [blame]798 a function. Parameter attributes are considered to be part of the function,
799 not of the function type, so functions with different parameter attributes
800 can have the same function type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]801
802 <p>Parameter attributes are simple keywords that follow the type specified. If
803 multiple parameter attributes are needed, they are space separated. For
804 example:</p>
805
806<div class="doc_code">
807<pre>
Duncan Sandsf5588dc2007-11-27 13:23:08 +0000[diff] [blame]808declare i32 @printf(i8* noalias , ...) nounwind
809declare i32 @atoi(i8*) nounwind readonly
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]810</pre>
811</div>
812
Duncan Sandsf5588dc2007-11-27 13:23:08 +0000[diff] [blame]813 <p>Note that any attributes for the function result (<tt>nounwind</tt>,
814 <tt>readonly</tt>) come immediately after the argument list.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]815
816 <p>Currently, only the following parameter attributes are defined:</p>
817 <dl>
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]818 <dt><tt>zeroext</tt></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]819 <dd>This indicates that the parameter should be zero extended just before
820 a call to this function.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]821
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]822 <dt><tt>signext</tt></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]823 <dd>This indicates that the parameter should be sign extended just before
824 a call to this function.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]825
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]826 <dt><tt>inreg</tt></dt>
827 <dd>This indicates that the parameter should be placed in register (if
828 possible) during assembling function call. Support for this attribute is
829 target-specific</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]830
831 <dt><tt>byval</tt></dt>
Chris Lattner04c86182008-01-15 04:34:22 +0000[diff] [blame]832 <dd>This indicates that the pointer parameter should really be passed by
833 value to the function. The attribute implies that a hidden copy of the
834 pointee is made between the caller and the callee, so the callee is unable
835 to modify the value in the callee. This attribute is only valid on llvm
836 pointer arguments. It is generally used to pass structs and arrays by
837 value, but is also valid on scalars (even though this is silly).</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]838
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]839 <dt><tt>sret</tt></dt>
Duncan Sands616cc032008-02-18 04:19:38 +0000[diff] [blame]840 <dd>This indicates that the pointer parameter specifies the address of a
841 structure that is the return value of the function in the source program.
Duncan Sandsef0e9e42008-03-17 12:17:41 +0000[diff] [blame]842 Loads and stores to the structure are assumed not to trap.
Duncan Sands616cc032008-02-18 04:19:38 +0000[diff] [blame]843 May only be applied to the first parameter.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]844
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]845 <dt><tt>noalias</tt></dt>
Owen Andersonc4fc4cd2008-02-18 04:09:01 +0000[diff] [blame]846 <dd>This indicates that the parameter does not alias any global or any other
847 parameter. The caller is responsible for ensuring that this is the case,
848 usually by placing the value in a stack allocation.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]849
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]850 <dt><tt>noreturn</tt></dt>
851 <dd>This function attribute indicates that the function never returns. This
852 indicates to LLVM that every call to this function should be treated as if
853 an <tt>unreachable</tt> instruction immediately followed the call.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]854
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]855 <dt><tt>nounwind</tt></dt>
Duncan Sandsef0e9e42008-03-17 12:17:41 +0000[diff] [blame]856 <dd>This function attribute indicates that no exceptions unwind out of the
857 function. Usually this is because the function makes no use of exceptions,
858 but it may also be that the function catches any exceptions thrown when
859 executing it.</dd>
860
Duncan Sands4ee46812007-07-27 19:57:41 +0000[diff] [blame]861 <dt><tt>nest</tt></dt>
862 <dd>This indicates that the parameter can be excised using the
863 <a href="#int_trampoline">trampoline intrinsics</a>.</dd>
Duncan Sands13e13f82007-11-22 20:23:04 +0000[diff] [blame]864 <dt><tt>readonly</tt></dt>
Duncan Sandsd69c0e62007-11-14 21:14:02 +0000[diff] [blame]865 <dd>This function attribute indicates that the function has no side-effects
Duncan Sands13e13f82007-11-22 20:23:04 +0000[diff] [blame]866 except for producing a return value or throwing an exception. The value
867 returned must only depend on the function arguments and/or global variables.
868 It may use values obtained by dereferencing pointers.</dd>
869 <dt><tt>readnone</tt></dt>
870 <dd>A <tt>readnone</tt> function has the same restrictions as a <tt>readonly</tt>
Duncan Sandsd69c0e62007-11-14 21:14:02 +0000[diff] [blame]871 function, but in addition it is not allowed to dereference any pointer arguments
872 or global variables.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]873 </dl>
874
875</div>
876
877<!-- ======================================================================= -->
878<div class="doc_subsection">
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]879 <a name="gc">Garbage Collector Names</a>
880</div>
881
882<div class="doc_text">
883<p>Each function may specify a garbage collector name, which is simply a
884string.</p>
885
886<div class="doc_code"><pre
887>define void @f() gc "name" { ...</pre></div>
888
889<p>The compiler declares the supported values of <i>name</i>. Specifying a
890collector which will cause the compiler to alter its output in order to support
891the named garbage collection algorithm.</p>
892</div>
893
894<!-- ======================================================================= -->
895<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]896 <a name="moduleasm">Module-Level Inline Assembly</a>
897</div>
898
899<div class="doc_text">
900<p>
901Modules may contain "module-level inline asm" blocks, which corresponds to the
902GCC "file scope inline asm" blocks. These blocks are internally concatenated by
903LLVM and treated as a single unit, but may be separated in the .ll file if
904desired. The syntax is very simple:
905</p>
906
907<div class="doc_code">
908<pre>
909module asm "inline asm code goes here"
910module asm "more can go here"
911</pre>
912</div>
913
914<p>The strings can contain any character by escaping non-printable characters.
915 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
916 for the number.
917</p>
918
919<p>
920 The inline asm code is simply printed to the machine code .s file when
921 assembly code is generated.
922</p>
923</div>
924
925<!-- ======================================================================= -->
926<div class="doc_subsection">
927 <a name="datalayout">Data Layout</a>
928</div>
929
930<div class="doc_text">
931<p>A module may specify a target specific data layout string that specifies how
932data is to be laid out in memory. The syntax for the data layout is simply:</p>
933<pre> target datalayout = "<i>layout specification</i>"</pre>
934<p>The <i>layout specification</i> consists of a list of specifications
935separated by the minus sign character ('-'). Each specification starts with a
936letter and may include other information after the letter to define some
937aspect of the data layout. The specifications accepted are as follows: </p>
938<dl>
939 <dt><tt>E</tt></dt>
940 <dd>Specifies that the target lays out data in big-endian form. That is, the
941 bits with the most significance have the lowest address location.</dd>
942 <dt><tt>e</tt></dt>
943 <dd>Specifies that hte target lays out data in little-endian form. That is,
944 the bits with the least significance have the lowest address location.</dd>
945 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
946 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
947 <i>preferred</i> alignments. All sizes are in bits. Specifying the <i>pref</i>
948 alignment is optional. If omitted, the preceding <tt>:</tt> should be omitted
949 too.</dd>
950 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
951 <dd>This specifies the alignment for an integer type of a given bit
952 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
953 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
954 <dd>This specifies the alignment for a vector type of a given bit
955 <i>size</i>.</dd>
956 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
957 <dd>This specifies the alignment for a floating point type of a given bit
958 <i>size</i>. The value of <i>size</i> must be either 32 (float) or 64
959 (double).</dd>
960 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
961 <dd>This specifies the alignment for an aggregate type of a given bit
962 <i>size</i>.</dd>
963</dl>
964<p>When constructing the data layout for a given target, LLVM starts with a
965default set of specifications which are then (possibly) overriden by the
966specifications in the <tt>datalayout</tt> keyword. The default specifications
967are given in this list:</p>
968<ul>
969 <li><tt>E</tt> - big endian</li>
970 <li><tt>p:32:64:64</tt> - 32-bit pointers with 64-bit alignment</li>
971 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
972 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
973 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
974 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
975 <li><tt>i64:32:64</tt> - i64 has abi alignment of 32-bits but preferred
976 alignment of 64-bits</li>
977 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
978 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
979 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
980 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
981 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
982</ul>
983<p>When llvm is determining the alignment for a given type, it uses the
984following rules:
985<ol>
986 <li>If the type sought is an exact match for one of the specifications, that
987 specification is used.</li>
988 <li>If no match is found, and the type sought is an integer type, then the
989 smallest integer type that is larger than the bitwidth of the sought type is
990 used. If none of the specifications are larger than the bitwidth then the the
991 largest integer type is used. For example, given the default specifications
992 above, the i7 type will use the alignment of i8 (next largest) while both
993 i65 and i256 will use the alignment of i64 (largest specified).</li>
994 <li>If no match is found, and the type sought is a vector type, then the
995 largest vector type that is smaller than the sought vector type will be used
996 as a fall back. This happens because <128 x double> can be implemented in
997 terms of 64 <2 x double>, for example.</li>
998</ol>
999</div>
1000
1001<!-- *********************************************************************** -->
1002<div class="doc_section"> <a name="typesystem">Type System</a> </div>
1003<!-- *********************************************************************** -->
1004
1005<div class="doc_text">
1006
1007<p>The LLVM type system is one of the most important features of the
1008intermediate representation. Being typed enables a number of
1009optimizations to be performed on the IR directly, without having to do
1010extra analyses on the side before the transformation. A strong type
1011system makes it easier to read the generated code and enables novel
1012analyses and transformations that are not feasible to perform on normal
1013three address code representations.</p>
1014
1015</div>
1016
1017<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1018<div class="doc_subsection"> <a name="t_classifications">Type
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1019Classifications</a> </div>
1020<div class="doc_text">
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1021<p>The types fall into a few useful
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1022classifications:</p>
1023
1024<table border="1" cellspacing="0" cellpadding="4">
1025 <tbody>
1026 <tr><th>Classification</th><th>Types</th></tr>
1027 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1028 <td><a href="#t_integer">integer</a></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1029 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
1030 </tr>
1031 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1032 <td><a href="#t_floating">floating point</a></td>
1033 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1034 </tr>
1035 <tr>
1036 <td><a name="t_firstclass">first class</a></td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1037 <td><a href="#t_integer">integer</a>,
1038 <a href="#t_floating">floating point</a>,
1039 <a href="#t_pointer">pointer</a>,
1040 <a href="#t_vector">vector</a>
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]1041 <a href="#t_struct">structure</a>,
1042 <a href="#t_array">array</a>,
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1043 </td>
1044 </tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1045 <tr>
1046 <td><a href="#t_primitive">primitive</a></td>
1047 <td><a href="#t_label">label</a>,
1048 <a href="#t_void">void</a>,
1049 <a href="#t_integer">integer</a>,
1050 <a href="#t_floating">floating point</a>.</td>
1051 </tr>
1052 <tr>
1053 <td><a href="#t_derived">derived</a></td>
1054 <td><a href="#t_integer">integer</a>,
1055 <a href="#t_array">array</a>,
1056 <a href="#t_function">function</a>,
1057 <a href="#t_pointer">pointer</a>,
1058 <a href="#t_struct">structure</a>,
1059 <a href="#t_pstruct">packed structure</a>,
1060 <a href="#t_vector">vector</a>,
1061 <a href="#t_opaque">opaque</a>.
1062 </tr>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1063 </tbody>
1064</table>
1065
1066<p>The <a href="#t_firstclass">first class</a> types are perhaps the
1067most important. Values of these types are the only ones which can be
1068produced by instructions, passed as arguments, or used as operands to
Dan Gohman4f29e422008-05-23 21:53:15 +0000[diff] [blame^]1069instructions.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1070</div>
1071
1072<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1073<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1074
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1075<div class="doc_text">
1076<p>The primitive types are the fundamental building blocks of the LLVM
1077system.</p>
1078
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1079</div>
1080
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1081<!-- _______________________________________________________________________ -->
1082<div class="doc_subsubsection"> <a name="t_floating">Floating Point Types</a> </div>
1083
1084<div class="doc_text">
1085 <table>
1086 <tbody>
1087 <tr><th>Type</th><th>Description</th></tr>
1088 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1089 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1090 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1091 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1092 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1093 </tbody>
1094 </table>
1095</div>
1096
1097<!-- _______________________________________________________________________ -->
1098<div class="doc_subsubsection"> <a name="t_void">Void Type</a> </div>
1099
1100<div class="doc_text">
1101<h5>Overview:</h5>
1102<p>The void type does not represent any value and has no size.</p>
1103
1104<h5>Syntax:</h5>
1105
1106<pre>
1107 void
1108</pre>
1109</div>
1110
1111<!-- _______________________________________________________________________ -->
1112<div class="doc_subsubsection"> <a name="t_label">Label Type</a> </div>
1113
1114<div class="doc_text">
1115<h5>Overview:</h5>
1116<p>The label type represents code labels.</p>
1117
1118<h5>Syntax:</h5>
1119
1120<pre>
1121 label
1122</pre>
1123</div>
1124
1125
1126<!-- ======================================================================= -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1127<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
1128
1129<div class="doc_text">
1130
1131<p>The real power in LLVM comes from the derived types in the system.
1132This is what allows a programmer to represent arrays, functions,
1133pointers, and other useful types. Note that these derived types may be
1134recursive: For example, it is possible to have a two dimensional array.</p>
1135
1136</div>
1137
1138<!-- _______________________________________________________________________ -->
1139<div class="doc_subsubsection"> <a name="t_integer">Integer Type</a> </div>
1140
1141<div class="doc_text">
1142
1143<h5>Overview:</h5>
1144<p>The integer type is a very simple derived type that simply specifies an
1145arbitrary bit width for the integer type desired. Any bit width from 1 bit to
11462^23-1 (about 8 million) can be specified.</p>
1147
1148<h5>Syntax:</h5>
1149
1150<pre>
1151 iN
1152</pre>
1153
1154<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1155value.</p>
1156
1157<h5>Examples:</h5>
1158<table class="layout">
Chris Lattner251ab812007-12-18 06:18:21 +0000[diff] [blame]1159 <tbody>
1160 <tr>
1161 <td><tt>i1</tt></td>
1162 <td>a single-bit integer.</td>
1163 </tr><tr>
1164 <td><tt>i32</tt></td>
1165 <td>a 32-bit integer.</td>
1166 </tr><tr>
1167 <td><tt>i1942652</tt></td>
1168 <td>a really big integer of over 1 million bits.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1169 </tr>
Chris Lattner251ab812007-12-18 06:18:21 +0000[diff] [blame]1170 </tbody>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1171</table>
1172</div>
1173
1174<!-- _______________________________________________________________________ -->
1175<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
1176
1177<div class="doc_text">
1178
1179<h5>Overview:</h5>
1180
1181<p>The array type is a very simple derived type that arranges elements
1182sequentially in memory. The array type requires a size (number of
1183elements) and an underlying data type.</p>
1184
1185<h5>Syntax:</h5>
1186
1187<pre>
1188 [&lt;# elements&gt; x &lt;elementtype&gt;]
1189</pre>
1190
1191<p>The number of elements is a constant integer value; elementtype may
1192be any type with a size.</p>
1193
1194<h5>Examples:</h5>
1195<table class="layout">
1196 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1197 <td class="left"><tt>[40 x i32]</tt></td>
1198 <td class="left">Array of 40 32-bit integer values.</td>
1199 </tr>
1200 <tr class="layout">
1201 <td class="left"><tt>[41 x i32]</tt></td>
1202 <td class="left">Array of 41 32-bit integer values.</td>
1203 </tr>
1204 <tr class="layout">
1205 <td class="left"><tt>[4 x i8]</tt></td>
1206 <td class="left">Array of 4 8-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1207 </tr>
1208</table>
1209<p>Here are some examples of multidimensional arrays:</p>
1210<table class="layout">
1211 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1212 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1213 <td class="left">3x4 array of 32-bit integer values.</td>
1214 </tr>
1215 <tr class="layout">
1216 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1217 <td class="left">12x10 array of single precision floating point values.</td>
1218 </tr>
1219 <tr class="layout">
1220 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1221 <td class="left">2x3x4 array of 16-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1222 </tr>
1223</table>
1224
1225<p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
1226length array. Normally, accesses past the end of an array are undefined in
1227LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
1228As a special case, however, zero length arrays are recognized to be variable
1229length. This allows implementation of 'pascal style arrays' with the LLVM
1230type "{ i32, [0 x float]}", for example.</p>
1231
1232</div>
1233
1234<!-- _______________________________________________________________________ -->
1235<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
1236<div class="doc_text">
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1237
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1238<h5>Overview:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1239
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1240<p>The function type can be thought of as a function signature. It
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1241consists of a return type and a list of formal parameter types. The
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1242return 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]1243If 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]1244class types, and the struct must have at least one element.</p>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1245
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1246<h5>Syntax:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1247
1248<pre>
1249 &lt;returntype list&gt; (&lt;parameter list&gt;)
1250</pre>
1251
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1252<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
1253specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1254which indicates that the function takes a variable number of arguments.
1255Variable argument functions can access their arguments with the <a
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1256 href="#int_varargs">variable argument handling intrinsic</a> functions.
1257'<tt>&lt;returntype list&gt;</tt>' is a comma-separated list of
1258<a href="#t_firstclass">first class</a> type specifiers.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1259
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1260<h5>Examples:</h5>
1261<table class="layout">
1262 <tr class="layout">
1263 <td class="left"><tt>i32 (i32)</tt></td>
1264 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
1265 </td>
1266 </tr><tr class="layout">
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]1267 <td class="left"><tt>float&nbsp;(i16&nbsp;signext,&nbsp;i32&nbsp;*)&nbsp;*
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1268 </tt></td>
1269 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
1270 an <tt>i16</tt> that should be sign extended and a
1271 <a href="#t_pointer">pointer</a> to <tt>i32</tt>, returning
1272 <tt>float</tt>.
1273 </td>
1274 </tr><tr class="layout">
1275 <td class="left"><tt>i32 (i8*, ...)</tt></td>
1276 <td class="left">A vararg function that takes at least one
1277 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1278 which returns an integer. This is the signature for <tt>printf</tt> in
1279 LLVM.
1280 </td>
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1281 </tr><tr class="layout">
1282 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Devang Patel6dddba22008-03-24 18:10:52 +0000[diff] [blame]1283 <td class="left">A function taking an <tt>i32></tt>, returning two
1284 <tt> i32 </tt> values as an aggregate of type <tt>{ i32, i32 }</tt>
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1285 </td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1286 </tr>
1287</table>
1288
1289</div>
1290<!-- _______________________________________________________________________ -->
1291<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
1292<div class="doc_text">
1293<h5>Overview:</h5>
1294<p>The structure type is used to represent a collection of data members
1295together in memory. The packing of the field types is defined to match
1296the ABI of the underlying processor. The elements of a structure may
1297be any type that has a size.</p>
1298<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1299and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1300field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1301instruction.</p>
1302<h5>Syntax:</h5>
1303<pre> { &lt;type list&gt; }<br></pre>
1304<h5>Examples:</h5>
1305<table class="layout">
1306 <tr class="layout">
1307 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1308 <td class="left">A triple of three <tt>i32</tt> values</td>
1309 </tr><tr class="layout">
1310 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1311 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1312 second element is a <a href="#t_pointer">pointer</a> to a
1313 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1314 an <tt>i32</tt>.</td>
1315 </tr>
1316</table>
1317</div>
1318
1319<!-- _______________________________________________________________________ -->
1320<div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
1321</div>
1322<div class="doc_text">
1323<h5>Overview:</h5>
1324<p>The packed structure type is used to represent a collection of data members
1325together in memory. There is no padding between fields. Further, the alignment
1326of a packed structure is 1 byte. The elements of a packed structure may
1327be any type that has a size.</p>
1328<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1329and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1330field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1331instruction.</p>
1332<h5>Syntax:</h5>
1333<pre> &lt; { &lt;type list&gt; } &gt; <br></pre>
1334<h5>Examples:</h5>
1335<table class="layout">
1336 <tr class="layout">
1337 <td class="left"><tt>&lt; { i32, i32, i32 } &gt;</tt></td>
1338 <td class="left">A triple of three <tt>i32</tt> values</td>
1339 </tr><tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1340 <td class="left"><tt>&lt; { float, i32 (i32)* } &gt;</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1341 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1342 second element is a <a href="#t_pointer">pointer</a> to a
1343 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1344 an <tt>i32</tt>.</td>
1345 </tr>
1346</table>
1347</div>
1348
1349<!-- _______________________________________________________________________ -->
1350<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
1351<div class="doc_text">
1352<h5>Overview:</h5>
1353<p>As in many languages, the pointer type represents a pointer or
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]1354reference to another object, which must live in memory. Pointer types may have
1355an optional address space attribute defining the target-specific numbered
1356address space where the pointed-to object resides. The default address space is
1357zero.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1358<h5>Syntax:</h5>
1359<pre> &lt;type&gt; *<br></pre>
1360<h5>Examples:</h5>
1361<table class="layout">
1362 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1363 <td class="left"><tt>[4x i32]*</tt></td>
1364 <td class="left">A <a href="#t_pointer">pointer</a> to <a
1365 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
1366 </tr>
1367 <tr class="layout">
1368 <td class="left"><tt>i32 (i32 *) *</tt></td>
1369 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1370 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1371 <tt>i32</tt>.</td>
1372 </tr>
1373 <tr class="layout">
1374 <td class="left"><tt>i32 addrspace(5)*</tt></td>
1375 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
1376 that resides in address space #5.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1377 </tr>
1378</table>
1379</div>
1380
1381<!-- _______________________________________________________________________ -->
1382<div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
1383<div class="doc_text">
1384
1385<h5>Overview:</h5>
1386
1387<p>A vector type is a simple derived type that represents a vector
1388of elements. Vector types are used when multiple primitive data
1389are operated in parallel using a single instruction (SIMD).
1390A vector type requires a size (number of
1391elements) and an underlying primitive data type. Vectors must have a power
1392of two length (1, 2, 4, 8, 16 ...). Vector types are
1393considered <a href="#t_firstclass">first class</a>.</p>
1394
1395<h5>Syntax:</h5>
1396
1397<pre>
1398 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1399</pre>
1400
1401<p>The number of elements is a constant integer value; elementtype may
1402be any integer or floating point type.</p>
1403
1404<h5>Examples:</h5>
1405
1406<table class="layout">
1407 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1408 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
1409 <td class="left">Vector of 4 32-bit integer values.</td>
1410 </tr>
1411 <tr class="layout">
1412 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
1413 <td class="left">Vector of 8 32-bit floating-point values.</td>
1414 </tr>
1415 <tr class="layout">
1416 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
1417 <td class="left">Vector of 2 64-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1418 </tr>
1419</table>
1420</div>
1421
1422<!-- _______________________________________________________________________ -->
1423<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1424<div class="doc_text">
1425
1426<h5>Overview:</h5>
1427
1428<p>Opaque types are used to represent unknown types in the system. This
Gordon Henriksenda0706e2007-10-14 00:34:53 +0000[diff] [blame]1429corresponds (for example) to the C notion of a forward declared structure type.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1430In LLVM, opaque types can eventually be resolved to any type (not just a
1431structure type).</p>
1432
1433<h5>Syntax:</h5>
1434
1435<pre>
1436 opaque
1437</pre>
1438
1439<h5>Examples:</h5>
1440
1441<table class="layout">
1442 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1443 <td class="left"><tt>opaque</tt></td>
1444 <td class="left">An opaque type.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1445 </tr>
1446</table>
1447</div>
1448
1449
1450<!-- *********************************************************************** -->
1451<div class="doc_section"> <a name="constants">Constants</a> </div>
1452<!-- *********************************************************************** -->
1453
1454<div class="doc_text">
1455
1456<p>LLVM has several different basic types of constants. This section describes
1457them all and their syntax.</p>
1458
1459</div>
1460
1461<!-- ======================================================================= -->
1462<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
1463
1464<div class="doc_text">
1465
1466<dl>
1467 <dt><b>Boolean constants</b></dt>
1468
1469 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
1470 constants of the <tt><a href="#t_primitive">i1</a></tt> type.
1471 </dd>
1472
1473 <dt><b>Integer constants</b></dt>
1474
1475 <dd>Standard integers (such as '4') are constants of the <a
1476 href="#t_integer">integer</a> type. Negative numbers may be used with
1477 integer types.
1478 </dd>
1479
1480 <dt><b>Floating point constants</b></dt>
1481
1482 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
1483 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]1484 notation (see below). The assembler requires the exact decimal value of
1485 a floating-point constant. For example, the assembler accepts 1.25 but
1486 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
1487 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1488
1489 <dt><b>Null pointer constants</b></dt>
1490
1491 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
1492 and must be of <a href="#t_pointer">pointer type</a>.</dd>
1493
1494</dl>
1495
1496<p>The one non-intuitive notation for constants is the optional hexadecimal form
1497of floating point constants. For example, the form '<tt>double
14980x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
14994.5e+15</tt>'. The only time hexadecimal floating point constants are required
1500(and the only time that they are generated by the disassembler) is when a
1501floating point constant must be emitted but it cannot be represented as a
1502decimal floating point number. For example, NaN's, infinities, and other
1503special values are represented in their IEEE hexadecimal format so that
1504assembly and disassembly do not cause any bits to change in the constants.</p>
1505
1506</div>
1507
1508<!-- ======================================================================= -->
1509<div class="doc_subsection"><a name="aggregateconstants">Aggregate Constants</a>
1510</div>
1511
1512<div class="doc_text">
1513<p>Aggregate constants arise from aggregation of simple constants
1514and smaller aggregate constants.</p>
1515
1516<dl>
1517 <dt><b>Structure constants</b></dt>
1518
1519 <dd>Structure constants are represented with notation similar to structure
1520 type definitions (a comma separated list of elements, surrounded by braces
Chris Lattner6c8de962007-12-25 20:34:52 +0000[diff] [blame]1521 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
1522 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]1523 must have <a href="#t_struct">structure type</a>, and the number and
1524 types of elements must match those specified by the type.
1525 </dd>
1526
1527 <dt><b>Array constants</b></dt>
1528
1529 <dd>Array constants are represented with notation similar to array type
1530 definitions (a comma separated list of elements, surrounded by square brackets
1531 (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74 ]</tt>". Array
1532 constants must have <a href="#t_array">array type</a>, and the number and
1533 types of elements must match those specified by the type.
1534 </dd>
1535
1536 <dt><b>Vector constants</b></dt>
1537
1538 <dd>Vector constants are represented with notation similar to vector type
1539 definitions (a comma separated list of elements, surrounded by
1540 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32 42,
1541 i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must have <a
1542 href="#t_vector">vector type</a>, and the number and types of elements must
1543 match those specified by the type.
1544 </dd>
1545
1546 <dt><b>Zero initialization</b></dt>
1547
1548 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1549 value to zero of <em>any</em> type, including scalar and aggregate types.
1550 This is often used to avoid having to print large zero initializers (e.g. for
1551 large arrays) and is always exactly equivalent to using explicit zero
1552 initializers.
1553 </dd>
1554</dl>
1555
1556</div>
1557
1558<!-- ======================================================================= -->
1559<div class="doc_subsection">
1560 <a name="globalconstants">Global Variable and Function Addresses</a>
1561</div>
1562
1563<div class="doc_text">
1564
1565<p>The addresses of <a href="#globalvars">global variables</a> and <a
1566href="#functionstructure">functions</a> are always implicitly valid (link-time)
1567constants. These constants are explicitly referenced when the <a
1568href="#identifiers">identifier for the global</a> is used and always have <a
1569href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1570file:</p>
1571
1572<div class="doc_code">
1573<pre>
1574@X = global i32 17
1575@Y = global i32 42
1576@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
1577</pre>
1578</div>
1579
1580</div>
1581
1582<!-- ======================================================================= -->
1583<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
1584<div class="doc_text">
1585 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
1586 no specific value. Undefined values may be of any type and be used anywhere
1587 a constant is permitted.</p>
1588
1589 <p>Undefined values indicate to the compiler that the program is well defined
1590 no matter what value is used, giving the compiler more freedom to optimize.
1591 </p>
1592</div>
1593
1594<!-- ======================================================================= -->
1595<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1596</div>
1597
1598<div class="doc_text">
1599
1600<p>Constant expressions are used to allow expressions involving other constants
1601to be used as constants. Constant expressions may be of any <a
1602href="#t_firstclass">first class</a> type and may involve any LLVM operation
1603that does not have side effects (e.g. load and call are not supported). The
1604following is the syntax for constant expressions:</p>
1605
1606<dl>
1607 <dt><b><tt>trunc ( CST to TYPE )</tt></b></dt>
1608 <dd>Truncate a constant to another type. The bit size of CST must be larger
1609 than the bit size of TYPE. Both types must be integers.</dd>
1610
1611 <dt><b><tt>zext ( CST to TYPE )</tt></b></dt>
1612 <dd>Zero extend a constant to another type. The bit size of CST must be
1613 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1614
1615 <dt><b><tt>sext ( CST to TYPE )</tt></b></dt>
1616 <dd>Sign extend a constant to another type. The bit size of CST must be
1617 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1618
1619 <dt><b><tt>fptrunc ( CST to TYPE )</tt></b></dt>
1620 <dd>Truncate a floating point constant to another floating point type. The
1621 size of CST must be larger than the size of TYPE. Both types must be
1622 floating point.</dd>
1623
1624 <dt><b><tt>fpext ( CST to TYPE )</tt></b></dt>
1625 <dd>Floating point extend a constant to another type. The size of CST must be
1626 smaller or equal to the size of TYPE. Both types must be floating point.</dd>
1627
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]1628 <dt><b><tt>fptoui ( CST to TYPE )</tt></b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1629 <dd>Convert a floating point constant to the corresponding unsigned integer
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1630 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1631 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1632 of the same number of elements. If the value won't fit in the integer type,
1633 the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1634
1635 <dt><b><tt>fptosi ( CST to TYPE )</tt></b></dt>
1636 <dd>Convert a floating point constant to the corresponding signed integer
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1637 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1638 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1639 of the same number of elements. If the value won't fit in the integer type,
1640 the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1641
1642 <dt><b><tt>uitofp ( CST to TYPE )</tt></b></dt>
1643 <dd>Convert an unsigned integer constant to the corresponding floating point
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1644 constant. TYPE must be a scalar or vector floating point type. CST must be of
1645 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1646 of the same number of elements. If the value won't fit in the floating point
1647 type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1648
1649 <dt><b><tt>sitofp ( CST to TYPE )</tt></b></dt>
1650 <dd>Convert a signed integer constant to the corresponding floating point
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1651 constant. TYPE must be a scalar or vector floating point type. CST must be of
1652 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1653 of the same number of elements. If the value won't fit in the floating point
1654 type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1655
1656 <dt><b><tt>ptrtoint ( CST to TYPE )</tt></b></dt>
1657 <dd>Convert a pointer typed constant to the corresponding integer constant
1658 TYPE must be an integer type. CST must be of pointer type. The CST value is
1659 zero extended, truncated, or unchanged to make it fit in TYPE.</dd>
1660
1661 <dt><b><tt>inttoptr ( CST to TYPE )</tt></b></dt>
1662 <dd>Convert a integer constant to a pointer constant. TYPE must be a
1663 pointer type. CST must be of integer type. The CST value is zero extended,
1664 truncated, or unchanged to make it fit in a pointer size. This one is
1665 <i>really</i> dangerous!</dd>
1666
1667 <dt><b><tt>bitcast ( CST to TYPE )</tt></b></dt>
1668 <dd>Convert a constant, CST, to another TYPE. The size of CST and TYPE must be
1669 identical (same number of bits). The conversion is done as if the CST value
1670 was stored to memory and read back as TYPE. In other words, no bits change
1671 with this operator, just the type. This can be used for conversion of
1672 vector types to any other type, as long as they have the same bit width. For
1673 pointers it is only valid to cast to another pointer type.
1674 </dd>
1675
1676 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
1677
1678 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
1679 constants. As with the <a href="#i_getelementptr">getelementptr</a>
1680 instruction, the index list may have zero or more indexes, which are required
1681 to make sense for the type of "CSTPTR".</dd>
1682
1683 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
1684
1685 <dd>Perform the <a href="#i_select">select operation</a> on
1686 constants.</dd>
1687
1688 <dt><b><tt>icmp COND ( VAL1, VAL2 )</tt></b></dt>
1689 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
1690
1691 <dt><b><tt>fcmp COND ( VAL1, VAL2 )</tt></b></dt>
1692 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
1693
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]1694 <dt><b><tt>vicmp COND ( VAL1, VAL2 )</tt></b></dt>
1695 <dd>Performs the <a href="#i_vicmp">vicmp operation</a> on constants.</dd>
1696
1697 <dt><b><tt>vfcmp COND ( VAL1, VAL2 )</tt></b></dt>
1698 <dd>Performs the <a href="#i_vfcmp">vfcmp operation</a> on constants.</dd>
1699
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1700 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
1701
1702 <dd>Perform the <a href="#i_extractelement">extractelement
1703 operation</a> on constants.
1704
1705 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
1706
1707 <dd>Perform the <a href="#i_insertelement">insertelement
1708 operation</a> on constants.</dd>
1709
1710
1711 <dt><b><tt>shufflevector ( VEC1, VEC2, IDXMASK )</tt></b></dt>
1712
1713 <dd>Perform the <a href="#i_shufflevector">shufflevector
1714 operation</a> on constants.</dd>
1715
1716 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
1717
1718 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
1719 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
1720 binary</a> operations. The constraints on operands are the same as those for
1721 the corresponding instruction (e.g. no bitwise operations on floating point
1722 values are allowed).</dd>
1723</dl>
1724</div>
1725
1726<!-- *********************************************************************** -->
1727<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
1728<!-- *********************************************************************** -->
1729
1730<!-- ======================================================================= -->
1731<div class="doc_subsection">
1732<a name="inlineasm">Inline Assembler Expressions</a>
1733</div>
1734
1735<div class="doc_text">
1736
1737<p>
1738LLVM supports inline assembler expressions (as opposed to <a href="#moduleasm">
1739Module-Level Inline Assembly</a>) through the use of a special value. This
1740value represents the inline assembler as a string (containing the instructions
1741to emit), a list of operand constraints (stored as a string), and a flag that
1742indicates whether or not the inline asm expression has side effects. An example
1743inline assembler expression is:
1744</p>
1745
1746<div class="doc_code">
1747<pre>
1748i32 (i32) asm "bswap $0", "=r,r"
1749</pre>
1750</div>
1751
1752<p>
1753Inline assembler expressions may <b>only</b> be used as the callee operand of
1754a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we have:
1755</p>
1756
1757<div class="doc_code">
1758<pre>
1759%X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
1760</pre>
1761</div>
1762
1763<p>
1764Inline asms with side effects not visible in the constraint list must be marked
1765as having side effects. This is done through the use of the
1766'<tt>sideeffect</tt>' keyword, like so:
1767</p>
1768
1769<div class="doc_code">
1770<pre>
1771call void asm sideeffect "eieio", ""()
1772</pre>
1773</div>
1774
1775<p>TODO: The format of the asm and constraints string still need to be
1776documented here. Constraints on what can be done (e.g. duplication, moving, etc
1777need to be documented).
1778</p>
1779
1780</div>
1781
1782<!-- *********************************************************************** -->
1783<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
1784<!-- *********************************************************************** -->
1785
1786<div class="doc_text">
1787
1788<p>The LLVM instruction set consists of several different
1789classifications of instructions: <a href="#terminators">terminator
1790instructions</a>, <a href="#binaryops">binary instructions</a>,
1791<a href="#bitwiseops">bitwise binary instructions</a>, <a
1792 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
1793instructions</a>.</p>
1794
1795</div>
1796
1797<!-- ======================================================================= -->
1798<div class="doc_subsection"> <a name="terminators">Terminator
1799Instructions</a> </div>
1800
1801<div class="doc_text">
1802
1803<p>As mentioned <a href="#functionstructure">previously</a>, every
1804basic block in a program ends with a "Terminator" instruction, which
1805indicates which block should be executed after the current block is
1806finished. These terminator instructions typically yield a '<tt>void</tt>'
1807value: they produce control flow, not values (the one exception being
1808the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
1809<p>There are six different terminator instructions: the '<a
1810 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
1811instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
1812the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
1813 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
1814 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
1815
1816</div>
1817
1818<!-- _______________________________________________________________________ -->
1819<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
1820Instruction</a> </div>
1821<div class="doc_text">
1822<h5>Syntax:</h5>
1823<pre> ret &lt;type&gt; &lt;value&gt; <i>; Return a value from a non-void function</i>
1824 ret void <i>; Return from void function</i>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1825 ret &lt;type&gt; &lt;value&gt;, &lt;type&gt; &lt;value&gt; <i>; Return two values from a non-void function </i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1826</pre>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1827
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1828<h5>Overview:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1829
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1830<p>The '<tt>ret</tt>' instruction is used to return control flow (and a
1831value) from a function back to the caller.</p>
1832<p>There are two forms of the '<tt>ret</tt>' instruction: one that
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1833returns value(s) and then causes control flow, and one that just causes
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1834control flow to occur.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1835
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1836<h5>Arguments:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1837
1838<p>The '<tt>ret</tt>' instruction may return zero, one or multiple values.
1839The type of each return value must be a '<a href="#t_firstclass">first
1840class</a>' type. Note that a function is not <a href="#wellformed">well
1841formed</a> if there exists a '<tt>ret</tt>' instruction inside of the
1842function that returns values that do not match the return type of the
1843function.</p>
1844
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1845<h5>Semantics:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1846
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1847<p>When the '<tt>ret</tt>' instruction is executed, control flow
1848returns back to the calling function's context. If the caller is a "<a
1849 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
1850the instruction after the call. If the caller was an "<a
1851 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
1852at the beginning of the "normal" destination block. If the instruction
1853returns a value, that value shall set the call or invoke instruction's
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1854return value. If the instruction returns multiple values then these
Devang Patelec8a5b02008-03-11 05:51:59 +0000[diff] [blame]1855values can only be accessed through a '<a href="#i_getresult"><tt>getresult</tt>
1856</a>' instruction.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1857
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1858<h5>Example:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1859
1860<pre>
1861 ret i32 5 <i>; Return an integer value of 5</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1862 ret void <i>; Return from a void function</i>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1863 ret i32 4, i8 2 <i>; Return two values 4 and 2 </i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1864</pre>
1865</div>
1866<!-- _______________________________________________________________________ -->
1867<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
1868<div class="doc_text">
1869<h5>Syntax:</h5>
1870<pre> br i1 &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;<br> br label &lt;dest&gt; <i>; Unconditional branch</i>
1871</pre>
1872<h5>Overview:</h5>
1873<p>The '<tt>br</tt>' instruction is used to cause control flow to
1874transfer to a different basic block in the current function. There are
1875two forms of this instruction, corresponding to a conditional branch
1876and an unconditional branch.</p>
1877<h5>Arguments:</h5>
1878<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
1879single '<tt>i1</tt>' value and two '<tt>label</tt>' values. The
1880unconditional form of the '<tt>br</tt>' instruction takes a single
1881'<tt>label</tt>' value as a target.</p>
1882<h5>Semantics:</h5>
1883<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
1884argument is evaluated. If the value is <tt>true</tt>, control flows
1885to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
1886control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
1887<h5>Example:</h5>
1888<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
1889 href="#i_ret">ret</a> i32 1<br>IfUnequal:<br> <a href="#i_ret">ret</a> i32 0<br></pre>
1890</div>
1891<!-- _______________________________________________________________________ -->
1892<div class="doc_subsubsection">
1893 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
1894</div>
1895
1896<div class="doc_text">
1897<h5>Syntax:</h5>
1898
1899<pre>
1900 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
1901</pre>
1902
1903<h5>Overview:</h5>
1904
1905<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
1906several different places. It is a generalization of the '<tt>br</tt>'
1907instruction, allowing a branch to occur to one of many possible
1908destinations.</p>
1909
1910
1911<h5>Arguments:</h5>
1912
1913<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
1914comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
1915an array of pairs of comparison value constants and '<tt>label</tt>'s. The
1916table is not allowed to contain duplicate constant entries.</p>
1917
1918<h5>Semantics:</h5>
1919
1920<p>The <tt>switch</tt> instruction specifies a table of values and
1921destinations. When the '<tt>switch</tt>' instruction is executed, this
1922table is searched for the given value. If the value is found, control flow is
1923transfered to the corresponding destination; otherwise, control flow is
1924transfered to the default destination.</p>
1925
1926<h5>Implementation:</h5>
1927
1928<p>Depending on properties of the target machine and the particular
1929<tt>switch</tt> instruction, this instruction may be code generated in different
1930ways. For example, it could be generated as a series of chained conditional
1931branches or with a lookup table.</p>
1932
1933<h5>Example:</h5>
1934
1935<pre>
1936 <i>; Emulate a conditional br instruction</i>
1937 %Val = <a href="#i_zext">zext</a> i1 %value to i32
1938 switch i32 %Val, label %truedest [i32 0, label %falsedest ]
1939
1940 <i>; Emulate an unconditional br instruction</i>
1941 switch i32 0, label %dest [ ]
1942
1943 <i>; Implement a jump table:</i>
1944 switch i32 %val, label %otherwise [ i32 0, label %onzero
1945 i32 1, label %onone
1946 i32 2, label %ontwo ]
1947</pre>
1948</div>
1949
1950<!-- _______________________________________________________________________ -->
1951<div class="doc_subsubsection">
1952 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
1953</div>
1954
1955<div class="doc_text">
1956
1957<h5>Syntax:</h5>
1958
1959<pre>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]1960 &lt;result&gt; = invoke [<a href="#callingconv">cconv</a>] &lt;ptr to function ty&gt; &lt;function ptr val&gt;(&lt;function args&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1961 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
1962</pre>
1963
1964<h5>Overview:</h5>
1965
1966<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
1967function, with the possibility of control flow transfer to either the
1968'<tt>normal</tt>' label or the
1969'<tt>exception</tt>' label. If the callee function returns with the
1970"<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
1971"normal" label. If the callee (or any indirect callees) returns with the "<a
1972href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1973continued at the dynamically nearest "exception" label. If the callee function
Devang Patelec8a5b02008-03-11 05:51:59 +0000[diff] [blame]1974returns multiple values then individual return values are only accessible through
1975a '<tt><a href="#i_getresult">getresult</a></tt>' instruction.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1976
1977<h5>Arguments:</h5>
1978
1979<p>This instruction requires several arguments:</p>
1980
1981<ol>
1982 <li>
1983 The optional "cconv" marker indicates which <a href="#callingconv">calling
1984 convention</a> the call should use. If none is specified, the call defaults
1985 to using C calling conventions.
1986 </li>
1987 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
1988 function value being invoked. In most cases, this is a direct function
1989 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
1990 an arbitrary pointer to function value.
1991 </li>
1992
1993 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
1994 function to be invoked. </li>
1995
1996 <li>'<tt>function args</tt>': argument list whose types match the function
1997 signature argument types. If the function signature indicates the function
1998 accepts a variable number of arguments, the extra arguments can be
1999 specified. </li>
2000
2001 <li>'<tt>normal label</tt>': the label reached when the called function
2002 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
2003
2004 <li>'<tt>exception label</tt>': the label reached when a callee returns with
2005 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
2006
2007</ol>
2008
2009<h5>Semantics:</h5>
2010
2011<p>This instruction is designed to operate as a standard '<tt><a
2012href="#i_call">call</a></tt>' instruction in most regards. The primary
2013difference is that it establishes an association with a label, which is used by
2014the runtime library to unwind the stack.</p>
2015
2016<p>This instruction is used in languages with destructors to ensure that proper
2017cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
2018exception. Additionally, this is important for implementation of
2019'<tt>catch</tt>' clauses in high-level languages that support them.</p>
2020
2021<h5>Example:</h5>
2022<pre>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2023 %retval = invoke i32 @Test(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2024 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2025 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2026 unwind label %TestCleanup <i>; {i32}:retval set</i>
2027</pre>
2028</div>
2029
2030
2031<!-- _______________________________________________________________________ -->
2032
2033<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
2034Instruction</a> </div>
2035
2036<div class="doc_text">
2037
2038<h5>Syntax:</h5>
2039<pre>
2040 unwind
2041</pre>
2042
2043<h5>Overview:</h5>
2044
2045<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
2046at the first callee in the dynamic call stack which used an <a
2047href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
2048primarily used to implement exception handling.</p>
2049
2050<h5>Semantics:</h5>
2051
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]2052<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2053immediately halt. The dynamic call stack is then searched for the first <a
2054href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
2055execution continues at the "exceptional" destination block specified by the
2056<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
2057dynamic call chain, undefined behavior results.</p>
2058</div>
2059
2060<!-- _______________________________________________________________________ -->
2061
2062<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
2063Instruction</a> </div>
2064
2065<div class="doc_text">
2066
2067<h5>Syntax:</h5>
2068<pre>
2069 unreachable
2070</pre>
2071
2072<h5>Overview:</h5>
2073
2074<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
2075instruction is used to inform the optimizer that a particular portion of the
2076code is not reachable. This can be used to indicate that the code after a
2077no-return function cannot be reached, and other facts.</p>
2078
2079<h5>Semantics:</h5>
2080
2081<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
2082</div>
2083
2084
2085
2086<!-- ======================================================================= -->
2087<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
2088<div class="doc_text">
2089<p>Binary operators are used to do most of the computation in a
Chris Lattnerab596d92008-04-01 18:47:32 +0000[diff] [blame]2090program. They require two operands of the same type, execute an operation on them, and
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2091produce a single value. The operands might represent
2092multiple 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]2093The result value has the same type as its operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2094<p>There are several different binary operators:</p>
2095</div>
2096<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2097<div class="doc_subsubsection">
2098 <a name="i_add">'<tt>add</tt>' Instruction</a>
2099</div>
2100
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2101<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2102
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2103<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2104
2105<pre>
2106 &lt;result&gt; = add &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2107</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2108
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2109<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2110
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2111<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2112
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2113<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2114
2115<p>The two arguments to the '<tt>add</tt>' instruction must be <a
2116 href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>, or
2117 <a href="#t_vector">vector</a> values. Both arguments must have identical
2118 types.</p>
2119
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2120<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2121
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2122<p>The value produced is the integer or floating point sum of the two
2123operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2124
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2125<p>If an integer sum has unsigned overflow, the result returned is the
2126mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2127the result.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2128
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2129<p>Because LLVM integers use a two's complement representation, this
2130instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2131
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2132<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2133
2134<pre>
2135 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2136</pre>
2137</div>
2138<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2139<div class="doc_subsubsection">
2140 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
2141</div>
2142
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2143<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2144
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2145<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2146
2147<pre>
2148 &lt;result&gt; = sub &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2149</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2150
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2151<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2152
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2153<p>The '<tt>sub</tt>' instruction returns the difference of its two
2154operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2155
2156<p>Note that the '<tt>sub</tt>' instruction is used to represent the
2157'<tt>neg</tt>' instruction present in most other intermediate
2158representations.</p>
2159
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2160<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2161
2162<p>The two arguments to the '<tt>sub</tt>' instruction must be <a
2163 href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>,
2164 or <a href="#t_vector">vector</a> values. Both arguments must have identical
2165 types.</p>
2166
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2167<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2168
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2169<p>The value produced is the integer or floating point difference of
2170the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2171
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2172<p>If an integer difference has unsigned overflow, the result returned is the
2173mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2174the result.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2175
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2176<p>Because LLVM integers use a two's complement representation, this
2177instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2178
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2179<h5>Example:</h5>
2180<pre>
2181 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
2182 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
2183</pre>
2184</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2185
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2186<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2187<div class="doc_subsubsection">
2188 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
2189</div>
2190
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2191<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2192
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2193<h5>Syntax:</h5>
2194<pre> &lt;result&gt; = mul &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2195</pre>
2196<h5>Overview:</h5>
2197<p>The '<tt>mul</tt>' instruction returns the product of its two
2198operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2199
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2200<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2201
2202<p>The two arguments to the '<tt>mul</tt>' instruction must be <a
2203href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>,
2204or <a href="#t_vector">vector</a> values. Both arguments must have identical
2205types.</p>
2206
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2207<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2208
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2209<p>The value produced is the integer or floating point product of the
2210two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2211
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2212<p>If the result of an integer multiplication has unsigned overflow,
2213the result returned is the mathematical result modulo
22142<sup>n</sup>, where n is the bit width of the result.</p>
2215<p>Because LLVM integers use a two's complement representation, and the
2216result is the same width as the operands, this instruction returns the
2217correct result for both signed and unsigned integers. If a full product
2218(e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands
2219should be sign-extended or zero-extended as appropriate to the
2220width of the full product.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2221<h5>Example:</h5>
2222<pre> &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
2223</pre>
2224</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2225
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2226<!-- _______________________________________________________________________ -->
2227<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
2228</a></div>
2229<div class="doc_text">
2230<h5>Syntax:</h5>
2231<pre> &lt;result&gt; = udiv &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2232</pre>
2233<h5>Overview:</h5>
2234<p>The '<tt>udiv</tt>' instruction returns the quotient of its two
2235operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2236
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2237<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2238
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2239<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2240<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2241values. Both arguments must have identical types.</p>
2242
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2243<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2244
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2245<p>The value produced is the unsigned integer quotient of the two operands.</p>
2246<p>Note that unsigned integer division and signed integer division are distinct
2247operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
2248<p>Division by zero leads to undefined behavior.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2249<h5>Example:</h5>
2250<pre> &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2251</pre>
2252</div>
2253<!-- _______________________________________________________________________ -->
2254<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
2255</a> </div>
2256<div class="doc_text">
2257<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2258<pre>
2259 &lt;result&gt; = sdiv &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2260</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2261
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2262<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2263
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2264<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two
2265operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2266
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2267<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2268
2269<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
2270<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2271values. Both arguments must have identical types.</p>
2272
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2273<h5>Semantics:</h5>
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2274<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]2275<p>Note that signed integer division and unsigned integer division are distinct
2276operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
2277<p>Division by zero leads to undefined behavior. Overflow also leads to
2278undefined behavior; this is a rare case, but can occur, for example,
2279by doing a 32-bit division of -2147483648 by -1.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2280<h5>Example:</h5>
2281<pre> &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2282</pre>
2283</div>
2284<!-- _______________________________________________________________________ -->
2285<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
2286Instruction</a> </div>
2287<div class="doc_text">
2288<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2289<pre>
2290 &lt;result&gt; = fdiv &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2291</pre>
2292<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2293
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2294<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two
2295operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2296
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2297<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2298
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2299<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2300<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2301of floating point values. Both arguments must have identical types.</p>
2302
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2303<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2304
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2305<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2306
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2307<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2308
2309<pre>
2310 &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]2311</pre>
2312</div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2313
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2314<!-- _______________________________________________________________________ -->
2315<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
2316</div>
2317<div class="doc_text">
2318<h5>Syntax:</h5>
2319<pre> &lt;result&gt; = urem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2320</pre>
2321<h5>Overview:</h5>
2322<p>The '<tt>urem</tt>' instruction returns the remainder from the
2323unsigned division of its two arguments.</p>
2324<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2325<p>The two arguments to the '<tt>urem</tt>' instruction must be
2326<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2327values. Both arguments must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2328<h5>Semantics:</h5>
2329<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2330This instruction always performs an unsigned division to get the remainder.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2331<p>Note that unsigned integer remainder and signed integer remainder are
2332distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
2333<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2334<h5>Example:</h5>
2335<pre> &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2336</pre>
2337
2338</div>
2339<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2340<div class="doc_subsubsection">
2341 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
2342</div>
2343
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2344<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2345
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2346<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2347
2348<pre>
2349 &lt;result&gt; = srem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2350</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2351
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2352<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2353
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2354<p>The '<tt>srem</tt>' instruction returns the remainder from the
Dan Gohman3e3fd8c2007-11-05 23:35:22 +0000[diff] [blame]2355signed division of its two operands. This instruction can also take
2356<a href="#t_vector">vector</a> versions of the values in which case
2357the elements must be integers.</p>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]2358
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2359<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2360
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2361<p>The two arguments to the '<tt>srem</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2362<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2363values. Both arguments must have identical types.</p>
2364
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2365<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2366
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2367<p>This instruction returns the <i>remainder</i> of a division (where the result
2368has the same sign as the dividend, <tt>var1</tt>), not the <i>modulo</i>
2369operator (where the result has the same sign as the divisor, <tt>var2</tt>) of
2370a value. For more information about the difference, see <a
2371 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
2372Math Forum</a>. For a table of how this is implemented in various languages,
2373please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
2374Wikipedia: modulo operation</a>.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2375<p>Note that signed integer remainder and unsigned integer remainder are
2376distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
2377<p>Taking the remainder of a division by zero leads to undefined behavior.
2378Overflow also leads to undefined behavior; this is a rare case, but can occur,
2379for example, by taking the remainder of a 32-bit division of -2147483648 by -1.
2380(The remainder doesn't actually overflow, but this rule lets srem be
2381implemented using instructions that return both the result of the division
2382and the remainder.)</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2383<h5>Example:</h5>
2384<pre> &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2385</pre>
2386
2387</div>
2388<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2389<div class="doc_subsubsection">
2390 <a name="i_frem">'<tt>frem</tt>' Instruction</a> </div>
2391
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2392<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2393
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2394<h5>Syntax:</h5>
2395<pre> &lt;result&gt; = frem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2396</pre>
2397<h5>Overview:</h5>
2398<p>The '<tt>frem</tt>' instruction returns the remainder from the
2399division of its two operands.</p>
2400<h5>Arguments:</h5>
2401<p>The two arguments to the '<tt>frem</tt>' instruction must be
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2402<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2403of floating point values. Both arguments must have identical types.</p>
2404
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2405<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2406
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2407<p>This instruction returns the <i>remainder</i> of a division.
2408The remainder has the same sign as the dividend.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2409
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2410<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2411
2412<pre>
2413 &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]2414</pre>
2415</div>
2416
2417<!-- ======================================================================= -->
2418<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
2419Operations</a> </div>
2420<div class="doc_text">
2421<p>Bitwise binary operators are used to do various forms of
2422bit-twiddling in a program. They are generally very efficient
2423instructions and can commonly be strength reduced from other
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2424instructions. They require two operands of the same type, execute an operation on them,
2425and 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]2426</div>
2427
2428<!-- _______________________________________________________________________ -->
2429<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
2430Instruction</a> </div>
2431<div class="doc_text">
2432<h5>Syntax:</h5>
2433<pre> &lt;result&gt; = shl &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2434</pre>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2435
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2436<h5>Overview:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2437
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2438<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
2439the left a specified number of bits.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2440
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2441<h5>Arguments:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2442
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2443<p>Both arguments to the '<tt>shl</tt>' instruction must be the same <a
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]2444 href="#t_integer">integer</a> type. '<tt>var2</tt>' is treated as an
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2445unsigned value. This instruction does not support
2446<a href="#t_vector">vector</a> operands.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2447
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2448<h5>Semantics:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2449
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2450<p>The value produced is <tt>var1</tt> * 2<sup><tt>var2</tt></sup> mod 2<sup>n</sup>,
2451where n is the width of the result. If <tt>var2</tt> is (statically or dynamically) negative or
2452equal to or larger than the number of bits in <tt>var1</tt>, the result is undefined.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2453
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2454<h5>Example:</h5><pre>
2455 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
2456 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
2457 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2458 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2459</pre>
2460</div>
2461<!-- _______________________________________________________________________ -->
2462<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
2463Instruction</a> </div>
2464<div class="doc_text">
2465<h5>Syntax:</h5>
2466<pre> &lt;result&gt; = lshr &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2467</pre>
2468
2469<h5>Overview:</h5>
2470<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
2471operand shifted to the right a specified number of bits with zero fill.</p>
2472
2473<h5>Arguments:</h5>
2474<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]2475<a href="#t_integer">integer</a> type. '<tt>var2</tt>' is treated as an
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2476unsigned value. This instruction does not support
2477<a href="#t_vector">vector</a> operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2478
2479<h5>Semantics:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2480
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2481<p>This instruction always performs a logical shift right operation. The most
2482significant bits of the result will be filled with zero bits after the
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2483shift. If <tt>var2</tt> is (statically or dynamically) equal to or larger than
2484the number of bits in <tt>var1</tt>, the result is undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2485
2486<h5>Example:</h5>
2487<pre>
2488 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
2489 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
2490 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
2491 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2492 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2493</pre>
2494</div>
2495
2496<!-- _______________________________________________________________________ -->
2497<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
2498Instruction</a> </div>
2499<div class="doc_text">
2500
2501<h5>Syntax:</h5>
2502<pre> &lt;result&gt; = ashr &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2503</pre>
2504
2505<h5>Overview:</h5>
2506<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
2507operand shifted to the right a specified number of bits with sign extension.</p>
2508
2509<h5>Arguments:</h5>
2510<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]2511<a href="#t_integer">integer</a> type. '<tt>var2</tt>' is treated as an
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2512unsigned value. This instruction does not support
2513<a href="#t_vector">vector</a> operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2514
2515<h5>Semantics:</h5>
2516<p>This instruction always performs an arithmetic shift right operation,
2517The most significant bits of the result will be filled with the sign bit
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2518of <tt>var1</tt>. If <tt>var2</tt> is (statically or dynamically) equal to or
2519larger than the number of bits in <tt>var1</tt>, the result is undefined.
2520</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2521
2522<h5>Example:</h5>
2523<pre>
2524 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
2525 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
2526 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
2527 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2528 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2529</pre>
2530</div>
2531
2532<!-- _______________________________________________________________________ -->
2533<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
2534Instruction</a> </div>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2535
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2536<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2537
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2538<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2539
2540<pre>
2541 &lt;result&gt; = and &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2542</pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2543
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2544<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2545
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2546<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
2547its two operands.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2548
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2549<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2550
2551<p>The two arguments to the '<tt>and</tt>' instruction must be
2552<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2553values. Both arguments must have identical types.</p>
2554
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2555<h5>Semantics:</h5>
2556<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
2557<p> </p>
2558<div style="align: center">
2559<table border="1" cellspacing="0" cellpadding="4">
2560 <tbody>
2561 <tr>
2562 <td>In0</td>
2563 <td>In1</td>
2564 <td>Out</td>
2565 </tr>
2566 <tr>
2567 <td>0</td>
2568 <td>0</td>
2569 <td>0</td>
2570 </tr>
2571 <tr>
2572 <td>0</td>
2573 <td>1</td>
2574 <td>0</td>
2575 </tr>
2576 <tr>
2577 <td>1</td>
2578 <td>0</td>
2579 <td>0</td>
2580 </tr>
2581 <tr>
2582 <td>1</td>
2583 <td>1</td>
2584 <td>1</td>
2585 </tr>
2586 </tbody>
2587</table>
2588</div>
2589<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2590<pre>
2591 &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]2592 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
2593 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
2594</pre>
2595</div>
2596<!-- _______________________________________________________________________ -->
2597<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
2598<div class="doc_text">
2599<h5>Syntax:</h5>
2600<pre> &lt;result&gt; = or &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2601</pre>
2602<h5>Overview:</h5>
2603<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
2604or of its two operands.</p>
2605<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2606
2607<p>The two arguments to the '<tt>or</tt>' instruction must be
2608<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2609values. Both arguments must have identical types.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2610<h5>Semantics:</h5>
2611<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
2612<p> </p>
2613<div style="align: center">
2614<table border="1" cellspacing="0" cellpadding="4">
2615 <tbody>
2616 <tr>
2617 <td>In0</td>
2618 <td>In1</td>
2619 <td>Out</td>
2620 </tr>
2621 <tr>
2622 <td>0</td>
2623 <td>0</td>
2624 <td>0</td>
2625 </tr>
2626 <tr>
2627 <td>0</td>
2628 <td>1</td>
2629 <td>1</td>
2630 </tr>
2631 <tr>
2632 <td>1</td>
2633 <td>0</td>
2634 <td>1</td>
2635 </tr>
2636 <tr>
2637 <td>1</td>
2638 <td>1</td>
2639 <td>1</td>
2640 </tr>
2641 </tbody>
2642</table>
2643</div>
2644<h5>Example:</h5>
2645<pre> &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
2646 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
2647 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
2648</pre>
2649</div>
2650<!-- _______________________________________________________________________ -->
2651<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
2652Instruction</a> </div>
2653<div class="doc_text">
2654<h5>Syntax:</h5>
2655<pre> &lt;result&gt; = xor &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2656</pre>
2657<h5>Overview:</h5>
2658<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
2659or of its two operands. The <tt>xor</tt> is used to implement the
2660"one's complement" operation, which is the "~" operator in C.</p>
2661<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2662<p>The two arguments to the '<tt>xor</tt>' instruction must be
2663<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2664values. Both arguments must have identical types.</p>
2665
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2666<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]2667
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2668<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
2669<p> </p>
2670<div style="align: center">
2671<table border="1" cellspacing="0" cellpadding="4">
2672 <tbody>
2673 <tr>
2674 <td>In0</td>
2675 <td>In1</td>
2676 <td>Out</td>
2677 </tr>
2678 <tr>
2679 <td>0</td>
2680 <td>0</td>
2681 <td>0</td>
2682 </tr>
2683 <tr>
2684 <td>0</td>
2685 <td>1</td>
2686 <td>1</td>
2687 </tr>
2688 <tr>
2689 <td>1</td>
2690 <td>0</td>
2691 <td>1</td>
2692 </tr>
2693 <tr>
2694 <td>1</td>
2695 <td>1</td>
2696 <td>0</td>
2697 </tr>
2698 </tbody>
2699</table>
2700</div>
2701<p> </p>
2702<h5>Example:</h5>
2703<pre> &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
2704 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
2705 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
2706 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
2707</pre>
2708</div>
2709
2710<!-- ======================================================================= -->
2711<div class="doc_subsection">
2712 <a name="vectorops">Vector Operations</a>
2713</div>
2714
2715<div class="doc_text">
2716
2717<p>LLVM supports several instructions to represent vector operations in a
2718target-independent manner. These instructions cover the element-access and
2719vector-specific operations needed to process vectors effectively. While LLVM
2720does directly support these vector operations, many sophisticated algorithms
2721will want to use target-specific intrinsics to take full advantage of a specific
2722target.</p>
2723
2724</div>
2725
2726<!-- _______________________________________________________________________ -->
2727<div class="doc_subsubsection">
2728 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
2729</div>
2730
2731<div class="doc_text">
2732
2733<h5>Syntax:</h5>
2734
2735<pre>
2736 &lt;result&gt; = extractelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, i32 &lt;idx&gt; <i>; yields &lt;ty&gt;</i>
2737</pre>
2738
2739<h5>Overview:</h5>
2740
2741<p>
2742The '<tt>extractelement</tt>' instruction extracts a single scalar
2743element from a vector at a specified index.
2744</p>
2745
2746
2747<h5>Arguments:</h5>
2748
2749<p>
2750The first operand of an '<tt>extractelement</tt>' instruction is a
2751value of <a href="#t_vector">vector</a> type. The second operand is
2752an index indicating the position from which to extract the element.
2753The index may be a variable.</p>
2754
2755<h5>Semantics:</h5>
2756
2757<p>
2758The result is a scalar of the same type as the element type of
2759<tt>val</tt>. Its value is the value at position <tt>idx</tt> of
2760<tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
2761results are undefined.
2762</p>
2763
2764<h5>Example:</h5>
2765
2766<pre>
2767 %result = extractelement &lt;4 x i32&gt; %vec, i32 0 <i>; yields i32</i>
2768</pre>
2769</div>
2770
2771
2772<!-- _______________________________________________________________________ -->
2773<div class="doc_subsubsection">
2774 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
2775</div>
2776
2777<div class="doc_text">
2778
2779<h5>Syntax:</h5>
2780
2781<pre>
Dan Gohmanbcc3c502008-05-12 23:38:42 +0000[diff] [blame]2782 &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]2783</pre>
2784
2785<h5>Overview:</h5>
2786
2787<p>
2788The '<tt>insertelement</tt>' instruction inserts a scalar
2789element into a vector at a specified index.
2790</p>
2791
2792
2793<h5>Arguments:</h5>
2794
2795<p>
2796The first operand of an '<tt>insertelement</tt>' instruction is a
2797value of <a href="#t_vector">vector</a> type. The second operand is a
2798scalar value whose type must equal the element type of the first
2799operand. The third operand is an index indicating the position at
2800which to insert the value. The index may be a variable.</p>
2801
2802<h5>Semantics:</h5>
2803
2804<p>
2805The result is a vector of the same type as <tt>val</tt>. Its
2806element values are those of <tt>val</tt> except at position
2807<tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
2808exceeds the length of <tt>val</tt>, the results are undefined.
2809</p>
2810
2811<h5>Example:</h5>
2812
2813<pre>
2814 %result = insertelement &lt;4 x i32&gt; %vec, i32 1, i32 0 <i>; yields &lt;4 x i32&gt;</i>
2815</pre>
2816</div>
2817
2818<!-- _______________________________________________________________________ -->
2819<div class="doc_subsubsection">
2820 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
2821</div>
2822
2823<div class="doc_text">
2824
2825<h5>Syntax:</h5>
2826
2827<pre>
2828 &lt;result&gt; = shufflevector &lt;n x &lt;ty&gt;&gt; &lt;v1&gt;, &lt;n x &lt;ty&gt;&gt; &lt;v2&gt;, &lt;n x i32&gt; &lt;mask&gt; <i>; yields &lt;n x &lt;ty&gt;&gt;</i>
2829</pre>
2830
2831<h5>Overview:</h5>
2832
2833<p>
2834The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
2835from two input vectors, returning a vector of the same type.
2836</p>
2837
2838<h5>Arguments:</h5>
2839
2840<p>
2841The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
2842with types that match each other and types that match the result of the
2843instruction. The third argument is a shuffle mask, which has the same number
2844of elements as the other vector type, but whose element type is always 'i32'.
2845</p>
2846
2847<p>
2848The shuffle mask operand is required to be a constant vector with either
2849constant integer or undef values.
2850</p>
2851
2852<h5>Semantics:</h5>
2853
2854<p>
2855The elements of the two input vectors are numbered from left to right across
2856both of the vectors. The shuffle mask operand specifies, for each element of
2857the result vector, which element of the two input registers the result element
2858gets. The element selector may be undef (meaning "don't care") and the second
2859operand may be undef if performing a shuffle from only one vector.
2860</p>
2861
2862<h5>Example:</h5>
2863
2864<pre>
2865 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
2866 &lt;4 x i32&gt; &lt;i32 0, i32 4, i32 1, i32 5&gt; <i>; yields &lt;4 x i32&gt;</i>
2867 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; undef,
2868 &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.
2869</pre>
2870</div>
2871
2872
2873<!-- ======================================================================= -->
2874<div class="doc_subsection">
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]2875 <a name="aggregateops">Aggregate Operations</a>
2876</div>
2877
2878<div class="doc_text">
2879
2880<p>LLVM supports several instructions for working with aggregate values.
2881</p>
2882
2883</div>
2884
2885<!-- _______________________________________________________________________ -->
2886<div class="doc_subsubsection">
2887 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
2888</div>
2889
2890<div class="doc_text">
2891
2892<h5>Syntax:</h5>
2893
2894<pre>
2895 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
2896</pre>
2897
2898<h5>Overview:</h5>
2899
2900<p>
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]2901The '<tt>extractvalue</tt>' instruction extracts the value of a struct field
2902or array element from an aggregate value.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]2903</p>
2904
2905
2906<h5>Arguments:</h5>
2907
2908<p>
2909The first operand of an '<tt>extractvalue</tt>' instruction is a
2910value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a>
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]2911type. The operands are constant indices to specify which value to extract
2912in the same manner as indices in a
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]2913'<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
2914</p>
2915
2916<h5>Semantics:</h5>
2917
2918<p>
2919The result is the value at the position in the aggregate specified by
2920the index operands.
2921</p>
2922
2923<h5>Example:</h5>
2924
2925<pre>
2926 %result = extractvalue {i32, float} %agg, i32 0 <i>; yields i32</i>
2927</pre>
2928</div>
2929
2930
2931<!-- _______________________________________________________________________ -->
2932<div class="doc_subsubsection">
2933 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
2934</div>
2935
2936<div class="doc_text">
2937
2938<h5>Syntax:</h5>
2939
2940<pre>
2941 &lt;result&gt; = insertvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;ty&gt; &lt;val&gt;, i32 &lt;idx&gt; <i>; yields &lt;n x &lt;ty&gt;&gt;</i>
2942</pre>
2943
2944<h5>Overview:</h5>
2945
2946<p>
2947The '<tt>insertvalue</tt>' instruction inserts a value
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]2948into a struct field or array element in an aggregate.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]2949</p>
2950
2951
2952<h5>Arguments:</h5>
2953
2954<p>
2955The first operand of an '<tt>insertvalue</tt>' instruction is a
2956value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a> type.
2957The second operand is a first-class value to insert.
Dan Gohman4f29e422008-05-23 21:53:15 +0000[diff] [blame^]2958The following operands are constant indices
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]2959indicating the position at which to insert the value in the same manner as
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]2960indices in a
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]2961'<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
2962The value to insert must have the same type as the value identified
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]2963by the indices.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]2964
2965<h5>Semantics:</h5>
2966
2967<p>
2968The result is an aggregate of the same type as <tt>val</tt>. Its
2969value is that of <tt>val</tt> except that the value at the position
Dan Gohman6c6dea02008-05-13 18:16:06 +0000[diff] [blame]2970specified by the indices is that of <tt>elt</tt>.
Dan Gohman74d6faf2008-05-12 23:51:09 +0000[diff] [blame]2971</p>
2972
2973<h5>Example:</h5>
2974
2975<pre>
2976 %result = insertvalue {i32, float} %agg, i32 1, i32 0 <i>; yields {i32, float}</i>
2977</pre>
2978</div>
2979
2980
2981<!-- ======================================================================= -->
2982<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2983 <a name="memoryops">Memory Access and Addressing Operations</a>
2984</div>
2985
2986<div class="doc_text">
2987
2988<p>A key design point of an SSA-based representation is how it
2989represents memory. In LLVM, no memory locations are in SSA form, which
2990makes things very simple. This section describes how to read, write,
2991allocate, and free memory in LLVM.</p>
2992
2993</div>
2994
2995<!-- _______________________________________________________________________ -->
2996<div class="doc_subsubsection">
2997 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
2998</div>
2999
3000<div class="doc_text">
3001
3002<h5>Syntax:</h5>
3003
3004<pre>
3005 &lt;result&gt; = malloc &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
3006</pre>
3007
3008<h5>Overview:</h5>
3009
3010<p>The '<tt>malloc</tt>' instruction allocates memory from the system
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]3011heap and returns a pointer to it. The object is always allocated in the generic
3012address space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3013
3014<h5>Arguments:</h5>
3015
3016<p>The '<tt>malloc</tt>' instruction allocates
3017<tt>sizeof(&lt;type&gt;)*NumElements</tt>
3018bytes of memory from the operating system and returns a pointer of the
3019appropriate type to the program. If "NumElements" is specified, it is the
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3020number of elements allocated, otherwise "NumElements" is defaulted to be one.
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3021If a constant alignment is specified, the value result of the allocation is guaranteed to
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3022be aligned to at least that boundary. If not specified, or if zero, the target can
3023choose to align the allocation on any convenient boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3024
3025<p>'<tt>type</tt>' must be a sized type.</p>
3026
3027<h5>Semantics:</h5>
3028
3029<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]3030a pointer is returned. The result of a zero byte allocattion is undefined. The
3031result is null if there is insufficient memory available.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3032
3033<h5>Example:</h5>
3034
3035<pre>
3036 %array = malloc [4 x i8 ] <i>; yields {[%4 x i8]*}:array</i>
3037
3038 %size = <a href="#i_add">add</a> i32 2, 2 <i>; yields {i32}:size = i32 4</i>
3039 %array1 = malloc i8, i32 4 <i>; yields {i8*}:array1</i>
3040 %array2 = malloc [12 x i8], i32 %size <i>; yields {[12 x i8]*}:array2</i>
3041 %array3 = malloc i32, i32 4, align 1024 <i>; yields {i32*}:array3</i>
3042 %array4 = malloc i32, align 1024 <i>; yields {i32*}:array4</i>
3043</pre>
3044</div>
3045
3046<!-- _______________________________________________________________________ -->
3047<div class="doc_subsubsection">
3048 <a name="i_free">'<tt>free</tt>' Instruction</a>
3049</div>
3050
3051<div class="doc_text">
3052
3053<h5>Syntax:</h5>
3054
3055<pre>
3056 free &lt;type&gt; &lt;value&gt; <i>; yields {void}</i>
3057</pre>
3058
3059<h5>Overview:</h5>
3060
3061<p>The '<tt>free</tt>' instruction returns memory back to the unused
3062memory heap to be reallocated in the future.</p>
3063
3064<h5>Arguments:</h5>
3065
3066<p>'<tt>value</tt>' shall be a pointer value that points to a value
3067that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
3068instruction.</p>
3069
3070<h5>Semantics:</h5>
3071
3072<p>Access to the memory pointed to by the pointer is no longer defined
Chris Lattner329d6532008-04-19 22:41:32 +0000[diff] [blame]3073after this instruction executes. If the pointer is null, the operation
3074is a noop.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3075
3076<h5>Example:</h5>
3077
3078<pre>
3079 %array = <a href="#i_malloc">malloc</a> [4 x i8] <i>; yields {[4 x i8]*}:array</i>
3080 free [4 x i8]* %array
3081</pre>
3082</div>
3083
3084<!-- _______________________________________________________________________ -->
3085<div class="doc_subsubsection">
3086 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
3087</div>
3088
3089<div class="doc_text">
3090
3091<h5>Syntax:</h5>
3092
3093<pre>
3094 &lt;result&gt; = alloca &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
3095</pre>
3096
3097<h5>Overview:</h5>
3098
3099<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
3100currently executing function, to be automatically released when this function
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]3101returns to its caller. The object is always allocated in the generic address
3102space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3103
3104<h5>Arguments:</h5>
3105
3106<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
3107bytes of memory on the runtime stack, returning a pointer of the
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3108appropriate type to the program. If "NumElements" is specified, it is the
3109number of elements allocated, otherwise "NumElements" is defaulted to be one.
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3110If a constant alignment is specified, the value result of the allocation is guaranteed
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]3111to be aligned to at least that boundary. If not specified, or if zero, the target
3112can choose to align the allocation on any convenient boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3113
3114<p>'<tt>type</tt>' may be any sized type.</p>
3115
3116<h5>Semantics:</h5>
3117
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]3118<p>Memory is allocated; a pointer is returned. The operation is undefiend if
3119there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3120memory is automatically released when the function returns. The '<tt>alloca</tt>'
3121instruction is commonly used to represent automatic variables that must
3122have an address available. When the function returns (either with the <tt><a
3123 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]3124instructions), the memory is reclaimed. Allocating zero bytes
3125is legal, but the result is undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3126
3127<h5>Example:</h5>
3128
3129<pre>
3130 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
3131 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
3132 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
3133 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
3134</pre>
3135</div>
3136
3137<!-- _______________________________________________________________________ -->
3138<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
3139Instruction</a> </div>
3140<div class="doc_text">
3141<h5>Syntax:</h5>
3142<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>
3143<h5>Overview:</h5>
3144<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
3145<h5>Arguments:</h5>
3146<p>The argument to the '<tt>load</tt>' instruction specifies the memory
3147address from which to load. The pointer must point to a <a
3148 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
3149marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
3150the number or order of execution of this <tt>load</tt> with other
3151volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
3152instructions. </p>
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3153<p>
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3154The optional constant "align" argument specifies the alignment of the operation
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3155(that is, the alignment of the memory address). A value of 0 or an
3156omitted "align" argument means that the operation has the preferential
3157alignment for the target. It is the responsibility of the code emitter
3158to ensure that the alignment information is correct. Overestimating
3159the alignment results in an undefined behavior. Underestimating the
3160alignment may produce less efficient code. An alignment of 1 is always
3161safe.
3162</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3163<h5>Semantics:</h5>
3164<p>The location of memory pointed to is loaded.</p>
3165<h5>Examples:</h5>
3166<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
3167 <a
3168 href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
3169 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
3170</pre>
3171</div>
3172<!-- _______________________________________________________________________ -->
3173<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
3174Instruction</a> </div>
3175<div class="doc_text">
3176<h5>Syntax:</h5>
3177<pre> store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
3178 volatile store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
3179</pre>
3180<h5>Overview:</h5>
3181<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
3182<h5>Arguments:</h5>
3183<p>There are two arguments to the '<tt>store</tt>' instruction: a value
3184to 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]3185operand must be a pointer to the <a href="#t_firstclass">first class</a> type
3186of the '<tt>&lt;value&gt;</tt>'
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3187operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
3188optimizer is not allowed to modify the number or order of execution of
3189this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
3190 href="#i_store">store</a></tt> instructions.</p>
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3191<p>
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3192The optional constant "align" argument specifies the alignment of the operation
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]3193(that is, the alignment of the memory address). A value of 0 or an
3194omitted "align" argument means that the operation has the preferential
3195alignment for the target. It is the responsibility of the code emitter
3196to ensure that the alignment information is correct. Overestimating
3197the alignment results in an undefined behavior. Underestimating the
3198alignment may produce less efficient code. An alignment of 1 is always
3199safe.
3200</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3201<h5>Semantics:</h5>
3202<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>'
3203at the location specified by the '<tt>&lt;pointer&gt;</tt>' operand.</p>
3204<h5>Example:</h5>
3205<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling63ffa142007-10-22 05:10:05 +0000[diff] [blame]3206 store i32 3, i32* %ptr <i>; yields {void}</i>
3207 %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]3208</pre>
3209</div>
3210
3211<!-- _______________________________________________________________________ -->
3212<div class="doc_subsubsection">
3213 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
3214</div>
3215
3216<div class="doc_text">
3217<h5>Syntax:</h5>
3218<pre>
3219 &lt;result&gt; = getelementptr &lt;ty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
3220</pre>
3221
3222<h5>Overview:</h5>
3223
3224<p>
3225The '<tt>getelementptr</tt>' instruction is used to get the address of a
3226subelement of an aggregate data structure.</p>
3227
3228<h5>Arguments:</h5>
3229
3230<p>This instruction takes a list of integer operands that indicate what
3231elements of the aggregate object to index to. The actual types of the arguments
3232provided depend on the type of the first pointer argument. The
3233'<tt>getelementptr</tt>' instruction is used to index down through the type
3234levels of a structure or to a specific index in an array. When indexing into a
3235structure, only <tt>i32</tt> integer constants are allowed. When indexing
Chris Lattnera7d94ba2008-04-24 05:59:56 +0000[diff] [blame]3236into an array or pointer, only integers of 32 or 64 bits are allowed; 32-bit
3237values will be sign extended to 64-bits if required.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3238
3239<p>For example, let's consider a C code fragment and how it gets
3240compiled to LLVM:</p>
3241
3242<div class="doc_code">
3243<pre>
3244struct RT {
3245 char A;
3246 int B[10][20];
3247 char C;
3248};
3249struct ST {
3250 int X;
3251 double Y;
3252 struct RT Z;
3253};
3254
3255int *foo(struct ST *s) {
3256 return &amp;s[1].Z.B[5][13];
3257}
3258</pre>
3259</div>
3260
3261<p>The LLVM code generated by the GCC frontend is:</p>
3262
3263<div class="doc_code">
3264<pre>
3265%RT = type { i8 , [10 x [20 x i32]], i8 }
3266%ST = type { i32, double, %RT }
3267
3268define i32* %foo(%ST* %s) {
3269entry:
3270 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
3271 ret i32* %reg
3272}
3273</pre>
3274</div>
3275
3276<h5>Semantics:</h5>
3277
3278<p>The index types specified for the '<tt>getelementptr</tt>' instruction depend
3279on the pointer type that is being indexed into. <a href="#t_pointer">Pointer</a>
3280and <a href="#t_array">array</a> types can use a 32-bit or 64-bit
3281<a href="#t_integer">integer</a> type but the value will always be sign extended
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3282to 64-bits. <a href="#t_struct">Structure</a> and <a href="#t_pstruct">packed
3283structure</a> types require <tt>i32</tt> <b>constants</b>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3284
3285<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
3286type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
3287}</tt>' type, a structure. The second index indexes into the third element of
3288the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
3289i8 }</tt>' type, another structure. The third index indexes into the second
3290element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
3291array. The two dimensions of the array are subscripted into, yielding an
3292'<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
3293to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
3294
3295<p>Note that it is perfectly legal to index partially through a
3296structure, returning a pointer to an inner element. Because of this,
3297the LLVM code for the given testcase is equivalent to:</p>
3298
3299<pre>
3300 define i32* %foo(%ST* %s) {
3301 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
3302 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
3303 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
3304 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
3305 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
3306 ret i32* %t5
3307 }
3308</pre>
3309
3310<p>Note that it is undefined to access an array out of bounds: array and
3311pointer indexes must always be within the defined bounds of the array type.
Chris Lattnera7d94ba2008-04-24 05:59:56 +0000[diff] [blame]3312The one exception for this rule is zero length arrays. These arrays are
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3313defined to be accessible as variable length arrays, which requires access
3314beyond the zero'th element.</p>
3315
3316<p>The getelementptr instruction is often confusing. For some more insight
3317into how it works, see <a href="GetElementPtr.html">the getelementptr
3318FAQ</a>.</p>
3319
3320<h5>Example:</h5>
3321
3322<pre>
3323 <i>; yields [12 x i8]*:aptr</i>
3324 %aptr = getelementptr {i32, [12 x i8]}* %sptr, i64 0, i32 1
3325</pre>
3326</div>
3327
3328<!-- ======================================================================= -->
3329<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
3330</div>
3331<div class="doc_text">
3332<p>The instructions in this category are the conversion instructions (casting)
3333which all take a single operand and a type. They perform various bit conversions
3334on the operand.</p>
3335</div>
3336
3337<!-- _______________________________________________________________________ -->
3338<div class="doc_subsubsection">
3339 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
3340</div>
3341<div class="doc_text">
3342
3343<h5>Syntax:</h5>
3344<pre>
3345 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3346</pre>
3347
3348<h5>Overview:</h5>
3349<p>
3350The '<tt>trunc</tt>' instruction truncates its operand to the type <tt>ty2</tt>.
3351</p>
3352
3353<h5>Arguments:</h5>
3354<p>
3355The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
3356be an <a href="#t_integer">integer</a> type, and a type that specifies the size
3357and type of the result, which must be an <a href="#t_integer">integer</a>
3358type. The bit size of <tt>value</tt> must be larger than the bit size of
3359<tt>ty2</tt>. Equal sized types are not allowed.</p>
3360
3361<h5>Semantics:</h5>
3362<p>
3363The '<tt>trunc</tt>' instruction truncates the high order bits in <tt>value</tt>
3364and converts the remaining bits to <tt>ty2</tt>. Since the source size must be
3365larger than the destination size, <tt>trunc</tt> cannot be a <i>no-op cast</i>.
3366It will always truncate bits.</p>
3367
3368<h5>Example:</h5>
3369<pre>
3370 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
3371 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
3372 %Y = trunc i32 122 to i1 <i>; yields i1:false</i>
3373</pre>
3374</div>
3375
3376<!-- _______________________________________________________________________ -->
3377<div class="doc_subsubsection">
3378 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
3379</div>
3380<div class="doc_text">
3381
3382<h5>Syntax:</h5>
3383<pre>
3384 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3385</pre>
3386
3387<h5>Overview:</h5>
3388<p>The '<tt>zext</tt>' instruction zero extends its operand to type
3389<tt>ty2</tt>.</p>
3390
3391
3392<h5>Arguments:</h5>
3393<p>The '<tt>zext</tt>' instruction takes a value to cast, which must be of
3394<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3395also be of <a href="#t_integer">integer</a> type. The bit size of the
3396<tt>value</tt> must be smaller than the bit size of the destination type,
3397<tt>ty2</tt>.</p>
3398
3399<h5>Semantics:</h5>
3400<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
3401bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
3402
3403<p>When zero extending from i1, the result will always be either 0 or 1.</p>
3404
3405<h5>Example:</h5>
3406<pre>
3407 %X = zext i32 257 to i64 <i>; yields i64:257</i>
3408 %Y = zext i1 true to i32 <i>; yields i32:1</i>
3409</pre>
3410</div>
3411
3412<!-- _______________________________________________________________________ -->
3413<div class="doc_subsubsection">
3414 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
3415</div>
3416<div class="doc_text">
3417
3418<h5>Syntax:</h5>
3419<pre>
3420 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3421</pre>
3422
3423<h5>Overview:</h5>
3424<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
3425
3426<h5>Arguments:</h5>
3427<p>
3428The '<tt>sext</tt>' instruction takes a value to cast, which must be of
3429<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3430also be of <a href="#t_integer">integer</a> type. The bit size of the
3431<tt>value</tt> must be smaller than the bit size of the destination type,
3432<tt>ty2</tt>.</p>
3433
3434<h5>Semantics:</h5>
3435<p>
3436The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
3437bit (highest order bit) of the <tt>value</tt> until it reaches the bit size of
3438the type <tt>ty2</tt>.</p>
3439
3440<p>When sign extending from i1, the extension always results in -1 or 0.</p>
3441
3442<h5>Example:</h5>
3443<pre>
3444 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
3445 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
3446</pre>
3447</div>
3448
3449<!-- _______________________________________________________________________ -->
3450<div class="doc_subsubsection">
3451 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
3452</div>
3453
3454<div class="doc_text">
3455
3456<h5>Syntax:</h5>
3457
3458<pre>
3459 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3460</pre>
3461
3462<h5>Overview:</h5>
3463<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
3464<tt>ty2</tt>.</p>
3465
3466
3467<h5>Arguments:</h5>
3468<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
3469 point</a> value to cast and a <a href="#t_floating">floating point</a> type to
3470cast it to. The size of <tt>value</tt> must be larger than the size of
3471<tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
3472<i>no-op cast</i>.</p>
3473
3474<h5>Semantics:</h5>
3475<p> The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
3476<a href="#t_floating">floating point</a> type to a smaller
3477<a href="#t_floating">floating point</a> type. If the value cannot fit within
3478the destination type, <tt>ty2</tt>, then the results are undefined.</p>
3479
3480<h5>Example:</h5>
3481<pre>
3482 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
3483 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
3484</pre>
3485</div>
3486
3487<!-- _______________________________________________________________________ -->
3488<div class="doc_subsubsection">
3489 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
3490</div>
3491<div class="doc_text">
3492
3493<h5>Syntax:</h5>
3494<pre>
3495 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3496</pre>
3497
3498<h5>Overview:</h5>
3499<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
3500floating point value.</p>
3501
3502<h5>Arguments:</h5>
3503<p>The '<tt>fpext</tt>' instruction takes a
3504<a href="#t_floating">floating point</a> <tt>value</tt> to cast,
3505and a <a href="#t_floating">floating point</a> type to cast it to. The source
3506type must be smaller than the destination type.</p>
3507
3508<h5>Semantics:</h5>
3509<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
3510<a href="#t_floating">floating point</a> type to a larger
3511<a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
3512used to make a <i>no-op cast</i> because it always changes bits. Use
3513<tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
3514
3515<h5>Example:</h5>
3516<pre>
3517 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
3518 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
3519</pre>
3520</div>
3521
3522<!-- _______________________________________________________________________ -->
3523<div class="doc_subsubsection">
3524 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
3525</div>
3526<div class="doc_text">
3527
3528<h5>Syntax:</h5>
3529<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3530 &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]3531</pre>
3532
3533<h5>Overview:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3534<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3535unsigned integer equivalent of type <tt>ty2</tt>.
3536</p>
3537
3538<h5>Arguments:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3539<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]3540scalar or vector <a href="#t_floating">floating point</a> value, and a type
3541to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
3542type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
3543vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3544
3545<h5>Semantics:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3546<p> The '<tt>fptoui</tt>' instruction converts its
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3547<a href="#t_floating">floating point</a> operand into the nearest (rounding
3548towards zero) unsigned integer value. If the value cannot fit in <tt>ty2</tt>,
3549the results are undefined.</p>
3550
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3551<h5>Example:</h5>
3552<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3553 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]3554 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3555 %X = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3556</pre>
3557</div>
3558
3559<!-- _______________________________________________________________________ -->
3560<div class="doc_subsubsection">
3561 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
3562</div>
3563<div class="doc_text">
3564
3565<h5>Syntax:</h5>
3566<pre>
3567 &lt;result&gt; = fptosi &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3568</pre>
3569
3570<h5>Overview:</h5>
3571<p>The '<tt>fptosi</tt>' instruction converts
3572<a href="#t_floating">floating point</a> <tt>value</tt> to type <tt>ty2</tt>.
3573</p>
3574
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3575<h5>Arguments:</h5>
3576<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]3577scalar or vector <a href="#t_floating">floating point</a> value, and a type
3578to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
3579type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
3580vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3581
3582<h5>Semantics:</h5>
3583<p>The '<tt>fptosi</tt>' instruction converts its
3584<a href="#t_floating">floating point</a> operand into the nearest (rounding
3585towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
3586the results are undefined.</p>
3587
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3588<h5>Example:</h5>
3589<pre>
3590 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]3591 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3592 %X = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
3593</pre>
3594</div>
3595
3596<!-- _______________________________________________________________________ -->
3597<div class="doc_subsubsection">
3598 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
3599</div>
3600<div class="doc_text">
3601
3602<h5>Syntax:</h5>
3603<pre>
3604 &lt;result&gt; = uitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3605</pre>
3606
3607<h5>Overview:</h5>
3608<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
3609integer and converts that value to the <tt>ty2</tt> type.</p>
3610
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3611<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]3612<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be a
3613scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
3614to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
3615type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
3616floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3617
3618<h5>Semantics:</h5>
3619<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
3620integer quantity and converts it to the corresponding floating point value. If
3621the value cannot fit in the floating point value, the results are undefined.</p>
3622
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3623<h5>Example:</h5>
3624<pre>
3625 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
3626 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
3627</pre>
3628</div>
3629
3630<!-- _______________________________________________________________________ -->
3631<div class="doc_subsubsection">
3632 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
3633</div>
3634<div class="doc_text">
3635
3636<h5>Syntax:</h5>
3637<pre>
3638 &lt;result&gt; = sitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3639</pre>
3640
3641<h5>Overview:</h5>
3642<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed
3643integer and converts that value to the <tt>ty2</tt> type.</p>
3644
3645<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]3646<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be a
3647scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
3648to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
3649type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
3650floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3651
3652<h5>Semantics:</h5>
3653<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed
3654integer quantity and converts it to the corresponding floating point value. If
3655the value cannot fit in the floating point value, the results are undefined.</p>
3656
3657<h5>Example:</h5>
3658<pre>
3659 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
3660 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
3661</pre>
3662</div>
3663
3664<!-- _______________________________________________________________________ -->
3665<div class="doc_subsubsection">
3666 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
3667</div>
3668<div class="doc_text">
3669
3670<h5>Syntax:</h5>
3671<pre>
3672 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3673</pre>
3674
3675<h5>Overview:</h5>
3676<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
3677the integer type <tt>ty2</tt>.</p>
3678
3679<h5>Arguments:</h5>
3680<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
3681must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
3682<tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> type.
3683
3684<h5>Semantics:</h5>
3685<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
3686<tt>ty2</tt> by interpreting the pointer value as an integer and either
3687truncating or zero extending that value to the size of the integer type. If
3688<tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
3689<tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
3690are the same size, then nothing is done (<i>no-op cast</i>) other than a type
3691change.</p>
3692
3693<h5>Example:</h5>
3694<pre>
3695 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
3696 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
3697</pre>
3698</div>
3699
3700<!-- _______________________________________________________________________ -->
3701<div class="doc_subsubsection">
3702 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
3703</div>
3704<div class="doc_text">
3705
3706<h5>Syntax:</h5>
3707<pre>
3708 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3709</pre>
3710
3711<h5>Overview:</h5>
3712<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to
3713a pointer type, <tt>ty2</tt>.</p>
3714
3715<h5>Arguments:</h5>
3716<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
3717value to cast, and a type to cast it to, which must be a
3718<a href="#t_pointer">pointer</a> type.
3719
3720<h5>Semantics:</h5>
3721<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
3722<tt>ty2</tt> by applying either a zero extension or a truncation depending on
3723the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
3724size of a pointer then a truncation is done. If <tt>value</tt> is smaller than
3725the size of a pointer then a zero extension is done. If they are the same size,
3726nothing is done (<i>no-op cast</i>).</p>
3727
3728<h5>Example:</h5>
3729<pre>
3730 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
3731 %X = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
3732 %Y = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
3733</pre>
3734</div>
3735
3736<!-- _______________________________________________________________________ -->
3737<div class="doc_subsubsection">
3738 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
3739</div>
3740<div class="doc_text">
3741
3742<h5>Syntax:</h5>
3743<pre>
3744 &lt;result&gt; = bitcast &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3745</pre>
3746
3747<h5>Overview:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3748
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3749<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
3750<tt>ty2</tt> without changing any bits.</p>
3751
3752<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3753
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3754<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be
3755a first class value, and a type to cast it to, which must also be a <a
3756 href="#t_firstclass">first class</a> type. The bit sizes of <tt>value</tt>
3757and the destination type, <tt>ty2</tt>, must be identical. If the source
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3758type is a pointer, the destination type must also be a pointer. This
3759instruction supports bitwise conversion of vectors to integers and to vectors
3760of other types (as long as they have the same size).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3761
3762<h5>Semantics:</h5>
3763<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
3764<tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
3765this conversion. The conversion is done as if the <tt>value</tt> had been
3766stored to memory and read back as type <tt>ty2</tt>. Pointer types may only be
3767converted to other pointer types with this instruction. To convert pointers to
3768other types, use the <a href="#i_inttoptr">inttoptr</a> or
3769<a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
3770
3771<h5>Example:</h5>
3772<pre>
3773 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
3774 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
3775 %Z = bitcast <2xint> %V to i64; <i>; yields i64: %V</i>
3776</pre>
3777</div>
3778
3779<!-- ======================================================================= -->
3780<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
3781<div class="doc_text">
3782<p>The instructions in this category are the "miscellaneous"
3783instructions, which defy better classification.</p>
3784</div>
3785
3786<!-- _______________________________________________________________________ -->
3787<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
3788</div>
3789<div class="doc_text">
3790<h5>Syntax:</h5>
3791<pre> &lt;result&gt; = icmp &lt;cond&gt; &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {i1}:result</i>
3792</pre>
3793<h5>Overview:</h5>
3794<p>The '<tt>icmp</tt>' instruction returns a boolean value based on comparison
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3795of its two integer or pointer operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3796<h5>Arguments:</h5>
3797<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
3798the condition code indicating the kind of comparison to perform. It is not
3799a value, just a keyword. The possible condition code are:
3800<ol>
3801 <li><tt>eq</tt>: equal</li>
3802 <li><tt>ne</tt>: not equal </li>
3803 <li><tt>ugt</tt>: unsigned greater than</li>
3804 <li><tt>uge</tt>: unsigned greater or equal</li>
3805 <li><tt>ult</tt>: unsigned less than</li>
3806 <li><tt>ule</tt>: unsigned less or equal</li>
3807 <li><tt>sgt</tt>: signed greater than</li>
3808 <li><tt>sge</tt>: signed greater or equal</li>
3809 <li><tt>slt</tt>: signed less than</li>
3810 <li><tt>sle</tt>: signed less or equal</li>
3811</ol>
3812<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
3813<a href="#t_pointer">pointer</a> typed. They must also be identical types.</p>
3814<h5>Semantics:</h5>
3815<p>The '<tt>icmp</tt>' compares <tt>var1</tt> and <tt>var2</tt> according to
3816the condition code given as <tt>cond</tt>. The comparison performed always
3817yields a <a href="#t_primitive">i1</a> result, as follows:
3818<ol>
3819 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
3820 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
3821 </li>
3822 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
3823 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
3824 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
3825 <tt>true</tt> if <tt>var1</tt> is greater than <tt>var2</tt>.</li>
3826 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
3827 <tt>true</tt> if <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3828 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
3829 <tt>true</tt> if <tt>var1</tt> is less than <tt>var2</tt>.</li>
3830 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
3831 <tt>true</tt> if <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
3832 <li><tt>sgt</tt>: interprets the operands as signed values and yields
3833 <tt>true</tt> if <tt>var1</tt> is greater than <tt>var2</tt>.</li>
3834 <li><tt>sge</tt>: interprets the operands as signed values and yields
3835 <tt>true</tt> if <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3836 <li><tt>slt</tt>: interprets the operands as signed values and yields
3837 <tt>true</tt> if <tt>var1</tt> is less than <tt>var2</tt>.</li>
3838 <li><tt>sle</tt>: interprets the operands as signed values and yields
3839 <tt>true</tt> if <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
3840</ol>
3841<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
3842values are compared as if they were integers.</p>
3843
3844<h5>Example:</h5>
3845<pre> &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
3846 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
3847 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
3848 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
3849 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
3850 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
3851</pre>
3852</div>
3853
3854<!-- _______________________________________________________________________ -->
3855<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
3856</div>
3857<div class="doc_text">
3858<h5>Syntax:</h5>
3859<pre> &lt;result&gt; = fcmp &lt;cond&gt; &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {i1}:result</i>
3860</pre>
3861<h5>Overview:</h5>
3862<p>The '<tt>fcmp</tt>' instruction returns a boolean value based on comparison
3863of its floating point operands.</p>
3864<h5>Arguments:</h5>
3865<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
3866the condition code indicating the kind of comparison to perform. It is not
3867a value, just a keyword. The possible condition code are:
3868<ol>
3869 <li><tt>false</tt>: no comparison, always returns false</li>
3870 <li><tt>oeq</tt>: ordered and equal</li>
3871 <li><tt>ogt</tt>: ordered and greater than </li>
3872 <li><tt>oge</tt>: ordered and greater than or equal</li>
3873 <li><tt>olt</tt>: ordered and less than </li>
3874 <li><tt>ole</tt>: ordered and less than or equal</li>
3875 <li><tt>one</tt>: ordered and not equal</li>
3876 <li><tt>ord</tt>: ordered (no nans)</li>
3877 <li><tt>ueq</tt>: unordered or equal</li>
3878 <li><tt>ugt</tt>: unordered or greater than </li>
3879 <li><tt>uge</tt>: unordered or greater than or equal</li>
3880 <li><tt>ult</tt>: unordered or less than </li>
3881 <li><tt>ule</tt>: unordered or less than or equal</li>
3882 <li><tt>une</tt>: unordered or not equal</li>
3883 <li><tt>uno</tt>: unordered (either nans)</li>
3884 <li><tt>true</tt>: no comparison, always returns true</li>
3885</ol>
3886<p><i>Ordered</i> means that neither operand is a QNAN while
3887<i>unordered</i> means that either operand may be a QNAN.</p>
3888<p>The <tt>val1</tt> and <tt>val2</tt> arguments must be
3889<a href="#t_floating">floating point</a> typed. They must have identical
3890types.</p>
3891<h5>Semantics:</h5>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]3892<p>The '<tt>fcmp</tt>' instruction compares <tt>var1</tt> and <tt>var2</tt>
3893according to the condition code given as <tt>cond</tt>. The comparison performed
3894always yields a <a href="#t_primitive">i1</a> result, as follows:
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3895<ol>
3896 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
3897 <li><tt>oeq</tt>: yields <tt>true</tt> if both operands are not a QNAN and
3898 <tt>var1</tt> is equal to <tt>var2</tt>.</li>
3899 <li><tt>ogt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
3900 <tt>var1</tt> is greather than <tt>var2</tt>.</li>
3901 <li><tt>oge</tt>: yields <tt>true</tt> if both operands are not a QNAN and
3902 <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3903 <li><tt>olt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
3904 <tt>var1</tt> is less than <tt>var2</tt>.</li>
3905 <li><tt>ole</tt>: yields <tt>true</tt> if both operands are not a QNAN and
3906 <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
3907 <li><tt>one</tt>: yields <tt>true</tt> if both operands are not a QNAN and
3908 <tt>var1</tt> is not equal to <tt>var2</tt>.</li>
3909 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
3910 <li><tt>ueq</tt>: yields <tt>true</tt> if either operand is a QNAN or
3911 <tt>var1</tt> is equal to <tt>var2</tt>.</li>
3912 <li><tt>ugt</tt>: yields <tt>true</tt> if either operand is a QNAN or
3913 <tt>var1</tt> is greater than <tt>var2</tt>.</li>
3914 <li><tt>uge</tt>: yields <tt>true</tt> if either operand is a QNAN or
3915 <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3916 <li><tt>ult</tt>: yields <tt>true</tt> if either operand is a QNAN or
3917 <tt>var1</tt> is less than <tt>var2</tt>.</li>
3918 <li><tt>ule</tt>: yields <tt>true</tt> if either operand is a QNAN or
3919 <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
3920 <li><tt>une</tt>: yields <tt>true</tt> if either operand is a QNAN or
3921 <tt>var1</tt> is not equal to <tt>var2</tt>.</li>
3922 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
3923 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
3924</ol>
3925
3926<h5>Example:</h5>
3927<pre> &lt;result&gt; = fcmp oeq float 4.0, 5.0 <i>; yields: result=false</i>
3928 &lt;result&gt; = icmp one float 4.0, 5.0 <i>; yields: result=true</i>
3929 &lt;result&gt; = icmp olt float 4.0, 5.0 <i>; yields: result=true</i>
3930 &lt;result&gt; = icmp ueq double 1.0, 2.0 <i>; yields: result=false</i>
3931</pre>
3932</div>
3933
3934<!-- _______________________________________________________________________ -->
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]3935<div class="doc_subsubsection">
3936 <a name="i_vicmp">'<tt>vicmp</tt>' Instruction</a>
3937</div>
3938<div class="doc_text">
3939<h5>Syntax:</h5>
3940<pre> &lt;result&gt; = vicmp &lt;cond&gt; &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
3941</pre>
3942<h5>Overview:</h5>
3943<p>The '<tt>vicmp</tt>' instruction returns an integer vector value based on
3944element-wise comparison of its two integer vector operands.</p>
3945<h5>Arguments:</h5>
3946<p>The '<tt>vicmp</tt>' instruction takes three operands. The first operand is
3947the condition code indicating the kind of comparison to perform. It is not
3948a value, just a keyword. The possible condition code are:
3949<ol>
3950 <li><tt>eq</tt>: equal</li>
3951 <li><tt>ne</tt>: not equal </li>
3952 <li><tt>ugt</tt>: unsigned greater than</li>
3953 <li><tt>uge</tt>: unsigned greater or equal</li>
3954 <li><tt>ult</tt>: unsigned less than</li>
3955 <li><tt>ule</tt>: unsigned less or equal</li>
3956 <li><tt>sgt</tt>: signed greater than</li>
3957 <li><tt>sge</tt>: signed greater or equal</li>
3958 <li><tt>slt</tt>: signed less than</li>
3959 <li><tt>sle</tt>: signed less or equal</li>
3960</ol>
3961<p>The remaining two arguments must be <a href="#t_vector">vector</a> of
3962<a href="#t_integer">integer</a> typed. They must also be identical types.</p>
3963<h5>Semantics:</h5>
3964<p>The '<tt>vicmp</tt>' instruction compares <tt>var1</tt> and <tt>var2</tt>
3965according to the condition code given as <tt>cond</tt>. The comparison yields a
3966<a href="#t_vector">vector</a> of <a href="#t_integer">integer</a> result, of
3967identical type as the values being compared. The most significant bit in each
3968element is 1 if the element-wise comparison evaluates to true, and is 0
3969otherwise. All other bits of the result are undefined. The condition codes
3970are evaluated identically to the <a href="#i_icmp">'<tt>icmp</tt>'
3971instruction</a>.
3972
3973<h5>Example:</h5>
3974<pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]3975 &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>
3976 &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]3977</pre>
3978</div>
3979
3980<!-- _______________________________________________________________________ -->
3981<div class="doc_subsubsection">
3982 <a name="i_vfcmp">'<tt>vfcmp</tt>' Instruction</a>
3983</div>
3984<div class="doc_text">
3985<h5>Syntax:</h5>
3986<pre> &lt;result&gt; = vfcmp &lt;cond&gt; &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt;</pre>
3987<h5>Overview:</h5>
3988<p>The '<tt>vfcmp</tt>' instruction returns an integer vector value based on
3989element-wise comparison of its two floating point vector operands. The output
3990elements have the same width as the input elements.</p>
3991<h5>Arguments:</h5>
3992<p>The '<tt>vfcmp</tt>' instruction takes three operands. The first operand is
3993the condition code indicating the kind of comparison to perform. It is not
3994a value, just a keyword. The possible condition code are:
3995<ol>
3996 <li><tt>false</tt>: no comparison, always returns false</li>
3997 <li><tt>oeq</tt>: ordered and equal</li>
3998 <li><tt>ogt</tt>: ordered and greater than </li>
3999 <li><tt>oge</tt>: ordered and greater than or equal</li>
4000 <li><tt>olt</tt>: ordered and less than </li>
4001 <li><tt>ole</tt>: ordered and less than or equal</li>
4002 <li><tt>one</tt>: ordered and not equal</li>
4003 <li><tt>ord</tt>: ordered (no nans)</li>
4004 <li><tt>ueq</tt>: unordered or equal</li>
4005 <li><tt>ugt</tt>: unordered or greater than </li>
4006 <li><tt>uge</tt>: unordered or greater than or equal</li>
4007 <li><tt>ult</tt>: unordered or less than </li>
4008 <li><tt>ule</tt>: unordered or less than or equal</li>
4009 <li><tt>une</tt>: unordered or not equal</li>
4010 <li><tt>uno</tt>: unordered (either nans)</li>
4011 <li><tt>true</tt>: no comparison, always returns true</li>
4012</ol>
4013<p>The remaining two arguments must be <a href="#t_vector">vector</a> of
4014<a href="#t_floating">floating point</a> typed. They must also be identical
4015types.</p>
4016<h5>Semantics:</h5>
4017<p>The '<tt>vfcmp</tt>' instruction compares <tt>var1</tt> and <tt>var2</tt>
4018according to the condition code given as <tt>cond</tt>. The comparison yields a
4019<a href="#t_vector">vector</a> of <a href="#t_integer">integer</a> result, with
4020an identical number of elements as the values being compared, and each element
4021having identical with to the width of the floating point elements. The most
4022significant bit in each element is 1 if the element-wise comparison evaluates to
4023true, and is 0 otherwise. All other bits of the result are undefined. The
4024condition codes are evaluated identically to the
4025<a href="#i_fcmp">'<tt>fcmp</tt>' instruction</a>.
4026
4027<h5>Example:</h5>
4028<pre>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4029 &lt;result&gt; = vfcmp oeq &lt;2 x float&gt; &lt; float 4, float 0 &gt;, &lt; float 5, float 0 &gt; <i>; yields: result=&lt;2 x i32&gt; &lt; i32 0, i32 -1 &gt;</i>
4030 &lt;result&gt; = vfcmp ult &lt;2 x double&gt; &lt; double 1, double 2 &gt;, &lt; double 2, double 2&gt; <i>; yields: result=&lt;2 x i64&gt; &lt; i64 -1, i64 0 &gt;</i>
Nate Begeman646fa482008-05-12 19:01:56 +0000[diff] [blame]4031</pre>
4032</div>
4033
4034<!-- _______________________________________________________________________ -->
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4035<div class="doc_subsubsection">
4036 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
4037</div>
4038
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4039<div class="doc_text">
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4040
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4041<h5>Syntax:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4042
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4043<pre> &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...<br></pre>
4044<h5>Overview:</h5>
4045<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in
4046the SSA graph representing the function.</p>
4047<h5>Arguments:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4048
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4049<p>The type of the incoming values is specified with the first type
4050field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
4051as arguments, with one pair for each predecessor basic block of the
4052current block. Only values of <a href="#t_firstclass">first class</a>
4053type may be used as the value arguments to the PHI node. Only labels
4054may be used as the label arguments.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4055
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4056<p>There must be no non-phi instructions between the start of a basic
4057block and the PHI instructions: i.e. PHI instructions must be first in
4058a basic block.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4059
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4060<h5>Semantics:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4061
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4062<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
4063specified by the pair corresponding to the predecessor basic block that executed
4064just prior to the current block.</p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4065
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4066<h5>Example:</h5>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4067<pre>
4068Loop: ; Infinite loop that counts from 0 on up...
4069 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
4070 %nextindvar = add i32 %indvar, 1
4071 br label %Loop
4072</pre>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4073</div>
4074
4075<!-- _______________________________________________________________________ -->
4076<div class="doc_subsubsection">
4077 <a name="i_select">'<tt>select</tt>' Instruction</a>
4078</div>
4079
4080<div class="doc_text">
4081
4082<h5>Syntax:</h5>
4083
4084<pre>
4085 &lt;result&gt; = select i1 &lt;cond&gt;, &lt;ty&gt; &lt;val1&gt;, &lt;ty&gt; &lt;val2&gt; <i>; yields ty</i>
4086</pre>
4087
4088<h5>Overview:</h5>
4089
4090<p>
4091The '<tt>select</tt>' instruction is used to choose one value based on a
4092condition, without branching.
4093</p>
4094
4095
4096<h5>Arguments:</h5>
4097
4098<p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4099The '<tt>select</tt>' instruction requires an 'i1' value indicating the
4100condition, and two values of the same <a href="#t_firstclass">first class</a>
4101type. If the val1/val2 are vectors, the entire vectors are selected, not
4102individual elements.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4103</p>
4104
4105<h5>Semantics:</h5>
4106
4107<p>
Chris Lattner6704c212008-05-20 20:48:21 +0000[diff] [blame]4108If the i1 condition evaluates is 1, the instruction returns the first
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4109value argument; otherwise, it returns the second value argument.
4110</p>
4111
4112<h5>Example:</h5>
4113
4114<pre>
4115 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
4116</pre>
4117</div>
4118
4119
4120<!-- _______________________________________________________________________ -->
4121<div class="doc_subsubsection">
4122 <a name="i_call">'<tt>call</tt>' Instruction</a>
4123</div>
4124
4125<div class="doc_text">
4126
4127<h5>Syntax:</h5>
4128<pre>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]4129 &lt;result&gt; = [tail] call [<a href="#callingconv">cconv</a>] &lt;ty&gt; [&lt;fnty&gt;*] &lt;fnptrval&gt;(&lt;param list&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4130</pre>
4131
4132<h5>Overview:</h5>
4133
4134<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
4135
4136<h5>Arguments:</h5>
4137
4138<p>This instruction requires several arguments:</p>
4139
4140<ol>
4141 <li>
4142 <p>The optional "tail" marker indicates whether the callee function accesses
4143 any allocas or varargs in the caller. If the "tail" marker is present, the
4144 function call is eligible for tail call optimization. Note that calls may
4145 be marked "tail" even if they do not occur before a <a
4146 href="#i_ret"><tt>ret</tt></a> instruction.
4147 </li>
4148 <li>
4149 <p>The optional "cconv" marker indicates which <a href="#callingconv">calling
4150 convention</a> the call should use. If none is specified, the call defaults
4151 to using C calling conventions.
4152 </li>
4153 <li>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]4154 <p>'<tt>ty</tt>': the type of the call instruction itself which is also
4155 the type of the return value. Functions that return no value are marked
4156 <tt><a href="#t_void">void</a></tt>.</p>
4157 </li>
4158 <li>
4159 <p>'<tt>fnty</tt>': shall be the signature of the pointer to function
4160 value being invoked. The argument types must match the types implied by
4161 this signature. This type can be omitted if the function is not varargs
4162 and if the function type does not return a pointer to a function.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4163 </li>
4164 <li>
4165 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
4166 be invoked. In most cases, this is a direct function invocation, but
4167 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
4168 to function value.</p>
4169 </li>
4170 <li>
4171 <p>'<tt>function args</tt>': argument list whose types match the
4172 function signature argument types. All arguments must be of
4173 <a href="#t_firstclass">first class</a> type. If the function signature
4174 indicates the function accepts a variable number of arguments, the extra
4175 arguments can be specified.</p>
4176 </li>
4177</ol>
4178
4179<h5>Semantics:</h5>
4180
4181<p>The '<tt>call</tt>' instruction is used to cause control flow to
4182transfer to a specified function, with its incoming arguments bound to
4183the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
4184instruction in the called function, control flow continues with the
4185instruction after the function call, and the return value of the
Chris Lattner5e893ef2008-03-21 17:24:17 +0000[diff] [blame]4186function is bound to the result argument. If the callee returns multiple
4187values then the return values of the function are only accessible through
4188the '<tt><a href="#i_getresult">getresult</a></tt>' instruction.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4189
4190<h5>Example:</h5>
4191
4192<pre>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]4193 %retval = call i32 @test(i32 %argc)
Chris Lattner5e893ef2008-03-21 17:24:17 +0000[diff] [blame]4194 call i32 (i8 *, ...)* @printf(i8 * %msg, i32 12, i8 42) <i>; yields i32</i>
4195 %X = tail call i32 @foo() <i>; yields i32</i>
4196 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
4197 call void %foo(i8 97 signext)
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]4198
4199 %struct.A = type { i32, i8 }
Chris Lattner5e893ef2008-03-21 17:24:17 +0000[diff] [blame]4200 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
4201 %gr = getresult %struct.A %r, 0 <i>; yields i32</i>
4202 %gr1 = getresult %struct.A %r, 1 <i>; yields i8</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4203</pre>
4204
4205</div>
4206
4207<!-- _______________________________________________________________________ -->
4208<div class="doc_subsubsection">
4209 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
4210</div>
4211
4212<div class="doc_text">
4213
4214<h5>Syntax:</h5>
4215
4216<pre>
4217 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
4218</pre>
4219
4220<h5>Overview:</h5>
4221
4222<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
4223the "variable argument" area of a function call. It is used to implement the
4224<tt>va_arg</tt> macro in C.</p>
4225
4226<h5>Arguments:</h5>
4227
4228<p>This instruction takes a <tt>va_list*</tt> value and the type of
4229the argument. It returns a value of the specified argument type and
4230increments the <tt>va_list</tt> to point to the next argument. The
4231actual type of <tt>va_list</tt> is target specific.</p>
4232
4233<h5>Semantics:</h5>
4234
4235<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
4236type from the specified <tt>va_list</tt> and causes the
4237<tt>va_list</tt> to point to the next argument. For more information,
4238see the variable argument handling <a href="#int_varargs">Intrinsic
4239Functions</a>.</p>
4240
4241<p>It is legal for this instruction to be called in a function which does not
4242take a variable number of arguments, for example, the <tt>vfprintf</tt>
4243function.</p>
4244
4245<p><tt>va_arg</tt> is an LLVM instruction instead of an <a
4246href="#intrinsics">intrinsic function</a> because it takes a type as an
4247argument.</p>
4248
4249<h5>Example:</h5>
4250
4251<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
4252
4253</div>
4254
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]4255<!-- _______________________________________________________________________ -->
4256<div class="doc_subsubsection">
4257 <a name="i_getresult">'<tt>getresult</tt>' Instruction</a>
4258</div>
4259
4260<div class="doc_text">
4261
4262<h5>Syntax:</h5>
4263<pre>
Chris Lattneree9da3f2008-03-21 17:20:51 +0000[diff] [blame]4264 &lt;resultval&gt; = getresult &lt;type&gt; &lt;retval&gt;, &lt;index&gt;
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]4265</pre>
Chris Lattneree9da3f2008-03-21 17:20:51 +0000[diff] [blame]4266
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]4267<h5>Overview:</h5>
4268
4269<p> The '<tt>getresult</tt>' instruction is used to extract individual values
Chris Lattneree9da3f2008-03-21 17:20:51 +0000[diff] [blame]4270from a '<tt><a href="#i_call">call</a></tt>'
4271or '<tt><a href="#i_invoke">invoke</a></tt>' instruction that returns multiple
4272results.</p>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]4273
4274<h5>Arguments:</h5>
4275
Chris Lattneree9da3f2008-03-21 17:20:51 +0000[diff] [blame]4276<p>The '<tt>getresult</tt>' instruction takes a call or invoke value as its
Chris Lattnerd8dd3522008-04-23 04:06:52 +0000[diff] [blame]4277first argument, or an undef value. The value must have <a
4278href="#t_struct">structure type</a>. The second argument is a constant
4279unsigned index value which must be in range for the number of values returned
4280by the call.</p>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]4281
4282<h5>Semantics:</h5>
4283
Chris Lattneree9da3f2008-03-21 17:20:51 +0000[diff] [blame]4284<p>The '<tt>getresult</tt>' instruction extracts the element identified by
4285'<tt>index</tt>' from the aggregate value.</p>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]4286
4287<h5>Example:</h5>
4288
4289<pre>
4290 %struct.A = type { i32, i8 }
4291
4292 %r = call %struct.A @foo()
Chris Lattneree9da3f2008-03-21 17:20:51 +0000[diff] [blame]4293 %gr = getresult %struct.A %r, 0 <i>; yields i32:%gr</i>
4294 %gr1 = getresult %struct.A %r, 1 <i>; yields i8:%gr1</i>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]4295 add i32 %gr, 42
4296 add i8 %gr1, 41
4297</pre>
4298
4299</div>
4300
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4301<!-- *********************************************************************** -->
4302<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
4303<!-- *********************************************************************** -->
4304
4305<div class="doc_text">
4306
4307<p>LLVM supports the notion of an "intrinsic function". These functions have
4308well known names and semantics and are required to follow certain restrictions.
4309Overall, these intrinsics represent an extension mechanism for the LLVM
4310language that does not require changing all of the transformations in LLVM when
4311adding to the language (or the bitcode reader/writer, the parser, etc...).</p>
4312
4313<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
4314prefix is reserved in LLVM for intrinsic names; thus, function names may not
4315begin with this prefix. Intrinsic functions must always be external functions:
4316you cannot define the body of intrinsic functions. Intrinsic functions may
4317only be used in call or invoke instructions: it is illegal to take the address
4318of an intrinsic function. Additionally, because intrinsic functions are part
4319of the LLVM language, it is required if any are added that they be documented
4320here.</p>
4321
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4322<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents
4323a family of functions that perform the same operation but on different data
4324types. Because LLVM can represent over 8 million different integer types,
4325overloading is used commonly to allow an intrinsic function to operate on any
4326integer type. One or more of the argument types or the result type can be
4327overloaded to accept any integer type. Argument types may also be defined as
4328exactly matching a previous argument's type or the result type. This allows an
4329intrinsic function which accepts multiple arguments, but needs all of them to
4330be of the same type, to only be overloaded with respect to a single argument or
4331the result.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4332
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4333<p>Overloaded intrinsics will have the names of its overloaded argument types
4334encoded into its function name, each preceded by a period. Only those types
4335which are overloaded result in a name suffix. Arguments whose type is matched
4336against another type do not. For example, the <tt>llvm.ctpop</tt> function can
4337take an integer of any width and returns an integer of exactly the same integer
4338width. This leads to a family of functions such as
4339<tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29 %val)</tt>.
4340Only one type, the return type, is overloaded, and only one type suffix is
4341required. Because the argument's type is matched against the return type, it
4342does not require its own name suffix.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4343
4344<p>To learn how to add an intrinsic function, please see the
4345<a href="ExtendingLLVM.html">Extending LLVM Guide</a>.
4346</p>
4347
4348</div>
4349
4350<!-- ======================================================================= -->
4351<div class="doc_subsection">
4352 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
4353</div>
4354
4355<div class="doc_text">
4356
4357<p>Variable argument support is defined in LLVM with the <a
4358 href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
4359intrinsic functions. These functions are related to the similarly
4360named macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
4361
4362<p>All of these functions operate on arguments that use a
4363target-specific value type "<tt>va_list</tt>". The LLVM assembly
4364language reference manual does not define what this type is, so all
4365transformations should be prepared to handle these functions regardless of
4366the type used.</p>
4367
4368<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
4369instruction and the variable argument handling intrinsic functions are
4370used.</p>
4371
4372<div class="doc_code">
4373<pre>
4374define i32 @test(i32 %X, ...) {
4375 ; Initialize variable argument processing
4376 %ap = alloca i8*
4377 %ap2 = bitcast i8** %ap to i8*
4378 call void @llvm.va_start(i8* %ap2)
4379
4380 ; Read a single integer argument
4381 %tmp = va_arg i8** %ap, i32
4382
4383 ; Demonstrate usage of llvm.va_copy and llvm.va_end
4384 %aq = alloca i8*
4385 %aq2 = bitcast i8** %aq to i8*
4386 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
4387 call void @llvm.va_end(i8* %aq2)
4388
4389 ; Stop processing of arguments.
4390 call void @llvm.va_end(i8* %ap2)
4391 ret i32 %tmp
4392}
4393
4394declare void @llvm.va_start(i8*)
4395declare void @llvm.va_copy(i8*, i8*)
4396declare void @llvm.va_end(i8*)
4397</pre>
4398</div>
4399
4400</div>
4401
4402<!-- _______________________________________________________________________ -->
4403<div class="doc_subsubsection">
4404 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
4405</div>
4406
4407
4408<div class="doc_text">
4409<h5>Syntax:</h5>
4410<pre> declare void %llvm.va_start(i8* &lt;arglist&gt;)<br></pre>
4411<h5>Overview:</h5>
4412<P>The '<tt>llvm.va_start</tt>' intrinsic initializes
4413<tt>*&lt;arglist&gt;</tt> for subsequent use by <tt><a
4414href="#i_va_arg">va_arg</a></tt>.</p>
4415
4416<h5>Arguments:</h5>
4417
4418<P>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
4419
4420<h5>Semantics:</h5>
4421
4422<P>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
4423macro available in C. In a target-dependent way, it initializes the
4424<tt>va_list</tt> element to which the argument points, so that the next call to
4425<tt>va_arg</tt> will produce the first variable argument passed to the function.
4426Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
4427last argument of the function as the compiler can figure that out.</p>
4428
4429</div>
4430
4431<!-- _______________________________________________________________________ -->
4432<div class="doc_subsubsection">
4433 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
4434</div>
4435
4436<div class="doc_text">
4437<h5>Syntax:</h5>
4438<pre> declare void @llvm.va_end(i8* &lt;arglist&gt;)<br></pre>
4439<h5>Overview:</h5>
4440
4441<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
4442which has been initialized previously with <tt><a href="#int_va_start">llvm.va_start</a></tt>
4443or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
4444
4445<h5>Arguments:</h5>
4446
4447<p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
4448
4449<h5>Semantics:</h5>
4450
4451<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
4452macro available in C. In a target-dependent way, it destroys the
4453<tt>va_list</tt> element to which the argument points. Calls to <a
4454href="#int_va_start"><tt>llvm.va_start</tt></a> and <a href="#int_va_copy">
4455<tt>llvm.va_copy</tt></a> must be matched exactly with calls to
4456<tt>llvm.va_end</tt>.</p>
4457
4458</div>
4459
4460<!-- _______________________________________________________________________ -->
4461<div class="doc_subsubsection">
4462 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
4463</div>
4464
4465<div class="doc_text">
4466
4467<h5>Syntax:</h5>
4468
4469<pre>
4470 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
4471</pre>
4472
4473<h5>Overview:</h5>
4474
4475<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
4476from the source argument list to the destination argument list.</p>
4477
4478<h5>Arguments:</h5>
4479
4480<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
4481The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
4482
4483
4484<h5>Semantics:</h5>
4485
4486<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
4487macro available in C. In a target-dependent way, it copies the source
4488<tt>va_list</tt> element into the destination <tt>va_list</tt> element. This
4489intrinsic is necessary because the <tt><a href="#int_va_start">
4490llvm.va_start</a></tt> intrinsic may be arbitrarily complex and require, for
4491example, memory allocation.</p>
4492
4493</div>
4494
4495<!-- ======================================================================= -->
4496<div class="doc_subsection">
4497 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
4498</div>
4499
4500<div class="doc_text">
4501
4502<p>
4503LLVM support for <a href="GarbageCollection.html">Accurate Garbage
4504Collection</a> requires the implementation and generation of these intrinsics.
4505These intrinsics allow identification of <a href="#int_gcroot">GC roots on the
4506stack</a>, as well as garbage collector implementations that require <a
4507href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a> barriers.
4508Front-ends for type-safe garbage collected languages should generate these
4509intrinsics to make use of the LLVM garbage collectors. For more details, see <a
4510href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
4511</p>
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]4512
4513<p>The garbage collection intrinsics only operate on objects in the generic
4514 address space (address space zero).</p>
4515
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4516</div>
4517
4518<!-- _______________________________________________________________________ -->
4519<div class="doc_subsubsection">
4520 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
4521</div>
4522
4523<div class="doc_text">
4524
4525<h5>Syntax:</h5>
4526
4527<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4528 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4529</pre>
4530
4531<h5>Overview:</h5>
4532
4533<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
4534the code generator, and allows some metadata to be associated with it.</p>
4535
4536<h5>Arguments:</h5>
4537
4538<p>The first argument specifies the address of a stack object that contains the
4539root pointer. The second pointer (which must be either a constant or a global
4540value address) contains the meta-data to be associated with the root.</p>
4541
4542<h5>Semantics:</h5>
4543
Chris Lattnera7d94ba2008-04-24 05:59:56 +0000[diff] [blame]4544<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]4545location. At compile-time, the code generator generates information to allow
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4546the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
4547intrinsic may only be used in a function which <a href="#gc">specifies a GC
4548algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4549
4550</div>
4551
4552
4553<!-- _______________________________________________________________________ -->
4554<div class="doc_subsubsection">
4555 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
4556</div>
4557
4558<div class="doc_text">
4559
4560<h5>Syntax:</h5>
4561
4562<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4563 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4564</pre>
4565
4566<h5>Overview:</h5>
4567
4568<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
4569locations, allowing garbage collector implementations that require read
4570barriers.</p>
4571
4572<h5>Arguments:</h5>
4573
4574<p>The second argument is the address to read from, which should be an address
4575allocated from the garbage collector. The first object is a pointer to the
4576start of the referenced object, if needed by the language runtime (otherwise
4577null).</p>
4578
4579<h5>Semantics:</h5>
4580
4581<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
4582instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4583garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
4584may only be used in a function which <a href="#gc">specifies a GC
4585algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4586
4587</div>
4588
4589
4590<!-- _______________________________________________________________________ -->
4591<div class="doc_subsubsection">
4592 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
4593</div>
4594
4595<div class="doc_text">
4596
4597<h5>Syntax:</h5>
4598
4599<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4600 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4601</pre>
4602
4603<h5>Overview:</h5>
4604
4605<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
4606locations, allowing garbage collector implementations that require write
4607barriers (such as generational or reference counting collectors).</p>
4608
4609<h5>Arguments:</h5>
4610
4611<p>The first argument is the reference to store, the second is the start of the
4612object to store it to, and the third is the address of the field of Obj to
4613store to. If the runtime does not require a pointer to the object, Obj may be
4614null.</p>
4615
4616<h5>Semantics:</h5>
4617
4618<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
4619instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4620garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
4621may only be used in a function which <a href="#gc">specifies a GC
4622algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4623
4624</div>
4625
4626
4627
4628<!-- ======================================================================= -->
4629<div class="doc_subsection">
4630 <a name="int_codegen">Code Generator Intrinsics</a>
4631</div>
4632
4633<div class="doc_text">
4634<p>
4635These intrinsics are provided by LLVM to expose special features that may only
4636be implemented with code generator support.
4637</p>
4638
4639</div>
4640
4641<!-- _______________________________________________________________________ -->
4642<div class="doc_subsubsection">
4643 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
4644</div>
4645
4646<div class="doc_text">
4647
4648<h5>Syntax:</h5>
4649<pre>
4650 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
4651</pre>
4652
4653<h5>Overview:</h5>
4654
4655<p>
4656The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
4657target-specific value indicating the return address of the current function
4658or one of its callers.
4659</p>
4660
4661<h5>Arguments:</h5>
4662
4663<p>
4664The argument to this intrinsic indicates which function to return the address
4665for. Zero indicates the calling function, one indicates its caller, etc. The
4666argument is <b>required</b> to be a constant integer value.
4667</p>
4668
4669<h5>Semantics:</h5>
4670
4671<p>
4672The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
4673the return address of the specified call frame, or zero if it cannot be
4674identified. The value returned by this intrinsic is likely to be incorrect or 0
4675for arguments other than zero, so it should only be used for debugging purposes.
4676</p>
4677
4678<p>
4679Note that calling this intrinsic does not prevent function inlining or other
4680aggressive transformations, so the value returned may not be that of the obvious
4681source-language caller.
4682</p>
4683</div>
4684
4685
4686<!-- _______________________________________________________________________ -->
4687<div class="doc_subsubsection">
4688 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
4689</div>
4690
4691<div class="doc_text">
4692
4693<h5>Syntax:</h5>
4694<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4695 declare i8 *@llvm.frameaddress(i32 &lt;level&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4696</pre>
4697
4698<h5>Overview:</h5>
4699
4700<p>
4701The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
4702target-specific frame pointer value for the specified stack frame.
4703</p>
4704
4705<h5>Arguments:</h5>
4706
4707<p>
4708The argument to this intrinsic indicates which function to return the frame
4709pointer for. Zero indicates the calling function, one indicates its caller,
4710etc. The argument is <b>required</b> to be a constant integer value.
4711</p>
4712
4713<h5>Semantics:</h5>
4714
4715<p>
4716The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
4717the frame address of the specified call frame, or zero if it cannot be
4718identified. The value returned by this intrinsic is likely to be incorrect or 0
4719for arguments other than zero, so it should only be used for debugging purposes.
4720</p>
4721
4722<p>
4723Note that calling this intrinsic does not prevent function inlining or other
4724aggressive transformations, so the value returned may not be that of the obvious
4725source-language caller.
4726</p>
4727</div>
4728
4729<!-- _______________________________________________________________________ -->
4730<div class="doc_subsubsection">
4731 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
4732</div>
4733
4734<div class="doc_text">
4735
4736<h5>Syntax:</h5>
4737<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4738 declare i8 *@llvm.stacksave()
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4739</pre>
4740
4741<h5>Overview:</h5>
4742
4743<p>
4744The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
4745the function stack, for use with <a href="#int_stackrestore">
4746<tt>llvm.stackrestore</tt></a>. This is useful for implementing language
4747features like scoped automatic variable sized arrays in C99.
4748</p>
4749
4750<h5>Semantics:</h5>
4751
4752<p>
4753This intrinsic returns a opaque pointer value that can be passed to <a
4754href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
4755<tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
4756<tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
4757state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
4758practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
4759that were allocated after the <tt>llvm.stacksave</tt> was executed.
4760</p>
4761
4762</div>
4763
4764<!-- _______________________________________________________________________ -->
4765<div class="doc_subsubsection">
4766 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
4767</div>
4768
4769<div class="doc_text">
4770
4771<h5>Syntax:</h5>
4772<pre>
4773 declare void @llvm.stackrestore(i8 * %ptr)
4774</pre>
4775
4776<h5>Overview:</h5>
4777
4778<p>
4779The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
4780the function stack to the state it was in when the corresponding <a
4781href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
4782useful for implementing language features like scoped automatic variable sized
4783arrays in C99.
4784</p>
4785
4786<h5>Semantics:</h5>
4787
4788<p>
4789See the description for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.
4790</p>
4791
4792</div>
4793
4794
4795<!-- _______________________________________________________________________ -->
4796<div class="doc_subsubsection">
4797 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
4798</div>
4799
4800<div class="doc_text">
4801
4802<h5>Syntax:</h5>
4803<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4804 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]4805</pre>
4806
4807<h5>Overview:</h5>
4808
4809
4810<p>
4811The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
4812a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
4813no
4814effect on the behavior of the program but can change its performance
4815characteristics.
4816</p>
4817
4818<h5>Arguments:</h5>
4819
4820<p>
4821<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
4822determining if the fetch should be for a read (0) or write (1), and
4823<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
4824locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
4825<tt>locality</tt> arguments must be constant integers.
4826</p>
4827
4828<h5>Semantics:</h5>
4829
4830<p>
4831This intrinsic does not modify the behavior of the program. In particular,
4832prefetches cannot trap and do not produce a value. On targets that support this
4833intrinsic, the prefetch can provide hints to the processor cache for better
4834performance.
4835</p>
4836
4837</div>
4838
4839<!-- _______________________________________________________________________ -->
4840<div class="doc_subsubsection">
4841 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
4842</div>
4843
4844<div class="doc_text">
4845
4846<h5>Syntax:</h5>
4847<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4848 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4849</pre>
4850
4851<h5>Overview:</h5>
4852
4853
4854<p>
4855The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
4856(PC) in a region of
4857code to simulators and other tools. The method is target specific, but it is
4858expected that the marker will use exported symbols to transmit the PC of the marker.
4859The marker makes no guarantees that it will remain with any specific instruction
4860after optimizations. It is possible that the presence of a marker will inhibit
4861optimizations. The intended use is to be inserted after optimizations to allow
4862correlations of simulation runs.
4863</p>
4864
4865<h5>Arguments:</h5>
4866
4867<p>
4868<tt>id</tt> is a numerical id identifying the marker.
4869</p>
4870
4871<h5>Semantics:</h5>
4872
4873<p>
4874This intrinsic does not modify the behavior of the program. Backends that do not
4875support this intrinisic may ignore it.
4876</p>
4877
4878</div>
4879
4880<!-- _______________________________________________________________________ -->
4881<div class="doc_subsubsection">
4882 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
4883</div>
4884
4885<div class="doc_text">
4886
4887<h5>Syntax:</h5>
4888<pre>
4889 declare i64 @llvm.readcyclecounter( )
4890</pre>
4891
4892<h5>Overview:</h5>
4893
4894
4895<p>
4896The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
4897counter register (or similar low latency, high accuracy clocks) on those targets
4898that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
4899As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
4900should only be used for small timings.
4901</p>
4902
4903<h5>Semantics:</h5>
4904
4905<p>
4906When directly supported, reading the cycle counter should not modify any memory.
4907Implementations are allowed to either return a application specific value or a
4908system wide value. On backends without support, this is lowered to a constant 0.
4909</p>
4910
4911</div>
4912
4913<!-- ======================================================================= -->
4914<div class="doc_subsection">
4915 <a name="int_libc">Standard C Library Intrinsics</a>
4916</div>
4917
4918<div class="doc_text">
4919<p>
4920LLVM provides intrinsics for a few important standard C library functions.
4921These intrinsics allow source-language front-ends to pass information about the
4922alignment of the pointer arguments to the code generator, providing opportunity
4923for more efficient code generation.
4924</p>
4925
4926</div>
4927
4928<!-- _______________________________________________________________________ -->
4929<div class="doc_subsubsection">
4930 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
4931</div>
4932
4933<div class="doc_text">
4934
4935<h5>Syntax:</h5>
4936<pre>
4937 declare void @llvm.memcpy.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
4938 i32 &lt;len&gt;, i32 &lt;align&gt;)
4939 declare void @llvm.memcpy.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
4940 i64 &lt;len&gt;, i32 &lt;align&gt;)
4941</pre>
4942
4943<h5>Overview:</h5>
4944
4945<p>
4946The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
4947location to the destination location.
4948</p>
4949
4950<p>
4951Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
4952intrinsics do not return a value, and takes an extra alignment argument.
4953</p>
4954
4955<h5>Arguments:</h5>
4956
4957<p>
4958The first argument is a pointer to the destination, the second is a pointer to
4959the source. The third argument is an integer argument
4960specifying the number of bytes to copy, and the fourth argument is the alignment
4961of the source and destination locations.
4962</p>
4963
4964<p>
4965If the call to this intrinisic has an alignment value that is not 0 or 1, then
4966the caller guarantees that both the source and destination pointers are aligned
4967to that boundary.
4968</p>
4969
4970<h5>Semantics:</h5>
4971
4972<p>
4973The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
4974location to the destination location, which are not allowed to overlap. It
4975copies "len" bytes of memory over. If the argument is known to be aligned to
4976some boundary, this can be specified as the fourth argument, otherwise it should
4977be set to 0 or 1.
4978</p>
4979</div>
4980
4981
4982<!-- _______________________________________________________________________ -->
4983<div class="doc_subsubsection">
4984 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
4985</div>
4986
4987<div class="doc_text">
4988
4989<h5>Syntax:</h5>
4990<pre>
4991 declare void @llvm.memmove.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
4992 i32 &lt;len&gt;, i32 &lt;align&gt;)
4993 declare void @llvm.memmove.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
4994 i64 &lt;len&gt;, i32 &lt;align&gt;)
4995</pre>
4996
4997<h5>Overview:</h5>
4998
4999<p>
5000The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the source
5001location to the destination location. It is similar to the
Chris Lattnerdba16ea2008-01-06 19:51:52 +0000[diff] [blame]5002'<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to overlap.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5003</p>
5004
5005<p>
5006Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
5007intrinsics do not return a value, and takes an extra alignment argument.
5008</p>
5009
5010<h5>Arguments:</h5>
5011
5012<p>
5013The first argument is a pointer to the destination, the second is a pointer to
5014the source. The third argument is an integer argument
5015specifying the number of bytes to copy, and the fourth argument is the alignment
5016of the source and destination locations.
5017</p>
5018
5019<p>
5020If the call to this intrinisic has an alignment value that is not 0 or 1, then
5021the caller guarantees that the source and destination pointers are aligned to
5022that boundary.
5023</p>
5024
5025<h5>Semantics:</h5>
5026
5027<p>
5028The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the source
5029location to the destination location, which may overlap. It
5030copies "len" bytes of memory over. If the argument is known to be aligned to
5031some boundary, this can be specified as the fourth argument, otherwise it should
5032be set to 0 or 1.
5033</p>
5034</div>
5035
5036
5037<!-- _______________________________________________________________________ -->
5038<div class="doc_subsubsection">
5039 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
5040</div>
5041
5042<div class="doc_text">
5043
5044<h5>Syntax:</h5>
5045<pre>
5046 declare void @llvm.memset.i32(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5047 i32 &lt;len&gt;, i32 &lt;align&gt;)
5048 declare void @llvm.memset.i64(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5049 i64 &lt;len&gt;, i32 &lt;align&gt;)
5050</pre>
5051
5052<h5>Overview:</h5>
5053
5054<p>
5055The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a particular
5056byte value.
5057</p>
5058
5059<p>
5060Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
5061does not return a value, and takes an extra alignment argument.
5062</p>
5063
5064<h5>Arguments:</h5>
5065
5066<p>
5067The first argument is a pointer to the destination to fill, the second is the
5068byte value to fill it with, the third argument is an integer
5069argument specifying the number of bytes to fill, and the fourth argument is the
5070known alignment of destination location.
5071</p>
5072
5073<p>
5074If the call to this intrinisic has an alignment value that is not 0 or 1, then
5075the caller guarantees that the destination pointer is aligned to that boundary.
5076</p>
5077
5078<h5>Semantics:</h5>
5079
5080<p>
5081The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting at
5082the
5083destination location. If the argument is known to be aligned to some boundary,
5084this can be specified as the fourth argument, otherwise it should be set to 0 or
50851.
5086</p>
5087</div>
5088
5089
5090<!-- _______________________________________________________________________ -->
5091<div class="doc_subsubsection">
5092 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
5093</div>
5094
5095<div class="doc_text">
5096
5097<h5>Syntax:</h5>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5098<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]5099floating point or vector of floating point type. Not all targets support all
5100types however.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5101<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5102 declare float @llvm.sqrt.f32(float %Val)
5103 declare double @llvm.sqrt.f64(double %Val)
5104 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
5105 declare fp128 @llvm.sqrt.f128(fp128 %Val)
5106 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5107</pre>
5108
5109<h5>Overview:</h5>
5110
5111<p>
5112The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5113returning the same value as the libm '<tt>sqrt</tt>' functions would. Unlike
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5114<tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
Chris Lattnerf914a002008-01-29 07:00:44 +0000[diff] [blame]5115negative numbers other than -0.0 (which allows for better optimization, because
5116there is no need to worry about errno being set). <tt>llvm.sqrt(-0.0)</tt> is
5117defined to return -0.0 like IEEE sqrt.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5118</p>
5119
5120<h5>Arguments:</h5>
5121
5122<p>
5123The argument and return value are floating point numbers of the same type.
5124</p>
5125
5126<h5>Semantics:</h5>
5127
5128<p>
5129This function returns the sqrt of the specified operand if it is a nonnegative
5130floating point number.
5131</p>
5132</div>
5133
5134<!-- _______________________________________________________________________ -->
5135<div class="doc_subsubsection">
5136 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
5137</div>
5138
5139<div class="doc_text">
5140
5141<h5>Syntax:</h5>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5142<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]5143floating point or vector of floating point type. Not all targets support all
5144types however.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5145<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]5146 declare float @llvm.powi.f32(float %Val, i32 %power)
5147 declare double @llvm.powi.f64(double %Val, i32 %power)
5148 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
5149 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
5150 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5151</pre>
5152
5153<h5>Overview:</h5>
5154
5155<p>
5156The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
5157specified (positive or negative) power. The order of evaluation of
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5158multiplications is not defined. When a vector of floating point type is
5159used, the second argument remains a scalar integer value.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5160</p>
5161
5162<h5>Arguments:</h5>
5163
5164<p>
5165The second argument is an integer power, and the first is a value to raise to
5166that power.
5167</p>
5168
5169<h5>Semantics:</h5>
5170
5171<p>
5172This function returns the first value raised to the second power with an
5173unspecified sequence of rounding operations.</p>
5174</div>
5175
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5176<!-- _______________________________________________________________________ -->
5177<div class="doc_subsubsection">
5178 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
5179</div>
5180
5181<div class="doc_text">
5182
5183<h5>Syntax:</h5>
5184<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
5185floating point or vector of floating point type. Not all targets support all
5186types however.
5187<pre>
5188 declare float @llvm.sin.f32(float %Val)
5189 declare double @llvm.sin.f64(double %Val)
5190 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
5191 declare fp128 @llvm.sin.f128(fp128 %Val)
5192 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
5193</pre>
5194
5195<h5>Overview:</h5>
5196
5197<p>
5198The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.
5199</p>
5200
5201<h5>Arguments:</h5>
5202
5203<p>
5204The argument and return value are floating point numbers of the same type.
5205</p>
5206
5207<h5>Semantics:</h5>
5208
5209<p>
5210This function returns the sine of the specified operand, returning the
5211same values as the libm <tt>sin</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5212conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5213</div>
5214
5215<!-- _______________________________________________________________________ -->
5216<div class="doc_subsubsection">
5217 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
5218</div>
5219
5220<div class="doc_text">
5221
5222<h5>Syntax:</h5>
5223<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
5224floating point or vector of floating point type. Not all targets support all
5225types however.
5226<pre>
5227 declare float @llvm.cos.f32(float %Val)
5228 declare double @llvm.cos.f64(double %Val)
5229 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
5230 declare fp128 @llvm.cos.f128(fp128 %Val)
5231 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
5232</pre>
5233
5234<h5>Overview:</h5>
5235
5236<p>
5237The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.
5238</p>
5239
5240<h5>Arguments:</h5>
5241
5242<p>
5243The argument and return value are floating point numbers of the same type.
5244</p>
5245
5246<h5>Semantics:</h5>
5247
5248<p>
5249This function returns the cosine of the specified operand, returning the
5250same values as the libm <tt>cos</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5251conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5252</div>
5253
5254<!-- _______________________________________________________________________ -->
5255<div class="doc_subsubsection">
5256 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
5257</div>
5258
5259<div class="doc_text">
5260
5261<h5>Syntax:</h5>
5262<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
5263floating point or vector of floating point type. Not all targets support all
5264types however.
5265<pre>
5266 declare float @llvm.pow.f32(float %Val, float %Power)
5267 declare double @llvm.pow.f64(double %Val, double %Power)
5268 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
5269 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
5270 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
5271</pre>
5272
5273<h5>Overview:</h5>
5274
5275<p>
5276The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
5277specified (positive or negative) power.
5278</p>
5279
5280<h5>Arguments:</h5>
5281
5282<p>
5283The second argument is a floating point power, and the first is a value to
5284raise to that power.
5285</p>
5286
5287<h5>Semantics:</h5>
5288
5289<p>
5290This function returns the first value raised to the second power,
5291returning the
5292same values as the libm <tt>pow</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]5293conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]5294</div>
5295
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5296
5297<!-- ======================================================================= -->
5298<div class="doc_subsection">
5299 <a name="int_manip">Bit Manipulation Intrinsics</a>
5300</div>
5301
5302<div class="doc_text">
5303<p>
5304LLVM provides intrinsics for a few important bit manipulation operations.
5305These allow efficient code generation for some algorithms.
5306</p>
5307
5308</div>
5309
5310<!-- _______________________________________________________________________ -->
5311<div class="doc_subsubsection">
5312 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
5313</div>
5314
5315<div class="doc_text">
5316
5317<h5>Syntax:</h5>
5318<p>This is an overloaded intrinsic function. You can use bswap on any integer
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5319type that is an even number of bytes (i.e. BitWidth % 16 == 0).
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5320<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5321 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
5322 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
5323 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5324</pre>
5325
5326<h5>Overview:</h5>
5327
5328<p>
5329The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
5330values with an even number of bytes (positive multiple of 16 bits). These are
5331useful for performing operations on data that is not in the target's native
5332byte order.
5333</p>
5334
5335<h5>Semantics:</h5>
5336
5337<p>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5338The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5339and low byte of the input i16 swapped. Similarly, the <tt>llvm.bswap.i32</tt>
5340intrinsic returns an i32 value that has the four bytes of the input i32
5341swapped, 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]5342i32 will have its bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i48</tt>,
5343<tt>llvm.bswap.i64</tt> and other intrinsics extend this concept to
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5344additional even-byte lengths (6 bytes, 8 bytes and more, respectively).
5345</p>
5346
5347</div>
5348
5349<!-- _______________________________________________________________________ -->
5350<div class="doc_subsubsection">
5351 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
5352</div>
5353
5354<div class="doc_text">
5355
5356<h5>Syntax:</h5>
5357<p>This is an overloaded intrinsic. You can use llvm.ctpop on any integer bit
5358width. Not all targets support all bit widths however.
5359<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5360 declare i8 @llvm.ctpop.i8 (i8 &lt;src&gt;)
5361 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5362 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5363 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
5364 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5365</pre>
5366
5367<h5>Overview:</h5>
5368
5369<p>
5370The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
5371value.
5372</p>
5373
5374<h5>Arguments:</h5>
5375
5376<p>
5377The only argument is the value to be counted. The argument may be of any
5378integer type. The return type must match the argument type.
5379</p>
5380
5381<h5>Semantics:</h5>
5382
5383<p>
5384The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
5385</p>
5386</div>
5387
5388<!-- _______________________________________________________________________ -->
5389<div class="doc_subsubsection">
5390 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
5391</div>
5392
5393<div class="doc_text">
5394
5395<h5>Syntax:</h5>
5396<p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
5397integer bit width. Not all targets support all bit widths however.
5398<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5399 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
5400 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5401 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5402 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
5403 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5404</pre>
5405
5406<h5>Overview:</h5>
5407
5408<p>
5409The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
5410leading zeros in a variable.
5411</p>
5412
5413<h5>Arguments:</h5>
5414
5415<p>
5416The only argument is the value to be counted. The argument may be of any
5417integer type. The return type must match the argument type.
5418</p>
5419
5420<h5>Semantics:</h5>
5421
5422<p>
5423The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
5424in a variable. If the src == 0 then the result is the size in bits of the type
5425of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.
5426</p>
5427</div>
5428
5429
5430
5431<!-- _______________________________________________________________________ -->
5432<div class="doc_subsubsection">
5433 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
5434</div>
5435
5436<div class="doc_text">
5437
5438<h5>Syntax:</h5>
5439<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
5440integer bit width. Not all targets support all bit widths however.
5441<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5442 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
5443 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5444 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5445 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
5446 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5447</pre>
5448
5449<h5>Overview:</h5>
5450
5451<p>
5452The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
5453trailing zeros.
5454</p>
5455
5456<h5>Arguments:</h5>
5457
5458<p>
5459The only argument is the value to be counted. The argument may be of any
5460integer type. The return type must match the argument type.
5461</p>
5462
5463<h5>Semantics:</h5>
5464
5465<p>
5466The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
5467in a variable. If the src == 0 then the result is the size in bits of the type
5468of src. For example, <tt>llvm.cttz(2) = 1</tt>.
5469</p>
5470</div>
5471
5472<!-- _______________________________________________________________________ -->
5473<div class="doc_subsubsection">
5474 <a name="int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic</a>
5475</div>
5476
5477<div class="doc_text">
5478
5479<h5>Syntax:</h5>
5480<p>This is an overloaded intrinsic. You can use <tt>llvm.part.select</tt>
5481on any integer bit width.
5482<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5483 declare i17 @llvm.part.select.i17 (i17 %val, i32 %loBit, i32 %hiBit)
5484 declare i29 @llvm.part.select.i29 (i29 %val, i32 %loBit, i32 %hiBit)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5485</pre>
5486
5487<h5>Overview:</h5>
5488<p>The '<tt>llvm.part.select</tt>' family of intrinsic functions selects a
5489range of bits from an integer value and returns them in the same bit width as
5490the original value.</p>
5491
5492<h5>Arguments:</h5>
5493<p>The first argument, <tt>%val</tt> and the result may be integer types of
5494any bit width but they must have the same bit width. The second and third
5495arguments must be <tt>i32</tt> type since they specify only a bit index.</p>
5496
5497<h5>Semantics:</h5>
5498<p>The operation of the '<tt>llvm.part.select</tt>' intrinsic has two modes
5499of operation: forwards and reverse. If <tt>%loBit</tt> is greater than
5500<tt>%hiBits</tt> then the intrinsic operates in reverse mode. Otherwise it
5501operates in forward mode.</p>
5502<p>In forward mode, this intrinsic is the equivalent of shifting <tt>%val</tt>
5503right by <tt>%loBit</tt> bits and then ANDing it with a mask with
5504only the <tt>%hiBit - %loBit</tt> bits set, as follows:</p>
5505<ol>
5506 <li>The <tt>%val</tt> is shifted right (LSHR) by the number of bits specified
5507 by <tt>%loBits</tt>. This normalizes the value to the low order bits.</li>
5508 <li>The <tt>%loBits</tt> value is subtracted from the <tt>%hiBits</tt> value
5509 to determine the number of bits to retain.</li>
5510 <li>A mask of the retained bits is created by shifting a -1 value.</li>
5511 <li>The mask is ANDed with <tt>%val</tt> to produce the result.
5512</ol>
5513<p>In reverse mode, a similar computation is made except that the bits are
5514returned in the reverse order. So, for example, if <tt>X</tt> has the value
5515<tt>i16 0x0ACF (101011001111)</tt> and we apply
5516<tt>part.select(i16 X, 8, 3)</tt> to it, we get back the value
5517<tt>i16 0x0026 (000000100110)</tt>.</p>
5518</div>
5519
5520<div class="doc_subsubsection">
5521 <a name="int_part_set">'<tt>llvm.part.set.*</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.part.set</tt>
5528on any integer bit width.
5529<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5530 declare i17 @llvm.part.set.i17.i9 (i17 %val, i9 %repl, i32 %lo, i32 %hi)
5531 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]5532</pre>
5533
5534<h5>Overview:</h5>
5535<p>The '<tt>llvm.part.set</tt>' family of intrinsic functions replaces a range
5536of bits in an integer value with another integer value. It returns the integer
5537with the replaced bits.</p>
5538
5539<h5>Arguments:</h5>
5540<p>The first argument, <tt>%val</tt> and the result may be integer types of
5541any bit width but they must have the same bit width. <tt>%val</tt> is the value
5542whose bits will be replaced. The second argument, <tt>%repl</tt> may be an
5543integer of any bit width. The third and fourth arguments must be <tt>i32</tt>
5544type since they specify only a bit index.</p>
5545
5546<h5>Semantics:</h5>
5547<p>The operation of the '<tt>llvm.part.set</tt>' intrinsic has two modes
5548of operation: forwards and reverse. If <tt>%lo</tt> is greater than
5549<tt>%hi</tt> then the intrinsic operates in reverse mode. Otherwise it
5550operates in forward mode.</p>
5551<p>For both modes, the <tt>%repl</tt> value is prepared for use by either
5552truncating it down to the size of the replacement area or zero extending it
5553up to that size.</p>
5554<p>In forward mode, the bits between <tt>%lo</tt> and <tt>%hi</tt> (inclusive)
5555are replaced with corresponding bits from <tt>%repl</tt>. That is the 0th bit
5556in <tt>%repl</tt> replaces the <tt>%lo</tt>th bit in <tt>%val</tt> and etc. up
5557to the <tt>%hi</tt>th bit.
5558<p>In reverse mode, a similar computation is made except that the bits are
5559reversed. That is, the <tt>0</tt>th bit in <tt>%repl</tt> replaces the
5560<tt>%hi</tt> bit in <tt>%val</tt> and etc. down to the <tt>%lo</tt>th bit.
5561<h5>Examples:</h5>
5562<pre>
5563 llvm.part.set(0xFFFF, 0, 4, 7) -&gt; 0xFF0F
5564 llvm.part.set(0xFFFF, 0, 7, 4) -&gt; 0xFF0F
5565 llvm.part.set(0xFFFF, 1, 7, 4) -&gt; 0xFF8F
5566 llvm.part.set(0xFFFF, F, 8, 3) -&gt; 0xFFE7
5567 llvm.part.set(0xFFFF, 0, 3, 8) -&gt; 0xFE07
5568</pre>
5569</div>
5570
5571<!-- ======================================================================= -->
5572<div class="doc_subsection">
5573 <a name="int_debugger">Debugger Intrinsics</a>
5574</div>
5575
5576<div class="doc_text">
5577<p>
5578The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
5579are described in the <a
5580href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
5581Debugging</a> document.
5582</p>
5583</div>
5584
5585
5586<!-- ======================================================================= -->
5587<div class="doc_subsection">
5588 <a name="int_eh">Exception Handling Intrinsics</a>
5589</div>
5590
5591<div class="doc_text">
5592<p> The LLVM exception handling intrinsics (which all start with
5593<tt>llvm.eh.</tt> prefix), are described in the <a
5594href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
5595Handling</a> document. </p>
5596</div>
5597
5598<!-- ======================================================================= -->
5599<div class="doc_subsection">
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5600 <a name="int_trampoline">Trampoline Intrinsic</a>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5601</div>
5602
5603<div class="doc_text">
5604<p>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5605 This intrinsic makes it possible to excise one parameter, marked with
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5606 the <tt>nest</tt> attribute, from a function. The result is a callable
5607 function pointer lacking the nest parameter - the caller does not need
5608 to provide a value for it. Instead, the value to use is stored in
5609 advance in a "trampoline", a block of memory usually allocated
5610 on the stack, which also contains code to splice the nest value into the
5611 argument list. This is used to implement the GCC nested function address
5612 extension.
5613</p>
5614<p>
5615 For example, if the function is
5616 <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]5617 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]5618<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5619 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
5620 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
5621 %p = call i8* @llvm.init.trampoline( i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8*), i8* %nval )
5622 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5623</pre>
Bill Wendlinge262b4c2007-09-22 09:23:55 +0000[diff] [blame]5624 <p>The call <tt>%val = call i32 %fp( i32 %x, i32 %y )</tt> is then equivalent
5625 to <tt>%val = call i32 %f( i8* %nval, i32 %x, i32 %y )</tt>.</p>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5626</div>
5627
5628<!-- _______________________________________________________________________ -->
5629<div class="doc_subsubsection">
5630 <a name="int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a>
5631</div>
5632<div class="doc_text">
5633<h5>Syntax:</h5>
5634<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5635declare 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]5636</pre>
5637<h5>Overview:</h5>
5638<p>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5639 This fills the memory pointed to by <tt>tramp</tt> with code
5640 and returns a function pointer suitable for executing it.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5641</p>
5642<h5>Arguments:</h5>
5643<p>
5644 The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
5645 pointers. The <tt>tramp</tt> argument must point to a sufficiently large
5646 and sufficiently aligned block of memory; this memory is written to by the
Duncan Sands35012212007-08-22 23:39:54 +0000[diff] [blame]5647 intrinsic. Note that the size and the alignment are target-specific - LLVM
5648 currently provides no portable way of determining them, so a front-end that
5649 generates this intrinsic needs to have some target-specific knowledge.
5650 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]5651</p>
5652<h5>Semantics:</h5>
5653<p>
5654 The block of memory pointed to by <tt>tramp</tt> is filled with target
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5655 dependent code, turning it into a function. A pointer to this function is
5656 returned, but needs to be bitcast to an
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5657 <a href="#int_trampoline">appropriate function pointer type</a>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5658 before being called. The new function's signature is the same as that of
5659 <tt>func</tt> with any arguments marked with the <tt>nest</tt> attribute
5660 removed. At most one such <tt>nest</tt> argument is allowed, and it must be
5661 of pointer type. Calling the new function is equivalent to calling
5662 <tt>func</tt> with the same argument list, but with <tt>nval</tt> used for the
5663 missing <tt>nest</tt> argument. If, after calling
5664 <tt>llvm.init.trampoline</tt>, the memory pointed to by <tt>tramp</tt> is
5665 modified, then the effect of any later call to the returned function pointer is
5666 undefined.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5667</p>
5668</div>
5669
5670<!-- ======================================================================= -->
5671<div class="doc_subsection">
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]5672 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
5673</div>
5674
5675<div class="doc_text">
5676<p>
5677 These intrinsic functions expand the "universal IR" of LLVM to represent
5678 hardware constructs for atomic operations and memory synchronization. This
5679 provides an interface to the hardware, not an interface to the programmer. It
5680 is aimed at a low enough level to allow any programming models or APIs which
5681 need atomic behaviors to map cleanly onto it. It is also modeled primarily on
5682 hardware behavior. Just as hardware provides a "universal IR" for source
5683 languages, it also provides a starting point for developing a "universal"
5684 atomic operation and synchronization IR.
5685</p>
5686<p>
5687 These do <em>not</em> form an API such as high-level threading libraries,
5688 software transaction memory systems, atomic primitives, and intrinsic
5689 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
5690 application libraries. The hardware interface provided by LLVM should allow
5691 a clean implementation of all of these APIs and parallel programming models.
5692 No one model or paradigm should be selected above others unless the hardware
5693 itself ubiquitously does so.
5694
5695</p>
5696</div>
5697
5698<!-- _______________________________________________________________________ -->
5699<div class="doc_subsubsection">
5700 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
5701</div>
5702<div class="doc_text">
5703<h5>Syntax:</h5>
5704<pre>
5705declare void @llvm.memory.barrier( i1 &lt;ll&gt;, i1 &lt;ls&gt;, i1 &lt;sl&gt;, i1 &lt;ss&gt;,
5706i1 &lt;device&gt; )
5707
5708</pre>
5709<h5>Overview:</h5>
5710<p>
5711 The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
5712 specific pairs of memory access types.
5713</p>
5714<h5>Arguments:</h5>
5715<p>
5716 The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
5717 The first four arguments enables a specific barrier as listed below. The fith
5718 argument specifies that the barrier applies to io or device or uncached memory.
5719
5720</p>
5721 <ul>
5722 <li><tt>ll</tt>: load-load barrier</li>
5723 <li><tt>ls</tt>: load-store barrier</li>
5724 <li><tt>sl</tt>: store-load barrier</li>
5725 <li><tt>ss</tt>: store-store barrier</li>
5726 <li><tt>device</tt>: barrier applies to device and uncached memory also.
5727 </ul>
5728<h5>Semantics:</h5>
5729<p>
5730 This intrinsic causes the system to enforce some ordering constraints upon
5731 the loads and stores of the program. This barrier does not indicate
5732 <em>when</em> any events will occur, it only enforces an <em>order</em> in
5733 which they occur. For any of the specified pairs of load and store operations
5734 (f.ex. load-load, or store-load), all of the first operations preceding the
5735 barrier will complete before any of the second operations succeeding the
5736 barrier begin. Specifically the semantics for each pairing is as follows:
5737</p>
5738 <ul>
5739 <li><tt>ll</tt>: All loads before the barrier must complete before any load
5740 after the barrier begins.</li>
5741
5742 <li><tt>ls</tt>: All loads before the barrier must complete before any
5743 store after the barrier begins.</li>
5744 <li><tt>ss</tt>: All stores before the barrier must complete before any
5745 store after the barrier begins.</li>
5746 <li><tt>sl</tt>: All stores before the barrier must complete before any
5747 load after the barrier begins.</li>
5748 </ul>
5749<p>
5750 These semantics are applied with a logical "and" behavior when more than one
5751 is enabled in a single memory barrier intrinsic.
5752</p>
5753<p>
5754 Backends may implement stronger barriers than those requested when they do not
5755 support as fine grained a barrier as requested. Some architectures do not
5756 need all types of barriers and on such architectures, these become noops.
5757</p>
5758<h5>Example:</h5>
5759<pre>
5760%ptr = malloc i32
5761 store i32 4, %ptr
5762
5763%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
5764 call void @llvm.memory.barrier( i1 false, i1 true, i1 false, i1 false )
5765 <i>; guarantee the above finishes</i>
5766 store i32 8, %ptr <i>; before this begins</i>
5767</pre>
5768</div>
5769
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5770<!-- _______________________________________________________________________ -->
5771<div class="doc_subsubsection">
5772 <a name="int_atomic_lcs">'<tt>llvm.atomic.lcs.*</tt>' Intrinsic</a>
5773</div>
5774<div class="doc_text">
5775<h5>Syntax:</h5>
5776<p>
5777 This is an overloaded intrinsic. You can use <tt>llvm.atomic.lcs</tt> on any
5778 integer bit width. Not all targets support all bit widths however.</p>
5779
5780<pre>
5781declare i8 @llvm.atomic.lcs.i8( i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt; )
5782declare i16 @llvm.atomic.lcs.i16( i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt; )
5783declare i32 @llvm.atomic.lcs.i32( i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt; )
5784declare i64 @llvm.atomic.lcs.i64( i64* &lt;ptr&gt;, i64 &lt;cmp&gt;, i64 &lt;val&gt; )
5785
5786</pre>
5787<h5>Overview:</h5>
5788<p>
5789 This loads a value in memory and compares it to a given value. If they are
5790 equal, it stores a new value into the memory.
5791</p>
5792<h5>Arguments:</h5>
5793<p>
5794 The <tt>llvm.atomic.lcs</tt> intrinsic takes three arguments. The result as
5795 well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
5796 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
5797 this integer type. While any bit width integer may be used, targets may only
5798 lower representations they support in hardware.
5799
5800</p>
5801<h5>Semantics:</h5>
5802<p>
5803 This entire intrinsic must be executed atomically. It first loads the value
5804 in memory pointed to by <tt>ptr</tt> and compares it with the value
5805 <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the memory. The
5806 loaded value is yielded in all cases. This provides the equivalent of an
5807 atomic compare-and-swap operation within the SSA framework.
5808</p>
5809<h5>Examples:</h5>
5810
5811<pre>
5812%ptr = malloc i32
5813 store i32 4, %ptr
5814
5815%val1 = add i32 4, 4
5816%result1 = call i32 @llvm.atomic.lcs.i32( i32* %ptr, i32 4, %val1 )
5817 <i>; yields {i32}:result1 = 4</i>
5818%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
5819%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
5820
5821%val2 = add i32 1, 1
5822%result2 = call i32 @llvm.atomic.lcs.i32( i32* %ptr, i32 5, %val2 )
5823 <i>; yields {i32}:result2 = 8</i>
5824%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
5825
5826%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
5827</pre>
5828</div>
5829
5830<!-- _______________________________________________________________________ -->
5831<div class="doc_subsubsection">
5832 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
5833</div>
5834<div class="doc_text">
5835<h5>Syntax:</h5>
5836
5837<p>
5838 This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
5839 integer bit width. Not all targets support all bit widths however.</p>
5840<pre>
5841declare i8 @llvm.atomic.swap.i8( i8* &lt;ptr&gt;, i8 &lt;val&gt; )
5842declare i16 @llvm.atomic.swap.i16( i16* &lt;ptr&gt;, i16 &lt;val&gt; )
5843declare i32 @llvm.atomic.swap.i32( i32* &lt;ptr&gt;, i32 &lt;val&gt; )
5844declare i64 @llvm.atomic.swap.i64( i64* &lt;ptr&gt;, i64 &lt;val&gt; )
5845
5846</pre>
5847<h5>Overview:</h5>
5848<p>
5849 This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
5850 the value from memory. It then stores the value in <tt>val</tt> in the memory
5851 at <tt>ptr</tt>.
5852</p>
5853<h5>Arguments:</h5>
5854
5855<p>
5856 The <tt>llvm.atomic.ls</tt> intrinsic takes two arguments. Both the
5857 <tt>val</tt> argument and the result must be integers of the same bit width.
5858 The first argument, <tt>ptr</tt>, must be a pointer to a value of this
5859 integer type. The targets may only lower integer representations they
5860 support.
5861</p>
5862<h5>Semantics:</h5>
5863<p>
5864 This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
5865 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
5866 equivalent of an atomic swap operation within the SSA framework.
5867
5868</p>
5869<h5>Examples:</h5>
5870<pre>
5871%ptr = malloc i32
5872 store i32 4, %ptr
5873
5874%val1 = add i32 4, 4
5875%result1 = call i32 @llvm.atomic.swap.i32( i32* %ptr, i32 %val1 )
5876 <i>; yields {i32}:result1 = 4</i>
5877%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
5878%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
5879
5880%val2 = add i32 1, 1
5881%result2 = call i32 @llvm.atomic.swap.i32( i32* %ptr, i32 %val2 )
5882 <i>; yields {i32}:result2 = 8</i>
5883
5884%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
5885%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
5886</pre>
5887</div>
5888
5889<!-- _______________________________________________________________________ -->
5890<div class="doc_subsubsection">
5891 <a name="int_atomic_las">'<tt>llvm.atomic.las.*</tt>' Intrinsic</a>
5892
5893</div>
5894<div class="doc_text">
5895<h5>Syntax:</h5>
5896<p>
5897 This is an overloaded intrinsic. You can use <tt>llvm.atomic.las</tt> on any
5898 integer bit width. Not all targets support all bit widths however.</p>
5899<pre>
5900declare i8 @llvm.atomic.las.i8.( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
5901declare i16 @llvm.atomic.las.i16.( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
5902declare i32 @llvm.atomic.las.i32.( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
5903declare i64 @llvm.atomic.las.i64.( i64* &lt;ptr&gt;, i64 &lt;delta&gt; )
5904
5905</pre>
5906<h5>Overview:</h5>
5907<p>
5908 This intrinsic adds <tt>delta</tt> to the value stored in memory at
5909 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
5910</p>
5911<h5>Arguments:</h5>
5912<p>
5913
5914 The intrinsic takes two arguments, the first a pointer to an integer value
5915 and the second an integer value. The result is also an integer value. These
5916 integer types can have any bit width, but they must all have the same bit
5917 width. The targets may only lower integer representations they support.
5918</p>
5919<h5>Semantics:</h5>
5920<p>
5921 This intrinsic does a series of operations atomically. It first loads the
5922 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
5923 to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
5924</p>
5925
5926<h5>Examples:</h5>
5927<pre>
5928%ptr = malloc i32
5929 store i32 4, %ptr
5930%result1 = call i32 @llvm.atomic.las.i32( i32* %ptr, i32 4 )
5931 <i>; yields {i32}:result1 = 4</i>
5932%result2 = call i32 @llvm.atomic.las.i32( i32* %ptr, i32 2 )
5933 <i>; yields {i32}:result2 = 8</i>
5934%result3 = call i32 @llvm.atomic.las.i32( i32* %ptr, i32 5 )
5935 <i>; yields {i32}:result3 = 10</i>
5936%memval = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
5937</pre>
5938</div>
5939
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]5940
5941<!-- ======================================================================= -->
5942<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5943 <a name="int_general">General Intrinsics</a>
5944</div>
5945
5946<div class="doc_text">
5947<p> This class of intrinsics is designed to be generic and has
5948no specific purpose. </p>
5949</div>
5950
5951<!-- _______________________________________________________________________ -->
5952<div class="doc_subsubsection">
5953 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
5954</div>
5955
5956<div class="doc_text">
5957
5958<h5>Syntax:</h5>
5959<pre>
5960 declare void @llvm.var.annotation(i8* &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
5961</pre>
5962
5963<h5>Overview:</h5>
5964
5965<p>
5966The '<tt>llvm.var.annotation</tt>' intrinsic
5967</p>
5968
5969<h5>Arguments:</h5>
5970
5971<p>
5972The first argument is a pointer to a value, the second is a pointer to a
5973global string, the third is a pointer to a global string which is the source
5974file name, and the last argument is the line number.
5975</p>
5976
5977<h5>Semantics:</h5>
5978
5979<p>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]5980This intrinsic allows annotation of local variables with arbitrary strings.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5981This can be useful for special purpose optimizations that want to look for these
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]5982annotations. These have no other defined use, they are ignored by code
5983generation and optimization.
5984</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5985</div>
5986
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]5987<!-- _______________________________________________________________________ -->
5988<div class="doc_subsubsection">
Tanya Lattnerc9869b12007-09-21 23:57:59 +0000[diff] [blame]5989 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]5990</div>
5991
5992<div class="doc_text">
5993
5994<h5>Syntax:</h5>
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]5995<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
5996any integer bit width.
5997</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]5998<pre>
Tanya Lattner09161fe2007-09-22 00:03:01 +0000[diff] [blame]5999 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6000 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6001 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6002 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6003 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]6004</pre>
6005
6006<h5>Overview:</h5>
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]6007
6008<p>
6009The '<tt>llvm.annotation</tt>' intrinsic.
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6010</p>
6011
6012<h5>Arguments:</h5>
6013
6014<p>
6015The first argument is an integer value (result of some expression),
6016the second is a pointer to a global string, the third is a pointer to a global
6017string which is the source file name, and the last argument is the line number.
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]6018It returns the value of the first argument.
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]6019</p>
6020
6021<h5>Semantics:</h5>
6022
6023<p>
6024This intrinsic allows annotations to be put on arbitrary expressions
6025with arbitrary strings. This can be useful for special purpose optimizations
6026that want to look for these annotations. These have no other defined use, they
6027are ignored by code generation and optimization.
6028</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6029
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]6030<!-- _______________________________________________________________________ -->
6031<div class="doc_subsubsection">
6032 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
6033</div>
6034
6035<div class="doc_text">
6036
6037<h5>Syntax:</h5>
6038<pre>
6039 declare void @llvm.trap()
6040</pre>
6041
6042<h5>Overview:</h5>
6043
6044<p>
6045The '<tt>llvm.trap</tt>' intrinsic
6046</p>
6047
6048<h5>Arguments:</h5>
6049
6050<p>
6051None
6052</p>
6053
6054<h5>Semantics:</h5>
6055
6056<p>
6057This intrinsics is lowered to the target dependent trap instruction. If the
6058target does not have a trap instruction, this intrinsic will be lowered to the
6059call of the abort() function.
6060</p>
6061</div>
6062
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6063<!-- *********************************************************************** -->
6064<hr>
6065<address>
6066 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
6067 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
6068 <a href="http://validator.w3.org/check/referer"><img
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]6069 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!"></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6070
6071 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
6072 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
6073 Last modified: $Date$
6074</address>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]6075
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]6076</body>
6077</html>