Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / ee9da3f641648da13668191e61a058ead015d5b3 / . / docs / LangRef.html
blob: dc37ec607665208c6417a6da154f9463857a4782 [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>
114 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
115 <ol>
116 <li><a href="#i_malloc">'<tt>malloc</tt>' Instruction</a></li>
117 <li><a href="#i_free">'<tt>free</tt>' Instruction</a></li>
118 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
119 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
120 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
121 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
122 </ol>
123 </li>
124 <li><a href="#convertops">Conversion Operations</a>
125 <ol>
126 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
127 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
128 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
129 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
130 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
131 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
132 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
133 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
134 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
135 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
136 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
137 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
138 </ol>
139 <li><a href="#otherops">Other Operations</a>
140 <ol>
141 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
142 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
143 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
144 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
145 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
146 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]147 <li><a href="#i_getresult">'<tt>getresult</tt>' Instruction</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]148 </ol>
149 </li>
150 </ol>
151 </li>
152 <li><a href="#intrinsics">Intrinsic Functions</a>
153 <ol>
154 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
155 <ol>
156 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
157 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
158 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
159 </ol>
160 </li>
161 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
162 <ol>
163 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
164 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
165 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
166 </ol>
167 </li>
168 <li><a href="#int_codegen">Code Generator Intrinsics</a>
169 <ol>
170 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
171 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
172 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
173 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
174 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
175 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
176 <li><a href="#int_readcyclecounter"><tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
177 </ol>
178 </li>
179 <li><a href="#int_libc">Standard C Library Intrinsics</a>
180 <ol>
181 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
182 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
183 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
184 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
185 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]186 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
187 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
188 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]189 </ol>
190 </li>
191 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
192 <ol>
193 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
194 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
195 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
196 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
197 <li><a href="#int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic </a></li>
198 <li><a href="#int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic </a></li>
199 </ol>
200 </li>
201 <li><a href="#int_debugger">Debugger intrinsics</a></li>
202 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]203 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]204 <ol>
205 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]206 </ol>
207 </li>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]208 <li><a href="#int_atomics">Atomic intrinsics</a>
209 <ol>
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]210 <li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
211 <li><a href="#int_atomic_lcs"><tt>llvm.atomic.lcs</tt></a></li>
212 <li><a href="#int_atomic_las"><tt>llvm.atomic.las</tt></a></li>
213 <li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]214 </ol>
215 </li>
Reid Spencerb043f672007-07-20 19:59:11 +0000[diff] [blame]216 <li><a href="#int_general">General intrinsics</a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]217 <ol>
Reid Spencerb043f672007-07-20 19:59:11 +0000[diff] [blame]218 <li><a href="#int_var_annotation">
Tanya Lattner51369f32007-09-22 00:01:26 +0000[diff] [blame]219 <tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]220 <li><a href="#int_annotation">
Tanya Lattner51369f32007-09-22 00:01:26 +0000[diff] [blame]221 <tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]222 <li><a href="#int_trap">
223 <tt>llvm.trap</tt>' Intrinsic</a></li>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]224 </ol>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]225 </li>
226 </ol>
227 </li>
228</ol>
229
230<div class="doc_author">
231 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
232 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
233</div>
234
235<!-- *********************************************************************** -->
236<div class="doc_section"> <a name="abstract">Abstract </a></div>
237<!-- *********************************************************************** -->
238
239<div class="doc_text">
240<p>This document is a reference manual for the LLVM assembly language.
241LLVM is an SSA based representation that provides type safety,
242low-level operations, flexibility, and the capability of representing
243'all' high-level languages cleanly. It is the common code
244representation used throughout all phases of the LLVM compilation
245strategy.</p>
246</div>
247
248<!-- *********************************************************************** -->
249<div class="doc_section"> <a name="introduction">Introduction</a> </div>
250<!-- *********************************************************************** -->
251
252<div class="doc_text">
253
254<p>The LLVM code representation is designed to be used in three
255different forms: as an in-memory compiler IR, as an on-disk bitcode
256representation (suitable for fast loading by a Just-In-Time compiler),
257and as a human readable assembly language representation. This allows
258LLVM to provide a powerful intermediate representation for efficient
259compiler transformations and analysis, while providing a natural means
260to debug and visualize the transformations. The three different forms
261of LLVM are all equivalent. This document describes the human readable
262representation and notation.</p>
263
264<p>The LLVM representation aims to be light-weight and low-level
265while being expressive, typed, and extensible at the same time. It
266aims to be a "universal IR" of sorts, by being at a low enough level
267that high-level ideas may be cleanly mapped to it (similar to how
268microprocessors are "universal IR's", allowing many source languages to
269be mapped to them). By providing type information, LLVM can be used as
270the target of optimizations: for example, through pointer analysis, it
271can be proven that a C automatic variable is never accessed outside of
272the current function... allowing it to be promoted to a simple SSA
273value instead of a memory location.</p>
274
275</div>
276
277<!-- _______________________________________________________________________ -->
278<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
279
280<div class="doc_text">
281
282<p>It is important to note that this document describes 'well formed'
283LLVM assembly language. There is a difference between what the parser
284accepts and what is considered 'well formed'. For example, the
285following instruction is syntactically okay, but not well formed:</p>
286
287<div class="doc_code">
288<pre>
289%x = <a href="#i_add">add</a> i32 1, %x
290</pre>
291</div>
292
293<p>...because the definition of <tt>%x</tt> does not dominate all of
294its uses. The LLVM infrastructure provides a verification pass that may
295be used to verify that an LLVM module is well formed. This pass is
296automatically run by the parser after parsing input assembly and by
297the optimizer before it outputs bitcode. The violations pointed out
298by the verifier pass indicate bugs in transformation passes or input to
299the parser.</p>
300</div>
301
Chris Lattnera83fdc02007-10-03 17:34:29 +0000[diff] [blame]302<!-- Describe the typesetting conventions here. -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]303
304<!-- *********************************************************************** -->
305<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
306<!-- *********************************************************************** -->
307
308<div class="doc_text">
309
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]310 <p>LLVM identifiers come in two basic types: global and local. Global
311 identifiers (functions, global variables) begin with the @ character. Local
312 identifiers (register names, types) begin with the % character. Additionally,
313 there are three different formats for identifiers, for different purposes:
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]314
315<ol>
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]316 <li>Named values are represented as a string of characters with their prefix.
317 For example, %foo, @DivisionByZero, %a.really.long.identifier. The actual
318 regular expression used is '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]319 Identifiers which require other characters in their names can be surrounded
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]320 with quotes. In this way, anything except a <tt>&quot;</tt> character can
321 be used in a named value.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]322
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]323 <li>Unnamed values are represented as an unsigned numeric value with their
324 prefix. For example, %12, @2, %44.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]325
326 <li>Constants, which are described in a <a href="#constants">section about
327 constants</a>, below.</li>
328</ol>
329
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]330<p>LLVM requires that values start with a prefix for two reasons: Compilers
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]331don't need to worry about name clashes with reserved words, and the set of
332reserved words may be expanded in the future without penalty. Additionally,
333unnamed identifiers allow a compiler to quickly come up with a temporary
334variable without having to avoid symbol table conflicts.</p>
335
336<p>Reserved words in LLVM are very similar to reserved words in other
337languages. There are keywords for different opcodes
338('<tt><a href="#i_add">add</a></tt>',
339 '<tt><a href="#i_bitcast">bitcast</a></tt>',
340 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
341href="#t_void">void</a></tt>', '<tt><a href="#t_primitive">i32</a></tt>', etc...),
342and others. These reserved words cannot conflict with variable names, because
Reid Spencerc8245b02007-08-07 14:34:28 +0000[diff] [blame]343none of them start with a prefix character ('%' or '@').</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]344
345<p>Here is an example of LLVM code to multiply the integer variable
346'<tt>%X</tt>' by 8:</p>
347
348<p>The easy way:</p>
349
350<div class="doc_code">
351<pre>
352%result = <a href="#i_mul">mul</a> i32 %X, 8
353</pre>
354</div>
355
356<p>After strength reduction:</p>
357
358<div class="doc_code">
359<pre>
360%result = <a href="#i_shl">shl</a> i32 %X, i8 3
361</pre>
362</div>
363
364<p>And the hard way:</p>
365
366<div class="doc_code">
367<pre>
368<a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
369<a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
370%result = <a href="#i_add">add</a> i32 %1, %1
371</pre>
372</div>
373
374<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
375important lexical features of LLVM:</p>
376
377<ol>
378
379 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
380 line.</li>
381
382 <li>Unnamed temporaries are created when the result of a computation is not
383 assigned to a named value.</li>
384
385 <li>Unnamed temporaries are numbered sequentially</li>
386
387</ol>
388
389<p>...and it also shows a convention that we follow in this document. When
390demonstrating instructions, we will follow an instruction with a comment that
391defines the type and name of value produced. Comments are shown in italic
392text.</p>
393
394</div>
395
396<!-- *********************************************************************** -->
397<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
398<!-- *********************************************************************** -->
399
400<!-- ======================================================================= -->
401<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
402</div>
403
404<div class="doc_text">
405
406<p>LLVM programs are composed of "Module"s, each of which is a
407translation unit of the input programs. Each module consists of
408functions, global variables, and symbol table entries. Modules may be
409combined together with the LLVM linker, which merges function (and
410global variable) definitions, resolves forward declarations, and merges
411symbol table entries. Here is an example of the "hello world" module:</p>
412
413<div class="doc_code">
414<pre><i>; Declare the string constant as a global constant...</i>
415<a href="#identifiers">@.LC0</a> = <a href="#linkage_internal">internal</a> <a
416 href="#globalvars">constant</a> <a href="#t_array">[13 x i8]</a> c"hello world\0A\00" <i>; [13 x i8]*</i>
417
418<i>; External declaration of the puts function</i>
419<a href="#functionstructure">declare</a> i32 @puts(i8 *) <i>; i32(i8 *)* </i>
420
421<i>; Definition of main function</i>
422define i32 @main() { <i>; i32()* </i>
423 <i>; Convert [13x i8 ]* to i8 *...</i>
424 %cast210 = <a
425 href="#i_getelementptr">getelementptr</a> [13 x i8 ]* @.LC0, i64 0, i64 0 <i>; i8 *</i>
426
427 <i>; Call puts function to write out the string to stdout...</i>
428 <a
429 href="#i_call">call</a> i32 @puts(i8 * %cast210) <i>; i32</i>
430 <a
431 href="#i_ret">ret</a> i32 0<br>}<br>
432</pre>
433</div>
434
435<p>This example is made up of a <a href="#globalvars">global variable</a>
436named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
437function, and a <a href="#functionstructure">function definition</a>
438for "<tt>main</tt>".</p>
439
440<p>In general, a module is made up of a list of global values,
441where both functions and global variables are global values. Global values are
442represented by a pointer to a memory location (in this case, a pointer to an
443array of char, and a pointer to a function), and have one of the following <a
444href="#linkage">linkage types</a>.</p>
445
446</div>
447
448<!-- ======================================================================= -->
449<div class="doc_subsection">
450 <a name="linkage">Linkage Types</a>
451</div>
452
453<div class="doc_text">
454
455<p>
456All Global Variables and Functions have one of the following types of linkage:
457</p>
458
459<dl>
460
461 <dt><tt><b><a name="linkage_internal">internal</a></b></tt> </dt>
462
463 <dd>Global values with internal linkage are only directly accessible by
464 objects in the current module. In particular, linking code into a module with
465 an internal global value may cause the internal to be renamed as necessary to
466 avoid collisions. Because the symbol is internal to the module, all
467 references can be updated. This corresponds to the notion of the
468 '<tt>static</tt>' keyword in C.
469 </dd>
470
471 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
472
473 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
474 the same name when linkage occurs. This is typically used to implement
475 inline functions, templates, or other code which must be generated in each
476 translation unit that uses it. Unreferenced <tt>linkonce</tt> globals are
477 allowed to be discarded.
478 </dd>
479
480 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
481
482 <dd>"<tt>weak</tt>" linkage is exactly the same as <tt>linkonce</tt> linkage,
483 except that unreferenced <tt>weak</tt> globals may not be discarded. This is
484 used for globals that may be emitted in multiple translation units, but that
485 are not guaranteed to be emitted into every translation unit that uses them.
486 One example of this are common globals in C, such as "<tt>int X;</tt>" at
487 global scope.
488 </dd>
489
490 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
491
492 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
493 pointer to array type. When two global variables with appending linkage are
494 linked together, the two global arrays are appended together. This is the
495 LLVM, typesafe, equivalent of having the system linker append together
496 "sections" with identical names when .o files are linked.
497 </dd>
498
499 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt>: </dt>
500 <dd>The semantics of this linkage follow the ELF model: the symbol is weak
501 until linked, if not linked, the symbol becomes null instead of being an
502 undefined reference.
503 </dd>
504
505 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
506
507 <dd>If none of the above identifiers are used, the global is externally
508 visible, meaning that it participates in linkage and can be used to resolve
509 external symbol references.
510 </dd>
511</dl>
512
513 <p>
514 The next two types of linkage are targeted for Microsoft Windows platform
515 only. They are designed to support importing (exporting) symbols from (to)
516 DLLs.
517 </p>
518
519 <dl>
520 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt>: </dt>
521
522 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
523 or variable via a global pointer to a pointer that is set up by the DLL
524 exporting the symbol. On Microsoft Windows targets, the pointer name is
525 formed by combining <code>_imp__</code> and the function or variable name.
526 </dd>
527
528 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt>: </dt>
529
530 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
531 pointer to a pointer in a DLL, so that it can be referenced with the
532 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
533 name is formed by combining <code>_imp__</code> and the function or variable
534 name.
535 </dd>
536
537</dl>
538
539<p><a name="linkage_external"></a>For example, since the "<tt>.LC0</tt>"
540variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
541variable and was linked with this one, one of the two would be renamed,
542preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
543external (i.e., lacking any linkage declarations), they are accessible
544outside of the current module.</p>
545<p>It is illegal for a function <i>declaration</i>
546to have any linkage type other than "externally visible", <tt>dllimport</tt>,
547or <tt>extern_weak</tt>.</p>
548<p>Aliases can have only <tt>external</tt>, <tt>internal</tt> and <tt>weak</tt>
549linkages.
550</div>
551
552<!-- ======================================================================= -->
553<div class="doc_subsection">
554 <a name="callingconv">Calling Conventions</a>
555</div>
556
557<div class="doc_text">
558
559<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
560and <a href="#i_invoke">invokes</a> can all have an optional calling convention
561specified for the call. The calling convention of any pair of dynamic
562caller/callee must match, or the behavior of the program is undefined. The
563following calling conventions are supported by LLVM, and more may be added in
564the future:</p>
565
566<dl>
567 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
568
569 <dd>This calling convention (the default if no other calling convention is
570 specified) matches the target C calling conventions. This calling convention
571 supports varargs function calls and tolerates some mismatch in the declared
572 prototype and implemented declaration of the function (as does normal C).
573 </dd>
574
575 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
576
577 <dd>This calling convention attempts to make calls as fast as possible
578 (e.g. by passing things in registers). This calling convention allows the
579 target to use whatever tricks it wants to produce fast code for the target,
580 without having to conform to an externally specified ABI. Implementations of
581 this convention should allow arbitrary tail call optimization to be supported.
582 This calling convention does not support varargs and requires the prototype of
583 all callees to exactly match the prototype of the function definition.
584 </dd>
585
586 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
587
588 <dd>This calling convention attempts to make code in the caller as efficient
589 as possible under the assumption that the call is not commonly executed. As
590 such, these calls often preserve all registers so that the call does not break
591 any live ranges in the caller side. This calling convention does not support
592 varargs and requires the prototype of all callees to exactly match the
593 prototype of the function definition.
594 </dd>
595
596 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
597
598 <dd>Any calling convention may be specified by number, allowing
599 target-specific calling conventions to be used. Target specific calling
600 conventions start at 64.
601 </dd>
602</dl>
603
604<p>More calling conventions can be added/defined on an as-needed basis, to
605support pascal conventions or any other well-known target-independent
606convention.</p>
607
608</div>
609
610<!-- ======================================================================= -->
611<div class="doc_subsection">
612 <a name="visibility">Visibility Styles</a>
613</div>
614
615<div class="doc_text">
616
617<p>
618All Global Variables and Functions have one of the following visibility styles:
619</p>
620
621<dl>
622 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
623
624 <dd>On ELF, default visibility means that the declaration is visible to other
625 modules and, in shared libraries, means that the declared entity may be
626 overridden. On Darwin, default visibility means that the declaration is
627 visible to other modules. Default visibility corresponds to "external
628 linkage" in the language.
629 </dd>
630
631 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
632
633 <dd>Two declarations of an object with hidden visibility refer to the same
634 object if they are in the same shared object. Usually, hidden visibility
635 indicates that the symbol will not be placed into the dynamic symbol table,
636 so no other module (executable or shared library) can reference it
637 directly.
638 </dd>
639
640 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
641
642 <dd>On ELF, protected visibility indicates that the symbol will be placed in
643 the dynamic symbol table, but that references within the defining module will
644 bind to the local symbol. That is, the symbol cannot be overridden by another
645 module.
646 </dd>
647</dl>
648
649</div>
650
651<!-- ======================================================================= -->
652<div class="doc_subsection">
653 <a name="globalvars">Global Variables</a>
654</div>
655
656<div class="doc_text">
657
658<p>Global variables define regions of memory allocated at compilation time
659instead of run-time. Global variables may optionally be initialized, may have
660an explicit section to be placed in, and may have an optional explicit alignment
661specified. A variable may be defined as "thread_local", which means that it
662will not be shared by threads (each thread will have a separated copy of the
663variable). A variable may be defined as a global "constant," which indicates
664that the contents of the variable will <b>never</b> be modified (enabling better
665optimization, allowing the global data to be placed in the read-only section of
666an executable, etc). Note that variables that need runtime initialization
667cannot be marked "constant" as there is a store to the variable.</p>
668
669<p>
670LLVM explicitly allows <em>declarations</em> of global variables to be marked
671constant, even if the final definition of the global is not. This capability
672can be used to enable slightly better optimization of the program, but requires
673the language definition to guarantee that optimizations based on the
674'constantness' are valid for the translation units that do not include the
675definition.
676</p>
677
678<p>As SSA values, global variables define pointer values that are in
679scope (i.e. they dominate) all basic blocks in the program. Global
680variables always define a pointer to their "content" type because they
681describe a region of memory, and all memory objects in LLVM are
682accessed through pointers.</p>
683
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]684<p>A global variable may be declared to reside in a target-specifc numbered
685address space. For targets that support them, address spaces may affect how
686optimizations are performed and/or what target instructions are used to access
Christopher Lamb20a39e92007-12-12 08:44:39 +0000[diff] [blame]687the variable. The default address space is zero. The address space qualifier
688must precede any other attributes.</p>
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]689
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]690<p>LLVM allows an explicit section to be specified for globals. If the target
691supports it, it will emit globals to the section specified.</p>
692
693<p>An explicit alignment may be specified for a global. If not present, or if
694the alignment is set to zero, the alignment of the global is set by the target
695to whatever it feels convenient. If an explicit alignment is specified, the
696global is forced to have at least that much alignment. All alignments must be
697a power of 2.</p>
698
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]699<p>For example, the following defines a global in a numbered address space with
700an initializer, section, and alignment:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]701
702<div class="doc_code">
703<pre>
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]704@G = constant float 1.0 addrspace(5), section "foo", align 4
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]705</pre>
706</div>
707
708</div>
709
710
711<!-- ======================================================================= -->
712<div class="doc_subsection">
713 <a name="functionstructure">Functions</a>
714</div>
715
716<div class="doc_text">
717
718<p>LLVM function definitions consist of the "<tt>define</tt>" keyord,
719an optional <a href="#linkage">linkage type</a>, an optional
720<a href="#visibility">visibility style</a>, an optional
721<a href="#callingconv">calling convention</a>, a return type, an optional
722<a href="#paramattrs">parameter attribute</a> for the return type, a function
723name, a (possibly empty) argument list (each with optional
724<a href="#paramattrs">parameter attributes</a>), an optional section, an
Gordon Henriksend606f5b2007-12-10 03:30:21 +0000[diff] [blame]725optional alignment, an optional <a href="#gc">garbage collector name</a>, an
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]726opening curly brace, a list of basic blocks, and a closing curly brace.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]727
728LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
729optional <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
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]733name, a possibly empty list of arguments, an optional alignment, and an optional
Gordon Henriksend606f5b2007-12-10 03:30:21 +0000[diff] [blame]734<a href="#gc">garbage collector name</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]735
736<p>A function definition contains a list of basic blocks, forming the CFG for
737the function. Each basic block may optionally start with a label (giving the
738basic block a symbol table entry), contains a list of instructions, and ends
739with a <a href="#terminators">terminator</a> instruction (such as a branch or
740function return).</p>
741
742<p>The first basic block in a function is special in two ways: it is immediately
743executed on entrance to the function, and it is not allowed to have predecessor
744basic blocks (i.e. there can not be any branches to the entry block of a
745function). Because the block can have no predecessors, it also cannot have any
746<a href="#i_phi">PHI nodes</a>.</p>
747
748<p>LLVM allows an explicit section to be specified for functions. If the target
749supports it, it will emit functions to the section specified.</p>
750
751<p>An explicit alignment may be specified for a function. If not present, or if
752the alignment is set to zero, the alignment of the function is set by the target
753to whatever it feels convenient. If an explicit alignment is specified, the
754function is forced to have at least that much alignment. All alignments must be
755a power of 2.</p>
756
757</div>
758
759
760<!-- ======================================================================= -->
761<div class="doc_subsection">
762 <a name="aliasstructure">Aliases</a>
763</div>
764<div class="doc_text">
765 <p>Aliases act as "second name" for the aliasee value (which can be either
766 function or global variable or bitcast of global value). Aliases may have an
767 optional <a href="#linkage">linkage type</a>, and an
768 optional <a href="#visibility">visibility style</a>.</p>
769
770 <h5>Syntax:</h5>
771
772<div class="doc_code">
773<pre>
774@&lt;Name&gt; = [Linkage] [Visibility] alias &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
775</pre>
776</div>
777
778</div>
779
780
781
782<!-- ======================================================================= -->
783<div class="doc_subsection"><a name="paramattrs">Parameter Attributes</a></div>
784<div class="doc_text">
785 <p>The return type and each parameter of a function type may have a set of
786 <i>parameter attributes</i> associated with them. Parameter attributes are
787 used to communicate additional information about the result or parameters of
Duncan Sandsf5588dc2007-11-27 13:23:08 +0000[diff] [blame]788 a function. Parameter attributes are considered to be part of the function,
789 not of the function type, so functions with different parameter attributes
790 can have the same function type.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]791
792 <p>Parameter attributes are simple keywords that follow the type specified. If
793 multiple parameter attributes are needed, they are space separated. For
794 example:</p>
795
796<div class="doc_code">
797<pre>
Duncan Sandsf5588dc2007-11-27 13:23:08 +0000[diff] [blame]798declare i32 @printf(i8* noalias , ...) nounwind
799declare i32 @atoi(i8*) nounwind readonly
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]800</pre>
801</div>
802
Duncan Sandsf5588dc2007-11-27 13:23:08 +0000[diff] [blame]803 <p>Note that any attributes for the function result (<tt>nounwind</tt>,
804 <tt>readonly</tt>) come immediately after the argument list.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]805
806 <p>Currently, only the following parameter attributes are defined:</p>
807 <dl>
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]808 <dt><tt>zeroext</tt></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]809 <dd>This indicates that the parameter should be zero extended just before
810 a call to this function.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]811
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]812 <dt><tt>signext</tt></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]813 <dd>This indicates that the parameter should be sign extended just before
814 a call to this function.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]815
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]816 <dt><tt>inreg</tt></dt>
817 <dd>This indicates that the parameter should be placed in register (if
818 possible) during assembling function call. Support for this attribute is
819 target-specific</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]820
821 <dt><tt>byval</tt></dt>
Chris Lattner04c86182008-01-15 04:34:22 +0000[diff] [blame]822 <dd>This indicates that the pointer parameter should really be passed by
823 value to the function. The attribute implies that a hidden copy of the
824 pointee is made between the caller and the callee, so the callee is unable
825 to modify the value in the callee. This attribute is only valid on llvm
826 pointer arguments. It is generally used to pass structs and arrays by
827 value, but is also valid on scalars (even though this is silly).</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]828
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]829 <dt><tt>sret</tt></dt>
Duncan Sands616cc032008-02-18 04:19:38 +0000[diff] [blame]830 <dd>This indicates that the pointer parameter specifies the address of a
831 structure that is the return value of the function in the source program.
Duncan Sandsef0e9e42008-03-17 12:17:41 +0000[diff] [blame]832 Loads and stores to the structure are assumed not to trap.
Duncan Sands616cc032008-02-18 04:19:38 +0000[diff] [blame]833 May only be applied to the first parameter.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]834
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]835 <dt><tt>noalias</tt></dt>
Owen Andersonc4fc4cd2008-02-18 04:09:01 +0000[diff] [blame]836 <dd>This indicates that the parameter does not alias any global or any other
837 parameter. The caller is responsible for ensuring that this is the case,
838 usually by placing the value in a stack allocation.</dd>
Chris Lattner275e6be2008-01-11 06:20:47 +0000[diff] [blame]839
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]840 <dt><tt>noreturn</tt></dt>
841 <dd>This function attribute indicates that the function never returns. This
842 indicates to LLVM that every call to this function should be treated as if
843 an <tt>unreachable</tt> instruction immediately followed the call.</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>nounwind</tt></dt>
Duncan Sandsef0e9e42008-03-17 12:17:41 +0000[diff] [blame]846 <dd>This function attribute indicates that no exceptions unwind out of the
847 function. Usually this is because the function makes no use of exceptions,
848 but it may also be that the function catches any exceptions thrown when
849 executing it.</dd>
850
Duncan Sands4ee46812007-07-27 19:57:41 +0000[diff] [blame]851 <dt><tt>nest</tt></dt>
852 <dd>This indicates that the parameter can be excised using the
853 <a href="#int_trampoline">trampoline intrinsics</a>.</dd>
Duncan Sands13e13f82007-11-22 20:23:04 +0000[diff] [blame]854 <dt><tt>readonly</tt></dt>
Duncan Sandsd69c0e62007-11-14 21:14:02 +0000[diff] [blame]855 <dd>This function attribute indicates that the function has no side-effects
Duncan Sands13e13f82007-11-22 20:23:04 +0000[diff] [blame]856 except for producing a return value or throwing an exception. The value
857 returned must only depend on the function arguments and/or global variables.
858 It may use values obtained by dereferencing pointers.</dd>
859 <dt><tt>readnone</tt></dt>
860 <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]861 function, but in addition it is not allowed to dereference any pointer arguments
862 or global variables.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]863 </dl>
864
865</div>
866
867<!-- ======================================================================= -->
868<div class="doc_subsection">
Gordon Henriksen13fe5e32007-12-10 03:18:06 +0000[diff] [blame]869 <a name="gc">Garbage Collector Names</a>
870</div>
871
872<div class="doc_text">
873<p>Each function may specify a garbage collector name, which is simply a
874string.</p>
875
876<div class="doc_code"><pre
877>define void @f() gc "name" { ...</pre></div>
878
879<p>The compiler declares the supported values of <i>name</i>. Specifying a
880collector which will cause the compiler to alter its output in order to support
881the named garbage collection algorithm.</p>
882</div>
883
884<!-- ======================================================================= -->
885<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]886 <a name="moduleasm">Module-Level Inline Assembly</a>
887</div>
888
889<div class="doc_text">
890<p>
891Modules may contain "module-level inline asm" blocks, which corresponds to the
892GCC "file scope inline asm" blocks. These blocks are internally concatenated by
893LLVM and treated as a single unit, but may be separated in the .ll file if
894desired. The syntax is very simple:
895</p>
896
897<div class="doc_code">
898<pre>
899module asm "inline asm code goes here"
900module asm "more can go here"
901</pre>
902</div>
903
904<p>The strings can contain any character by escaping non-printable characters.
905 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
906 for the number.
907</p>
908
909<p>
910 The inline asm code is simply printed to the machine code .s file when
911 assembly code is generated.
912</p>
913</div>
914
915<!-- ======================================================================= -->
916<div class="doc_subsection">
917 <a name="datalayout">Data Layout</a>
918</div>
919
920<div class="doc_text">
921<p>A module may specify a target specific data layout string that specifies how
922data is to be laid out in memory. The syntax for the data layout is simply:</p>
923<pre> target datalayout = "<i>layout specification</i>"</pre>
924<p>The <i>layout specification</i> consists of a list of specifications
925separated by the minus sign character ('-'). Each specification starts with a
926letter and may include other information after the letter to define some
927aspect of the data layout. The specifications accepted are as follows: </p>
928<dl>
929 <dt><tt>E</tt></dt>
930 <dd>Specifies that the target lays out data in big-endian form. That is, the
931 bits with the most significance have the lowest address location.</dd>
932 <dt><tt>e</tt></dt>
933 <dd>Specifies that hte target lays out data in little-endian form. That is,
934 the bits with the least significance have the lowest address location.</dd>
935 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
936 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
937 <i>preferred</i> alignments. All sizes are in bits. Specifying the <i>pref</i>
938 alignment is optional. If omitted, the preceding <tt>:</tt> should be omitted
939 too.</dd>
940 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
941 <dd>This specifies the alignment for an integer type of a given bit
942 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
943 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
944 <dd>This specifies the alignment for a vector type of a given bit
945 <i>size</i>.</dd>
946 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
947 <dd>This specifies the alignment for a floating point type of a given bit
948 <i>size</i>. The value of <i>size</i> must be either 32 (float) or 64
949 (double).</dd>
950 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
951 <dd>This specifies the alignment for an aggregate type of a given bit
952 <i>size</i>.</dd>
953</dl>
954<p>When constructing the data layout for a given target, LLVM starts with a
955default set of specifications which are then (possibly) overriden by the
956specifications in the <tt>datalayout</tt> keyword. The default specifications
957are given in this list:</p>
958<ul>
959 <li><tt>E</tt> - big endian</li>
960 <li><tt>p:32:64:64</tt> - 32-bit pointers with 64-bit alignment</li>
961 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
962 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
963 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
964 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
965 <li><tt>i64:32:64</tt> - i64 has abi alignment of 32-bits but preferred
966 alignment of 64-bits</li>
967 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
968 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
969 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
970 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
971 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
972</ul>
973<p>When llvm is determining the alignment for a given type, it uses the
974following rules:
975<ol>
976 <li>If the type sought is an exact match for one of the specifications, that
977 specification is used.</li>
978 <li>If no match is found, and the type sought is an integer type, then the
979 smallest integer type that is larger than the bitwidth of the sought type is
980 used. If none of the specifications are larger than the bitwidth then the the
981 largest integer type is used. For example, given the default specifications
982 above, the i7 type will use the alignment of i8 (next largest) while both
983 i65 and i256 will use the alignment of i64 (largest specified).</li>
984 <li>If no match is found, and the type sought is a vector type, then the
985 largest vector type that is smaller than the sought vector type will be used
986 as a fall back. This happens because <128 x double> can be implemented in
987 terms of 64 <2 x double>, for example.</li>
988</ol>
989</div>
990
991<!-- *********************************************************************** -->
992<div class="doc_section"> <a name="typesystem">Type System</a> </div>
993<!-- *********************************************************************** -->
994
995<div class="doc_text">
996
997<p>The LLVM type system is one of the most important features of the
998intermediate representation. Being typed enables a number of
999optimizations to be performed on the IR directly, without having to do
1000extra analyses on the side before the transformation. A strong type
1001system makes it easier to read the generated code and enables novel
1002analyses and transformations that are not feasible to perform on normal
1003three address code representations.</p>
1004
1005</div>
1006
1007<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1008<div class="doc_subsection"> <a name="t_classifications">Type
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1009Classifications</a> </div>
1010<div class="doc_text">
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1011<p>The types fall into a few useful
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1012classifications:</p>
1013
1014<table border="1" cellspacing="0" cellpadding="4">
1015 <tbody>
1016 <tr><th>Classification</th><th>Types</th></tr>
1017 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1018 <td><a href="#t_integer">integer</a></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1019 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
1020 </tr>
1021 <tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1022 <td><a href="#t_floating">floating point</a></td>
1023 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1024 </tr>
1025 <tr>
1026 <td><a name="t_firstclass">first class</a></td>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1027 <td><a href="#t_integer">integer</a>,
1028 <a href="#t_floating">floating point</a>,
1029 <a href="#t_pointer">pointer</a>,
1030 <a href="#t_vector">vector</a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1031 </td>
1032 </tr>
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1033 <tr>
1034 <td><a href="#t_primitive">primitive</a></td>
1035 <td><a href="#t_label">label</a>,
1036 <a href="#t_void">void</a>,
1037 <a href="#t_integer">integer</a>,
1038 <a href="#t_floating">floating point</a>.</td>
1039 </tr>
1040 <tr>
1041 <td><a href="#t_derived">derived</a></td>
1042 <td><a href="#t_integer">integer</a>,
1043 <a href="#t_array">array</a>,
1044 <a href="#t_function">function</a>,
1045 <a href="#t_pointer">pointer</a>,
1046 <a href="#t_struct">structure</a>,
1047 <a href="#t_pstruct">packed structure</a>,
1048 <a href="#t_vector">vector</a>,
1049 <a href="#t_opaque">opaque</a>.
1050 </tr>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1051 </tbody>
1052</table>
1053
1054<p>The <a href="#t_firstclass">first class</a> types are perhaps the
1055most important. Values of these types are the only ones which can be
1056produced by instructions, passed as arguments, or used as operands to
1057instructions. This means that all structures and arrays must be
1058manipulated either by pointer or by component.</p>
1059</div>
1060
1061<!-- ======================================================================= -->
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1062<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1063
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1064<div class="doc_text">
1065<p>The primitive types are the fundamental building blocks of the LLVM
1066system.</p>
1067
Chris Lattner86437612008-01-04 04:34:14 +0000[diff] [blame]1068</div>
1069
Chris Lattner488772f2008-01-04 04:32:38 +0000[diff] [blame]1070<!-- _______________________________________________________________________ -->
1071<div class="doc_subsubsection"> <a name="t_floating">Floating Point Types</a> </div>
1072
1073<div class="doc_text">
1074 <table>
1075 <tbody>
1076 <tr><th>Type</th><th>Description</th></tr>
1077 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1078 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1079 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1080 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1081 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1082 </tbody>
1083 </table>
1084</div>
1085
1086<!-- _______________________________________________________________________ -->
1087<div class="doc_subsubsection"> <a name="t_void">Void Type</a> </div>
1088
1089<div class="doc_text">
1090<h5>Overview:</h5>
1091<p>The void type does not represent any value and has no size.</p>
1092
1093<h5>Syntax:</h5>
1094
1095<pre>
1096 void
1097</pre>
1098</div>
1099
1100<!-- _______________________________________________________________________ -->
1101<div class="doc_subsubsection"> <a name="t_label">Label Type</a> </div>
1102
1103<div class="doc_text">
1104<h5>Overview:</h5>
1105<p>The label type represents code labels.</p>
1106
1107<h5>Syntax:</h5>
1108
1109<pre>
1110 label
1111</pre>
1112</div>
1113
1114
1115<!-- ======================================================================= -->
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1116<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
1117
1118<div class="doc_text">
1119
1120<p>The real power in LLVM comes from the derived types in the system.
1121This is what allows a programmer to represent arrays, functions,
1122pointers, and other useful types. Note that these derived types may be
1123recursive: For example, it is possible to have a two dimensional array.</p>
1124
1125</div>
1126
1127<!-- _______________________________________________________________________ -->
1128<div class="doc_subsubsection"> <a name="t_integer">Integer Type</a> </div>
1129
1130<div class="doc_text">
1131
1132<h5>Overview:</h5>
1133<p>The integer type is a very simple derived type that simply specifies an
1134arbitrary bit width for the integer type desired. Any bit width from 1 bit to
11352^23-1 (about 8 million) can be specified.</p>
1136
1137<h5>Syntax:</h5>
1138
1139<pre>
1140 iN
1141</pre>
1142
1143<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1144value.</p>
1145
1146<h5>Examples:</h5>
1147<table class="layout">
Chris Lattner251ab812007-12-18 06:18:21 +0000[diff] [blame]1148 <tbody>
1149 <tr>
1150 <td><tt>i1</tt></td>
1151 <td>a single-bit integer.</td>
1152 </tr><tr>
1153 <td><tt>i32</tt></td>
1154 <td>a 32-bit integer.</td>
1155 </tr><tr>
1156 <td><tt>i1942652</tt></td>
1157 <td>a really big integer of over 1 million bits.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1158 </tr>
Chris Lattner251ab812007-12-18 06:18:21 +0000[diff] [blame]1159 </tbody>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1160</table>
1161</div>
1162
1163<!-- _______________________________________________________________________ -->
1164<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
1165
1166<div class="doc_text">
1167
1168<h5>Overview:</h5>
1169
1170<p>The array type is a very simple derived type that arranges elements
1171sequentially in memory. The array type requires a size (number of
1172elements) and an underlying data type.</p>
1173
1174<h5>Syntax:</h5>
1175
1176<pre>
1177 [&lt;# elements&gt; x &lt;elementtype&gt;]
1178</pre>
1179
1180<p>The number of elements is a constant integer value; elementtype may
1181be any type with a size.</p>
1182
1183<h5>Examples:</h5>
1184<table class="layout">
1185 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1186 <td class="left"><tt>[40 x i32]</tt></td>
1187 <td class="left">Array of 40 32-bit integer values.</td>
1188 </tr>
1189 <tr class="layout">
1190 <td class="left"><tt>[41 x i32]</tt></td>
1191 <td class="left">Array of 41 32-bit integer values.</td>
1192 </tr>
1193 <tr class="layout">
1194 <td class="left"><tt>[4 x i8]</tt></td>
1195 <td class="left">Array of 4 8-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1196 </tr>
1197</table>
1198<p>Here are some examples of multidimensional arrays:</p>
1199<table class="layout">
1200 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1201 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1202 <td class="left">3x4 array of 32-bit integer values.</td>
1203 </tr>
1204 <tr class="layout">
1205 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1206 <td class="left">12x10 array of single precision floating point values.</td>
1207 </tr>
1208 <tr class="layout">
1209 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1210 <td class="left">2x3x4 array of 16-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1211 </tr>
1212</table>
1213
1214<p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
1215length array. Normally, accesses past the end of an array are undefined in
1216LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
1217As a special case, however, zero length arrays are recognized to be variable
1218length. This allows implementation of 'pascal style arrays' with the LLVM
1219type "{ i32, [0 x float]}", for example.</p>
1220
1221</div>
1222
1223<!-- _______________________________________________________________________ -->
1224<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
1225<div class="doc_text">
1226<h5>Overview:</h5>
1227<p>The function type can be thought of as a function signature. It
1228consists of a return type and a list of formal parameter types.
1229Function types are usually used to build virtual function tables
1230(which are structures of pointers to functions), for indirect function
1231calls, and when defining a function.</p>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1232
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1233<h5>Syntax:</h5>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1234<pre> &lt;returntype list&gt; (&lt;parameter list&gt;)<br></pre>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1235<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
1236specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1237which indicates that the function takes a variable number of arguments.
1238Variable argument functions can access their arguments with the <a
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1239 href="#int_varargs">variable argument handling intrinsic</a> functions.
1240'<tt>&lt;returntype list&gt;</tt>' is a comma-separated list of
1241<a href="#t_firstclass">first class</a> type specifiers.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1242<h5>Examples:</h5>
1243<table class="layout">
1244 <tr class="layout">
1245 <td class="left"><tt>i32 (i32)</tt></td>
1246 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
1247 </td>
1248 </tr><tr class="layout">
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]1249 <td class="left"><tt>float&nbsp;(i16&nbsp;signext,&nbsp;i32&nbsp;*)&nbsp;*
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1250 </tt></td>
1251 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
1252 an <tt>i16</tt> that should be sign extended and a
1253 <a href="#t_pointer">pointer</a> to <tt>i32</tt>, returning
1254 <tt>float</tt>.
1255 </td>
1256 </tr><tr class="layout">
1257 <td class="left"><tt>i32 (i8*, ...)</tt></td>
1258 <td class="left">A vararg function that takes at least one
1259 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1260 which returns an integer. This is the signature for <tt>printf</tt> in
1261 LLVM.
1262 </td>
1263 </tr>
1264</table>
1265
1266</div>
1267<!-- _______________________________________________________________________ -->
1268<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
1269<div class="doc_text">
1270<h5>Overview:</h5>
1271<p>The structure type is used to represent a collection of data members
1272together in memory. The packing of the field types is defined to match
1273the ABI of the underlying processor. The elements of a structure may
1274be any type that has a size.</p>
1275<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1276and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1277field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1278instruction.</p>
1279<h5>Syntax:</h5>
1280<pre> { &lt;type list&gt; }<br></pre>
1281<h5>Examples:</h5>
1282<table class="layout">
1283 <tr class="layout">
1284 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1285 <td class="left">A triple of three <tt>i32</tt> values</td>
1286 </tr><tr class="layout">
1287 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1288 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1289 second element is a <a href="#t_pointer">pointer</a> to a
1290 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1291 an <tt>i32</tt>.</td>
1292 </tr>
1293</table>
1294</div>
1295
1296<!-- _______________________________________________________________________ -->
1297<div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
1298</div>
1299<div class="doc_text">
1300<h5>Overview:</h5>
1301<p>The packed structure type is used to represent a collection of data members
1302together in memory. There is no padding between fields. Further, the alignment
1303of a packed structure is 1 byte. The elements of a packed structure may
1304be any type that has a size.</p>
1305<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1306and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1307field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1308instruction.</p>
1309<h5>Syntax:</h5>
1310<pre> &lt; { &lt;type list&gt; } &gt; <br></pre>
1311<h5>Examples:</h5>
1312<table class="layout">
1313 <tr class="layout">
1314 <td class="left"><tt>&lt; { i32, i32, i32 } &gt;</tt></td>
1315 <td class="left">A triple of three <tt>i32</tt> values</td>
1316 </tr><tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1317 <td class="left"><tt>&lt; { float, i32 (i32)* } &gt;</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1318 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1319 second element is a <a href="#t_pointer">pointer</a> to a
1320 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1321 an <tt>i32</tt>.</td>
1322 </tr>
1323</table>
1324</div>
1325
1326<!-- _______________________________________________________________________ -->
1327<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
1328<div class="doc_text">
1329<h5>Overview:</h5>
1330<p>As in many languages, the pointer type represents a pointer or
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]1331reference to another object, which must live in memory. Pointer types may have
1332an optional address space attribute defining the target-specific numbered
1333address space where the pointed-to object resides. The default address space is
1334zero.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1335<h5>Syntax:</h5>
1336<pre> &lt;type&gt; *<br></pre>
1337<h5>Examples:</h5>
1338<table class="layout">
1339 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1340 <td class="left"><tt>[4x i32]*</tt></td>
1341 <td class="left">A <a href="#t_pointer">pointer</a> to <a
1342 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
1343 </tr>
1344 <tr class="layout">
1345 <td class="left"><tt>i32 (i32 *) *</tt></td>
1346 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1347 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1348 <tt>i32</tt>.</td>
1349 </tr>
1350 <tr class="layout">
1351 <td class="left"><tt>i32 addrspace(5)*</tt></td>
1352 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
1353 that resides in address space #5.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1354 </tr>
1355</table>
1356</div>
1357
1358<!-- _______________________________________________________________________ -->
1359<div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
1360<div class="doc_text">
1361
1362<h5>Overview:</h5>
1363
1364<p>A vector type is a simple derived type that represents a vector
1365of elements. Vector types are used when multiple primitive data
1366are operated in parallel using a single instruction (SIMD).
1367A vector type requires a size (number of
1368elements) and an underlying primitive data type. Vectors must have a power
1369of two length (1, 2, 4, 8, 16 ...). Vector types are
1370considered <a href="#t_firstclass">first class</a>.</p>
1371
1372<h5>Syntax:</h5>
1373
1374<pre>
1375 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1376</pre>
1377
1378<p>The number of elements is a constant integer value; elementtype may
1379be any integer or floating point type.</p>
1380
1381<h5>Examples:</h5>
1382
1383<table class="layout">
1384 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1385 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
1386 <td class="left">Vector of 4 32-bit integer values.</td>
1387 </tr>
1388 <tr class="layout">
1389 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
1390 <td class="left">Vector of 8 32-bit floating-point values.</td>
1391 </tr>
1392 <tr class="layout">
1393 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
1394 <td class="left">Vector of 2 64-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1395 </tr>
1396</table>
1397</div>
1398
1399<!-- _______________________________________________________________________ -->
1400<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1401<div class="doc_text">
1402
1403<h5>Overview:</h5>
1404
1405<p>Opaque types are used to represent unknown types in the system. This
Gordon Henriksenda0706e2007-10-14 00:34:53 +0000[diff] [blame]1406corresponds (for example) to the C notion of a forward declared structure type.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1407In LLVM, opaque types can eventually be resolved to any type (not just a
1408structure type).</p>
1409
1410<h5>Syntax:</h5>
1411
1412<pre>
1413 opaque
1414</pre>
1415
1416<h5>Examples:</h5>
1417
1418<table class="layout">
1419 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1420 <td class="left"><tt>opaque</tt></td>
1421 <td class="left">An opaque type.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1422 </tr>
1423</table>
1424</div>
1425
1426
1427<!-- *********************************************************************** -->
1428<div class="doc_section"> <a name="constants">Constants</a> </div>
1429<!-- *********************************************************************** -->
1430
1431<div class="doc_text">
1432
1433<p>LLVM has several different basic types of constants. This section describes
1434them all and their syntax.</p>
1435
1436</div>
1437
1438<!-- ======================================================================= -->
1439<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
1440
1441<div class="doc_text">
1442
1443<dl>
1444 <dt><b>Boolean constants</b></dt>
1445
1446 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
1447 constants of the <tt><a href="#t_primitive">i1</a></tt> type.
1448 </dd>
1449
1450 <dt><b>Integer constants</b></dt>
1451
1452 <dd>Standard integers (such as '4') are constants of the <a
1453 href="#t_integer">integer</a> type. Negative numbers may be used with
1454 integer types.
1455 </dd>
1456
1457 <dt><b>Floating point constants</b></dt>
1458
1459 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
1460 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
1461 notation (see below). Floating point constants must have a <a
1462 href="#t_floating">floating point</a> type. </dd>
1463
1464 <dt><b>Null pointer constants</b></dt>
1465
1466 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
1467 and must be of <a href="#t_pointer">pointer type</a>.</dd>
1468
1469</dl>
1470
1471<p>The one non-intuitive notation for constants is the optional hexadecimal form
1472of floating point constants. For example, the form '<tt>double
14730x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
14744.5e+15</tt>'. The only time hexadecimal floating point constants are required
1475(and the only time that they are generated by the disassembler) is when a
1476floating point constant must be emitted but it cannot be represented as a
1477decimal floating point number. For example, NaN's, infinities, and other
1478special values are represented in their IEEE hexadecimal format so that
1479assembly and disassembly do not cause any bits to change in the constants.</p>
1480
1481</div>
1482
1483<!-- ======================================================================= -->
1484<div class="doc_subsection"><a name="aggregateconstants">Aggregate Constants</a>
1485</div>
1486
1487<div class="doc_text">
1488<p>Aggregate constants arise from aggregation of simple constants
1489and smaller aggregate constants.</p>
1490
1491<dl>
1492 <dt><b>Structure constants</b></dt>
1493
1494 <dd>Structure constants are represented with notation similar to structure
1495 type definitions (a comma separated list of elements, surrounded by braces
Chris Lattner6c8de962007-12-25 20:34:52 +0000[diff] [blame]1496 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
1497 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]1498 must have <a href="#t_struct">structure type</a>, and the number and
1499 types of elements must match those specified by the type.
1500 </dd>
1501
1502 <dt><b>Array constants</b></dt>
1503
1504 <dd>Array constants are represented with notation similar to array type
1505 definitions (a comma separated list of elements, surrounded by square brackets
1506 (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74 ]</tt>". Array
1507 constants must have <a href="#t_array">array type</a>, and the number and
1508 types of elements must match those specified by the type.
1509 </dd>
1510
1511 <dt><b>Vector constants</b></dt>
1512
1513 <dd>Vector constants are represented with notation similar to vector type
1514 definitions (a comma separated list of elements, surrounded by
1515 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32 42,
1516 i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must have <a
1517 href="#t_vector">vector type</a>, and the number and types of elements must
1518 match those specified by the type.
1519 </dd>
1520
1521 <dt><b>Zero initialization</b></dt>
1522
1523 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1524 value to zero of <em>any</em> type, including scalar and aggregate types.
1525 This is often used to avoid having to print large zero initializers (e.g. for
1526 large arrays) and is always exactly equivalent to using explicit zero
1527 initializers.
1528 </dd>
1529</dl>
1530
1531</div>
1532
1533<!-- ======================================================================= -->
1534<div class="doc_subsection">
1535 <a name="globalconstants">Global Variable and Function Addresses</a>
1536</div>
1537
1538<div class="doc_text">
1539
1540<p>The addresses of <a href="#globalvars">global variables</a> and <a
1541href="#functionstructure">functions</a> are always implicitly valid (link-time)
1542constants. These constants are explicitly referenced when the <a
1543href="#identifiers">identifier for the global</a> is used and always have <a
1544href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1545file:</p>
1546
1547<div class="doc_code">
1548<pre>
1549@X = global i32 17
1550@Y = global i32 42
1551@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
1552</pre>
1553</div>
1554
1555</div>
1556
1557<!-- ======================================================================= -->
1558<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
1559<div class="doc_text">
1560 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
1561 no specific value. Undefined values may be of any type and be used anywhere
1562 a constant is permitted.</p>
1563
1564 <p>Undefined values indicate to the compiler that the program is well defined
1565 no matter what value is used, giving the compiler more freedom to optimize.
1566 </p>
1567</div>
1568
1569<!-- ======================================================================= -->
1570<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1571</div>
1572
1573<div class="doc_text">
1574
1575<p>Constant expressions are used to allow expressions involving other constants
1576to be used as constants. Constant expressions may be of any <a
1577href="#t_firstclass">first class</a> type and may involve any LLVM operation
1578that does not have side effects (e.g. load and call are not supported). The
1579following is the syntax for constant expressions:</p>
1580
1581<dl>
1582 <dt><b><tt>trunc ( CST to TYPE )</tt></b></dt>
1583 <dd>Truncate a constant to another type. The bit size of CST must be larger
1584 than the bit size of TYPE. Both types must be integers.</dd>
1585
1586 <dt><b><tt>zext ( CST to TYPE )</tt></b></dt>
1587 <dd>Zero extend a constant to another type. The bit size of CST must be
1588 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1589
1590 <dt><b><tt>sext ( CST to TYPE )</tt></b></dt>
1591 <dd>Sign extend a constant to another type. The bit size of CST must be
1592 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1593
1594 <dt><b><tt>fptrunc ( CST to TYPE )</tt></b></dt>
1595 <dd>Truncate a floating point constant to another floating point type. The
1596 size of CST must be larger than the size of TYPE. Both types must be
1597 floating point.</dd>
1598
1599 <dt><b><tt>fpext ( CST to TYPE )</tt></b></dt>
1600 <dd>Floating point extend a constant to another type. The size of CST must be
1601 smaller or equal to the size of TYPE. Both types must be floating point.</dd>
1602
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]1603 <dt><b><tt>fptoui ( CST to TYPE )</tt></b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1604 <dd>Convert a floating point constant to the corresponding unsigned integer
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1605 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1606 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1607 of the same number of elements. If the value won't fit in the integer type,
1608 the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1609
1610 <dt><b><tt>fptosi ( CST to TYPE )</tt></b></dt>
1611 <dd>Convert a floating point constant to the corresponding signed integer
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1612 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1613 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1614 of the same number of elements. If the value won't fit in the integer type,
1615 the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1616
1617 <dt><b><tt>uitofp ( CST to TYPE )</tt></b></dt>
1618 <dd>Convert an unsigned integer constant to the corresponding floating point
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1619 constant. TYPE must be a scalar or vector floating point type. CST must be of
1620 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1621 of the same number of elements. If the value won't fit in the floating point
1622 type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1623
1624 <dt><b><tt>sitofp ( CST to TYPE )</tt></b></dt>
1625 <dd>Convert a signed integer constant to the corresponding floating point
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1626 constant. TYPE must be a scalar or vector floating point type. CST must be of
1627 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1628 of the same number of elements. If the value won't fit in the floating point
1629 type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1630
1631 <dt><b><tt>ptrtoint ( CST to TYPE )</tt></b></dt>
1632 <dd>Convert a pointer typed constant to the corresponding integer constant
1633 TYPE must be an integer type. CST must be of pointer type. The CST value is
1634 zero extended, truncated, or unchanged to make it fit in TYPE.</dd>
1635
1636 <dt><b><tt>inttoptr ( CST to TYPE )</tt></b></dt>
1637 <dd>Convert a integer constant to a pointer constant. TYPE must be a
1638 pointer type. CST must be of integer type. The CST value is zero extended,
1639 truncated, or unchanged to make it fit in a pointer size. This one is
1640 <i>really</i> dangerous!</dd>
1641
1642 <dt><b><tt>bitcast ( CST to TYPE )</tt></b></dt>
1643 <dd>Convert a constant, CST, to another TYPE. The size of CST and TYPE must be
1644 identical (same number of bits). The conversion is done as if the CST value
1645 was stored to memory and read back as TYPE. In other words, no bits change
1646 with this operator, just the type. This can be used for conversion of
1647 vector types to any other type, as long as they have the same bit width. For
1648 pointers it is only valid to cast to another pointer type.
1649 </dd>
1650
1651 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
1652
1653 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
1654 constants. As with the <a href="#i_getelementptr">getelementptr</a>
1655 instruction, the index list may have zero or more indexes, which are required
1656 to make sense for the type of "CSTPTR".</dd>
1657
1658 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
1659
1660 <dd>Perform the <a href="#i_select">select operation</a> on
1661 constants.</dd>
1662
1663 <dt><b><tt>icmp COND ( VAL1, VAL2 )</tt></b></dt>
1664 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
1665
1666 <dt><b><tt>fcmp COND ( VAL1, VAL2 )</tt></b></dt>
1667 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
1668
1669 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
1670
1671 <dd>Perform the <a href="#i_extractelement">extractelement
1672 operation</a> on constants.
1673
1674 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
1675
1676 <dd>Perform the <a href="#i_insertelement">insertelement
1677 operation</a> on constants.</dd>
1678
1679
1680 <dt><b><tt>shufflevector ( VEC1, VEC2, IDXMASK )</tt></b></dt>
1681
1682 <dd>Perform the <a href="#i_shufflevector">shufflevector
1683 operation</a> on constants.</dd>
1684
1685 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
1686
1687 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
1688 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
1689 binary</a> operations. The constraints on operands are the same as those for
1690 the corresponding instruction (e.g. no bitwise operations on floating point
1691 values are allowed).</dd>
1692</dl>
1693</div>
1694
1695<!-- *********************************************************************** -->
1696<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
1697<!-- *********************************************************************** -->
1698
1699<!-- ======================================================================= -->
1700<div class="doc_subsection">
1701<a name="inlineasm">Inline Assembler Expressions</a>
1702</div>
1703
1704<div class="doc_text">
1705
1706<p>
1707LLVM supports inline assembler expressions (as opposed to <a href="#moduleasm">
1708Module-Level Inline Assembly</a>) through the use of a special value. This
1709value represents the inline assembler as a string (containing the instructions
1710to emit), a list of operand constraints (stored as a string), and a flag that
1711indicates whether or not the inline asm expression has side effects. An example
1712inline assembler expression is:
1713</p>
1714
1715<div class="doc_code">
1716<pre>
1717i32 (i32) asm "bswap $0", "=r,r"
1718</pre>
1719</div>
1720
1721<p>
1722Inline assembler expressions may <b>only</b> be used as the callee operand of
1723a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we have:
1724</p>
1725
1726<div class="doc_code">
1727<pre>
1728%X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
1729</pre>
1730</div>
1731
1732<p>
1733Inline asms with side effects not visible in the constraint list must be marked
1734as having side effects. This is done through the use of the
1735'<tt>sideeffect</tt>' keyword, like so:
1736</p>
1737
1738<div class="doc_code">
1739<pre>
1740call void asm sideeffect "eieio", ""()
1741</pre>
1742</div>
1743
1744<p>TODO: The format of the asm and constraints string still need to be
1745documented here. Constraints on what can be done (e.g. duplication, moving, etc
1746need to be documented).
1747</p>
1748
1749</div>
1750
1751<!-- *********************************************************************** -->
1752<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
1753<!-- *********************************************************************** -->
1754
1755<div class="doc_text">
1756
1757<p>The LLVM instruction set consists of several different
1758classifications of instructions: <a href="#terminators">terminator
1759instructions</a>, <a href="#binaryops">binary instructions</a>,
1760<a href="#bitwiseops">bitwise binary instructions</a>, <a
1761 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
1762instructions</a>.</p>
1763
1764</div>
1765
1766<!-- ======================================================================= -->
1767<div class="doc_subsection"> <a name="terminators">Terminator
1768Instructions</a> </div>
1769
1770<div class="doc_text">
1771
1772<p>As mentioned <a href="#functionstructure">previously</a>, every
1773basic block in a program ends with a "Terminator" instruction, which
1774indicates which block should be executed after the current block is
1775finished. These terminator instructions typically yield a '<tt>void</tt>'
1776value: they produce control flow, not values (the one exception being
1777the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
1778<p>There are six different terminator instructions: the '<a
1779 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
1780instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
1781the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
1782 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
1783 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
1784
1785</div>
1786
1787<!-- _______________________________________________________________________ -->
1788<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
1789Instruction</a> </div>
1790<div class="doc_text">
1791<h5>Syntax:</h5>
1792<pre> ret &lt;type&gt; &lt;value&gt; <i>; Return a value from a non-void function</i>
1793 ret void <i>; Return from void function</i>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1794 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]1795</pre>
1796<h5>Overview:</h5>
1797<p>The '<tt>ret</tt>' instruction is used to return control flow (and a
1798value) from a function back to the caller.</p>
1799<p>There are two forms of the '<tt>ret</tt>' instruction: one that
1800returns a value and then causes control flow, and one that just causes
1801control flow to occur.</p>
1802<h5>Arguments:</h5>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1803<p>The '<tt>ret</tt>' instruction may return one or multiple values. The
Devang Patelec8a5b02008-03-11 05:51:59 +0000[diff] [blame]1804type of each return value must be a '<a href="#t_firstclass">first class</a>'
1805 type. Note that a function is not <a href="#wellformed">well formed</a>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1806if there exists a '<tt>ret</tt>' instruction inside of the function that
Devang Patelec8a5b02008-03-11 05:51:59 +0000[diff] [blame]1807returns values that do not match the return type of the function.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1808<h5>Semantics:</h5>
1809<p>When the '<tt>ret</tt>' instruction is executed, control flow
1810returns back to the calling function's context. If the caller is a "<a
1811 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
1812the instruction after the call. If the caller was an "<a
1813 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
1814at the beginning of the "normal" destination block. If the instruction
1815returns a value, that value shall set the call or invoke instruction's
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1816return value. If the instruction returns multiple values then these
Devang Patelec8a5b02008-03-11 05:51:59 +0000[diff] [blame]1817values can only be accessed through a '<a href="#i_getresult"><tt>getresult</tt>
1818</a>' instruction.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1819<h5>Example:</h5>
1820<pre> ret i32 5 <i>; Return an integer value of 5</i>
1821 ret void <i>; Return from a void function</i>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1822 ret i32 4, i8 2 <i>; Return two values 4 and 2 </i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1823</pre>
1824</div>
1825<!-- _______________________________________________________________________ -->
1826<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
1827<div class="doc_text">
1828<h5>Syntax:</h5>
1829<pre> br i1 &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;<br> br label &lt;dest&gt; <i>; Unconditional branch</i>
1830</pre>
1831<h5>Overview:</h5>
1832<p>The '<tt>br</tt>' instruction is used to cause control flow to
1833transfer to a different basic block in the current function. There are
1834two forms of this instruction, corresponding to a conditional branch
1835and an unconditional branch.</p>
1836<h5>Arguments:</h5>
1837<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
1838single '<tt>i1</tt>' value and two '<tt>label</tt>' values. The
1839unconditional form of the '<tt>br</tt>' instruction takes a single
1840'<tt>label</tt>' value as a target.</p>
1841<h5>Semantics:</h5>
1842<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
1843argument is evaluated. If the value is <tt>true</tt>, control flows
1844to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
1845control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
1846<h5>Example:</h5>
1847<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
1848 href="#i_ret">ret</a> i32 1<br>IfUnequal:<br> <a href="#i_ret">ret</a> i32 0<br></pre>
1849</div>
1850<!-- _______________________________________________________________________ -->
1851<div class="doc_subsubsection">
1852 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
1853</div>
1854
1855<div class="doc_text">
1856<h5>Syntax:</h5>
1857
1858<pre>
1859 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
1860</pre>
1861
1862<h5>Overview:</h5>
1863
1864<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
1865several different places. It is a generalization of the '<tt>br</tt>'
1866instruction, allowing a branch to occur to one of many possible
1867destinations.</p>
1868
1869
1870<h5>Arguments:</h5>
1871
1872<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
1873comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
1874an array of pairs of comparison value constants and '<tt>label</tt>'s. The
1875table is not allowed to contain duplicate constant entries.</p>
1876
1877<h5>Semantics:</h5>
1878
1879<p>The <tt>switch</tt> instruction specifies a table of values and
1880destinations. When the '<tt>switch</tt>' instruction is executed, this
1881table is searched for the given value. If the value is found, control flow is
1882transfered to the corresponding destination; otherwise, control flow is
1883transfered to the default destination.</p>
1884
1885<h5>Implementation:</h5>
1886
1887<p>Depending on properties of the target machine and the particular
1888<tt>switch</tt> instruction, this instruction may be code generated in different
1889ways. For example, it could be generated as a series of chained conditional
1890branches or with a lookup table.</p>
1891
1892<h5>Example:</h5>
1893
1894<pre>
1895 <i>; Emulate a conditional br instruction</i>
1896 %Val = <a href="#i_zext">zext</a> i1 %value to i32
1897 switch i32 %Val, label %truedest [i32 0, label %falsedest ]
1898
1899 <i>; Emulate an unconditional br instruction</i>
1900 switch i32 0, label %dest [ ]
1901
1902 <i>; Implement a jump table:</i>
1903 switch i32 %val, label %otherwise [ i32 0, label %onzero
1904 i32 1, label %onone
1905 i32 2, label %ontwo ]
1906</pre>
1907</div>
1908
1909<!-- _______________________________________________________________________ -->
1910<div class="doc_subsubsection">
1911 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
1912</div>
1913
1914<div class="doc_text">
1915
1916<h5>Syntax:</h5>
1917
1918<pre>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]1919 &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]1920 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
1921</pre>
1922
1923<h5>Overview:</h5>
1924
1925<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
1926function, with the possibility of control flow transfer to either the
1927'<tt>normal</tt>' label or the
1928'<tt>exception</tt>' label. If the callee function returns with the
1929"<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
1930"normal" label. If the callee (or any indirect callees) returns with the "<a
1931href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1932continued at the dynamically nearest "exception" label. If the callee function
Devang Patelec8a5b02008-03-11 05:51:59 +0000[diff] [blame]1933returns multiple values then individual return values are only accessible through
1934a '<tt><a href="#i_getresult">getresult</a></tt>' instruction.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1935
1936<h5>Arguments:</h5>
1937
1938<p>This instruction requires several arguments:</p>
1939
1940<ol>
1941 <li>
1942 The optional "cconv" marker indicates which <a href="#callingconv">calling
1943 convention</a> the call should use. If none is specified, the call defaults
1944 to using C calling conventions.
1945 </li>
1946 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
1947 function value being invoked. In most cases, this is a direct function
1948 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
1949 an arbitrary pointer to function value.
1950 </li>
1951
1952 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
1953 function to be invoked. </li>
1954
1955 <li>'<tt>function args</tt>': argument list whose types match the function
1956 signature argument types. If the function signature indicates the function
1957 accepts a variable number of arguments, the extra arguments can be
1958 specified. </li>
1959
1960 <li>'<tt>normal label</tt>': the label reached when the called function
1961 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
1962
1963 <li>'<tt>exception label</tt>': the label reached when a callee returns with
1964 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
1965
1966</ol>
1967
1968<h5>Semantics:</h5>
1969
1970<p>This instruction is designed to operate as a standard '<tt><a
1971href="#i_call">call</a></tt>' instruction in most regards. The primary
1972difference is that it establishes an association with a label, which is used by
1973the runtime library to unwind the stack.</p>
1974
1975<p>This instruction is used in languages with destructors to ensure that proper
1976cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
1977exception. Additionally, this is important for implementation of
1978'<tt>catch</tt>' clauses in high-level languages that support them.</p>
1979
1980<h5>Example:</h5>
1981<pre>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]1982 %retval = invoke i32 @Test(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1983 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]1984 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1985 unwind label %TestCleanup <i>; {i32}:retval set</i>
1986</pre>
1987</div>
1988
1989
1990<!-- _______________________________________________________________________ -->
1991
1992<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
1993Instruction</a> </div>
1994
1995<div class="doc_text">
1996
1997<h5>Syntax:</h5>
1998<pre>
1999 unwind
2000</pre>
2001
2002<h5>Overview:</h5>
2003
2004<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
2005at the first callee in the dynamic call stack which used an <a
2006href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
2007primarily used to implement exception handling.</p>
2008
2009<h5>Semantics:</h5>
2010
2011<p>The '<tt>unwind</tt>' intrinsic causes execution of the current function to
2012immediately halt. The dynamic call stack is then searched for the first <a
2013href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
2014execution continues at the "exceptional" destination block specified by the
2015<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
2016dynamic call chain, undefined behavior results.</p>
2017</div>
2018
2019<!-- _______________________________________________________________________ -->
2020
2021<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
2022Instruction</a> </div>
2023
2024<div class="doc_text">
2025
2026<h5>Syntax:</h5>
2027<pre>
2028 unreachable
2029</pre>
2030
2031<h5>Overview:</h5>
2032
2033<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
2034instruction is used to inform the optimizer that a particular portion of the
2035code is not reachable. This can be used to indicate that the code after a
2036no-return function cannot be reached, and other facts.</p>
2037
2038<h5>Semantics:</h5>
2039
2040<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
2041</div>
2042
2043
2044
2045<!-- ======================================================================= -->
2046<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
2047<div class="doc_text">
2048<p>Binary operators are used to do most of the computation in a
2049program. They require two operands, execute an operation on them, and
2050produce a single value. The operands might represent
2051multiple data, as is the case with the <a href="#t_vector">vector</a> data type.
2052The result value of a binary operator is not
2053necessarily the same type as its operands.</p>
2054<p>There are several different binary operators:</p>
2055</div>
2056<!-- _______________________________________________________________________ -->
2057<div class="doc_subsubsection"> <a name="i_add">'<tt>add</tt>'
2058Instruction</a> </div>
2059<div class="doc_text">
2060<h5>Syntax:</h5>
2061<pre> &lt;result&gt; = add &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2062</pre>
2063<h5>Overview:</h5>
2064<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
2065<h5>Arguments:</h5>
2066<p>The two arguments to the '<tt>add</tt>' instruction must be either <a
2067 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a> values.
2068 This instruction can also take <a href="#t_vector">vector</a> versions of the values.
2069Both arguments must have identical types.</p>
2070<h5>Semantics:</h5>
2071<p>The value produced is the integer or floating point sum of the two
2072operands.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2073<p>If an integer sum has unsigned overflow, the result returned is the
2074mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2075the result.</p>
2076<p>Because LLVM integers use a two's complement representation, this
2077instruction is appropriate for both signed and unsigned integers.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2078<h5>Example:</h5>
2079<pre> &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
2080</pre>
2081</div>
2082<!-- _______________________________________________________________________ -->
2083<div class="doc_subsubsection"> <a name="i_sub">'<tt>sub</tt>'
2084Instruction</a> </div>
2085<div class="doc_text">
2086<h5>Syntax:</h5>
2087<pre> &lt;result&gt; = sub &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2088</pre>
2089<h5>Overview:</h5>
2090<p>The '<tt>sub</tt>' instruction returns the difference of its two
2091operands.</p>
2092<p>Note that the '<tt>sub</tt>' instruction is used to represent the '<tt>neg</tt>'
2093instruction present in most other intermediate representations.</p>
2094<h5>Arguments:</h5>
2095<p>The two arguments to the '<tt>sub</tt>' instruction must be either <a
2096 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
2097values.
2098This instruction can also take <a href="#t_vector">vector</a> versions of the values.
2099Both arguments must have identical types.</p>
2100<h5>Semantics:</h5>
2101<p>The value produced is the integer or floating point difference of
2102the two operands.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2103<p>If an integer difference has unsigned overflow, the result returned is the
2104mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2105the result.</p>
2106<p>Because LLVM integers use a two's complement representation, this
2107instruction is appropriate for both signed and unsigned integers.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2108<h5>Example:</h5>
2109<pre>
2110 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
2111 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
2112</pre>
2113</div>
2114<!-- _______________________________________________________________________ -->
2115<div class="doc_subsubsection"> <a name="i_mul">'<tt>mul</tt>'
2116Instruction</a> </div>
2117<div class="doc_text">
2118<h5>Syntax:</h5>
2119<pre> &lt;result&gt; = mul &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2120</pre>
2121<h5>Overview:</h5>
2122<p>The '<tt>mul</tt>' instruction returns the product of its two
2123operands.</p>
2124<h5>Arguments:</h5>
2125<p>The two arguments to the '<tt>mul</tt>' instruction must be either <a
2126 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
2127values.
2128This instruction can also take <a href="#t_vector">vector</a> versions of the values.
2129Both arguments must have identical types.</p>
2130<h5>Semantics:</h5>
2131<p>The value produced is the integer or floating point product of the
2132two operands.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2133<p>If the result of an integer multiplication has unsigned overflow,
2134the result returned is the mathematical result modulo
21352<sup>n</sup>, where n is the bit width of the result.</p>
2136<p>Because LLVM integers use a two's complement representation, and the
2137result is the same width as the operands, this instruction returns the
2138correct result for both signed and unsigned integers. If a full product
2139(e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands
2140should be sign-extended or zero-extended as appropriate to the
2141width of the full product.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2142<h5>Example:</h5>
2143<pre> &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
2144</pre>
2145</div>
2146<!-- _______________________________________________________________________ -->
2147<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
2148</a></div>
2149<div class="doc_text">
2150<h5>Syntax:</h5>
2151<pre> &lt;result&gt; = udiv &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2152</pre>
2153<h5>Overview:</h5>
2154<p>The '<tt>udiv</tt>' instruction returns the quotient of its two
2155operands.</p>
2156<h5>Arguments:</h5>
2157<p>The two arguments to the '<tt>udiv</tt>' instruction must be
2158<a href="#t_integer">integer</a> values. Both arguments must have identical
2159types. This instruction can also take <a href="#t_vector">vector</a> versions
2160of the values in which case the elements must be integers.</p>
2161<h5>Semantics:</h5>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2162<p>The value produced is the unsigned integer quotient of the two operands.</p>
2163<p>Note that unsigned integer division and signed integer division are distinct
2164operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
2165<p>Division by zero leads to undefined behavior.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2166<h5>Example:</h5>
2167<pre> &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2168</pre>
2169</div>
2170<!-- _______________________________________________________________________ -->
2171<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
2172</a> </div>
2173<div class="doc_text">
2174<h5>Syntax:</h5>
2175<pre> &lt;result&gt; = sdiv &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2176</pre>
2177<h5>Overview:</h5>
2178<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two
2179operands.</p>
2180<h5>Arguments:</h5>
2181<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
2182<a href="#t_integer">integer</a> values. Both arguments must have identical
2183types. This instruction can also take <a href="#t_vector">vector</a> versions
2184of the values in which case the elements must be integers.</p>
2185<h5>Semantics:</h5>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2186<p>The value produced is the signed integer quotient of the two operands.</p>
2187<p>Note that signed integer division and unsigned integer division are distinct
2188operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
2189<p>Division by zero leads to undefined behavior. Overflow also leads to
2190undefined behavior; this is a rare case, but can occur, for example,
2191by doing a 32-bit division of -2147483648 by -1.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2192<h5>Example:</h5>
2193<pre> &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2194</pre>
2195</div>
2196<!-- _______________________________________________________________________ -->
2197<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
2198Instruction</a> </div>
2199<div class="doc_text">
2200<h5>Syntax:</h5>
2201<pre> &lt;result&gt; = fdiv &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2202</pre>
2203<h5>Overview:</h5>
2204<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two
2205operands.</p>
2206<h5>Arguments:</h5>
2207<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
2208<a href="#t_floating">floating point</a> values. Both arguments must have
2209identical types. This instruction can also take <a href="#t_vector">vector</a>
2210versions of floating point values.</p>
2211<h5>Semantics:</h5>
2212<p>The value produced is the floating point quotient of the two operands.</p>
2213<h5>Example:</h5>
2214<pre> &lt;result&gt; = fdiv float 4.0, %var <i>; yields {float}:result = 4.0 / %var</i>
2215</pre>
2216</div>
2217<!-- _______________________________________________________________________ -->
2218<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
2219</div>
2220<div class="doc_text">
2221<h5>Syntax:</h5>
2222<pre> &lt;result&gt; = urem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2223</pre>
2224<h5>Overview:</h5>
2225<p>The '<tt>urem</tt>' instruction returns the remainder from the
2226unsigned division of its two arguments.</p>
2227<h5>Arguments:</h5>
2228<p>The two arguments to the '<tt>urem</tt>' instruction must be
2229<a href="#t_integer">integer</a> values. Both arguments must have identical
Dan Gohman3e3fd8c2007-11-05 23:35:22 +0000[diff] [blame]2230types. This instruction can also take <a href="#t_vector">vector</a> versions
2231of the values in which case the elements must be integers.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2232<h5>Semantics:</h5>
2233<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
2234This instruction always performs an unsigned division to get the remainder,
2235regardless of whether the arguments are unsigned or not.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2236<p>Note that unsigned integer remainder and signed integer remainder are
2237distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
2238<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2239<h5>Example:</h5>
2240<pre> &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2241</pre>
2242
2243</div>
2244<!-- _______________________________________________________________________ -->
2245<div class="doc_subsubsection"> <a name="i_srem">'<tt>srem</tt>'
2246Instruction</a> </div>
2247<div class="doc_text">
2248<h5>Syntax:</h5>
2249<pre> &lt;result&gt; = srem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2250</pre>
2251<h5>Overview:</h5>
2252<p>The '<tt>srem</tt>' instruction returns the remainder from the
Dan Gohman3e3fd8c2007-11-05 23:35:22 +0000[diff] [blame]2253signed division of its two operands. This instruction can also take
2254<a href="#t_vector">vector</a> versions of the values in which case
2255the elements must be integers.</p>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]2256
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2257<h5>Arguments:</h5>
2258<p>The two arguments to the '<tt>srem</tt>' instruction must be
2259<a href="#t_integer">integer</a> values. Both arguments must have identical
2260types.</p>
2261<h5>Semantics:</h5>
2262<p>This instruction returns the <i>remainder</i> of a division (where the result
2263has the same sign as the dividend, <tt>var1</tt>), not the <i>modulo</i>
2264operator (where the result has the same sign as the divisor, <tt>var2</tt>) of
2265a value. For more information about the difference, see <a
2266 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
2267Math Forum</a>. For a table of how this is implemented in various languages,
2268please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
2269Wikipedia: modulo operation</a>.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2270<p>Note that signed integer remainder and unsigned integer remainder are
2271distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
2272<p>Taking the remainder of a division by zero leads to undefined behavior.
2273Overflow also leads to undefined behavior; this is a rare case, but can occur,
2274for example, by taking the remainder of a 32-bit division of -2147483648 by -1.
2275(The remainder doesn't actually overflow, but this rule lets srem be
2276implemented using instructions that return both the result of the division
2277and the remainder.)</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2278<h5>Example:</h5>
2279<pre> &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2280</pre>
2281
2282</div>
2283<!-- _______________________________________________________________________ -->
2284<div class="doc_subsubsection"> <a name="i_frem">'<tt>frem</tt>'
2285Instruction</a> </div>
2286<div class="doc_text">
2287<h5>Syntax:</h5>
2288<pre> &lt;result&gt; = frem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2289</pre>
2290<h5>Overview:</h5>
2291<p>The '<tt>frem</tt>' instruction returns the remainder from the
2292division of its two operands.</p>
2293<h5>Arguments:</h5>
2294<p>The two arguments to the '<tt>frem</tt>' instruction must be
2295<a href="#t_floating">floating point</a> values. Both arguments must have
Dan Gohman3e3fd8c2007-11-05 23:35:22 +0000[diff] [blame]2296identical types. This instruction can also take <a href="#t_vector">vector</a>
2297versions of floating point values.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2298<h5>Semantics:</h5>
2299<p>This instruction returns the <i>remainder</i> of a division.</p>
2300<h5>Example:</h5>
2301<pre> &lt;result&gt; = frem float 4.0, %var <i>; yields {float}:result = 4.0 % %var</i>
2302</pre>
2303</div>
2304
2305<!-- ======================================================================= -->
2306<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
2307Operations</a> </div>
2308<div class="doc_text">
2309<p>Bitwise binary operators are used to do various forms of
2310bit-twiddling in a program. They are generally very efficient
2311instructions and can commonly be strength reduced from other
2312instructions. They require two operands, execute an operation on them,
2313and produce a single value. The resulting value of the bitwise binary
2314operators is always the same type as its first operand.</p>
2315</div>
2316
2317<!-- _______________________________________________________________________ -->
2318<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
2319Instruction</a> </div>
2320<div class="doc_text">
2321<h5>Syntax:</h5>
2322<pre> &lt;result&gt; = shl &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2323</pre>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2324
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2325<h5>Overview:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2326
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2327<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
2328the left a specified number of bits.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2329
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2330<h5>Arguments:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2331
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2332<p>Both arguments to the '<tt>shl</tt>' instruction must be the same <a
2333 href="#t_integer">integer</a> type.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2334
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2335<h5>Semantics:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2336
2337<p>The value produced is <tt>var1</tt> * 2<sup><tt>var2</tt></sup>. If
2338<tt>var2</tt> is (statically or dynamically) equal to or larger than the number
2339of bits in <tt>var1</tt>, the result is undefined.</p>
2340
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2341<h5>Example:</h5><pre>
2342 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
2343 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
2344 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2345 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2346</pre>
2347</div>
2348<!-- _______________________________________________________________________ -->
2349<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
2350Instruction</a> </div>
2351<div class="doc_text">
2352<h5>Syntax:</h5>
2353<pre> &lt;result&gt; = lshr &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2354</pre>
2355
2356<h5>Overview:</h5>
2357<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
2358operand shifted to the right a specified number of bits with zero fill.</p>
2359
2360<h5>Arguments:</h5>
2361<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
2362<a href="#t_integer">integer</a> type.</p>
2363
2364<h5>Semantics:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2365
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2366<p>This instruction always performs a logical shift right operation. The most
2367significant bits of the result will be filled with zero bits after the
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2368shift. If <tt>var2</tt> is (statically or dynamically) equal to or larger than
2369the number of bits in <tt>var1</tt>, the result is undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2370
2371<h5>Example:</h5>
2372<pre>
2373 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
2374 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
2375 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
2376 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2377 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2378</pre>
2379</div>
2380
2381<!-- _______________________________________________________________________ -->
2382<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
2383Instruction</a> </div>
2384<div class="doc_text">
2385
2386<h5>Syntax:</h5>
2387<pre> &lt;result&gt; = ashr &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2388</pre>
2389
2390<h5>Overview:</h5>
2391<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
2392operand shifted to the right a specified number of bits with sign extension.</p>
2393
2394<h5>Arguments:</h5>
2395<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
2396<a href="#t_integer">integer</a> type.</p>
2397
2398<h5>Semantics:</h5>
2399<p>This instruction always performs an arithmetic shift right operation,
2400The most significant bits of the result will be filled with the sign bit
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2401of <tt>var1</tt>. If <tt>var2</tt> is (statically or dynamically) equal to or
2402larger than the number of bits in <tt>var1</tt>, the result is undefined.
2403</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2404
2405<h5>Example:</h5>
2406<pre>
2407 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
2408 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
2409 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
2410 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2411 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2412</pre>
2413</div>
2414
2415<!-- _______________________________________________________________________ -->
2416<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
2417Instruction</a> </div>
2418<div class="doc_text">
2419<h5>Syntax:</h5>
2420<pre> &lt;result&gt; = and &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2421</pre>
2422<h5>Overview:</h5>
2423<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
2424its two operands.</p>
2425<h5>Arguments:</h5>
2426<p>The two arguments to the '<tt>and</tt>' instruction must be <a
2427 href="#t_integer">integer</a> values. Both arguments must have
2428identical types.</p>
2429<h5>Semantics:</h5>
2430<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
2431<p> </p>
2432<div style="align: center">
2433<table border="1" cellspacing="0" cellpadding="4">
2434 <tbody>
2435 <tr>
2436 <td>In0</td>
2437 <td>In1</td>
2438 <td>Out</td>
2439 </tr>
2440 <tr>
2441 <td>0</td>
2442 <td>0</td>
2443 <td>0</td>
2444 </tr>
2445 <tr>
2446 <td>0</td>
2447 <td>1</td>
2448 <td>0</td>
2449 </tr>
2450 <tr>
2451 <td>1</td>
2452 <td>0</td>
2453 <td>0</td>
2454 </tr>
2455 <tr>
2456 <td>1</td>
2457 <td>1</td>
2458 <td>1</td>
2459 </tr>
2460 </tbody>
2461</table>
2462</div>
2463<h5>Example:</h5>
2464<pre> &lt;result&gt; = and i32 4, %var <i>; yields {i32}:result = 4 &amp; %var</i>
2465 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
2466 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
2467</pre>
2468</div>
2469<!-- _______________________________________________________________________ -->
2470<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
2471<div class="doc_text">
2472<h5>Syntax:</h5>
2473<pre> &lt;result&gt; = or &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2474</pre>
2475<h5>Overview:</h5>
2476<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
2477or of its two operands.</p>
2478<h5>Arguments:</h5>
2479<p>The two arguments to the '<tt>or</tt>' instruction must be <a
2480 href="#t_integer">integer</a> values. Both arguments must have
2481identical types.</p>
2482<h5>Semantics:</h5>
2483<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
2484<p> </p>
2485<div style="align: center">
2486<table border="1" cellspacing="0" cellpadding="4">
2487 <tbody>
2488 <tr>
2489 <td>In0</td>
2490 <td>In1</td>
2491 <td>Out</td>
2492 </tr>
2493 <tr>
2494 <td>0</td>
2495 <td>0</td>
2496 <td>0</td>
2497 </tr>
2498 <tr>
2499 <td>0</td>
2500 <td>1</td>
2501 <td>1</td>
2502 </tr>
2503 <tr>
2504 <td>1</td>
2505 <td>0</td>
2506 <td>1</td>
2507 </tr>
2508 <tr>
2509 <td>1</td>
2510 <td>1</td>
2511 <td>1</td>
2512 </tr>
2513 </tbody>
2514</table>
2515</div>
2516<h5>Example:</h5>
2517<pre> &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
2518 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
2519 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
2520</pre>
2521</div>
2522<!-- _______________________________________________________________________ -->
2523<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
2524Instruction</a> </div>
2525<div class="doc_text">
2526<h5>Syntax:</h5>
2527<pre> &lt;result&gt; = xor &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2528</pre>
2529<h5>Overview:</h5>
2530<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
2531or of its two operands. The <tt>xor</tt> is used to implement the
2532"one's complement" operation, which is the "~" operator in C.</p>
2533<h5>Arguments:</h5>
2534<p>The two arguments to the '<tt>xor</tt>' instruction must be <a
2535 href="#t_integer">integer</a> values. Both arguments must have
2536identical types.</p>
2537<h5>Semantics:</h5>
2538<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
2539<p> </p>
2540<div style="align: center">
2541<table border="1" cellspacing="0" cellpadding="4">
2542 <tbody>
2543 <tr>
2544 <td>In0</td>
2545 <td>In1</td>
2546 <td>Out</td>
2547 </tr>
2548 <tr>
2549 <td>0</td>
2550 <td>0</td>
2551 <td>0</td>
2552 </tr>
2553 <tr>
2554 <td>0</td>
2555 <td>1</td>
2556 <td>1</td>
2557 </tr>
2558 <tr>
2559 <td>1</td>
2560 <td>0</td>
2561 <td>1</td>
2562 </tr>
2563 <tr>
2564 <td>1</td>
2565 <td>1</td>
2566 <td>0</td>
2567 </tr>
2568 </tbody>
2569</table>
2570</div>
2571<p> </p>
2572<h5>Example:</h5>
2573<pre> &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
2574 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
2575 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
2576 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
2577</pre>
2578</div>
2579
2580<!-- ======================================================================= -->
2581<div class="doc_subsection">
2582 <a name="vectorops">Vector Operations</a>
2583</div>
2584
2585<div class="doc_text">
2586
2587<p>LLVM supports several instructions to represent vector operations in a
2588target-independent manner. These instructions cover the element-access and
2589vector-specific operations needed to process vectors effectively. While LLVM
2590does directly support these vector operations, many sophisticated algorithms
2591will want to use target-specific intrinsics to take full advantage of a specific
2592target.</p>
2593
2594</div>
2595
2596<!-- _______________________________________________________________________ -->
2597<div class="doc_subsubsection">
2598 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
2599</div>
2600
2601<div class="doc_text">
2602
2603<h5>Syntax:</h5>
2604
2605<pre>
2606 &lt;result&gt; = extractelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, i32 &lt;idx&gt; <i>; yields &lt;ty&gt;</i>
2607</pre>
2608
2609<h5>Overview:</h5>
2610
2611<p>
2612The '<tt>extractelement</tt>' instruction extracts a single scalar
2613element from a vector at a specified index.
2614</p>
2615
2616
2617<h5>Arguments:</h5>
2618
2619<p>
2620The first operand of an '<tt>extractelement</tt>' instruction is a
2621value of <a href="#t_vector">vector</a> type. The second operand is
2622an index indicating the position from which to extract the element.
2623The index may be a variable.</p>
2624
2625<h5>Semantics:</h5>
2626
2627<p>
2628The result is a scalar of the same type as the element type of
2629<tt>val</tt>. Its value is the value at position <tt>idx</tt> of
2630<tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
2631results are undefined.
2632</p>
2633
2634<h5>Example:</h5>
2635
2636<pre>
2637 %result = extractelement &lt;4 x i32&gt; %vec, i32 0 <i>; yields i32</i>
2638</pre>
2639</div>
2640
2641
2642<!-- _______________________________________________________________________ -->
2643<div class="doc_subsubsection">
2644 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
2645</div>
2646
2647<div class="doc_text">
2648
2649<h5>Syntax:</h5>
2650
2651<pre>
2652 &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>
2653</pre>
2654
2655<h5>Overview:</h5>
2656
2657<p>
2658The '<tt>insertelement</tt>' instruction inserts a scalar
2659element into a vector at a specified index.
2660</p>
2661
2662
2663<h5>Arguments:</h5>
2664
2665<p>
2666The first operand of an '<tt>insertelement</tt>' instruction is a
2667value of <a href="#t_vector">vector</a> type. The second operand is a
2668scalar value whose type must equal the element type of the first
2669operand. The third operand is an index indicating the position at
2670which to insert the value. The index may be a variable.</p>
2671
2672<h5>Semantics:</h5>
2673
2674<p>
2675The result is a vector of the same type as <tt>val</tt>. Its
2676element values are those of <tt>val</tt> except at position
2677<tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
2678exceeds the length of <tt>val</tt>, the results are undefined.
2679</p>
2680
2681<h5>Example:</h5>
2682
2683<pre>
2684 %result = insertelement &lt;4 x i32&gt; %vec, i32 1, i32 0 <i>; yields &lt;4 x i32&gt;</i>
2685</pre>
2686</div>
2687
2688<!-- _______________________________________________________________________ -->
2689<div class="doc_subsubsection">
2690 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
2691</div>
2692
2693<div class="doc_text">
2694
2695<h5>Syntax:</h5>
2696
2697<pre>
2698 &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>
2699</pre>
2700
2701<h5>Overview:</h5>
2702
2703<p>
2704The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
2705from two input vectors, returning a vector of the same type.
2706</p>
2707
2708<h5>Arguments:</h5>
2709
2710<p>
2711The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
2712with types that match each other and types that match the result of the
2713instruction. The third argument is a shuffle mask, which has the same number
2714of elements as the other vector type, but whose element type is always 'i32'.
2715</p>
2716
2717<p>
2718The shuffle mask operand is required to be a constant vector with either
2719constant integer or undef values.
2720</p>
2721
2722<h5>Semantics:</h5>
2723
2724<p>
2725The elements of the two input vectors are numbered from left to right across
2726both of the vectors. The shuffle mask operand specifies, for each element of
2727the result vector, which element of the two input registers the result element
2728gets. The element selector may be undef (meaning "don't care") and the second
2729operand may be undef if performing a shuffle from only one vector.
2730</p>
2731
2732<h5>Example:</h5>
2733
2734<pre>
2735 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
2736 &lt;4 x i32&gt; &lt;i32 0, i32 4, i32 1, i32 5&gt; <i>; yields &lt;4 x i32&gt;</i>
2737 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; undef,
2738 &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.
2739</pre>
2740</div>
2741
2742
2743<!-- ======================================================================= -->
2744<div class="doc_subsection">
2745 <a name="memoryops">Memory Access and Addressing Operations</a>
2746</div>
2747
2748<div class="doc_text">
2749
2750<p>A key design point of an SSA-based representation is how it
2751represents memory. In LLVM, no memory locations are in SSA form, which
2752makes things very simple. This section describes how to read, write,
2753allocate, and free memory in LLVM.</p>
2754
2755</div>
2756
2757<!-- _______________________________________________________________________ -->
2758<div class="doc_subsubsection">
2759 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
2760</div>
2761
2762<div class="doc_text">
2763
2764<h5>Syntax:</h5>
2765
2766<pre>
2767 &lt;result&gt; = malloc &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
2768</pre>
2769
2770<h5>Overview:</h5>
2771
2772<p>The '<tt>malloc</tt>' instruction allocates memory from the system
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]2773heap and returns a pointer to it. The object is always allocated in the generic
2774address space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2775
2776<h5>Arguments:</h5>
2777
2778<p>The '<tt>malloc</tt>' instruction allocates
2779<tt>sizeof(&lt;type&gt;)*NumElements</tt>
2780bytes of memory from the operating system and returns a pointer of the
2781appropriate type to the program. If "NumElements" is specified, it is the
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]2782number of elements allocated, otherwise "NumElements" is defaulted to be one.
2783If an alignment is specified, the value result of the allocation is guaranteed to
2784be aligned to at least that boundary. If not specified, or if zero, the target can
2785choose to align the allocation on any convenient boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2786
2787<p>'<tt>type</tt>' must be a sized type.</p>
2788
2789<h5>Semantics:</h5>
2790
2791<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
2792a pointer is returned.</p>
2793
2794<h5>Example:</h5>
2795
2796<pre>
2797 %array = malloc [4 x i8 ] <i>; yields {[%4 x i8]*}:array</i>
2798
2799 %size = <a href="#i_add">add</a> i32 2, 2 <i>; yields {i32}:size = i32 4</i>
2800 %array1 = malloc i8, i32 4 <i>; yields {i8*}:array1</i>
2801 %array2 = malloc [12 x i8], i32 %size <i>; yields {[12 x i8]*}:array2</i>
2802 %array3 = malloc i32, i32 4, align 1024 <i>; yields {i32*}:array3</i>
2803 %array4 = malloc i32, align 1024 <i>; yields {i32*}:array4</i>
2804</pre>
2805</div>
2806
2807<!-- _______________________________________________________________________ -->
2808<div class="doc_subsubsection">
2809 <a name="i_free">'<tt>free</tt>' Instruction</a>
2810</div>
2811
2812<div class="doc_text">
2813
2814<h5>Syntax:</h5>
2815
2816<pre>
2817 free &lt;type&gt; &lt;value&gt; <i>; yields {void}</i>
2818</pre>
2819
2820<h5>Overview:</h5>
2821
2822<p>The '<tt>free</tt>' instruction returns memory back to the unused
2823memory heap to be reallocated in the future.</p>
2824
2825<h5>Arguments:</h5>
2826
2827<p>'<tt>value</tt>' shall be a pointer value that points to a value
2828that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
2829instruction.</p>
2830
2831<h5>Semantics:</h5>
2832
2833<p>Access to the memory pointed to by the pointer is no longer defined
2834after this instruction executes.</p>
2835
2836<h5>Example:</h5>
2837
2838<pre>
2839 %array = <a href="#i_malloc">malloc</a> [4 x i8] <i>; yields {[4 x i8]*}:array</i>
2840 free [4 x i8]* %array
2841</pre>
2842</div>
2843
2844<!-- _______________________________________________________________________ -->
2845<div class="doc_subsubsection">
2846 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
2847</div>
2848
2849<div class="doc_text">
2850
2851<h5>Syntax:</h5>
2852
2853<pre>
2854 &lt;result&gt; = alloca &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
2855</pre>
2856
2857<h5>Overview:</h5>
2858
2859<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
2860currently executing function, to be automatically released when this function
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]2861returns to its caller. The object is always allocated in the generic address
2862space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2863
2864<h5>Arguments:</h5>
2865
2866<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
2867bytes of memory on the runtime stack, returning a pointer of the
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]2868appropriate type to the program. If "NumElements" is specified, it is the
2869number of elements allocated, otherwise "NumElements" is defaulted to be one.
2870If an alignment is specified, the value result of the allocation is guaranteed
2871to be aligned to at least that boundary. If not specified, or if zero, the target
2872can choose to align the allocation on any convenient boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2873
2874<p>'<tt>type</tt>' may be any sized type.</p>
2875
2876<h5>Semantics:</h5>
2877
2878<p>Memory is allocated; a pointer is returned. '<tt>alloca</tt>'d
2879memory is automatically released when the function returns. The '<tt>alloca</tt>'
2880instruction is commonly used to represent automatic variables that must
2881have an address available. When the function returns (either with the <tt><a
2882 href="#i_ret">ret</a></tt> or <tt><a href="#i_unwind">unwind</a></tt>
2883instructions), the memory is reclaimed.</p>
2884
2885<h5>Example:</h5>
2886
2887<pre>
2888 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
2889 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
2890 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
2891 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
2892</pre>
2893</div>
2894
2895<!-- _______________________________________________________________________ -->
2896<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
2897Instruction</a> </div>
2898<div class="doc_text">
2899<h5>Syntax:</h5>
2900<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>
2901<h5>Overview:</h5>
2902<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
2903<h5>Arguments:</h5>
2904<p>The argument to the '<tt>load</tt>' instruction specifies the memory
2905address from which to load. The pointer must point to a <a
2906 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
2907marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
2908the number or order of execution of this <tt>load</tt> with other
2909volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
2910instructions. </p>
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]2911<p>
2912The optional "align" argument specifies the alignment of the operation
2913(that is, the alignment of the memory address). A value of 0 or an
2914omitted "align" argument means that the operation has the preferential
2915alignment for the target. It is the responsibility of the code emitter
2916to ensure that the alignment information is correct. Overestimating
2917the alignment results in an undefined behavior. Underestimating the
2918alignment may produce less efficient code. An alignment of 1 is always
2919safe.
2920</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2921<h5>Semantics:</h5>
2922<p>The location of memory pointed to is loaded.</p>
2923<h5>Examples:</h5>
2924<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
2925 <a
2926 href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
2927 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
2928</pre>
2929</div>
2930<!-- _______________________________________________________________________ -->
2931<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
2932Instruction</a> </div>
2933<div class="doc_text">
2934<h5>Syntax:</h5>
2935<pre> store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
2936 volatile store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
2937</pre>
2938<h5>Overview:</h5>
2939<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
2940<h5>Arguments:</h5>
2941<p>There are two arguments to the '<tt>store</tt>' instruction: a value
2942to store and an address at which to store it. The type of the '<tt>&lt;pointer&gt;</tt>'
2943operand must be a pointer to the type of the '<tt>&lt;value&gt;</tt>'
2944operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
2945optimizer is not allowed to modify the number or order of execution of
2946this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
2947 href="#i_store">store</a></tt> instructions.</p>
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]2948<p>
2949The optional "align" argument specifies the alignment of the operation
2950(that is, the alignment of the memory address). A value of 0 or an
2951omitted "align" argument means that the operation has the preferential
2952alignment for the target. It is the responsibility of the code emitter
2953to ensure that the alignment information is correct. Overestimating
2954the alignment results in an undefined behavior. Underestimating the
2955alignment may produce less efficient code. An alignment of 1 is always
2956safe.
2957</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2958<h5>Semantics:</h5>
2959<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>'
2960at the location specified by the '<tt>&lt;pointer&gt;</tt>' operand.</p>
2961<h5>Example:</h5>
2962<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling63ffa142007-10-22 05:10:05 +0000[diff] [blame]2963 store i32 3, i32* %ptr <i>; yields {void}</i>
2964 %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]2965</pre>
2966</div>
2967
2968<!-- _______________________________________________________________________ -->
2969<div class="doc_subsubsection">
2970 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
2971</div>
2972
2973<div class="doc_text">
2974<h5>Syntax:</h5>
2975<pre>
2976 &lt;result&gt; = getelementptr &lt;ty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
2977</pre>
2978
2979<h5>Overview:</h5>
2980
2981<p>
2982The '<tt>getelementptr</tt>' instruction is used to get the address of a
2983subelement of an aggregate data structure.</p>
2984
2985<h5>Arguments:</h5>
2986
2987<p>This instruction takes a list of integer operands that indicate what
2988elements of the aggregate object to index to. The actual types of the arguments
2989provided depend on the type of the first pointer argument. The
2990'<tt>getelementptr</tt>' instruction is used to index down through the type
2991levels of a structure or to a specific index in an array. When indexing into a
2992structure, only <tt>i32</tt> integer constants are allowed. When indexing
2993into an array or pointer, only integers of 32 or 64 bits are allowed, and will
2994be sign extended to 64-bit values.</p>
2995
2996<p>For example, let's consider a C code fragment and how it gets
2997compiled to LLVM:</p>
2998
2999<div class="doc_code">
3000<pre>
3001struct RT {
3002 char A;
3003 int B[10][20];
3004 char C;
3005};
3006struct ST {
3007 int X;
3008 double Y;
3009 struct RT Z;
3010};
3011
3012int *foo(struct ST *s) {
3013 return &amp;s[1].Z.B[5][13];
3014}
3015</pre>
3016</div>
3017
3018<p>The LLVM code generated by the GCC frontend is:</p>
3019
3020<div class="doc_code">
3021<pre>
3022%RT = type { i8 , [10 x [20 x i32]], i8 }
3023%ST = type { i32, double, %RT }
3024
3025define i32* %foo(%ST* %s) {
3026entry:
3027 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
3028 ret i32* %reg
3029}
3030</pre>
3031</div>
3032
3033<h5>Semantics:</h5>
3034
3035<p>The index types specified for the '<tt>getelementptr</tt>' instruction depend
3036on the pointer type that is being indexed into. <a href="#t_pointer">Pointer</a>
3037and <a href="#t_array">array</a> types can use a 32-bit or 64-bit
3038<a href="#t_integer">integer</a> type but the value will always be sign extended
3039to 64-bits. <a href="#t_struct">Structure</a> types require <tt>i32</tt>
3040<b>constants</b>.</p>
3041
3042<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
3043type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
3044}</tt>' type, a structure. The second index indexes into the third element of
3045the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
3046i8 }</tt>' type, another structure. The third index indexes into the second
3047element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
3048array. The two dimensions of the array are subscripted into, yielding an
3049'<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
3050to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
3051
3052<p>Note that it is perfectly legal to index partially through a
3053structure, returning a pointer to an inner element. Because of this,
3054the LLVM code for the given testcase is equivalent to:</p>
3055
3056<pre>
3057 define i32* %foo(%ST* %s) {
3058 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
3059 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
3060 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
3061 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
3062 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
3063 ret i32* %t5
3064 }
3065</pre>
3066
3067<p>Note that it is undefined to access an array out of bounds: array and
3068pointer indexes must always be within the defined bounds of the array type.
3069The one exception for this rules is zero length arrays. These arrays are
3070defined to be accessible as variable length arrays, which requires access
3071beyond the zero'th element.</p>
3072
3073<p>The getelementptr instruction is often confusing. For some more insight
3074into how it works, see <a href="GetElementPtr.html">the getelementptr
3075FAQ</a>.</p>
3076
3077<h5>Example:</h5>
3078
3079<pre>
3080 <i>; yields [12 x i8]*:aptr</i>
3081 %aptr = getelementptr {i32, [12 x i8]}* %sptr, i64 0, i32 1
3082</pre>
3083</div>
3084
3085<!-- ======================================================================= -->
3086<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
3087</div>
3088<div class="doc_text">
3089<p>The instructions in this category are the conversion instructions (casting)
3090which all take a single operand and a type. They perform various bit conversions
3091on the operand.</p>
3092</div>
3093
3094<!-- _______________________________________________________________________ -->
3095<div class="doc_subsubsection">
3096 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
3097</div>
3098<div class="doc_text">
3099
3100<h5>Syntax:</h5>
3101<pre>
3102 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3103</pre>
3104
3105<h5>Overview:</h5>
3106<p>
3107The '<tt>trunc</tt>' instruction truncates its operand to the type <tt>ty2</tt>.
3108</p>
3109
3110<h5>Arguments:</h5>
3111<p>
3112The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
3113be an <a href="#t_integer">integer</a> type, and a type that specifies the size
3114and type of the result, which must be an <a href="#t_integer">integer</a>
3115type. The bit size of <tt>value</tt> must be larger than the bit size of
3116<tt>ty2</tt>. Equal sized types are not allowed.</p>
3117
3118<h5>Semantics:</h5>
3119<p>
3120The '<tt>trunc</tt>' instruction truncates the high order bits in <tt>value</tt>
3121and converts the remaining bits to <tt>ty2</tt>. Since the source size must be
3122larger than the destination size, <tt>trunc</tt> cannot be a <i>no-op cast</i>.
3123It will always truncate bits.</p>
3124
3125<h5>Example:</h5>
3126<pre>
3127 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
3128 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
3129 %Y = trunc i32 122 to i1 <i>; yields i1:false</i>
3130</pre>
3131</div>
3132
3133<!-- _______________________________________________________________________ -->
3134<div class="doc_subsubsection">
3135 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
3136</div>
3137<div class="doc_text">
3138
3139<h5>Syntax:</h5>
3140<pre>
3141 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3142</pre>
3143
3144<h5>Overview:</h5>
3145<p>The '<tt>zext</tt>' instruction zero extends its operand to type
3146<tt>ty2</tt>.</p>
3147
3148
3149<h5>Arguments:</h5>
3150<p>The '<tt>zext</tt>' instruction takes a value to cast, which must be of
3151<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3152also be of <a href="#t_integer">integer</a> type. The bit size of the
3153<tt>value</tt> must be smaller than the bit size of the destination type,
3154<tt>ty2</tt>.</p>
3155
3156<h5>Semantics:</h5>
3157<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
3158bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
3159
3160<p>When zero extending from i1, the result will always be either 0 or 1.</p>
3161
3162<h5>Example:</h5>
3163<pre>
3164 %X = zext i32 257 to i64 <i>; yields i64:257</i>
3165 %Y = zext i1 true to i32 <i>; yields i32:1</i>
3166</pre>
3167</div>
3168
3169<!-- _______________________________________________________________________ -->
3170<div class="doc_subsubsection">
3171 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
3172</div>
3173<div class="doc_text">
3174
3175<h5>Syntax:</h5>
3176<pre>
3177 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3178</pre>
3179
3180<h5>Overview:</h5>
3181<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
3182
3183<h5>Arguments:</h5>
3184<p>
3185The '<tt>sext</tt>' instruction takes a value to cast, which must be of
3186<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3187also be of <a href="#t_integer">integer</a> type. The bit size of the
3188<tt>value</tt> must be smaller than the bit size of the destination type,
3189<tt>ty2</tt>.</p>
3190
3191<h5>Semantics:</h5>
3192<p>
3193The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
3194bit (highest order bit) of the <tt>value</tt> until it reaches the bit size of
3195the type <tt>ty2</tt>.</p>
3196
3197<p>When sign extending from i1, the extension always results in -1 or 0.</p>
3198
3199<h5>Example:</h5>
3200<pre>
3201 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
3202 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
3203</pre>
3204</div>
3205
3206<!-- _______________________________________________________________________ -->
3207<div class="doc_subsubsection">
3208 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
3209</div>
3210
3211<div class="doc_text">
3212
3213<h5>Syntax:</h5>
3214
3215<pre>
3216 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3217</pre>
3218
3219<h5>Overview:</h5>
3220<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
3221<tt>ty2</tt>.</p>
3222
3223
3224<h5>Arguments:</h5>
3225<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
3226 point</a> value to cast and a <a href="#t_floating">floating point</a> type to
3227cast it to. The size of <tt>value</tt> must be larger than the size of
3228<tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
3229<i>no-op cast</i>.</p>
3230
3231<h5>Semantics:</h5>
3232<p> The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
3233<a href="#t_floating">floating point</a> type to a smaller
3234<a href="#t_floating">floating point</a> type. If the value cannot fit within
3235the destination type, <tt>ty2</tt>, then the results are undefined.</p>
3236
3237<h5>Example:</h5>
3238<pre>
3239 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
3240 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
3241</pre>
3242</div>
3243
3244<!-- _______________________________________________________________________ -->
3245<div class="doc_subsubsection">
3246 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
3247</div>
3248<div class="doc_text">
3249
3250<h5>Syntax:</h5>
3251<pre>
3252 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3253</pre>
3254
3255<h5>Overview:</h5>
3256<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
3257floating point value.</p>
3258
3259<h5>Arguments:</h5>
3260<p>The '<tt>fpext</tt>' instruction takes a
3261<a href="#t_floating">floating point</a> <tt>value</tt> to cast,
3262and a <a href="#t_floating">floating point</a> type to cast it to. The source
3263type must be smaller than the destination type.</p>
3264
3265<h5>Semantics:</h5>
3266<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
3267<a href="#t_floating">floating point</a> type to a larger
3268<a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
3269used to make a <i>no-op cast</i> because it always changes bits. Use
3270<tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
3271
3272<h5>Example:</h5>
3273<pre>
3274 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
3275 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
3276</pre>
3277</div>
3278
3279<!-- _______________________________________________________________________ -->
3280<div class="doc_subsubsection">
3281 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
3282</div>
3283<div class="doc_text">
3284
3285<h5>Syntax:</h5>
3286<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3287 &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]3288</pre>
3289
3290<h5>Overview:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3291<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3292unsigned integer equivalent of type <tt>ty2</tt>.
3293</p>
3294
3295<h5>Arguments:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3296<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]3297scalar or vector <a href="#t_floating">floating point</a> value, and a type
3298to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
3299type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
3300vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3301
3302<h5>Semantics:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3303<p> The '<tt>fptoui</tt>' instruction converts its
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3304<a href="#t_floating">floating point</a> operand into the nearest (rounding
3305towards zero) unsigned integer value. If the value cannot fit in <tt>ty2</tt>,
3306the results are undefined.</p>
3307
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3308<h5>Example:</h5>
3309<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3310 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]3311 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3312 %X = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3313</pre>
3314</div>
3315
3316<!-- _______________________________________________________________________ -->
3317<div class="doc_subsubsection">
3318 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
3319</div>
3320<div class="doc_text">
3321
3322<h5>Syntax:</h5>
3323<pre>
3324 &lt;result&gt; = fptosi &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3325</pre>
3326
3327<h5>Overview:</h5>
3328<p>The '<tt>fptosi</tt>' instruction converts
3329<a href="#t_floating">floating point</a> <tt>value</tt> to type <tt>ty2</tt>.
3330</p>
3331
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3332<h5>Arguments:</h5>
3333<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]3334scalar or vector <a href="#t_floating">floating point</a> value, and a type
3335to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
3336type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
3337vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3338
3339<h5>Semantics:</h5>
3340<p>The '<tt>fptosi</tt>' instruction converts its
3341<a href="#t_floating">floating point</a> operand into the nearest (rounding
3342towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
3343the results are undefined.</p>
3344
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3345<h5>Example:</h5>
3346<pre>
3347 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]3348 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3349 %X = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
3350</pre>
3351</div>
3352
3353<!-- _______________________________________________________________________ -->
3354<div class="doc_subsubsection">
3355 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
3356</div>
3357<div class="doc_text">
3358
3359<h5>Syntax:</h5>
3360<pre>
3361 &lt;result&gt; = uitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3362</pre>
3363
3364<h5>Overview:</h5>
3365<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
3366integer and converts that value to the <tt>ty2</tt> type.</p>
3367
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3368<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]3369<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be a
3370scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
3371to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
3372type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
3373floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3374
3375<h5>Semantics:</h5>
3376<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
3377integer quantity and converts it to the corresponding floating point value. If
3378the value cannot fit in the floating point value, the results are undefined.</p>
3379
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3380<h5>Example:</h5>
3381<pre>
3382 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
3383 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
3384</pre>
3385</div>
3386
3387<!-- _______________________________________________________________________ -->
3388<div class="doc_subsubsection">
3389 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
3390</div>
3391<div class="doc_text">
3392
3393<h5>Syntax:</h5>
3394<pre>
3395 &lt;result&gt; = sitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3396</pre>
3397
3398<h5>Overview:</h5>
3399<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed
3400integer and converts that value to the <tt>ty2</tt> type.</p>
3401
3402<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]3403<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be a
3404scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
3405to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
3406type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
3407floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3408
3409<h5>Semantics:</h5>
3410<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed
3411integer quantity and converts it to the corresponding floating point value. If
3412the value cannot fit in the floating point value, the results are undefined.</p>
3413
3414<h5>Example:</h5>
3415<pre>
3416 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
3417 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
3418</pre>
3419</div>
3420
3421<!-- _______________________________________________________________________ -->
3422<div class="doc_subsubsection">
3423 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
3424</div>
3425<div class="doc_text">
3426
3427<h5>Syntax:</h5>
3428<pre>
3429 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3430</pre>
3431
3432<h5>Overview:</h5>
3433<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
3434the integer type <tt>ty2</tt>.</p>
3435
3436<h5>Arguments:</h5>
3437<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
3438must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
3439<tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> type.
3440
3441<h5>Semantics:</h5>
3442<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
3443<tt>ty2</tt> by interpreting the pointer value as an integer and either
3444truncating or zero extending that value to the size of the integer type. If
3445<tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
3446<tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
3447are the same size, then nothing is done (<i>no-op cast</i>) other than a type
3448change.</p>
3449
3450<h5>Example:</h5>
3451<pre>
3452 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
3453 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
3454</pre>
3455</div>
3456
3457<!-- _______________________________________________________________________ -->
3458<div class="doc_subsubsection">
3459 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
3460</div>
3461<div class="doc_text">
3462
3463<h5>Syntax:</h5>
3464<pre>
3465 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3466</pre>
3467
3468<h5>Overview:</h5>
3469<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to
3470a pointer type, <tt>ty2</tt>.</p>
3471
3472<h5>Arguments:</h5>
3473<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
3474value to cast, and a type to cast it to, which must be a
3475<a href="#t_pointer">pointer</a> type.
3476
3477<h5>Semantics:</h5>
3478<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
3479<tt>ty2</tt> by applying either a zero extension or a truncation depending on
3480the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
3481size of a pointer then a truncation is done. If <tt>value</tt> is smaller than
3482the size of a pointer then a zero extension is done. If they are the same size,
3483nothing is done (<i>no-op cast</i>).</p>
3484
3485<h5>Example:</h5>
3486<pre>
3487 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
3488 %X = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
3489 %Y = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
3490</pre>
3491</div>
3492
3493<!-- _______________________________________________________________________ -->
3494<div class="doc_subsubsection">
3495 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
3496</div>
3497<div class="doc_text">
3498
3499<h5>Syntax:</h5>
3500<pre>
3501 &lt;result&gt; = bitcast &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3502</pre>
3503
3504<h5>Overview:</h5>
3505<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
3506<tt>ty2</tt> without changing any bits.</p>
3507
3508<h5>Arguments:</h5>
3509<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be
3510a first class value, and a type to cast it to, which must also be a <a
3511 href="#t_firstclass">first class</a> type. The bit sizes of <tt>value</tt>
3512and the destination type, <tt>ty2</tt>, must be identical. If the source
3513type is a pointer, the destination type must also be a pointer.</p>
3514
3515<h5>Semantics:</h5>
3516<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
3517<tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
3518this conversion. The conversion is done as if the <tt>value</tt> had been
3519stored to memory and read back as type <tt>ty2</tt>. Pointer types may only be
3520converted to other pointer types with this instruction. To convert pointers to
3521other types, use the <a href="#i_inttoptr">inttoptr</a> or
3522<a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
3523
3524<h5>Example:</h5>
3525<pre>
3526 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
3527 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
3528 %Z = bitcast <2xint> %V to i64; <i>; yields i64: %V</i>
3529</pre>
3530</div>
3531
3532<!-- ======================================================================= -->
3533<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
3534<div class="doc_text">
3535<p>The instructions in this category are the "miscellaneous"
3536instructions, which defy better classification.</p>
3537</div>
3538
3539<!-- _______________________________________________________________________ -->
3540<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
3541</div>
3542<div class="doc_text">
3543<h5>Syntax:</h5>
3544<pre> &lt;result&gt; = icmp &lt;cond&gt; &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {i1}:result</i>
3545</pre>
3546<h5>Overview:</h5>
3547<p>The '<tt>icmp</tt>' instruction returns a boolean value based on comparison
3548of its two integer operands.</p>
3549<h5>Arguments:</h5>
3550<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
3551the condition code indicating the kind of comparison to perform. It is not
3552a value, just a keyword. The possible condition code are:
3553<ol>
3554 <li><tt>eq</tt>: equal</li>
3555 <li><tt>ne</tt>: not equal </li>
3556 <li><tt>ugt</tt>: unsigned greater than</li>
3557 <li><tt>uge</tt>: unsigned greater or equal</li>
3558 <li><tt>ult</tt>: unsigned less than</li>
3559 <li><tt>ule</tt>: unsigned less or equal</li>
3560 <li><tt>sgt</tt>: signed greater than</li>
3561 <li><tt>sge</tt>: signed greater or equal</li>
3562 <li><tt>slt</tt>: signed less than</li>
3563 <li><tt>sle</tt>: signed less or equal</li>
3564</ol>
3565<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
3566<a href="#t_pointer">pointer</a> typed. They must also be identical types.</p>
3567<h5>Semantics:</h5>
3568<p>The '<tt>icmp</tt>' compares <tt>var1</tt> and <tt>var2</tt> according to
3569the condition code given as <tt>cond</tt>. The comparison performed always
3570yields a <a href="#t_primitive">i1</a> result, as follows:
3571<ol>
3572 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
3573 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
3574 </li>
3575 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
3576 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
3577 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
3578 <tt>true</tt> if <tt>var1</tt> is greater than <tt>var2</tt>.</li>
3579 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
3580 <tt>true</tt> if <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3581 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
3582 <tt>true</tt> if <tt>var1</tt> is less than <tt>var2</tt>.</li>
3583 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
3584 <tt>true</tt> if <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
3585 <li><tt>sgt</tt>: interprets the operands as signed values and yields
3586 <tt>true</tt> if <tt>var1</tt> is greater than <tt>var2</tt>.</li>
3587 <li><tt>sge</tt>: interprets the operands as signed values and yields
3588 <tt>true</tt> if <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3589 <li><tt>slt</tt>: interprets the operands as signed values and yields
3590 <tt>true</tt> if <tt>var1</tt> is less than <tt>var2</tt>.</li>
3591 <li><tt>sle</tt>: interprets the operands as signed values and yields
3592 <tt>true</tt> if <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
3593</ol>
3594<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
3595values are compared as if they were integers.</p>
3596
3597<h5>Example:</h5>
3598<pre> &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
3599 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
3600 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
3601 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
3602 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
3603 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
3604</pre>
3605</div>
3606
3607<!-- _______________________________________________________________________ -->
3608<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
3609</div>
3610<div class="doc_text">
3611<h5>Syntax:</h5>
3612<pre> &lt;result&gt; = fcmp &lt;cond&gt; &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {i1}:result</i>
3613</pre>
3614<h5>Overview:</h5>
3615<p>The '<tt>fcmp</tt>' instruction returns a boolean value based on comparison
3616of its floating point operands.</p>
3617<h5>Arguments:</h5>
3618<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
3619the condition code indicating the kind of comparison to perform. It is not
3620a value, just a keyword. The possible condition code are:
3621<ol>
3622 <li><tt>false</tt>: no comparison, always returns false</li>
3623 <li><tt>oeq</tt>: ordered and equal</li>
3624 <li><tt>ogt</tt>: ordered and greater than </li>
3625 <li><tt>oge</tt>: ordered and greater than or equal</li>
3626 <li><tt>olt</tt>: ordered and less than </li>
3627 <li><tt>ole</tt>: ordered and less than or equal</li>
3628 <li><tt>one</tt>: ordered and not equal</li>
3629 <li><tt>ord</tt>: ordered (no nans)</li>
3630 <li><tt>ueq</tt>: unordered or equal</li>
3631 <li><tt>ugt</tt>: unordered or greater than </li>
3632 <li><tt>uge</tt>: unordered or greater than or equal</li>
3633 <li><tt>ult</tt>: unordered or less than </li>
3634 <li><tt>ule</tt>: unordered or less than or equal</li>
3635 <li><tt>une</tt>: unordered or not equal</li>
3636 <li><tt>uno</tt>: unordered (either nans)</li>
3637 <li><tt>true</tt>: no comparison, always returns true</li>
3638</ol>
3639<p><i>Ordered</i> means that neither operand is a QNAN while
3640<i>unordered</i> means that either operand may be a QNAN.</p>
3641<p>The <tt>val1</tt> and <tt>val2</tt> arguments must be
3642<a href="#t_floating">floating point</a> typed. They must have identical
3643types.</p>
3644<h5>Semantics:</h5>
3645<p>The '<tt>fcmp</tt>' compares <tt>var1</tt> and <tt>var2</tt> according to
3646the condition code given as <tt>cond</tt>. The comparison performed always
3647yields a <a href="#t_primitive">i1</a> result, as follows:
3648<ol>
3649 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
3650 <li><tt>oeq</tt>: yields <tt>true</tt> if both operands are not a QNAN and
3651 <tt>var1</tt> is equal to <tt>var2</tt>.</li>
3652 <li><tt>ogt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
3653 <tt>var1</tt> is greather than <tt>var2</tt>.</li>
3654 <li><tt>oge</tt>: yields <tt>true</tt> if both operands are not a QNAN and
3655 <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3656 <li><tt>olt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
3657 <tt>var1</tt> is less than <tt>var2</tt>.</li>
3658 <li><tt>ole</tt>: yields <tt>true</tt> if both operands are not a QNAN and
3659 <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
3660 <li><tt>one</tt>: yields <tt>true</tt> if both operands are not a QNAN and
3661 <tt>var1</tt> is not equal to <tt>var2</tt>.</li>
3662 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
3663 <li><tt>ueq</tt>: yields <tt>true</tt> if either operand is a QNAN or
3664 <tt>var1</tt> is equal to <tt>var2</tt>.</li>
3665 <li><tt>ugt</tt>: yields <tt>true</tt> if either operand is a QNAN or
3666 <tt>var1</tt> is greater than <tt>var2</tt>.</li>
3667 <li><tt>uge</tt>: yields <tt>true</tt> if either operand is a QNAN or
3668 <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3669 <li><tt>ult</tt>: yields <tt>true</tt> if either operand is a QNAN or
3670 <tt>var1</tt> is less than <tt>var2</tt>.</li>
3671 <li><tt>ule</tt>: yields <tt>true</tt> if either operand is a QNAN or
3672 <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
3673 <li><tt>une</tt>: yields <tt>true</tt> if either operand is a QNAN or
3674 <tt>var1</tt> is not equal to <tt>var2</tt>.</li>
3675 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
3676 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
3677</ol>
3678
3679<h5>Example:</h5>
3680<pre> &lt;result&gt; = fcmp oeq float 4.0, 5.0 <i>; yields: result=false</i>
3681 &lt;result&gt; = icmp one float 4.0, 5.0 <i>; yields: result=true</i>
3682 &lt;result&gt; = icmp olt float 4.0, 5.0 <i>; yields: result=true</i>
3683 &lt;result&gt; = icmp ueq double 1.0, 2.0 <i>; yields: result=false</i>
3684</pre>
3685</div>
3686
3687<!-- _______________________________________________________________________ -->
3688<div class="doc_subsubsection"> <a name="i_phi">'<tt>phi</tt>'
3689Instruction</a> </div>
3690<div class="doc_text">
3691<h5>Syntax:</h5>
3692<pre> &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...<br></pre>
3693<h5>Overview:</h5>
3694<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in
3695the SSA graph representing the function.</p>
3696<h5>Arguments:</h5>
3697<p>The type of the incoming values is specified with the first type
3698field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
3699as arguments, with one pair for each predecessor basic block of the
3700current block. Only values of <a href="#t_firstclass">first class</a>
3701type may be used as the value arguments to the PHI node. Only labels
3702may be used as the label arguments.</p>
3703<p>There must be no non-phi instructions between the start of a basic
3704block and the PHI instructions: i.e. PHI instructions must be first in
3705a basic block.</p>
3706<h5>Semantics:</h5>
3707<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
3708specified by the pair corresponding to the predecessor basic block that executed
3709just prior to the current block.</p>
3710<h5>Example:</h5>
3711<pre>Loop: ; Infinite loop that counts from 0 on up...<br> %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]<br> %nextindvar = add i32 %indvar, 1<br> br label %Loop<br></pre>
3712</div>
3713
3714<!-- _______________________________________________________________________ -->
3715<div class="doc_subsubsection">
3716 <a name="i_select">'<tt>select</tt>' Instruction</a>
3717</div>
3718
3719<div class="doc_text">
3720
3721<h5>Syntax:</h5>
3722
3723<pre>
3724 &lt;result&gt; = select i1 &lt;cond&gt;, &lt;ty&gt; &lt;val1&gt;, &lt;ty&gt; &lt;val2&gt; <i>; yields ty</i>
3725</pre>
3726
3727<h5>Overview:</h5>
3728
3729<p>
3730The '<tt>select</tt>' instruction is used to choose one value based on a
3731condition, without branching.
3732</p>
3733
3734
3735<h5>Arguments:</h5>
3736
3737<p>
3738The '<tt>select</tt>' instruction requires a boolean value indicating the condition, and two values of the same <a href="#t_firstclass">first class</a> type.
3739</p>
3740
3741<h5>Semantics:</h5>
3742
3743<p>
3744If the boolean condition evaluates to true, the instruction returns the first
3745value argument; otherwise, it returns the second value argument.
3746</p>
3747
3748<h5>Example:</h5>
3749
3750<pre>
3751 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
3752</pre>
3753</div>
3754
3755
3756<!-- _______________________________________________________________________ -->
3757<div class="doc_subsubsection">
3758 <a name="i_call">'<tt>call</tt>' Instruction</a>
3759</div>
3760
3761<div class="doc_text">
3762
3763<h5>Syntax:</h5>
3764<pre>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]3765 &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]3766</pre>
3767
3768<h5>Overview:</h5>
3769
3770<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
3771
3772<h5>Arguments:</h5>
3773
3774<p>This instruction requires several arguments:</p>
3775
3776<ol>
3777 <li>
3778 <p>The optional "tail" marker indicates whether the callee function accesses
3779 any allocas or varargs in the caller. If the "tail" marker is present, the
3780 function call is eligible for tail call optimization. Note that calls may
3781 be marked "tail" even if they do not occur before a <a
3782 href="#i_ret"><tt>ret</tt></a> instruction.
3783 </li>
3784 <li>
3785 <p>The optional "cconv" marker indicates which <a href="#callingconv">calling
3786 convention</a> the call should use. If none is specified, the call defaults
3787 to using C calling conventions.
3788 </li>
3789 <li>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]3790 <p>'<tt>ty</tt>': the type of the call instruction itself which is also
3791 the type of the return value. Functions that return no value are marked
3792 <tt><a href="#t_void">void</a></tt>.</p>
3793 </li>
3794 <li>
3795 <p>'<tt>fnty</tt>': shall be the signature of the pointer to function
3796 value being invoked. The argument types must match the types implied by
3797 this signature. This type can be omitted if the function is not varargs
3798 and if the function type does not return a pointer to a function.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3799 </li>
3800 <li>
3801 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
3802 be invoked. In most cases, this is a direct function invocation, but
3803 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
3804 to function value.</p>
3805 </li>
3806 <li>
3807 <p>'<tt>function args</tt>': argument list whose types match the
3808 function signature argument types. All arguments must be of
3809 <a href="#t_firstclass">first class</a> type. If the function signature
3810 indicates the function accepts a variable number of arguments, the extra
3811 arguments can be specified.</p>
3812 </li>
3813</ol>
3814
3815<h5>Semantics:</h5>
3816
3817<p>The '<tt>call</tt>' instruction is used to cause control flow to
3818transfer to a specified function, with its incoming arguments bound to
3819the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
3820instruction in the called function, control flow continues with the
3821instruction after the function call, and the return value of the
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]3822function is bound to the result argument. If the '<tt><a href="#i_ret">ret</a>
Devang Patelec8a5b02008-03-11 05:51:59 +0000[diff] [blame]3823</tt>' instruction returns multiple values then the return values of the
3824function are only accessible through a '<tt><a href="#i_getresult">getresult</a>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]3825</tt>' instruction. This is a simpler case of
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3826the <a href="#i_invoke">invoke</a> instruction.</p>
3827
3828<h5>Example:</h5>
3829
3830<pre>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]3831 %retval = call i32 @test(i32 %argc)
3832 call i32 (i8 *, ...)* @printf(i8 * %msg, i32 12, i8 42);
3833 %X = tail call i32 @foo()
3834 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo()
3835 %Z = call void %foo(i8 97 signext)
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]3836
3837 %struct.A = type { i32, i8 }
3838 %r = call %struct.A @foo()
3839 %gr = getresult %struct.A %r, 0
3840 %gr1 = getresult %struct.A %r, 1
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3841</pre>
3842
3843</div>
3844
3845<!-- _______________________________________________________________________ -->
3846<div class="doc_subsubsection">
3847 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
3848</div>
3849
3850<div class="doc_text">
3851
3852<h5>Syntax:</h5>
3853
3854<pre>
3855 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
3856</pre>
3857
3858<h5>Overview:</h5>
3859
3860<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
3861the "variable argument" area of a function call. It is used to implement the
3862<tt>va_arg</tt> macro in C.</p>
3863
3864<h5>Arguments:</h5>
3865
3866<p>This instruction takes a <tt>va_list*</tt> value and the type of
3867the argument. It returns a value of the specified argument type and
3868increments the <tt>va_list</tt> to point to the next argument. The
3869actual type of <tt>va_list</tt> is target specific.</p>
3870
3871<h5>Semantics:</h5>
3872
3873<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
3874type from the specified <tt>va_list</tt> and causes the
3875<tt>va_list</tt> to point to the next argument. For more information,
3876see the variable argument handling <a href="#int_varargs">Intrinsic
3877Functions</a>.</p>
3878
3879<p>It is legal for this instruction to be called in a function which does not
3880take a variable number of arguments, for example, the <tt>vfprintf</tt>
3881function.</p>
3882
3883<p><tt>va_arg</tt> is an LLVM instruction instead of an <a
3884href="#intrinsics">intrinsic function</a> because it takes a type as an
3885argument.</p>
3886
3887<h5>Example:</h5>
3888
3889<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
3890
3891</div>
3892
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]3893<!-- _______________________________________________________________________ -->
3894<div class="doc_subsubsection">
3895 <a name="i_getresult">'<tt>getresult</tt>' Instruction</a>
3896</div>
3897
3898<div class="doc_text">
3899
3900<h5>Syntax:</h5>
3901<pre>
Chris Lattneree9da3f2008-03-21 17:20:51 +0000[diff] [blame^]3902 &lt;resultval&gt; = getresult &lt;type&gt; &lt;retval&gt;, &lt;index&gt;
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]3903</pre>
Chris Lattneree9da3f2008-03-21 17:20:51 +0000[diff] [blame^]3904
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]3905<h5>Overview:</h5>
3906
3907<p> The '<tt>getresult</tt>' instruction is used to extract individual values
Chris Lattneree9da3f2008-03-21 17:20:51 +0000[diff] [blame^]3908from a '<tt><a href="#i_call">call</a></tt>'
3909or '<tt><a href="#i_invoke">invoke</a></tt>' instruction that returns multiple
3910results.</p>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]3911
3912<h5>Arguments:</h5>
3913
Chris Lattneree9da3f2008-03-21 17:20:51 +0000[diff] [blame^]3914<p>The '<tt>getresult</tt>' instruction takes a call or invoke value as its
3915first argument. The value must have <a href="#t_struct">structure type</a>.
3916The second argument is an unsigned index value which must be in range for
3917the number of values returned by the call.</p>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]3918
3919<h5>Semantics:</h5>
3920
Chris Lattneree9da3f2008-03-21 17:20:51 +0000[diff] [blame^]3921<p>The '<tt>getresult</tt>' instruction extracts the element identified by
3922'<tt>index</tt>' from the aggregate value.</p>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]3923
3924<h5>Example:</h5>
3925
3926<pre>
3927 %struct.A = type { i32, i8 }
3928
3929 %r = call %struct.A @foo()
Chris Lattneree9da3f2008-03-21 17:20:51 +0000[diff] [blame^]3930 %gr = getresult %struct.A %r, 0 <i>; yields i32:%gr</i>
3931 %gr1 = getresult %struct.A %r, 1 <i>; yields i8:%gr1</i>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]3932 add i32 %gr, 42
3933 add i8 %gr1, 41
3934</pre>
3935
3936</div>
3937
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3938<!-- *********************************************************************** -->
3939<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
3940<!-- *********************************************************************** -->
3941
3942<div class="doc_text">
3943
3944<p>LLVM supports the notion of an "intrinsic function". These functions have
3945well known names and semantics and are required to follow certain restrictions.
3946Overall, these intrinsics represent an extension mechanism for the LLVM
3947language that does not require changing all of the transformations in LLVM when
3948adding to the language (or the bitcode reader/writer, the parser, etc...).</p>
3949
3950<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
3951prefix is reserved in LLVM for intrinsic names; thus, function names may not
3952begin with this prefix. Intrinsic functions must always be external functions:
3953you cannot define the body of intrinsic functions. Intrinsic functions may
3954only be used in call or invoke instructions: it is illegal to take the address
3955of an intrinsic function. Additionally, because intrinsic functions are part
3956of the LLVM language, it is required if any are added that they be documented
3957here.</p>
3958
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]3959<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents
3960a family of functions that perform the same operation but on different data
3961types. Because LLVM can represent over 8 million different integer types,
3962overloading is used commonly to allow an intrinsic function to operate on any
3963integer type. One or more of the argument types or the result type can be
3964overloaded to accept any integer type. Argument types may also be defined as
3965exactly matching a previous argument's type or the result type. This allows an
3966intrinsic function which accepts multiple arguments, but needs all of them to
3967be of the same type, to only be overloaded with respect to a single argument or
3968the result.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3969
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]3970<p>Overloaded intrinsics will have the names of its overloaded argument types
3971encoded into its function name, each preceded by a period. Only those types
3972which are overloaded result in a name suffix. Arguments whose type is matched
3973against another type do not. For example, the <tt>llvm.ctpop</tt> function can
3974take an integer of any width and returns an integer of exactly the same integer
3975width. This leads to a family of functions such as
3976<tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29 %val)</tt>.
3977Only one type, the return type, is overloaded, and only one type suffix is
3978required. Because the argument's type is matched against the return type, it
3979does not require its own name suffix.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3980
3981<p>To learn how to add an intrinsic function, please see the
3982<a href="ExtendingLLVM.html">Extending LLVM Guide</a>.
3983</p>
3984
3985</div>
3986
3987<!-- ======================================================================= -->
3988<div class="doc_subsection">
3989 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
3990</div>
3991
3992<div class="doc_text">
3993
3994<p>Variable argument support is defined in LLVM with the <a
3995 href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
3996intrinsic functions. These functions are related to the similarly
3997named macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
3998
3999<p>All of these functions operate on arguments that use a
4000target-specific value type "<tt>va_list</tt>". The LLVM assembly
4001language reference manual does not define what this type is, so all
4002transformations should be prepared to handle these functions regardless of
4003the type used.</p>
4004
4005<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
4006instruction and the variable argument handling intrinsic functions are
4007used.</p>
4008
4009<div class="doc_code">
4010<pre>
4011define i32 @test(i32 %X, ...) {
4012 ; Initialize variable argument processing
4013 %ap = alloca i8*
4014 %ap2 = bitcast i8** %ap to i8*
4015 call void @llvm.va_start(i8* %ap2)
4016
4017 ; Read a single integer argument
4018 %tmp = va_arg i8** %ap, i32
4019
4020 ; Demonstrate usage of llvm.va_copy and llvm.va_end
4021 %aq = alloca i8*
4022 %aq2 = bitcast i8** %aq to i8*
4023 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
4024 call void @llvm.va_end(i8* %aq2)
4025
4026 ; Stop processing of arguments.
4027 call void @llvm.va_end(i8* %ap2)
4028 ret i32 %tmp
4029}
4030
4031declare void @llvm.va_start(i8*)
4032declare void @llvm.va_copy(i8*, i8*)
4033declare void @llvm.va_end(i8*)
4034</pre>
4035</div>
4036
4037</div>
4038
4039<!-- _______________________________________________________________________ -->
4040<div class="doc_subsubsection">
4041 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
4042</div>
4043
4044
4045<div class="doc_text">
4046<h5>Syntax:</h5>
4047<pre> declare void %llvm.va_start(i8* &lt;arglist&gt;)<br></pre>
4048<h5>Overview:</h5>
4049<P>The '<tt>llvm.va_start</tt>' intrinsic initializes
4050<tt>*&lt;arglist&gt;</tt> for subsequent use by <tt><a
4051href="#i_va_arg">va_arg</a></tt>.</p>
4052
4053<h5>Arguments:</h5>
4054
4055<P>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
4056
4057<h5>Semantics:</h5>
4058
4059<P>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
4060macro available in C. In a target-dependent way, it initializes the
4061<tt>va_list</tt> element to which the argument points, so that the next call to
4062<tt>va_arg</tt> will produce the first variable argument passed to the function.
4063Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
4064last argument of the function as the compiler can figure that out.</p>
4065
4066</div>
4067
4068<!-- _______________________________________________________________________ -->
4069<div class="doc_subsubsection">
4070 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
4071</div>
4072
4073<div class="doc_text">
4074<h5>Syntax:</h5>
4075<pre> declare void @llvm.va_end(i8* &lt;arglist&gt;)<br></pre>
4076<h5>Overview:</h5>
4077
4078<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
4079which has been initialized previously with <tt><a href="#int_va_start">llvm.va_start</a></tt>
4080or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
4081
4082<h5>Arguments:</h5>
4083
4084<p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
4085
4086<h5>Semantics:</h5>
4087
4088<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
4089macro available in C. In a target-dependent way, it destroys the
4090<tt>va_list</tt> element to which the argument points. Calls to <a
4091href="#int_va_start"><tt>llvm.va_start</tt></a> and <a href="#int_va_copy">
4092<tt>llvm.va_copy</tt></a> must be matched exactly with calls to
4093<tt>llvm.va_end</tt>.</p>
4094
4095</div>
4096
4097<!-- _______________________________________________________________________ -->
4098<div class="doc_subsubsection">
4099 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
4100</div>
4101
4102<div class="doc_text">
4103
4104<h5>Syntax:</h5>
4105
4106<pre>
4107 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
4108</pre>
4109
4110<h5>Overview:</h5>
4111
4112<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
4113from the source argument list to the destination argument list.</p>
4114
4115<h5>Arguments:</h5>
4116
4117<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
4118The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
4119
4120
4121<h5>Semantics:</h5>
4122
4123<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
4124macro available in C. In a target-dependent way, it copies the source
4125<tt>va_list</tt> element into the destination <tt>va_list</tt> element. This
4126intrinsic is necessary because the <tt><a href="#int_va_start">
4127llvm.va_start</a></tt> intrinsic may be arbitrarily complex and require, for
4128example, memory allocation.</p>
4129
4130</div>
4131
4132<!-- ======================================================================= -->
4133<div class="doc_subsection">
4134 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
4135</div>
4136
4137<div class="doc_text">
4138
4139<p>
4140LLVM support for <a href="GarbageCollection.html">Accurate Garbage
4141Collection</a> requires the implementation and generation of these intrinsics.
4142These intrinsics allow identification of <a href="#int_gcroot">GC roots on the
4143stack</a>, as well as garbage collector implementations that require <a
4144href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a> barriers.
4145Front-ends for type-safe garbage collected languages should generate these
4146intrinsics to make use of the LLVM garbage collectors. For more details, see <a
4147href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
4148</p>
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]4149
4150<p>The garbage collection intrinsics only operate on objects in the generic
4151 address space (address space zero).</p>
4152
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4153</div>
4154
4155<!-- _______________________________________________________________________ -->
4156<div class="doc_subsubsection">
4157 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
4158</div>
4159
4160<div class="doc_text">
4161
4162<h5>Syntax:</h5>
4163
4164<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4165 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4166</pre>
4167
4168<h5>Overview:</h5>
4169
4170<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
4171the code generator, and allows some metadata to be associated with it.</p>
4172
4173<h5>Arguments:</h5>
4174
4175<p>The first argument specifies the address of a stack object that contains the
4176root pointer. The second pointer (which must be either a constant or a global
4177value address) contains the meta-data to be associated with the root.</p>
4178
4179<h5>Semantics:</h5>
4180
4181<p>At runtime, a call to this intrinsics stores a null pointer into the "ptrloc"
4182location. At compile-time, the code generator generates information to allow
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4183the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
4184intrinsic may only be used in a function which <a href="#gc">specifies a GC
4185algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4186
4187</div>
4188
4189
4190<!-- _______________________________________________________________________ -->
4191<div class="doc_subsubsection">
4192 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
4193</div>
4194
4195<div class="doc_text">
4196
4197<h5>Syntax:</h5>
4198
4199<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4200 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4201</pre>
4202
4203<h5>Overview:</h5>
4204
4205<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
4206locations, allowing garbage collector implementations that require read
4207barriers.</p>
4208
4209<h5>Arguments:</h5>
4210
4211<p>The second argument is the address to read from, which should be an address
4212allocated from the garbage collector. The first object is a pointer to the
4213start of the referenced object, if needed by the language runtime (otherwise
4214null).</p>
4215
4216<h5>Semantics:</h5>
4217
4218<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
4219instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4220garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
4221may only be used in a function which <a href="#gc">specifies a GC
4222algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4223
4224</div>
4225
4226
4227<!-- _______________________________________________________________________ -->
4228<div class="doc_subsubsection">
4229 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
4230</div>
4231
4232<div class="doc_text">
4233
4234<h5>Syntax:</h5>
4235
4236<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4237 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4238</pre>
4239
4240<h5>Overview:</h5>
4241
4242<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
4243locations, allowing garbage collector implementations that require write
4244barriers (such as generational or reference counting collectors).</p>
4245
4246<h5>Arguments:</h5>
4247
4248<p>The first argument is the reference to store, the second is the start of the
4249object to store it to, and the third is the address of the field of Obj to
4250store to. If the runtime does not require a pointer to the object, Obj may be
4251null.</p>
4252
4253<h5>Semantics:</h5>
4254
4255<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
4256instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4257garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
4258may only be used in a function which <a href="#gc">specifies a GC
4259algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4260
4261</div>
4262
4263
4264
4265<!-- ======================================================================= -->
4266<div class="doc_subsection">
4267 <a name="int_codegen">Code Generator Intrinsics</a>
4268</div>
4269
4270<div class="doc_text">
4271<p>
4272These intrinsics are provided by LLVM to expose special features that may only
4273be implemented with code generator support.
4274</p>
4275
4276</div>
4277
4278<!-- _______________________________________________________________________ -->
4279<div class="doc_subsubsection">
4280 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
4281</div>
4282
4283<div class="doc_text">
4284
4285<h5>Syntax:</h5>
4286<pre>
4287 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
4288</pre>
4289
4290<h5>Overview:</h5>
4291
4292<p>
4293The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
4294target-specific value indicating the return address of the current function
4295or one of its callers.
4296</p>
4297
4298<h5>Arguments:</h5>
4299
4300<p>
4301The argument to this intrinsic indicates which function to return the address
4302for. Zero indicates the calling function, one indicates its caller, etc. The
4303argument is <b>required</b> to be a constant integer value.
4304</p>
4305
4306<h5>Semantics:</h5>
4307
4308<p>
4309The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
4310the return address of the specified call frame, or zero if it cannot be
4311identified. The value returned by this intrinsic is likely to be incorrect or 0
4312for arguments other than zero, so it should only be used for debugging purposes.
4313</p>
4314
4315<p>
4316Note that calling this intrinsic does not prevent function inlining or other
4317aggressive transformations, so the value returned may not be that of the obvious
4318source-language caller.
4319</p>
4320</div>
4321
4322
4323<!-- _______________________________________________________________________ -->
4324<div class="doc_subsubsection">
4325 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
4326</div>
4327
4328<div class="doc_text">
4329
4330<h5>Syntax:</h5>
4331<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4332 declare i8 *@llvm.frameaddress(i32 &lt;level&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4333</pre>
4334
4335<h5>Overview:</h5>
4336
4337<p>
4338The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
4339target-specific frame pointer value for the specified stack frame.
4340</p>
4341
4342<h5>Arguments:</h5>
4343
4344<p>
4345The argument to this intrinsic indicates which function to return the frame
4346pointer for. Zero indicates the calling function, one indicates its caller,
4347etc. The argument is <b>required</b> to be a constant integer value.
4348</p>
4349
4350<h5>Semantics:</h5>
4351
4352<p>
4353The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
4354the frame address of the specified call frame, or zero if it cannot be
4355identified. The value returned by this intrinsic is likely to be incorrect or 0
4356for arguments other than zero, so it should only be used for debugging purposes.
4357</p>
4358
4359<p>
4360Note that calling this intrinsic does not prevent function inlining or other
4361aggressive transformations, so the value returned may not be that of the obvious
4362source-language caller.
4363</p>
4364</div>
4365
4366<!-- _______________________________________________________________________ -->
4367<div class="doc_subsubsection">
4368 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
4369</div>
4370
4371<div class="doc_text">
4372
4373<h5>Syntax:</h5>
4374<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4375 declare i8 *@llvm.stacksave()
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4376</pre>
4377
4378<h5>Overview:</h5>
4379
4380<p>
4381The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
4382the function stack, for use with <a href="#int_stackrestore">
4383<tt>llvm.stackrestore</tt></a>. This is useful for implementing language
4384features like scoped automatic variable sized arrays in C99.
4385</p>
4386
4387<h5>Semantics:</h5>
4388
4389<p>
4390This intrinsic returns a opaque pointer value that can be passed to <a
4391href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
4392<tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
4393<tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
4394state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
4395practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
4396that were allocated after the <tt>llvm.stacksave</tt> was executed.
4397</p>
4398
4399</div>
4400
4401<!-- _______________________________________________________________________ -->
4402<div class="doc_subsubsection">
4403 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
4404</div>
4405
4406<div class="doc_text">
4407
4408<h5>Syntax:</h5>
4409<pre>
4410 declare void @llvm.stackrestore(i8 * %ptr)
4411</pre>
4412
4413<h5>Overview:</h5>
4414
4415<p>
4416The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
4417the function stack to the state it was in when the corresponding <a
4418href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
4419useful for implementing language features like scoped automatic variable sized
4420arrays in C99.
4421</p>
4422
4423<h5>Semantics:</h5>
4424
4425<p>
4426See the description for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.
4427</p>
4428
4429</div>
4430
4431
4432<!-- _______________________________________________________________________ -->
4433<div class="doc_subsubsection">
4434 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
4435</div>
4436
4437<div class="doc_text">
4438
4439<h5>Syntax:</h5>
4440<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4441 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]4442</pre>
4443
4444<h5>Overview:</h5>
4445
4446
4447<p>
4448The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
4449a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
4450no
4451effect on the behavior of the program but can change its performance
4452characteristics.
4453</p>
4454
4455<h5>Arguments:</h5>
4456
4457<p>
4458<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
4459determining if the fetch should be for a read (0) or write (1), and
4460<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
4461locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
4462<tt>locality</tt> arguments must be constant integers.
4463</p>
4464
4465<h5>Semantics:</h5>
4466
4467<p>
4468This intrinsic does not modify the behavior of the program. In particular,
4469prefetches cannot trap and do not produce a value. On targets that support this
4470intrinsic, the prefetch can provide hints to the processor cache for better
4471performance.
4472</p>
4473
4474</div>
4475
4476<!-- _______________________________________________________________________ -->
4477<div class="doc_subsubsection">
4478 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
4479</div>
4480
4481<div class="doc_text">
4482
4483<h5>Syntax:</h5>
4484<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4485 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4486</pre>
4487
4488<h5>Overview:</h5>
4489
4490
4491<p>
4492The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
4493(PC) in a region of
4494code to simulators and other tools. The method is target specific, but it is
4495expected that the marker will use exported symbols to transmit the PC of the marker.
4496The marker makes no guarantees that it will remain with any specific instruction
4497after optimizations. It is possible that the presence of a marker will inhibit
4498optimizations. The intended use is to be inserted after optimizations to allow
4499correlations of simulation runs.
4500</p>
4501
4502<h5>Arguments:</h5>
4503
4504<p>
4505<tt>id</tt> is a numerical id identifying the marker.
4506</p>
4507
4508<h5>Semantics:</h5>
4509
4510<p>
4511This intrinsic does not modify the behavior of the program. Backends that do not
4512support this intrinisic may ignore it.
4513</p>
4514
4515</div>
4516
4517<!-- _______________________________________________________________________ -->
4518<div class="doc_subsubsection">
4519 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
4520</div>
4521
4522<div class="doc_text">
4523
4524<h5>Syntax:</h5>
4525<pre>
4526 declare i64 @llvm.readcyclecounter( )
4527</pre>
4528
4529<h5>Overview:</h5>
4530
4531
4532<p>
4533The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
4534counter register (or similar low latency, high accuracy clocks) on those targets
4535that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
4536As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
4537should only be used for small timings.
4538</p>
4539
4540<h5>Semantics:</h5>
4541
4542<p>
4543When directly supported, reading the cycle counter should not modify any memory.
4544Implementations are allowed to either return a application specific value or a
4545system wide value. On backends without support, this is lowered to a constant 0.
4546</p>
4547
4548</div>
4549
4550<!-- ======================================================================= -->
4551<div class="doc_subsection">
4552 <a name="int_libc">Standard C Library Intrinsics</a>
4553</div>
4554
4555<div class="doc_text">
4556<p>
4557LLVM provides intrinsics for a few important standard C library functions.
4558These intrinsics allow source-language front-ends to pass information about the
4559alignment of the pointer arguments to the code generator, providing opportunity
4560for more efficient code generation.
4561</p>
4562
4563</div>
4564
4565<!-- _______________________________________________________________________ -->
4566<div class="doc_subsubsection">
4567 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
4568</div>
4569
4570<div class="doc_text">
4571
4572<h5>Syntax:</h5>
4573<pre>
4574 declare void @llvm.memcpy.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
4575 i32 &lt;len&gt;, i32 &lt;align&gt;)
4576 declare void @llvm.memcpy.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
4577 i64 &lt;len&gt;, i32 &lt;align&gt;)
4578</pre>
4579
4580<h5>Overview:</h5>
4581
4582<p>
4583The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
4584location to the destination location.
4585</p>
4586
4587<p>
4588Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
4589intrinsics do not return a value, and takes an extra alignment argument.
4590</p>
4591
4592<h5>Arguments:</h5>
4593
4594<p>
4595The first argument is a pointer to the destination, the second is a pointer to
4596the source. The third argument is an integer argument
4597specifying the number of bytes to copy, and the fourth argument is the alignment
4598of the source and destination locations.
4599</p>
4600
4601<p>
4602If the call to this intrinisic has an alignment value that is not 0 or 1, then
4603the caller guarantees that both the source and destination pointers are aligned
4604to that boundary.
4605</p>
4606
4607<h5>Semantics:</h5>
4608
4609<p>
4610The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
4611location to the destination location, which are not allowed to overlap. It
4612copies "len" bytes of memory over. If the argument is known to be aligned to
4613some boundary, this can be specified as the fourth argument, otherwise it should
4614be set to 0 or 1.
4615</p>
4616</div>
4617
4618
4619<!-- _______________________________________________________________________ -->
4620<div class="doc_subsubsection">
4621 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
4622</div>
4623
4624<div class="doc_text">
4625
4626<h5>Syntax:</h5>
4627<pre>
4628 declare void @llvm.memmove.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
4629 i32 &lt;len&gt;, i32 &lt;align&gt;)
4630 declare void @llvm.memmove.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
4631 i64 &lt;len&gt;, i32 &lt;align&gt;)
4632</pre>
4633
4634<h5>Overview:</h5>
4635
4636<p>
4637The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the source
4638location to the destination location. It is similar to the
Chris Lattnerdba16ea2008-01-06 19:51:52 +0000[diff] [blame]4639'<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to overlap.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4640</p>
4641
4642<p>
4643Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
4644intrinsics do not return a value, and takes an extra alignment argument.
4645</p>
4646
4647<h5>Arguments:</h5>
4648
4649<p>
4650The first argument is a pointer to the destination, the second is a pointer to
4651the source. The third argument is an integer argument
4652specifying the number of bytes to copy, and the fourth argument is the alignment
4653of the source and destination locations.
4654</p>
4655
4656<p>
4657If the call to this intrinisic has an alignment value that is not 0 or 1, then
4658the caller guarantees that the source and destination pointers are aligned to
4659that boundary.
4660</p>
4661
4662<h5>Semantics:</h5>
4663
4664<p>
4665The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the source
4666location to the destination location, which may overlap. It
4667copies "len" bytes of memory over. If the argument is known to be aligned to
4668some boundary, this can be specified as the fourth argument, otherwise it should
4669be set to 0 or 1.
4670</p>
4671</div>
4672
4673
4674<!-- _______________________________________________________________________ -->
4675<div class="doc_subsubsection">
4676 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
4677</div>
4678
4679<div class="doc_text">
4680
4681<h5>Syntax:</h5>
4682<pre>
4683 declare void @llvm.memset.i32(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
4684 i32 &lt;len&gt;, i32 &lt;align&gt;)
4685 declare void @llvm.memset.i64(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
4686 i64 &lt;len&gt;, i32 &lt;align&gt;)
4687</pre>
4688
4689<h5>Overview:</h5>
4690
4691<p>
4692The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a particular
4693byte value.
4694</p>
4695
4696<p>
4697Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
4698does not return a value, and takes an extra alignment argument.
4699</p>
4700
4701<h5>Arguments:</h5>
4702
4703<p>
4704The first argument is a pointer to the destination to fill, the second is the
4705byte value to fill it with, the third argument is an integer
4706argument specifying the number of bytes to fill, and the fourth argument is the
4707known alignment of destination location.
4708</p>
4709
4710<p>
4711If the call to this intrinisic has an alignment value that is not 0 or 1, then
4712the caller guarantees that the destination pointer is aligned to that boundary.
4713</p>
4714
4715<h5>Semantics:</h5>
4716
4717<p>
4718The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting at
4719the
4720destination location. If the argument is known to be aligned to some boundary,
4721this can be specified as the fourth argument, otherwise it should be set to 0 or
47221.
4723</p>
4724</div>
4725
4726
4727<!-- _______________________________________________________________________ -->
4728<div class="doc_subsubsection">
4729 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
4730</div>
4731
4732<div class="doc_text">
4733
4734<h5>Syntax:</h5>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]4735<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]4736floating point or vector of floating point type. Not all targets support all
4737types however.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4738<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]4739 declare float @llvm.sqrt.f32(float %Val)
4740 declare double @llvm.sqrt.f64(double %Val)
4741 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
4742 declare fp128 @llvm.sqrt.f128(fp128 %Val)
4743 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4744</pre>
4745
4746<h5>Overview:</h5>
4747
4748<p>
4749The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]4750returning the same value as the libm '<tt>sqrt</tt>' functions would. Unlike
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4751<tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
Chris Lattnerf914a002008-01-29 07:00:44 +0000[diff] [blame]4752negative numbers other than -0.0 (which allows for better optimization, because
4753there is no need to worry about errno being set). <tt>llvm.sqrt(-0.0)</tt> is
4754defined to return -0.0 like IEEE sqrt.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4755</p>
4756
4757<h5>Arguments:</h5>
4758
4759<p>
4760The argument and return value are floating point numbers of the same type.
4761</p>
4762
4763<h5>Semantics:</h5>
4764
4765<p>
4766This function returns the sqrt of the specified operand if it is a nonnegative
4767floating point number.
4768</p>
4769</div>
4770
4771<!-- _______________________________________________________________________ -->
4772<div class="doc_subsubsection">
4773 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
4774</div>
4775
4776<div class="doc_text">
4777
4778<h5>Syntax:</h5>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]4779<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]4780floating point or vector of floating point type. Not all targets support all
4781types however.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4782<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]4783 declare float @llvm.powi.f32(float %Val, i32 %power)
4784 declare double @llvm.powi.f64(double %Val, i32 %power)
4785 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
4786 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
4787 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4788</pre>
4789
4790<h5>Overview:</h5>
4791
4792<p>
4793The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
4794specified (positive or negative) power. The order of evaluation of
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]4795multiplications is not defined. When a vector of floating point type is
4796used, the second argument remains a scalar integer value.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4797</p>
4798
4799<h5>Arguments:</h5>
4800
4801<p>
4802The second argument is an integer power, and the first is a value to raise to
4803that power.
4804</p>
4805
4806<h5>Semantics:</h5>
4807
4808<p>
4809This function returns the first value raised to the second power with an
4810unspecified sequence of rounding operations.</p>
4811</div>
4812
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]4813<!-- _______________________________________________________________________ -->
4814<div class="doc_subsubsection">
4815 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
4816</div>
4817
4818<div class="doc_text">
4819
4820<h5>Syntax:</h5>
4821<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
4822floating point or vector of floating point type. Not all targets support all
4823types however.
4824<pre>
4825 declare float @llvm.sin.f32(float %Val)
4826 declare double @llvm.sin.f64(double %Val)
4827 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
4828 declare fp128 @llvm.sin.f128(fp128 %Val)
4829 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
4830</pre>
4831
4832<h5>Overview:</h5>
4833
4834<p>
4835The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.
4836</p>
4837
4838<h5>Arguments:</h5>
4839
4840<p>
4841The argument and return value are floating point numbers of the same type.
4842</p>
4843
4844<h5>Semantics:</h5>
4845
4846<p>
4847This function returns the sine of the specified operand, returning the
4848same values as the libm <tt>sin</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]4849conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]4850</div>
4851
4852<!-- _______________________________________________________________________ -->
4853<div class="doc_subsubsection">
4854 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
4855</div>
4856
4857<div class="doc_text">
4858
4859<h5>Syntax:</h5>
4860<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
4861floating point or vector of floating point type. Not all targets support all
4862types however.
4863<pre>
4864 declare float @llvm.cos.f32(float %Val)
4865 declare double @llvm.cos.f64(double %Val)
4866 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
4867 declare fp128 @llvm.cos.f128(fp128 %Val)
4868 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
4869</pre>
4870
4871<h5>Overview:</h5>
4872
4873<p>
4874The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.
4875</p>
4876
4877<h5>Arguments:</h5>
4878
4879<p>
4880The argument and return value are floating point numbers of the same type.
4881</p>
4882
4883<h5>Semantics:</h5>
4884
4885<p>
4886This function returns the cosine of the specified operand, returning the
4887same values as the libm <tt>cos</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]4888conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]4889</div>
4890
4891<!-- _______________________________________________________________________ -->
4892<div class="doc_subsubsection">
4893 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
4894</div>
4895
4896<div class="doc_text">
4897
4898<h5>Syntax:</h5>
4899<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
4900floating point or vector of floating point type. Not all targets support all
4901types however.
4902<pre>
4903 declare float @llvm.pow.f32(float %Val, float %Power)
4904 declare double @llvm.pow.f64(double %Val, double %Power)
4905 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
4906 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
4907 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
4908</pre>
4909
4910<h5>Overview:</h5>
4911
4912<p>
4913The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
4914specified (positive or negative) power.
4915</p>
4916
4917<h5>Arguments:</h5>
4918
4919<p>
4920The second argument is a floating point power, and the first is a value to
4921raise to that power.
4922</p>
4923
4924<h5>Semantics:</h5>
4925
4926<p>
4927This function returns the first value raised to the second power,
4928returning the
4929same values as the libm <tt>pow</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]4930conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]4931</div>
4932
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4933
4934<!-- ======================================================================= -->
4935<div class="doc_subsection">
4936 <a name="int_manip">Bit Manipulation Intrinsics</a>
4937</div>
4938
4939<div class="doc_text">
4940<p>
4941LLVM provides intrinsics for a few important bit manipulation operations.
4942These allow efficient code generation for some algorithms.
4943</p>
4944
4945</div>
4946
4947<!-- _______________________________________________________________________ -->
4948<div class="doc_subsubsection">
4949 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
4950</div>
4951
4952<div class="doc_text">
4953
4954<h5>Syntax:</h5>
4955<p>This is an overloaded intrinsic function. You can use bswap on any integer
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4956type that is an even number of bytes (i.e. BitWidth % 16 == 0).
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4957<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4958 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
4959 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
4960 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4961</pre>
4962
4963<h5>Overview:</h5>
4964
4965<p>
4966The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
4967values with an even number of bytes (positive multiple of 16 bits). These are
4968useful for performing operations on data that is not in the target's native
4969byte order.
4970</p>
4971
4972<h5>Semantics:</h5>
4973
4974<p>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4975The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4976and low byte of the input i16 swapped. Similarly, the <tt>llvm.bswap.i32</tt>
4977intrinsic returns an i32 value that has the four bytes of the input i32
4978swapped, 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]4979i32 will have its bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i48</tt>,
4980<tt>llvm.bswap.i64</tt> and other intrinsics extend this concept to
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4981additional even-byte lengths (6 bytes, 8 bytes and more, respectively).
4982</p>
4983
4984</div>
4985
4986<!-- _______________________________________________________________________ -->
4987<div class="doc_subsubsection">
4988 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
4989</div>
4990
4991<div class="doc_text">
4992
4993<h5>Syntax:</h5>
4994<p>This is an overloaded intrinsic. You can use llvm.ctpop on any integer bit
4995width. Not all targets support all bit widths however.
4996<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4997 declare i8 @llvm.ctpop.i8 (i8 &lt;src&gt;)
4998 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4999 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5000 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
5001 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5002</pre>
5003
5004<h5>Overview:</h5>
5005
5006<p>
5007The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
5008value.
5009</p>
5010
5011<h5>Arguments:</h5>
5012
5013<p>
5014The only argument is the value to be counted. The argument may be of any
5015integer type. The return type must match the argument type.
5016</p>
5017
5018<h5>Semantics:</h5>
5019
5020<p>
5021The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
5022</p>
5023</div>
5024
5025<!-- _______________________________________________________________________ -->
5026<div class="doc_subsubsection">
5027 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
5028</div>
5029
5030<div class="doc_text">
5031
5032<h5>Syntax:</h5>
5033<p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
5034integer bit width. Not all targets support all bit widths however.
5035<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5036 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
5037 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5038 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5039 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
5040 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5041</pre>
5042
5043<h5>Overview:</h5>
5044
5045<p>
5046The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
5047leading zeros in a variable.
5048</p>
5049
5050<h5>Arguments:</h5>
5051
5052<p>
5053The only argument is the value to be counted. The argument may be of any
5054integer type. The return type must match the argument type.
5055</p>
5056
5057<h5>Semantics:</h5>
5058
5059<p>
5060The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
5061in a variable. If the src == 0 then the result is the size in bits of the type
5062of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.
5063</p>
5064</div>
5065
5066
5067
5068<!-- _______________________________________________________________________ -->
5069<div class="doc_subsubsection">
5070 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
5071</div>
5072
5073<div class="doc_text">
5074
5075<h5>Syntax:</h5>
5076<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
5077integer bit width. Not all targets support all bit widths however.
5078<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5079 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
5080 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5081 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5082 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
5083 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5084</pre>
5085
5086<h5>Overview:</h5>
5087
5088<p>
5089The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
5090trailing zeros.
5091</p>
5092
5093<h5>Arguments:</h5>
5094
5095<p>
5096The only argument is the value to be counted. The argument may be of any
5097integer type. The return type must match the argument type.
5098</p>
5099
5100<h5>Semantics:</h5>
5101
5102<p>
5103The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
5104in a variable. If the src == 0 then the result is the size in bits of the type
5105of src. For example, <tt>llvm.cttz(2) = 1</tt>.
5106</p>
5107</div>
5108
5109<!-- _______________________________________________________________________ -->
5110<div class="doc_subsubsection">
5111 <a name="int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic</a>
5112</div>
5113
5114<div class="doc_text">
5115
5116<h5>Syntax:</h5>
5117<p>This is an overloaded intrinsic. You can use <tt>llvm.part.select</tt>
5118on any integer bit width.
5119<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5120 declare i17 @llvm.part.select.i17 (i17 %val, i32 %loBit, i32 %hiBit)
5121 declare i29 @llvm.part.select.i29 (i29 %val, i32 %loBit, i32 %hiBit)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5122</pre>
5123
5124<h5>Overview:</h5>
5125<p>The '<tt>llvm.part.select</tt>' family of intrinsic functions selects a
5126range of bits from an integer value and returns them in the same bit width as
5127the original value.</p>
5128
5129<h5>Arguments:</h5>
5130<p>The first argument, <tt>%val</tt> and the result may be integer types of
5131any bit width but they must have the same bit width. The second and third
5132arguments must be <tt>i32</tt> type since they specify only a bit index.</p>
5133
5134<h5>Semantics:</h5>
5135<p>The operation of the '<tt>llvm.part.select</tt>' intrinsic has two modes
5136of operation: forwards and reverse. If <tt>%loBit</tt> is greater than
5137<tt>%hiBits</tt> then the intrinsic operates in reverse mode. Otherwise it
5138operates in forward mode.</p>
5139<p>In forward mode, this intrinsic is the equivalent of shifting <tt>%val</tt>
5140right by <tt>%loBit</tt> bits and then ANDing it with a mask with
5141only the <tt>%hiBit - %loBit</tt> bits set, as follows:</p>
5142<ol>
5143 <li>The <tt>%val</tt> is shifted right (LSHR) by the number of bits specified
5144 by <tt>%loBits</tt>. This normalizes the value to the low order bits.</li>
5145 <li>The <tt>%loBits</tt> value is subtracted from the <tt>%hiBits</tt> value
5146 to determine the number of bits to retain.</li>
5147 <li>A mask of the retained bits is created by shifting a -1 value.</li>
5148 <li>The mask is ANDed with <tt>%val</tt> to produce the result.
5149</ol>
5150<p>In reverse mode, a similar computation is made except that the bits are
5151returned in the reverse order. So, for example, if <tt>X</tt> has the value
5152<tt>i16 0x0ACF (101011001111)</tt> and we apply
5153<tt>part.select(i16 X, 8, 3)</tt> to it, we get back the value
5154<tt>i16 0x0026 (000000100110)</tt>.</p>
5155</div>
5156
5157<div class="doc_subsubsection">
5158 <a name="int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic</a>
5159</div>
5160
5161<div class="doc_text">
5162
5163<h5>Syntax:</h5>
5164<p>This is an overloaded intrinsic. You can use <tt>llvm.part.set</tt>
5165on any integer bit width.
5166<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5167 declare i17 @llvm.part.set.i17.i9 (i17 %val, i9 %repl, i32 %lo, i32 %hi)
5168 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]5169</pre>
5170
5171<h5>Overview:</h5>
5172<p>The '<tt>llvm.part.set</tt>' family of intrinsic functions replaces a range
5173of bits in an integer value with another integer value. It returns the integer
5174with the replaced bits.</p>
5175
5176<h5>Arguments:</h5>
5177<p>The first argument, <tt>%val</tt> and the result may be integer types of
5178any bit width but they must have the same bit width. <tt>%val</tt> is the value
5179whose bits will be replaced. The second argument, <tt>%repl</tt> may be an
5180integer of any bit width. The third and fourth arguments must be <tt>i32</tt>
5181type since they specify only a bit index.</p>
5182
5183<h5>Semantics:</h5>
5184<p>The operation of the '<tt>llvm.part.set</tt>' intrinsic has two modes
5185of operation: forwards and reverse. If <tt>%lo</tt> is greater than
5186<tt>%hi</tt> then the intrinsic operates in reverse mode. Otherwise it
5187operates in forward mode.</p>
5188<p>For both modes, the <tt>%repl</tt> value is prepared for use by either
5189truncating it down to the size of the replacement area or zero extending it
5190up to that size.</p>
5191<p>In forward mode, the bits between <tt>%lo</tt> and <tt>%hi</tt> (inclusive)
5192are replaced with corresponding bits from <tt>%repl</tt>. That is the 0th bit
5193in <tt>%repl</tt> replaces the <tt>%lo</tt>th bit in <tt>%val</tt> and etc. up
5194to the <tt>%hi</tt>th bit.
5195<p>In reverse mode, a similar computation is made except that the bits are
5196reversed. That is, the <tt>0</tt>th bit in <tt>%repl</tt> replaces the
5197<tt>%hi</tt> bit in <tt>%val</tt> and etc. down to the <tt>%lo</tt>th bit.
5198<h5>Examples:</h5>
5199<pre>
5200 llvm.part.set(0xFFFF, 0, 4, 7) -&gt; 0xFF0F
5201 llvm.part.set(0xFFFF, 0, 7, 4) -&gt; 0xFF0F
5202 llvm.part.set(0xFFFF, 1, 7, 4) -&gt; 0xFF8F
5203 llvm.part.set(0xFFFF, F, 8, 3) -&gt; 0xFFE7
5204 llvm.part.set(0xFFFF, 0, 3, 8) -&gt; 0xFE07
5205</pre>
5206</div>
5207
5208<!-- ======================================================================= -->
5209<div class="doc_subsection">
5210 <a name="int_debugger">Debugger Intrinsics</a>
5211</div>
5212
5213<div class="doc_text">
5214<p>
5215The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
5216are described in the <a
5217href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
5218Debugging</a> document.
5219</p>
5220</div>
5221
5222
5223<!-- ======================================================================= -->
5224<div class="doc_subsection">
5225 <a name="int_eh">Exception Handling Intrinsics</a>
5226</div>
5227
5228<div class="doc_text">
5229<p> The LLVM exception handling intrinsics (which all start with
5230<tt>llvm.eh.</tt> prefix), are described in the <a
5231href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
5232Handling</a> document. </p>
5233</div>
5234
5235<!-- ======================================================================= -->
5236<div class="doc_subsection">
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5237 <a name="int_trampoline">Trampoline Intrinsic</a>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5238</div>
5239
5240<div class="doc_text">
5241<p>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5242 This intrinsic makes it possible to excise one parameter, marked with
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5243 the <tt>nest</tt> attribute, from a function. The result is a callable
5244 function pointer lacking the nest parameter - the caller does not need
5245 to provide a value for it. Instead, the value to use is stored in
5246 advance in a "trampoline", a block of memory usually allocated
5247 on the stack, which also contains code to splice the nest value into the
5248 argument list. This is used to implement the GCC nested function address
5249 extension.
5250</p>
5251<p>
5252 For example, if the function is
5253 <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]5254 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]5255<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5256 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
5257 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
5258 %p = call i8* @llvm.init.trampoline( i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8*), i8* %nval )
5259 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5260</pre>
Bill Wendlinge262b4c2007-09-22 09:23:55 +0000[diff] [blame]5261 <p>The call <tt>%val = call i32 %fp( i32 %x, i32 %y )</tt> is then equivalent
5262 to <tt>%val = call i32 %f( i8* %nval, i32 %x, i32 %y )</tt>.</p>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5263</div>
5264
5265<!-- _______________________________________________________________________ -->
5266<div class="doc_subsubsection">
5267 <a name="int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a>
5268</div>
5269<div class="doc_text">
5270<h5>Syntax:</h5>
5271<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5272declare 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]5273</pre>
5274<h5>Overview:</h5>
5275<p>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5276 This fills the memory pointed to by <tt>tramp</tt> with code
5277 and returns a function pointer suitable for executing it.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5278</p>
5279<h5>Arguments:</h5>
5280<p>
5281 The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
5282 pointers. The <tt>tramp</tt> argument must point to a sufficiently large
5283 and sufficiently aligned block of memory; this memory is written to by the
Duncan Sands35012212007-08-22 23:39:54 +0000[diff] [blame]5284 intrinsic. Note that the size and the alignment are target-specific - LLVM
5285 currently provides no portable way of determining them, so a front-end that
5286 generates this intrinsic needs to have some target-specific knowledge.
5287 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]5288</p>
5289<h5>Semantics:</h5>
5290<p>
5291 The block of memory pointed to by <tt>tramp</tt> is filled with target
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5292 dependent code, turning it into a function. A pointer to this function is
5293 returned, but needs to be bitcast to an
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5294 <a href="#int_trampoline">appropriate function pointer type</a>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5295 before being called. The new function's signature is the same as that of
5296 <tt>func</tt> with any arguments marked with the <tt>nest</tt> attribute
5297 removed. At most one such <tt>nest</tt> argument is allowed, and it must be
5298 of pointer type. Calling the new function is equivalent to calling
5299 <tt>func</tt> with the same argument list, but with <tt>nval</tt> used for the
5300 missing <tt>nest</tt> argument. If, after calling
5301 <tt>llvm.init.trampoline</tt>, the memory pointed to by <tt>tramp</tt> is
5302 modified, then the effect of any later call to the returned function pointer is
5303 undefined.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5304</p>
5305</div>
5306
5307<!-- ======================================================================= -->
5308<div class="doc_subsection">
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]5309 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
5310</div>
5311
5312<div class="doc_text">
5313<p>
5314 These intrinsic functions expand the "universal IR" of LLVM to represent
5315 hardware constructs for atomic operations and memory synchronization. This
5316 provides an interface to the hardware, not an interface to the programmer. It
5317 is aimed at a low enough level to allow any programming models or APIs which
5318 need atomic behaviors to map cleanly onto it. It is also modeled primarily on
5319 hardware behavior. Just as hardware provides a "universal IR" for source
5320 languages, it also provides a starting point for developing a "universal"
5321 atomic operation and synchronization IR.
5322</p>
5323<p>
5324 These do <em>not</em> form an API such as high-level threading libraries,
5325 software transaction memory systems, atomic primitives, and intrinsic
5326 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
5327 application libraries. The hardware interface provided by LLVM should allow
5328 a clean implementation of all of these APIs and parallel programming models.
5329 No one model or paradigm should be selected above others unless the hardware
5330 itself ubiquitously does so.
5331
5332</p>
5333</div>
5334
5335<!-- _______________________________________________________________________ -->
5336<div class="doc_subsubsection">
5337 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
5338</div>
5339<div class="doc_text">
5340<h5>Syntax:</h5>
5341<pre>
5342declare void @llvm.memory.barrier( i1 &lt;ll&gt;, i1 &lt;ls&gt;, i1 &lt;sl&gt;, i1 &lt;ss&gt;,
5343i1 &lt;device&gt; )
5344
5345</pre>
5346<h5>Overview:</h5>
5347<p>
5348 The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
5349 specific pairs of memory access types.
5350</p>
5351<h5>Arguments:</h5>
5352<p>
5353 The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
5354 The first four arguments enables a specific barrier as listed below. The fith
5355 argument specifies that the barrier applies to io or device or uncached memory.
5356
5357</p>
5358 <ul>
5359 <li><tt>ll</tt>: load-load barrier</li>
5360 <li><tt>ls</tt>: load-store barrier</li>
5361 <li><tt>sl</tt>: store-load barrier</li>
5362 <li><tt>ss</tt>: store-store barrier</li>
5363 <li><tt>device</tt>: barrier applies to device and uncached memory also.
5364 </ul>
5365<h5>Semantics:</h5>
5366<p>
5367 This intrinsic causes the system to enforce some ordering constraints upon
5368 the loads and stores of the program. This barrier does not indicate
5369 <em>when</em> any events will occur, it only enforces an <em>order</em> in
5370 which they occur. For any of the specified pairs of load and store operations
5371 (f.ex. load-load, or store-load), all of the first operations preceding the
5372 barrier will complete before any of the second operations succeeding the
5373 barrier begin. Specifically the semantics for each pairing is as follows:
5374</p>
5375 <ul>
5376 <li><tt>ll</tt>: All loads before the barrier must complete before any load
5377 after the barrier begins.</li>
5378
5379 <li><tt>ls</tt>: All loads before the barrier must complete before any
5380 store after the barrier begins.</li>
5381 <li><tt>ss</tt>: All stores before the barrier must complete before any
5382 store after the barrier begins.</li>
5383 <li><tt>sl</tt>: All stores before the barrier must complete before any
5384 load after the barrier begins.</li>
5385 </ul>
5386<p>
5387 These semantics are applied with a logical "and" behavior when more than one
5388 is enabled in a single memory barrier intrinsic.
5389</p>
5390<p>
5391 Backends may implement stronger barriers than those requested when they do not
5392 support as fine grained a barrier as requested. Some architectures do not
5393 need all types of barriers and on such architectures, these become noops.
5394</p>
5395<h5>Example:</h5>
5396<pre>
5397%ptr = malloc i32
5398 store i32 4, %ptr
5399
5400%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
5401 call void @llvm.memory.barrier( i1 false, i1 true, i1 false, i1 false )
5402 <i>; guarantee the above finishes</i>
5403 store i32 8, %ptr <i>; before this begins</i>
5404</pre>
5405</div>
5406
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5407<!-- _______________________________________________________________________ -->
5408<div class="doc_subsubsection">
5409 <a name="int_atomic_lcs">'<tt>llvm.atomic.lcs.*</tt>' Intrinsic</a>
5410</div>
5411<div class="doc_text">
5412<h5>Syntax:</h5>
5413<p>
5414 This is an overloaded intrinsic. You can use <tt>llvm.atomic.lcs</tt> on any
5415 integer bit width. Not all targets support all bit widths however.</p>
5416
5417<pre>
5418declare i8 @llvm.atomic.lcs.i8( i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt; )
5419declare i16 @llvm.atomic.lcs.i16( i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt; )
5420declare i32 @llvm.atomic.lcs.i32( i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt; )
5421declare i64 @llvm.atomic.lcs.i64( i64* &lt;ptr&gt;, i64 &lt;cmp&gt;, i64 &lt;val&gt; )
5422
5423</pre>
5424<h5>Overview:</h5>
5425<p>
5426 This loads a value in memory and compares it to a given value. If they are
5427 equal, it stores a new value into the memory.
5428</p>
5429<h5>Arguments:</h5>
5430<p>
5431 The <tt>llvm.atomic.lcs</tt> intrinsic takes three arguments. The result as
5432 well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
5433 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
5434 this integer type. While any bit width integer may be used, targets may only
5435 lower representations they support in hardware.
5436
5437</p>
5438<h5>Semantics:</h5>
5439<p>
5440 This entire intrinsic must be executed atomically. It first loads the value
5441 in memory pointed to by <tt>ptr</tt> and compares it with the value
5442 <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the memory. The
5443 loaded value is yielded in all cases. This provides the equivalent of an
5444 atomic compare-and-swap operation within the SSA framework.
5445</p>
5446<h5>Examples:</h5>
5447
5448<pre>
5449%ptr = malloc i32
5450 store i32 4, %ptr
5451
5452%val1 = add i32 4, 4
5453%result1 = call i32 @llvm.atomic.lcs.i32( i32* %ptr, i32 4, %val1 )
5454 <i>; yields {i32}:result1 = 4</i>
5455%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
5456%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
5457
5458%val2 = add i32 1, 1
5459%result2 = call i32 @llvm.atomic.lcs.i32( i32* %ptr, i32 5, %val2 )
5460 <i>; yields {i32}:result2 = 8</i>
5461%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
5462
5463%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
5464</pre>
5465</div>
5466
5467<!-- _______________________________________________________________________ -->
5468<div class="doc_subsubsection">
5469 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
5470</div>
5471<div class="doc_text">
5472<h5>Syntax:</h5>
5473
5474<p>
5475 This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
5476 integer bit width. Not all targets support all bit widths however.</p>
5477<pre>
5478declare i8 @llvm.atomic.swap.i8( i8* &lt;ptr&gt;, i8 &lt;val&gt; )
5479declare i16 @llvm.atomic.swap.i16( i16* &lt;ptr&gt;, i16 &lt;val&gt; )
5480declare i32 @llvm.atomic.swap.i32( i32* &lt;ptr&gt;, i32 &lt;val&gt; )
5481declare i64 @llvm.atomic.swap.i64( i64* &lt;ptr&gt;, i64 &lt;val&gt; )
5482
5483</pre>
5484<h5>Overview:</h5>
5485<p>
5486 This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
5487 the value from memory. It then stores the value in <tt>val</tt> in the memory
5488 at <tt>ptr</tt>.
5489</p>
5490<h5>Arguments:</h5>
5491
5492<p>
5493 The <tt>llvm.atomic.ls</tt> intrinsic takes two arguments. Both the
5494 <tt>val</tt> argument and the result must be integers of the same bit width.
5495 The first argument, <tt>ptr</tt>, must be a pointer to a value of this
5496 integer type. The targets may only lower integer representations they
5497 support.
5498</p>
5499<h5>Semantics:</h5>
5500<p>
5501 This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
5502 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
5503 equivalent of an atomic swap operation within the SSA framework.
5504
5505</p>
5506<h5>Examples:</h5>
5507<pre>
5508%ptr = malloc i32
5509 store i32 4, %ptr
5510
5511%val1 = add i32 4, 4
5512%result1 = call i32 @llvm.atomic.swap.i32( i32* %ptr, i32 %val1 )
5513 <i>; yields {i32}:result1 = 4</i>
5514%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
5515%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
5516
5517%val2 = add i32 1, 1
5518%result2 = call i32 @llvm.atomic.swap.i32( i32* %ptr, i32 %val2 )
5519 <i>; yields {i32}:result2 = 8</i>
5520
5521%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
5522%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
5523</pre>
5524</div>
5525
5526<!-- _______________________________________________________________________ -->
5527<div class="doc_subsubsection">
5528 <a name="int_atomic_las">'<tt>llvm.atomic.las.*</tt>' Intrinsic</a>
5529
5530</div>
5531<div class="doc_text">
5532<h5>Syntax:</h5>
5533<p>
5534 This is an overloaded intrinsic. You can use <tt>llvm.atomic.las</tt> on any
5535 integer bit width. Not all targets support all bit widths however.</p>
5536<pre>
5537declare i8 @llvm.atomic.las.i8.( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
5538declare i16 @llvm.atomic.las.i16.( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
5539declare i32 @llvm.atomic.las.i32.( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
5540declare i64 @llvm.atomic.las.i64.( i64* &lt;ptr&gt;, i64 &lt;delta&gt; )
5541
5542</pre>
5543<h5>Overview:</h5>
5544<p>
5545 This intrinsic adds <tt>delta</tt> to the value stored in memory at
5546 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
5547</p>
5548<h5>Arguments:</h5>
5549<p>
5550
5551 The intrinsic takes two arguments, the first a pointer to an integer value
5552 and the second an integer value. The result is also an integer value. These
5553 integer types can have any bit width, but they must all have the same bit
5554 width. The targets may only lower integer representations they support.
5555</p>
5556<h5>Semantics:</h5>
5557<p>
5558 This intrinsic does a series of operations atomically. It first loads the
5559 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
5560 to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
5561</p>
5562
5563<h5>Examples:</h5>
5564<pre>
5565%ptr = malloc i32
5566 store i32 4, %ptr
5567%result1 = call i32 @llvm.atomic.las.i32( i32* %ptr, i32 4 )
5568 <i>; yields {i32}:result1 = 4</i>
5569%result2 = call i32 @llvm.atomic.las.i32( i32* %ptr, i32 2 )
5570 <i>; yields {i32}:result2 = 8</i>
5571%result3 = call i32 @llvm.atomic.las.i32( i32* %ptr, i32 5 )
5572 <i>; yields {i32}:result3 = 10</i>
5573%memval = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
5574</pre>
5575</div>
5576
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]5577
5578<!-- ======================================================================= -->
5579<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5580 <a name="int_general">General Intrinsics</a>
5581</div>
5582
5583<div class="doc_text">
5584<p> This class of intrinsics is designed to be generic and has
5585no specific purpose. </p>
5586</div>
5587
5588<!-- _______________________________________________________________________ -->
5589<div class="doc_subsubsection">
5590 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
5591</div>
5592
5593<div class="doc_text">
5594
5595<h5>Syntax:</h5>
5596<pre>
5597 declare void @llvm.var.annotation(i8* &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
5598</pre>
5599
5600<h5>Overview:</h5>
5601
5602<p>
5603The '<tt>llvm.var.annotation</tt>' intrinsic
5604</p>
5605
5606<h5>Arguments:</h5>
5607
5608<p>
5609The first argument is a pointer to a value, the second is a pointer to a
5610global string, the third is a pointer to a global string which is the source
5611file name, and the last argument is the line number.
5612</p>
5613
5614<h5>Semantics:</h5>
5615
5616<p>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]5617This intrinsic allows annotation of local variables with arbitrary strings.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5618This can be useful for special purpose optimizations that want to look for these
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]5619annotations. These have no other defined use, they are ignored by code
5620generation and optimization.
5621</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5622</div>
5623
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]5624<!-- _______________________________________________________________________ -->
5625<div class="doc_subsubsection">
Tanya Lattnerc9869b12007-09-21 23:57:59 +0000[diff] [blame]5626 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]5627</div>
5628
5629<div class="doc_text">
5630
5631<h5>Syntax:</h5>
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]5632<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
5633any integer bit width.
5634</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]5635<pre>
Tanya Lattner09161fe2007-09-22 00:03:01 +0000[diff] [blame]5636 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
5637 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
5638 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
5639 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
5640 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]5641</pre>
5642
5643<h5>Overview:</h5>
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]5644
5645<p>
5646The '<tt>llvm.annotation</tt>' intrinsic.
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]5647</p>
5648
5649<h5>Arguments:</h5>
5650
5651<p>
5652The first argument is an integer value (result of some expression),
5653the second is a pointer to a global string, the third is a pointer to a global
5654string which is the source file name, and the last argument is the line number.
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]5655It returns the value of the first argument.
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]5656</p>
5657
5658<h5>Semantics:</h5>
5659
5660<p>
5661This intrinsic allows annotations to be put on arbitrary expressions
5662with arbitrary strings. This can be useful for special purpose optimizations
5663that want to look for these annotations. These have no other defined use, they
5664are ignored by code generation and optimization.
5665</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5666
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]5667<!-- _______________________________________________________________________ -->
5668<div class="doc_subsubsection">
5669 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
5670</div>
5671
5672<div class="doc_text">
5673
5674<h5>Syntax:</h5>
5675<pre>
5676 declare void @llvm.trap()
5677</pre>
5678
5679<h5>Overview:</h5>
5680
5681<p>
5682The '<tt>llvm.trap</tt>' intrinsic
5683</p>
5684
5685<h5>Arguments:</h5>
5686
5687<p>
5688None
5689</p>
5690
5691<h5>Semantics:</h5>
5692
5693<p>
5694This intrinsics is lowered to the target dependent trap instruction. If the
5695target does not have a trap instruction, this intrinsic will be lowered to the
5696call of the abort() function.
5697</p>
5698</div>
5699
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5700<!-- *********************************************************************** -->
5701<hr>
5702<address>
5703 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
5704 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
5705 <a href="http://validator.w3.org/check/referer"><img
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]5706 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!"></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5707
5708 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
5709 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
5710 Last modified: $Date$
5711</address>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]5712
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5713</body>
5714</html>