Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / 50846cf9df1fe97c0be8e2eb74bc3114aa8a1385 / . / docs / LangRef.html
blob: 9679291125abf35680ab1958d465f2efcda7bc11 [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
Anton Korobeynikov96822812008-03-22 08:36:14 +0000[diff] [blame]766 function, global variable, another alias or bitcast of global value). Aliases
767 may have an optional <a href="#linkage">linkage type</a>, and an
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]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">
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1226
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1227<h5>Overview:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1228
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1229<p>The function type can be thought of as a function signature. It
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1230consists of a return type and a list of formal parameter types. The
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1231return type of a function type is a scalar type, a void type, or a struct type.
Devang Pateld5404c02008-03-24 20:52:42 +0000[diff] [blame]1232If the return type is a struct type then all struct elements must be of first
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1233class types, and the struct must have at least one element.</p>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1234
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1235<h5>Syntax:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1236
1237<pre>
1238 &lt;returntype list&gt; (&lt;parameter list&gt;)
1239</pre>
1240
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1241<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
1242specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1243which indicates that the function takes a variable number of arguments.
1244Variable argument functions can access their arguments with the <a
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1245 href="#int_varargs">variable argument handling intrinsic</a> functions.
1246'<tt>&lt;returntype list&gt;</tt>' is a comma-separated list of
1247<a href="#t_firstclass">first class</a> type specifiers.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1248
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1249<h5>Examples:</h5>
1250<table class="layout">
1251 <tr class="layout">
1252 <td class="left"><tt>i32 (i32)</tt></td>
1253 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
1254 </td>
1255 </tr><tr class="layout">
Reid Spencerf234bed2007-07-19 23:13:04 +0000[diff] [blame]1256 <td class="left"><tt>float&nbsp;(i16&nbsp;signext,&nbsp;i32&nbsp;*)&nbsp;*
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1257 </tt></td>
1258 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
1259 an <tt>i16</tt> that should be sign extended and a
1260 <a href="#t_pointer">pointer</a> to <tt>i32</tt>, returning
1261 <tt>float</tt>.
1262 </td>
1263 </tr><tr class="layout">
1264 <td class="left"><tt>i32 (i8*, ...)</tt></td>
1265 <td class="left">A vararg function that takes at least one
1266 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1267 which returns an integer. This is the signature for <tt>printf</tt> in
1268 LLVM.
1269 </td>
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1270 </tr><tr class="layout">
1271 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Devang Patel6dddba22008-03-24 18:10:52 +0000[diff] [blame]1272 <td class="left">A function taking an <tt>i32></tt>, returning two
1273 <tt> i32 </tt> values as an aggregate of type <tt>{ i32, i32 }</tt>
Devang Pateld4ba41d2008-03-24 05:35:41 +0000[diff] [blame]1274 </td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1275 </tr>
1276</table>
1277
1278</div>
1279<!-- _______________________________________________________________________ -->
1280<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
1281<div class="doc_text">
1282<h5>Overview:</h5>
1283<p>The structure type is used to represent a collection of data members
1284together in memory. The packing of the field types is defined to match
1285the ABI of the underlying processor. The elements of a structure may
1286be any type that has a size.</p>
1287<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1288and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1289field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1290instruction.</p>
1291<h5>Syntax:</h5>
1292<pre> { &lt;type list&gt; }<br></pre>
1293<h5>Examples:</h5>
1294<table class="layout">
1295 <tr class="layout">
1296 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1297 <td class="left">A triple of three <tt>i32</tt> values</td>
1298 </tr><tr class="layout">
1299 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1300 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1301 second element is a <a href="#t_pointer">pointer</a> to a
1302 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1303 an <tt>i32</tt>.</td>
1304 </tr>
1305</table>
1306</div>
1307
1308<!-- _______________________________________________________________________ -->
1309<div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
1310</div>
1311<div class="doc_text">
1312<h5>Overview:</h5>
1313<p>The packed structure type is used to represent a collection of data members
1314together in memory. There is no padding between fields. Further, the alignment
1315of a packed structure is 1 byte. The elements of a packed structure may
1316be any type that has a size.</p>
1317<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1318and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1319field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1320instruction.</p>
1321<h5>Syntax:</h5>
1322<pre> &lt; { &lt;type list&gt; } &gt; <br></pre>
1323<h5>Examples:</h5>
1324<table class="layout">
1325 <tr class="layout">
1326 <td class="left"><tt>&lt; { i32, i32, i32 } &gt;</tt></td>
1327 <td class="left">A triple of three <tt>i32</tt> values</td>
1328 </tr><tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1329 <td class="left"><tt>&lt; { float, i32 (i32)* } &gt;</tt></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1330 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1331 second element is a <a href="#t_pointer">pointer</a> to a
1332 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1333 an <tt>i32</tt>.</td>
1334 </tr>
1335</table>
1336</div>
1337
1338<!-- _______________________________________________________________________ -->
1339<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
1340<div class="doc_text">
1341<h5>Overview:</h5>
1342<p>As in many languages, the pointer type represents a pointer or
Christopher Lambdd0049d2007-12-11 09:31:00 +0000[diff] [blame]1343reference to another object, which must live in memory. Pointer types may have
1344an optional address space attribute defining the target-specific numbered
1345address space where the pointed-to object resides. The default address space is
1346zero.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1347<h5>Syntax:</h5>
1348<pre> &lt;type&gt; *<br></pre>
1349<h5>Examples:</h5>
1350<table class="layout">
1351 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1352 <td class="left"><tt>[4x i32]*</tt></td>
1353 <td class="left">A <a href="#t_pointer">pointer</a> to <a
1354 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
1355 </tr>
1356 <tr class="layout">
1357 <td class="left"><tt>i32 (i32 *) *</tt></td>
1358 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1359 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1360 <tt>i32</tt>.</td>
1361 </tr>
1362 <tr class="layout">
1363 <td class="left"><tt>i32 addrspace(5)*</tt></td>
1364 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
1365 that resides in address space #5.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1366 </tr>
1367</table>
1368</div>
1369
1370<!-- _______________________________________________________________________ -->
1371<div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
1372<div class="doc_text">
1373
1374<h5>Overview:</h5>
1375
1376<p>A vector type is a simple derived type that represents a vector
1377of elements. Vector types are used when multiple primitive data
1378are operated in parallel using a single instruction (SIMD).
1379A vector type requires a size (number of
1380elements) and an underlying primitive data type. Vectors must have a power
1381of two length (1, 2, 4, 8, 16 ...). Vector types are
1382considered <a href="#t_firstclass">first class</a>.</p>
1383
1384<h5>Syntax:</h5>
1385
1386<pre>
1387 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1388</pre>
1389
1390<p>The number of elements is a constant integer value; elementtype may
1391be any integer or floating point type.</p>
1392
1393<h5>Examples:</h5>
1394
1395<table class="layout">
1396 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1397 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
1398 <td class="left">Vector of 4 32-bit integer values.</td>
1399 </tr>
1400 <tr class="layout">
1401 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
1402 <td class="left">Vector of 8 32-bit floating-point values.</td>
1403 </tr>
1404 <tr class="layout">
1405 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
1406 <td class="left">Vector of 2 64-bit integer values.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1407 </tr>
1408</table>
1409</div>
1410
1411<!-- _______________________________________________________________________ -->
1412<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1413<div class="doc_text">
1414
1415<h5>Overview:</h5>
1416
1417<p>Opaque types are used to represent unknown types in the system. This
Gordon Henriksenda0706e2007-10-14 00:34:53 +0000[diff] [blame]1418corresponds (for example) to the C notion of a forward declared structure type.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1419In LLVM, opaque types can eventually be resolved to any type (not just a
1420structure type).</p>
1421
1422<h5>Syntax:</h5>
1423
1424<pre>
1425 opaque
1426</pre>
1427
1428<h5>Examples:</h5>
1429
1430<table class="layout">
1431 <tr class="layout">
Chris Lattner7311d222007-12-19 05:04:11 +0000[diff] [blame]1432 <td class="left"><tt>opaque</tt></td>
1433 <td class="left">An opaque type.</td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1434 </tr>
1435</table>
1436</div>
1437
1438
1439<!-- *********************************************************************** -->
1440<div class="doc_section"> <a name="constants">Constants</a> </div>
1441<!-- *********************************************************************** -->
1442
1443<div class="doc_text">
1444
1445<p>LLVM has several different basic types of constants. This section describes
1446them all and their syntax.</p>
1447
1448</div>
1449
1450<!-- ======================================================================= -->
1451<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
1452
1453<div class="doc_text">
1454
1455<dl>
1456 <dt><b>Boolean constants</b></dt>
1457
1458 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
1459 constants of the <tt><a href="#t_primitive">i1</a></tt> type.
1460 </dd>
1461
1462 <dt><b>Integer constants</b></dt>
1463
1464 <dd>Standard integers (such as '4') are constants of the <a
1465 href="#t_integer">integer</a> type. Negative numbers may be used with
1466 integer types.
1467 </dd>
1468
1469 <dt><b>Floating point constants</b></dt>
1470
1471 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
1472 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]1473 notation (see below). The assembler requires the exact decimal value of
1474 a floating-point constant. For example, the assembler accepts 1.25 but
1475 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
1476 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1477
1478 <dt><b>Null pointer constants</b></dt>
1479
1480 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
1481 and must be of <a href="#t_pointer">pointer type</a>.</dd>
1482
1483</dl>
1484
1485<p>The one non-intuitive notation for constants is the optional hexadecimal form
1486of floating point constants. For example, the form '<tt>double
14870x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
14884.5e+15</tt>'. The only time hexadecimal floating point constants are required
1489(and the only time that they are generated by the disassembler) is when a
1490floating point constant must be emitted but it cannot be represented as a
1491decimal floating point number. For example, NaN's, infinities, and other
1492special values are represented in their IEEE hexadecimal format so that
1493assembly and disassembly do not cause any bits to change in the constants.</p>
1494
1495</div>
1496
1497<!-- ======================================================================= -->
1498<div class="doc_subsection"><a name="aggregateconstants">Aggregate Constants</a>
1499</div>
1500
1501<div class="doc_text">
1502<p>Aggregate constants arise from aggregation of simple constants
1503and smaller aggregate constants.</p>
1504
1505<dl>
1506 <dt><b>Structure constants</b></dt>
1507
1508 <dd>Structure constants are represented with notation similar to structure
1509 type definitions (a comma separated list of elements, surrounded by braces
Chris Lattner6c8de962007-12-25 20:34:52 +0000[diff] [blame]1510 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
1511 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]1512 must have <a href="#t_struct">structure type</a>, and the number and
1513 types of elements must match those specified by the type.
1514 </dd>
1515
1516 <dt><b>Array constants</b></dt>
1517
1518 <dd>Array constants are represented with notation similar to array type
1519 definitions (a comma separated list of elements, surrounded by square brackets
1520 (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74 ]</tt>". Array
1521 constants must have <a href="#t_array">array type</a>, and the number and
1522 types of elements must match those specified by the type.
1523 </dd>
1524
1525 <dt><b>Vector constants</b></dt>
1526
1527 <dd>Vector constants are represented with notation similar to vector type
1528 definitions (a comma separated list of elements, surrounded by
1529 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32 42,
1530 i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must have <a
1531 href="#t_vector">vector type</a>, and the number and types of elements must
1532 match those specified by the type.
1533 </dd>
1534
1535 <dt><b>Zero initialization</b></dt>
1536
1537 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1538 value to zero of <em>any</em> type, including scalar and aggregate types.
1539 This is often used to avoid having to print large zero initializers (e.g. for
1540 large arrays) and is always exactly equivalent to using explicit zero
1541 initializers.
1542 </dd>
1543</dl>
1544
1545</div>
1546
1547<!-- ======================================================================= -->
1548<div class="doc_subsection">
1549 <a name="globalconstants">Global Variable and Function Addresses</a>
1550</div>
1551
1552<div class="doc_text">
1553
1554<p>The addresses of <a href="#globalvars">global variables</a> and <a
1555href="#functionstructure">functions</a> are always implicitly valid (link-time)
1556constants. These constants are explicitly referenced when the <a
1557href="#identifiers">identifier for the global</a> is used and always have <a
1558href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1559file:</p>
1560
1561<div class="doc_code">
1562<pre>
1563@X = global i32 17
1564@Y = global i32 42
1565@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
1566</pre>
1567</div>
1568
1569</div>
1570
1571<!-- ======================================================================= -->
1572<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
1573<div class="doc_text">
1574 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
1575 no specific value. Undefined values may be of any type and be used anywhere
1576 a constant is permitted.</p>
1577
1578 <p>Undefined values indicate to the compiler that the program is well defined
1579 no matter what value is used, giving the compiler more freedom to optimize.
1580 </p>
1581</div>
1582
1583<!-- ======================================================================= -->
1584<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1585</div>
1586
1587<div class="doc_text">
1588
1589<p>Constant expressions are used to allow expressions involving other constants
1590to be used as constants. Constant expressions may be of any <a
1591href="#t_firstclass">first class</a> type and may involve any LLVM operation
1592that does not have side effects (e.g. load and call are not supported). The
1593following is the syntax for constant expressions:</p>
1594
1595<dl>
1596 <dt><b><tt>trunc ( CST to TYPE )</tt></b></dt>
1597 <dd>Truncate a constant to another type. The bit size of CST must be larger
1598 than the bit size of TYPE. Both types must be integers.</dd>
1599
1600 <dt><b><tt>zext ( CST to TYPE )</tt></b></dt>
1601 <dd>Zero extend a constant to another type. The bit size of CST must be
1602 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1603
1604 <dt><b><tt>sext ( CST to TYPE )</tt></b></dt>
1605 <dd>Sign extend a constant to another type. The bit size of CST must be
1606 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1607
1608 <dt><b><tt>fptrunc ( CST to TYPE )</tt></b></dt>
1609 <dd>Truncate a floating point constant to another floating point type. The
1610 size of CST must be larger than the size of TYPE. Both types must be
1611 floating point.</dd>
1612
1613 <dt><b><tt>fpext ( CST to TYPE )</tt></b></dt>
1614 <dd>Floating point extend a constant to another type. The size of CST must be
1615 smaller or equal to the size of TYPE. Both types must be floating point.</dd>
1616
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]1617 <dt><b><tt>fptoui ( CST to TYPE )</tt></b></dt>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1618 <dd>Convert a floating point constant to the corresponding unsigned integer
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1619 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1620 or vector floating point 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 integer type,
1622 the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1623
1624 <dt><b><tt>fptosi ( CST to TYPE )</tt></b></dt>
1625 <dd>Convert a floating point constant to the corresponding signed integer
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1626 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1627 or vector floating point 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 integer type,
1629 the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1630
1631 <dt><b><tt>uitofp ( CST to TYPE )</tt></b></dt>
1632 <dd>Convert an unsigned integer constant to the corresponding floating point
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1633 constant. TYPE must be a scalar or vector floating point type. CST must be of
1634 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1635 of the same number of elements. If the value won't fit in the floating point
1636 type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1637
1638 <dt><b><tt>sitofp ( CST to TYPE )</tt></b></dt>
1639 <dd>Convert a signed integer constant to the corresponding floating point
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]1640 constant. TYPE must be a scalar or vector floating point type. CST must be of
1641 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1642 of the same number of elements. If the value won't fit in the floating point
1643 type, the results are undefined.</dd>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1644
1645 <dt><b><tt>ptrtoint ( CST to TYPE )</tt></b></dt>
1646 <dd>Convert a pointer typed constant to the corresponding integer constant
1647 TYPE must be an integer type. CST must be of pointer type. The CST value is
1648 zero extended, truncated, or unchanged to make it fit in TYPE.</dd>
1649
1650 <dt><b><tt>inttoptr ( CST to TYPE )</tt></b></dt>
1651 <dd>Convert a integer constant to a pointer constant. TYPE must be a
1652 pointer type. CST must be of integer type. The CST value is zero extended,
1653 truncated, or unchanged to make it fit in a pointer size. This one is
1654 <i>really</i> dangerous!</dd>
1655
1656 <dt><b><tt>bitcast ( CST to TYPE )</tt></b></dt>
1657 <dd>Convert a constant, CST, to another TYPE. The size of CST and TYPE must be
1658 identical (same number of bits). The conversion is done as if the CST value
1659 was stored to memory and read back as TYPE. In other words, no bits change
1660 with this operator, just the type. This can be used for conversion of
1661 vector types to any other type, as long as they have the same bit width. For
1662 pointers it is only valid to cast to another pointer type.
1663 </dd>
1664
1665 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
1666
1667 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
1668 constants. As with the <a href="#i_getelementptr">getelementptr</a>
1669 instruction, the index list may have zero or more indexes, which are required
1670 to make sense for the type of "CSTPTR".</dd>
1671
1672 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
1673
1674 <dd>Perform the <a href="#i_select">select operation</a> on
1675 constants.</dd>
1676
1677 <dt><b><tt>icmp COND ( VAL1, VAL2 )</tt></b></dt>
1678 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
1679
1680 <dt><b><tt>fcmp COND ( VAL1, VAL2 )</tt></b></dt>
1681 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
1682
1683 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
1684
1685 <dd>Perform the <a href="#i_extractelement">extractelement
1686 operation</a> on constants.
1687
1688 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
1689
1690 <dd>Perform the <a href="#i_insertelement">insertelement
1691 operation</a> on constants.</dd>
1692
1693
1694 <dt><b><tt>shufflevector ( VEC1, VEC2, IDXMASK )</tt></b></dt>
1695
1696 <dd>Perform the <a href="#i_shufflevector">shufflevector
1697 operation</a> on constants.</dd>
1698
1699 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
1700
1701 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
1702 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
1703 binary</a> operations. The constraints on operands are the same as those for
1704 the corresponding instruction (e.g. no bitwise operations on floating point
1705 values are allowed).</dd>
1706</dl>
1707</div>
1708
1709<!-- *********************************************************************** -->
1710<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
1711<!-- *********************************************************************** -->
1712
1713<!-- ======================================================================= -->
1714<div class="doc_subsection">
1715<a name="inlineasm">Inline Assembler Expressions</a>
1716</div>
1717
1718<div class="doc_text">
1719
1720<p>
1721LLVM supports inline assembler expressions (as opposed to <a href="#moduleasm">
1722Module-Level Inline Assembly</a>) through the use of a special value. This
1723value represents the inline assembler as a string (containing the instructions
1724to emit), a list of operand constraints (stored as a string), and a flag that
1725indicates whether or not the inline asm expression has side effects. An example
1726inline assembler expression is:
1727</p>
1728
1729<div class="doc_code">
1730<pre>
1731i32 (i32) asm "bswap $0", "=r,r"
1732</pre>
1733</div>
1734
1735<p>
1736Inline assembler expressions may <b>only</b> be used as the callee operand of
1737a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we have:
1738</p>
1739
1740<div class="doc_code">
1741<pre>
1742%X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
1743</pre>
1744</div>
1745
1746<p>
1747Inline asms with side effects not visible in the constraint list must be marked
1748as having side effects. This is done through the use of the
1749'<tt>sideeffect</tt>' keyword, like so:
1750</p>
1751
1752<div class="doc_code">
1753<pre>
1754call void asm sideeffect "eieio", ""()
1755</pre>
1756</div>
1757
1758<p>TODO: The format of the asm and constraints string still need to be
1759documented here. Constraints on what can be done (e.g. duplication, moving, etc
1760need to be documented).
1761</p>
1762
1763</div>
1764
1765<!-- *********************************************************************** -->
1766<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
1767<!-- *********************************************************************** -->
1768
1769<div class="doc_text">
1770
1771<p>The LLVM instruction set consists of several different
1772classifications of instructions: <a href="#terminators">terminator
1773instructions</a>, <a href="#binaryops">binary instructions</a>,
1774<a href="#bitwiseops">bitwise binary instructions</a>, <a
1775 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
1776instructions</a>.</p>
1777
1778</div>
1779
1780<!-- ======================================================================= -->
1781<div class="doc_subsection"> <a name="terminators">Terminator
1782Instructions</a> </div>
1783
1784<div class="doc_text">
1785
1786<p>As mentioned <a href="#functionstructure">previously</a>, every
1787basic block in a program ends with a "Terminator" instruction, which
1788indicates which block should be executed after the current block is
1789finished. These terminator instructions typically yield a '<tt>void</tt>'
1790value: they produce control flow, not values (the one exception being
1791the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
1792<p>There are six different terminator instructions: the '<a
1793 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
1794instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
1795the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
1796 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
1797 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
1798
1799</div>
1800
1801<!-- _______________________________________________________________________ -->
1802<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
1803Instruction</a> </div>
1804<div class="doc_text">
1805<h5>Syntax:</h5>
1806<pre> ret &lt;type&gt; &lt;value&gt; <i>; Return a value from a non-void function</i>
1807 ret void <i>; Return from void function</i>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1808 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]1809</pre>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1810
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1811<h5>Overview:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1812
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1813<p>The '<tt>ret</tt>' instruction is used to return control flow (and a
1814value) from a function back to the caller.</p>
1815<p>There are two forms of the '<tt>ret</tt>' instruction: one that
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1816returns value(s) and then causes control flow, and one that just causes
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1817control flow to occur.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1818
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1819<h5>Arguments:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1820
1821<p>The '<tt>ret</tt>' instruction may return zero, one or multiple values.
1822The type of each return value must be a '<a href="#t_firstclass">first
1823class</a>' type. Note that a function is not <a href="#wellformed">well
1824formed</a> if there exists a '<tt>ret</tt>' instruction inside of the
1825function that returns values that do not match the return type of the
1826function.</p>
1827
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1828<h5>Semantics:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1829
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1830<p>When the '<tt>ret</tt>' instruction is executed, control flow
1831returns back to the calling function's context. If the caller is a "<a
1832 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
1833the instruction after the call. If the caller was an "<a
1834 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
1835at the beginning of the "normal" destination block. If the instruction
1836returns a value, that value shall set the call or invoke instruction's
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1837return value. If the instruction returns multiple values then these
Devang Patelec8a5b02008-03-11 05:51:59 +0000[diff] [blame]1838values can only be accessed through a '<a href="#i_getresult"><tt>getresult</tt>
1839</a>' instruction.</p>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1840
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1841<h5>Example:</h5>
Chris Lattner43030e72008-04-23 04:59:35 +0000[diff] [blame]1842
1843<pre>
1844 ret i32 5 <i>; Return an integer value of 5</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1845 ret void <i>; Return from a void function</i>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1846 ret i32 4, i8 2 <i>; Return two values 4 and 2 </i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1847</pre>
1848</div>
1849<!-- _______________________________________________________________________ -->
1850<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
1851<div class="doc_text">
1852<h5>Syntax:</h5>
1853<pre> br i1 &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;<br> br label &lt;dest&gt; <i>; Unconditional branch</i>
1854</pre>
1855<h5>Overview:</h5>
1856<p>The '<tt>br</tt>' instruction is used to cause control flow to
1857transfer to a different basic block in the current function. There are
1858two forms of this instruction, corresponding to a conditional branch
1859and an unconditional branch.</p>
1860<h5>Arguments:</h5>
1861<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
1862single '<tt>i1</tt>' value and two '<tt>label</tt>' values. The
1863unconditional form of the '<tt>br</tt>' instruction takes a single
1864'<tt>label</tt>' value as a target.</p>
1865<h5>Semantics:</h5>
1866<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
1867argument is evaluated. If the value is <tt>true</tt>, control flows
1868to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
1869control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
1870<h5>Example:</h5>
1871<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
1872 href="#i_ret">ret</a> i32 1<br>IfUnequal:<br> <a href="#i_ret">ret</a> i32 0<br></pre>
1873</div>
1874<!-- _______________________________________________________________________ -->
1875<div class="doc_subsubsection">
1876 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
1877</div>
1878
1879<div class="doc_text">
1880<h5>Syntax:</h5>
1881
1882<pre>
1883 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
1884</pre>
1885
1886<h5>Overview:</h5>
1887
1888<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
1889several different places. It is a generalization of the '<tt>br</tt>'
1890instruction, allowing a branch to occur to one of many possible
1891destinations.</p>
1892
1893
1894<h5>Arguments:</h5>
1895
1896<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
1897comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
1898an array of pairs of comparison value constants and '<tt>label</tt>'s. The
1899table is not allowed to contain duplicate constant entries.</p>
1900
1901<h5>Semantics:</h5>
1902
1903<p>The <tt>switch</tt> instruction specifies a table of values and
1904destinations. When the '<tt>switch</tt>' instruction is executed, this
1905table is searched for the given value. If the value is found, control flow is
1906transfered to the corresponding destination; otherwise, control flow is
1907transfered to the default destination.</p>
1908
1909<h5>Implementation:</h5>
1910
1911<p>Depending on properties of the target machine and the particular
1912<tt>switch</tt> instruction, this instruction may be code generated in different
1913ways. For example, it could be generated as a series of chained conditional
1914branches or with a lookup table.</p>
1915
1916<h5>Example:</h5>
1917
1918<pre>
1919 <i>; Emulate a conditional br instruction</i>
1920 %Val = <a href="#i_zext">zext</a> i1 %value to i32
1921 switch i32 %Val, label %truedest [i32 0, label %falsedest ]
1922
1923 <i>; Emulate an unconditional br instruction</i>
1924 switch i32 0, label %dest [ ]
1925
1926 <i>; Implement a jump table:</i>
1927 switch i32 %val, label %otherwise [ i32 0, label %onzero
1928 i32 1, label %onone
1929 i32 2, label %ontwo ]
1930</pre>
1931</div>
1932
1933<!-- _______________________________________________________________________ -->
1934<div class="doc_subsubsection">
1935 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
1936</div>
1937
1938<div class="doc_text">
1939
1940<h5>Syntax:</h5>
1941
1942<pre>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]1943 &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]1944 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
1945</pre>
1946
1947<h5>Overview:</h5>
1948
1949<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
1950function, with the possibility of control flow transfer to either the
1951'<tt>normal</tt>' label or the
1952'<tt>exception</tt>' label. If the callee function returns with the
1953"<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
1954"normal" label. If the callee (or any indirect callees) returns with the "<a
1955href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]1956continued at the dynamically nearest "exception" label. If the callee function
Devang Patelec8a5b02008-03-11 05:51:59 +0000[diff] [blame]1957returns multiple values then individual return values are only accessible through
1958a '<tt><a href="#i_getresult">getresult</a></tt>' instruction.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1959
1960<h5>Arguments:</h5>
1961
1962<p>This instruction requires several arguments:</p>
1963
1964<ol>
1965 <li>
1966 The optional "cconv" marker indicates which <a href="#callingconv">calling
1967 convention</a> the call should use. If none is specified, the call defaults
1968 to using C calling conventions.
1969 </li>
1970 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
1971 function value being invoked. In most cases, this is a direct function
1972 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
1973 an arbitrary pointer to function value.
1974 </li>
1975
1976 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
1977 function to be invoked. </li>
1978
1979 <li>'<tt>function args</tt>': argument list whose types match the function
1980 signature argument types. If the function signature indicates the function
1981 accepts a variable number of arguments, the extra arguments can be
1982 specified. </li>
1983
1984 <li>'<tt>normal label</tt>': the label reached when the called function
1985 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
1986
1987 <li>'<tt>exception label</tt>': the label reached when a callee returns with
1988 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
1989
1990</ol>
1991
1992<h5>Semantics:</h5>
1993
1994<p>This instruction is designed to operate as a standard '<tt><a
1995href="#i_call">call</a></tt>' instruction in most regards. The primary
1996difference is that it establishes an association with a label, which is used by
1997the runtime library to unwind the stack.</p>
1998
1999<p>This instruction is used in languages with destructors to ensure that proper
2000cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
2001exception. Additionally, this is important for implementation of
2002'<tt>catch</tt>' clauses in high-level languages that support them.</p>
2003
2004<h5>Example:</h5>
2005<pre>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2006 %retval = invoke i32 @Test(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2007 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewyckya1c11a12008-03-16 07:18:12 +0000[diff] [blame]2008 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2009 unwind label %TestCleanup <i>; {i32}:retval set</i>
2010</pre>
2011</div>
2012
2013
2014<!-- _______________________________________________________________________ -->
2015
2016<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
2017Instruction</a> </div>
2018
2019<div class="doc_text">
2020
2021<h5>Syntax:</h5>
2022<pre>
2023 unwind
2024</pre>
2025
2026<h5>Overview:</h5>
2027
2028<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
2029at the first callee in the dynamic call stack which used an <a
2030href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
2031primarily used to implement exception handling.</p>
2032
2033<h5>Semantics:</h5>
2034
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]2035<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2036immediately halt. The dynamic call stack is then searched for the first <a
2037href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
2038execution continues at the "exceptional" destination block specified by the
2039<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
2040dynamic call chain, undefined behavior results.</p>
2041</div>
2042
2043<!-- _______________________________________________________________________ -->
2044
2045<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
2046Instruction</a> </div>
2047
2048<div class="doc_text">
2049
2050<h5>Syntax:</h5>
2051<pre>
2052 unreachable
2053</pre>
2054
2055<h5>Overview:</h5>
2056
2057<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
2058instruction is used to inform the optimizer that a particular portion of the
2059code is not reachable. This can be used to indicate that the code after a
2060no-return function cannot be reached, and other facts.</p>
2061
2062<h5>Semantics:</h5>
2063
2064<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
2065</div>
2066
2067
2068
2069<!-- ======================================================================= -->
2070<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
2071<div class="doc_text">
2072<p>Binary operators are used to do most of the computation in a
Chris Lattnerab596d92008-04-01 18:47:32 +0000[diff] [blame]2073program. They require two operands of the same type, execute an operation on them, and
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2074produce a single value. The operands might represent
2075multiple data, as is the case with the <a href="#t_vector">vector</a> data type.
Chris Lattnerab596d92008-04-01 18:47:32 +0000[diff] [blame]2076The result value has the same type as its operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2077<p>There are several different binary operators:</p>
2078</div>
2079<!-- _______________________________________________________________________ -->
2080<div class="doc_subsubsection"> <a name="i_add">'<tt>add</tt>'
2081Instruction</a> </div>
2082<div class="doc_text">
2083<h5>Syntax:</h5>
2084<pre> &lt;result&gt; = add &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2085</pre>
2086<h5>Overview:</h5>
2087<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
2088<h5>Arguments:</h5>
2089<p>The two arguments to the '<tt>add</tt>' instruction must be either <a
2090 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a> values.
2091 This instruction can also take <a href="#t_vector">vector</a> versions of the values.
2092Both arguments must have identical types.</p>
2093<h5>Semantics:</h5>
2094<p>The value produced is the integer or floating point sum of the two
2095operands.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2096<p>If an integer sum has unsigned overflow, the result returned is the
2097mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2098the result.</p>
2099<p>Because LLVM integers use a two's complement representation, this
2100instruction is appropriate for both signed and unsigned integers.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2101<h5>Example:</h5>
2102<pre> &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
2103</pre>
2104</div>
2105<!-- _______________________________________________________________________ -->
2106<div class="doc_subsubsection"> <a name="i_sub">'<tt>sub</tt>'
2107Instruction</a> </div>
2108<div class="doc_text">
2109<h5>Syntax:</h5>
2110<pre> &lt;result&gt; = sub &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2111</pre>
2112<h5>Overview:</h5>
2113<p>The '<tt>sub</tt>' instruction returns the difference of its two
2114operands.</p>
2115<p>Note that the '<tt>sub</tt>' instruction is used to represent the '<tt>neg</tt>'
2116instruction present in most other intermediate representations.</p>
2117<h5>Arguments:</h5>
2118<p>The two arguments to the '<tt>sub</tt>' instruction must be either <a
2119 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
2120values.
2121This instruction can also take <a href="#t_vector">vector</a> versions of the values.
2122Both arguments must have identical types.</p>
2123<h5>Semantics:</h5>
2124<p>The value produced is the integer or floating point difference of
2125the two operands.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2126<p>If an integer difference has unsigned overflow, the result returned is the
2127mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2128the result.</p>
2129<p>Because LLVM integers use a two's complement representation, this
2130instruction is appropriate for both signed and unsigned integers.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2131<h5>Example:</h5>
2132<pre>
2133 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
2134 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
2135</pre>
2136</div>
2137<!-- _______________________________________________________________________ -->
2138<div class="doc_subsubsection"> <a name="i_mul">'<tt>mul</tt>'
2139Instruction</a> </div>
2140<div class="doc_text">
2141<h5>Syntax:</h5>
2142<pre> &lt;result&gt; = mul &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2143</pre>
2144<h5>Overview:</h5>
2145<p>The '<tt>mul</tt>' instruction returns the product of its two
2146operands.</p>
2147<h5>Arguments:</h5>
2148<p>The two arguments to the '<tt>mul</tt>' instruction must be either <a
2149 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
2150values.
2151This instruction can also take <a href="#t_vector">vector</a> versions of the values.
2152Both arguments must have identical types.</p>
2153<h5>Semantics:</h5>
2154<p>The value produced is the integer or floating point product of the
2155two operands.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2156<p>If the result of an integer multiplication has unsigned overflow,
2157the result returned is the mathematical result modulo
21582<sup>n</sup>, where n is the bit width of the result.</p>
2159<p>Because LLVM integers use a two's complement representation, and the
2160result is the same width as the operands, this instruction returns the
2161correct result for both signed and unsigned integers. If a full product
2162(e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands
2163should be sign-extended or zero-extended as appropriate to the
2164width of the full product.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2165<h5>Example:</h5>
2166<pre> &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
2167</pre>
2168</div>
2169<!-- _______________________________________________________________________ -->
2170<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
2171</a></div>
2172<div class="doc_text">
2173<h5>Syntax:</h5>
2174<pre> &lt;result&gt; = udiv &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2175</pre>
2176<h5>Overview:</h5>
2177<p>The '<tt>udiv</tt>' instruction returns the quotient of its two
2178operands.</p>
2179<h5>Arguments:</h5>
2180<p>The two arguments to the '<tt>udiv</tt>' instruction must be
2181<a href="#t_integer">integer</a> values. Both arguments must have identical
2182types. This instruction can also take <a href="#t_vector">vector</a> versions
2183of the values in which case the elements must be integers.</p>
2184<h5>Semantics:</h5>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2185<p>The value produced is the unsigned integer quotient of the two operands.</p>
2186<p>Note that unsigned integer division and signed integer division are distinct
2187operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
2188<p>Division by zero leads to undefined behavior.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2189<h5>Example:</h5>
2190<pre> &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2191</pre>
2192</div>
2193<!-- _______________________________________________________________________ -->
2194<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
2195</a> </div>
2196<div class="doc_text">
2197<h5>Syntax:</h5>
2198<pre> &lt;result&gt; = sdiv &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2199</pre>
2200<h5>Overview:</h5>
2201<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two
2202operands.</p>
2203<h5>Arguments:</h5>
2204<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
2205<a href="#t_integer">integer</a> values. Both arguments must have identical
2206types. This instruction can also take <a href="#t_vector">vector</a> versions
2207of the values in which case the elements must be integers.</p>
2208<h5>Semantics:</h5>
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2209<p>The value produced is the signed integer quotient of the two operands rounded towards zero.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2210<p>Note that signed integer division and unsigned integer division are distinct
2211operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
2212<p>Division by zero leads to undefined behavior. Overflow also leads to
2213undefined behavior; this is a rare case, but can occur, for example,
2214by doing a 32-bit division of -2147483648 by -1.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2215<h5>Example:</h5>
2216<pre> &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2217</pre>
2218</div>
2219<!-- _______________________________________________________________________ -->
2220<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
2221Instruction</a> </div>
2222<div class="doc_text">
2223<h5>Syntax:</h5>
2224<pre> &lt;result&gt; = fdiv &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2225</pre>
2226<h5>Overview:</h5>
2227<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two
2228operands.</p>
2229<h5>Arguments:</h5>
2230<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
2231<a href="#t_floating">floating point</a> values. Both arguments must have
2232identical types. This instruction can also take <a href="#t_vector">vector</a>
2233versions of floating point values.</p>
2234<h5>Semantics:</h5>
2235<p>The value produced is the floating point quotient of the two operands.</p>
2236<h5>Example:</h5>
2237<pre> &lt;result&gt; = fdiv float 4.0, %var <i>; yields {float}:result = 4.0 / %var</i>
2238</pre>
2239</div>
2240<!-- _______________________________________________________________________ -->
2241<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
2242</div>
2243<div class="doc_text">
2244<h5>Syntax:</h5>
2245<pre> &lt;result&gt; = urem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2246</pre>
2247<h5>Overview:</h5>
2248<p>The '<tt>urem</tt>' instruction returns the remainder from the
2249unsigned division of its two arguments.</p>
2250<h5>Arguments:</h5>
2251<p>The two arguments to the '<tt>urem</tt>' instruction must be
2252<a href="#t_integer">integer</a> values. Both arguments must have identical
Dan Gohman3e3fd8c2007-11-05 23:35:22 +0000[diff] [blame]2253types. This instruction can also take <a href="#t_vector">vector</a> versions
2254of the values in which case the elements must be integers.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2255<h5>Semantics:</h5>
2256<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2257This instruction always performs an unsigned division to get the remainder.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2258<p>Note that unsigned integer remainder and signed integer remainder are
2259distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
2260<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2261<h5>Example:</h5>
2262<pre> &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2263</pre>
2264
2265</div>
2266<!-- _______________________________________________________________________ -->
2267<div class="doc_subsubsection"> <a name="i_srem">'<tt>srem</tt>'
2268Instruction</a> </div>
2269<div class="doc_text">
2270<h5>Syntax:</h5>
2271<pre> &lt;result&gt; = srem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2272</pre>
2273<h5>Overview:</h5>
2274<p>The '<tt>srem</tt>' instruction returns the remainder from the
Dan Gohman3e3fd8c2007-11-05 23:35:22 +0000[diff] [blame]2275signed division of its two operands. This instruction can also take
2276<a href="#t_vector">vector</a> versions of the values in which case
2277the elements must be integers.</p>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]2278
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2279<h5>Arguments:</h5>
2280<p>The two arguments to the '<tt>srem</tt>' instruction must be
2281<a href="#t_integer">integer</a> values. Both arguments must have identical
2282types.</p>
2283<h5>Semantics:</h5>
2284<p>This instruction returns the <i>remainder</i> of a division (where the result
2285has the same sign as the dividend, <tt>var1</tt>), not the <i>modulo</i>
2286operator (where the result has the same sign as the divisor, <tt>var2</tt>) of
2287a value. For more information about the difference, see <a
2288 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
2289Math Forum</a>. For a table of how this is implemented in various languages,
2290please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
2291Wikipedia: modulo operation</a>.</p>
Chris Lattner9aba1e22008-01-28 00:36:27 +0000[diff] [blame]2292<p>Note that signed integer remainder and unsigned integer remainder are
2293distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
2294<p>Taking the remainder of a division by zero leads to undefined behavior.
2295Overflow also leads to undefined behavior; this is a rare case, but can occur,
2296for example, by taking the remainder of a 32-bit division of -2147483648 by -1.
2297(The remainder doesn't actually overflow, but this rule lets srem be
2298implemented using instructions that return both the result of the division
2299and the remainder.)</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2300<h5>Example:</h5>
2301<pre> &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2302</pre>
2303
2304</div>
2305<!-- _______________________________________________________________________ -->
2306<div class="doc_subsubsection"> <a name="i_frem">'<tt>frem</tt>'
2307Instruction</a> </div>
2308<div class="doc_text">
2309<h5>Syntax:</h5>
2310<pre> &lt;result&gt; = frem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2311</pre>
2312<h5>Overview:</h5>
2313<p>The '<tt>frem</tt>' instruction returns the remainder from the
2314division of its two operands.</p>
2315<h5>Arguments:</h5>
2316<p>The two arguments to the '<tt>frem</tt>' instruction must be
2317<a href="#t_floating">floating point</a> values. Both arguments must have
Dan Gohman3e3fd8c2007-11-05 23:35:22 +0000[diff] [blame]2318identical types. This instruction can also take <a href="#t_vector">vector</a>
2319versions of floating point values.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2320<h5>Semantics:</h5>
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2321<p>This instruction returns the <i>remainder</i> of a division.
2322The remainder has the same sign as the dividend.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2323<h5>Example:</h5>
2324<pre> &lt;result&gt; = frem float 4.0, %var <i>; yields {float}:result = 4.0 % %var</i>
2325</pre>
2326</div>
2327
2328<!-- ======================================================================= -->
2329<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
2330Operations</a> </div>
2331<div class="doc_text">
2332<p>Bitwise binary operators are used to do various forms of
2333bit-twiddling in a program. They are generally very efficient
2334instructions and can commonly be strength reduced from other
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2335instructions. They require two operands of the same type, execute an operation on them,
2336and produce a single value. The resulting value is the same type as its operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2337</div>
2338
2339<!-- _______________________________________________________________________ -->
2340<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
2341Instruction</a> </div>
2342<div class="doc_text">
2343<h5>Syntax:</h5>
2344<pre> &lt;result&gt; = shl &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2345</pre>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2346
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2347<h5>Overview:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2348
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2349<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
2350the left a specified number of bits.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2351
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2352<h5>Arguments:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2353
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2354<p>Both arguments to the '<tt>shl</tt>' instruction must be the same <a
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]2355 href="#t_integer">integer</a> type. '<tt>var2</tt>' is treated as an
2356unsigned value.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2357
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2358<h5>Semantics:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2359
Chris Lattnerdc15b1d2008-04-01 18:45:27 +0000[diff] [blame]2360<p>The value produced is <tt>var1</tt> * 2<sup><tt>var2</tt></sup> mod 2<sup>n</sup>,
2361where n is the width of the result. If <tt>var2</tt> is (statically or dynamically) negative or
2362equal to or larger than the number of bits in <tt>var1</tt>, the result is undefined.</p>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2363
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2364<h5>Example:</h5><pre>
2365 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
2366 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
2367 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2368 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2369</pre>
2370</div>
2371<!-- _______________________________________________________________________ -->
2372<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
2373Instruction</a> </div>
2374<div class="doc_text">
2375<h5>Syntax:</h5>
2376<pre> &lt;result&gt; = lshr &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2377</pre>
2378
2379<h5>Overview:</h5>
2380<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
2381operand shifted to the right a specified number of bits with zero fill.</p>
2382
2383<h5>Arguments:</h5>
2384<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]2385<a href="#t_integer">integer</a> type. '<tt>var2</tt>' is treated as an
2386unsigned value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2387
2388<h5>Semantics:</h5>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2389
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2390<p>This instruction always performs a logical shift right operation. The most
2391significant bits of the result will be filled with zero bits after the
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2392shift. If <tt>var2</tt> is (statically or dynamically) equal to or larger than
2393the number of bits in <tt>var1</tt>, the result is undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2394
2395<h5>Example:</h5>
2396<pre>
2397 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
2398 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
2399 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
2400 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2401 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2402</pre>
2403</div>
2404
2405<!-- _______________________________________________________________________ -->
2406<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
2407Instruction</a> </div>
2408<div class="doc_text">
2409
2410<h5>Syntax:</h5>
2411<pre> &lt;result&gt; = ashr &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2412</pre>
2413
2414<h5>Overview:</h5>
2415<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
2416operand shifted to the right a specified number of bits with sign extension.</p>
2417
2418<h5>Arguments:</h5>
2419<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]2420<a href="#t_integer">integer</a> type. '<tt>var2</tt>' is treated as an
2421unsigned value.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2422
2423<h5>Semantics:</h5>
2424<p>This instruction always performs an arithmetic shift right operation,
2425The most significant bits of the result will be filled with the sign bit
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2426of <tt>var1</tt>. If <tt>var2</tt> is (statically or dynamically) equal to or
2427larger than the number of bits in <tt>var1</tt>, the result is undefined.
2428</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2429
2430<h5>Example:</h5>
2431<pre>
2432 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
2433 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
2434 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
2435 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattnerd939d9f2007-10-03 21:01:14 +0000[diff] [blame]2436 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2437</pre>
2438</div>
2439
2440<!-- _______________________________________________________________________ -->
2441<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
2442Instruction</a> </div>
2443<div class="doc_text">
2444<h5>Syntax:</h5>
2445<pre> &lt;result&gt; = and &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2446</pre>
2447<h5>Overview:</h5>
2448<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
2449its two operands.</p>
2450<h5>Arguments:</h5>
2451<p>The two arguments to the '<tt>and</tt>' instruction must be <a
2452 href="#t_integer">integer</a> values. Both arguments must have
2453identical types.</p>
2454<h5>Semantics:</h5>
2455<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
2456<p> </p>
2457<div style="align: center">
2458<table border="1" cellspacing="0" cellpadding="4">
2459 <tbody>
2460 <tr>
2461 <td>In0</td>
2462 <td>In1</td>
2463 <td>Out</td>
2464 </tr>
2465 <tr>
2466 <td>0</td>
2467 <td>0</td>
2468 <td>0</td>
2469 </tr>
2470 <tr>
2471 <td>0</td>
2472 <td>1</td>
2473 <td>0</td>
2474 </tr>
2475 <tr>
2476 <td>1</td>
2477 <td>0</td>
2478 <td>0</td>
2479 </tr>
2480 <tr>
2481 <td>1</td>
2482 <td>1</td>
2483 <td>1</td>
2484 </tr>
2485 </tbody>
2486</table>
2487</div>
2488<h5>Example:</h5>
2489<pre> &lt;result&gt; = and i32 4, %var <i>; yields {i32}:result = 4 &amp; %var</i>
2490 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
2491 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
2492</pre>
2493</div>
2494<!-- _______________________________________________________________________ -->
2495<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
2496<div class="doc_text">
2497<h5>Syntax:</h5>
2498<pre> &lt;result&gt; = or &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2499</pre>
2500<h5>Overview:</h5>
2501<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
2502or of its two operands.</p>
2503<h5>Arguments:</h5>
2504<p>The two arguments to the '<tt>or</tt>' instruction must be <a
2505 href="#t_integer">integer</a> values. Both arguments must have
2506identical types.</p>
2507<h5>Semantics:</h5>
2508<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
2509<p> </p>
2510<div style="align: center">
2511<table border="1" cellspacing="0" cellpadding="4">
2512 <tbody>
2513 <tr>
2514 <td>In0</td>
2515 <td>In1</td>
2516 <td>Out</td>
2517 </tr>
2518 <tr>
2519 <td>0</td>
2520 <td>0</td>
2521 <td>0</td>
2522 </tr>
2523 <tr>
2524 <td>0</td>
2525 <td>1</td>
2526 <td>1</td>
2527 </tr>
2528 <tr>
2529 <td>1</td>
2530 <td>0</td>
2531 <td>1</td>
2532 </tr>
2533 <tr>
2534 <td>1</td>
2535 <td>1</td>
2536 <td>1</td>
2537 </tr>
2538 </tbody>
2539</table>
2540</div>
2541<h5>Example:</h5>
2542<pre> &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
2543 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
2544 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
2545</pre>
2546</div>
2547<!-- _______________________________________________________________________ -->
2548<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
2549Instruction</a> </div>
2550<div class="doc_text">
2551<h5>Syntax:</h5>
2552<pre> &lt;result&gt; = xor &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2553</pre>
2554<h5>Overview:</h5>
2555<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
2556or of its two operands. The <tt>xor</tt> is used to implement the
2557"one's complement" operation, which is the "~" operator in C.</p>
2558<h5>Arguments:</h5>
2559<p>The two arguments to the '<tt>xor</tt>' instruction must be <a
2560 href="#t_integer">integer</a> values. Both arguments must have
2561identical types.</p>
2562<h5>Semantics:</h5>
2563<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
2564<p> </p>
2565<div style="align: center">
2566<table border="1" cellspacing="0" cellpadding="4">
2567 <tbody>
2568 <tr>
2569 <td>In0</td>
2570 <td>In1</td>
2571 <td>Out</td>
2572 </tr>
2573 <tr>
2574 <td>0</td>
2575 <td>0</td>
2576 <td>0</td>
2577 </tr>
2578 <tr>
2579 <td>0</td>
2580 <td>1</td>
2581 <td>1</td>
2582 </tr>
2583 <tr>
2584 <td>1</td>
2585 <td>0</td>
2586 <td>1</td>
2587 </tr>
2588 <tr>
2589 <td>1</td>
2590 <td>1</td>
2591 <td>0</td>
2592 </tr>
2593 </tbody>
2594</table>
2595</div>
2596<p> </p>
2597<h5>Example:</h5>
2598<pre> &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
2599 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
2600 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
2601 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
2602</pre>
2603</div>
2604
2605<!-- ======================================================================= -->
2606<div class="doc_subsection">
2607 <a name="vectorops">Vector Operations</a>
2608</div>
2609
2610<div class="doc_text">
2611
2612<p>LLVM supports several instructions to represent vector operations in a
2613target-independent manner. These instructions cover the element-access and
2614vector-specific operations needed to process vectors effectively. While LLVM
2615does directly support these vector operations, many sophisticated algorithms
2616will want to use target-specific intrinsics to take full advantage of a specific
2617target.</p>
2618
2619</div>
2620
2621<!-- _______________________________________________________________________ -->
2622<div class="doc_subsubsection">
2623 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
2624</div>
2625
2626<div class="doc_text">
2627
2628<h5>Syntax:</h5>
2629
2630<pre>
2631 &lt;result&gt; = extractelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, i32 &lt;idx&gt; <i>; yields &lt;ty&gt;</i>
2632</pre>
2633
2634<h5>Overview:</h5>
2635
2636<p>
2637The '<tt>extractelement</tt>' instruction extracts a single scalar
2638element from a vector at a specified index.
2639</p>
2640
2641
2642<h5>Arguments:</h5>
2643
2644<p>
2645The first operand of an '<tt>extractelement</tt>' instruction is a
2646value of <a href="#t_vector">vector</a> type. The second operand is
2647an index indicating the position from which to extract the element.
2648The index may be a variable.</p>
2649
2650<h5>Semantics:</h5>
2651
2652<p>
2653The result is a scalar of the same type as the element type of
2654<tt>val</tt>. Its value is the value at position <tt>idx</tt> of
2655<tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
2656results are undefined.
2657</p>
2658
2659<h5>Example:</h5>
2660
2661<pre>
2662 %result = extractelement &lt;4 x i32&gt; %vec, i32 0 <i>; yields i32</i>
2663</pre>
2664</div>
2665
2666
2667<!-- _______________________________________________________________________ -->
2668<div class="doc_subsubsection">
2669 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
2670</div>
2671
2672<div class="doc_text">
2673
2674<h5>Syntax:</h5>
2675
2676<pre>
2677 &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>
2678</pre>
2679
2680<h5>Overview:</h5>
2681
2682<p>
2683The '<tt>insertelement</tt>' instruction inserts a scalar
2684element into a vector at a specified index.
2685</p>
2686
2687
2688<h5>Arguments:</h5>
2689
2690<p>
2691The first operand of an '<tt>insertelement</tt>' instruction is a
2692value of <a href="#t_vector">vector</a> type. The second operand is a
2693scalar value whose type must equal the element type of the first
2694operand. The third operand is an index indicating the position at
2695which to insert the value. The index may be a variable.</p>
2696
2697<h5>Semantics:</h5>
2698
2699<p>
2700The result is a vector of the same type as <tt>val</tt>. Its
2701element values are those of <tt>val</tt> except at position
2702<tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
2703exceeds the length of <tt>val</tt>, the results are undefined.
2704</p>
2705
2706<h5>Example:</h5>
2707
2708<pre>
2709 %result = insertelement &lt;4 x i32&gt; %vec, i32 1, i32 0 <i>; yields &lt;4 x i32&gt;</i>
2710</pre>
2711</div>
2712
2713<!-- _______________________________________________________________________ -->
2714<div class="doc_subsubsection">
2715 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
2716</div>
2717
2718<div class="doc_text">
2719
2720<h5>Syntax:</h5>
2721
2722<pre>
2723 &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>
2724</pre>
2725
2726<h5>Overview:</h5>
2727
2728<p>
2729The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
2730from two input vectors, returning a vector of the same type.
2731</p>
2732
2733<h5>Arguments:</h5>
2734
2735<p>
2736The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
2737with types that match each other and types that match the result of the
2738instruction. The third argument is a shuffle mask, which has the same number
2739of elements as the other vector type, but whose element type is always 'i32'.
2740</p>
2741
2742<p>
2743The shuffle mask operand is required to be a constant vector with either
2744constant integer or undef values.
2745</p>
2746
2747<h5>Semantics:</h5>
2748
2749<p>
2750The elements of the two input vectors are numbered from left to right across
2751both of the vectors. The shuffle mask operand specifies, for each element of
2752the result vector, which element of the two input registers the result element
2753gets. The element selector may be undef (meaning "don't care") and the second
2754operand may be undef if performing a shuffle from only one vector.
2755</p>
2756
2757<h5>Example:</h5>
2758
2759<pre>
2760 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
2761 &lt;4 x i32&gt; &lt;i32 0, i32 4, i32 1, i32 5&gt; <i>; yields &lt;4 x i32&gt;</i>
2762 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; undef,
2763 &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.
2764</pre>
2765</div>
2766
2767
2768<!-- ======================================================================= -->
2769<div class="doc_subsection">
2770 <a name="memoryops">Memory Access and Addressing Operations</a>
2771</div>
2772
2773<div class="doc_text">
2774
2775<p>A key design point of an SSA-based representation is how it
2776represents memory. In LLVM, no memory locations are in SSA form, which
2777makes things very simple. This section describes how to read, write,
2778allocate, and free memory in LLVM.</p>
2779
2780</div>
2781
2782<!-- _______________________________________________________________________ -->
2783<div class="doc_subsubsection">
2784 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
2785</div>
2786
2787<div class="doc_text">
2788
2789<h5>Syntax:</h5>
2790
2791<pre>
2792 &lt;result&gt; = malloc &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
2793</pre>
2794
2795<h5>Overview:</h5>
2796
2797<p>The '<tt>malloc</tt>' instruction allocates memory from the system
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]2798heap and returns a pointer to it. The object is always allocated in the generic
2799address space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2800
2801<h5>Arguments:</h5>
2802
2803<p>The '<tt>malloc</tt>' instruction allocates
2804<tt>sizeof(&lt;type&gt;)*NumElements</tt>
2805bytes of memory from the operating system and returns a pointer of the
2806appropriate type to the program. If "NumElements" is specified, it is the
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]2807number of elements allocated, otherwise "NumElements" is defaulted to be one.
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]2808If a constant alignment is specified, the value result of the allocation is guaranteed to
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]2809be aligned to at least that boundary. If not specified, or if zero, the target can
2810choose to align the allocation on any convenient boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2811
2812<p>'<tt>type</tt>' must be a sized type.</p>
2813
2814<h5>Semantics:</h5>
2815
2816<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]2817a pointer is returned. The result of a zero byte allocattion is undefined. The
2818result is null if there is insufficient memory available.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2819
2820<h5>Example:</h5>
2821
2822<pre>
2823 %array = malloc [4 x i8 ] <i>; yields {[%4 x i8]*}:array</i>
2824
2825 %size = <a href="#i_add">add</a> i32 2, 2 <i>; yields {i32}:size = i32 4</i>
2826 %array1 = malloc i8, i32 4 <i>; yields {i8*}:array1</i>
2827 %array2 = malloc [12 x i8], i32 %size <i>; yields {[12 x i8]*}:array2</i>
2828 %array3 = malloc i32, i32 4, align 1024 <i>; yields {i32*}:array3</i>
2829 %array4 = malloc i32, align 1024 <i>; yields {i32*}:array4</i>
2830</pre>
2831</div>
2832
2833<!-- _______________________________________________________________________ -->
2834<div class="doc_subsubsection">
2835 <a name="i_free">'<tt>free</tt>' Instruction</a>
2836</div>
2837
2838<div class="doc_text">
2839
2840<h5>Syntax:</h5>
2841
2842<pre>
2843 free &lt;type&gt; &lt;value&gt; <i>; yields {void}</i>
2844</pre>
2845
2846<h5>Overview:</h5>
2847
2848<p>The '<tt>free</tt>' instruction returns memory back to the unused
2849memory heap to be reallocated in the future.</p>
2850
2851<h5>Arguments:</h5>
2852
2853<p>'<tt>value</tt>' shall be a pointer value that points to a value
2854that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
2855instruction.</p>
2856
2857<h5>Semantics:</h5>
2858
2859<p>Access to the memory pointed to by the pointer is no longer defined
Chris Lattner329d6532008-04-19 22:41:32 +0000[diff] [blame]2860after this instruction executes. If the pointer is null, the operation
2861is a noop.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2862
2863<h5>Example:</h5>
2864
2865<pre>
2866 %array = <a href="#i_malloc">malloc</a> [4 x i8] <i>; yields {[4 x i8]*}:array</i>
2867 free [4 x i8]* %array
2868</pre>
2869</div>
2870
2871<!-- _______________________________________________________________________ -->
2872<div class="doc_subsubsection">
2873 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
2874</div>
2875
2876<div class="doc_text">
2877
2878<h5>Syntax:</h5>
2879
2880<pre>
2881 &lt;result&gt; = alloca &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
2882</pre>
2883
2884<h5>Overview:</h5>
2885
2886<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
2887currently executing function, to be automatically released when this function
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]2888returns to its caller. The object is always allocated in the generic address
2889space (address space zero).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2890
2891<h5>Arguments:</h5>
2892
2893<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
2894bytes of memory on the runtime stack, returning a pointer of the
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]2895appropriate type to the program. If "NumElements" is specified, it is the
2896number of elements allocated, otherwise "NumElements" is defaulted to be one.
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]2897If a constant alignment is specified, the value result of the allocation is guaranteed
Gabor Greif5082cf42008-02-09 22:24:34 +0000[diff] [blame]2898to be aligned to at least that boundary. If not specified, or if zero, the target
2899can choose to align the allocation on any convenient boundary.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2900
2901<p>'<tt>type</tt>' may be any sized type.</p>
2902
2903<h5>Semantics:</h5>
2904
Chris Lattner8b094fc2008-04-19 21:01:16 +0000[diff] [blame]2905<p>Memory is allocated; a pointer is returned. The operation is undefiend if
2906there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2907memory is automatically released when the function returns. The '<tt>alloca</tt>'
2908instruction is commonly used to represent automatic variables that must
2909have an address available. When the function returns (either with the <tt><a
2910 href="#i_ret">ret</a></tt> or <tt><a href="#i_unwind">unwind</a></tt>
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]2911instructions), the memory is reclaimed. Allocating zero bytes
2912is legal, but the result is undefined.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2913
2914<h5>Example:</h5>
2915
2916<pre>
2917 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
2918 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
2919 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
2920 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
2921</pre>
2922</div>
2923
2924<!-- _______________________________________________________________________ -->
2925<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
2926Instruction</a> </div>
2927<div class="doc_text">
2928<h5>Syntax:</h5>
2929<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>
2930<h5>Overview:</h5>
2931<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
2932<h5>Arguments:</h5>
2933<p>The argument to the '<tt>load</tt>' instruction specifies the memory
2934address from which to load. The pointer must point to a <a
2935 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
2936marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
2937the number or order of execution of this <tt>load</tt> with other
2938volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
2939instructions. </p>
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]2940<p>
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]2941The optional constant "align" argument specifies the alignment of the operation
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]2942(that is, the alignment of the memory address). A value of 0 or an
2943omitted "align" argument means that the operation has the preferential
2944alignment for the target. It is the responsibility of the code emitter
2945to ensure that the alignment information is correct. Overestimating
2946the alignment results in an undefined behavior. Underestimating the
2947alignment may produce less efficient code. An alignment of 1 is always
2948safe.
2949</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2950<h5>Semantics:</h5>
2951<p>The location of memory pointed to is loaded.</p>
2952<h5>Examples:</h5>
2953<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
2954 <a
2955 href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
2956 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
2957</pre>
2958</div>
2959<!-- _______________________________________________________________________ -->
2960<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
2961Instruction</a> </div>
2962<div class="doc_text">
2963<h5>Syntax:</h5>
2964<pre> store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
2965 volatile store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
2966</pre>
2967<h5>Overview:</h5>
2968<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
2969<h5>Arguments:</h5>
2970<p>There are two arguments to the '<tt>store</tt>' instruction: a value
2971to store and an address at which to store it. The type of the '<tt>&lt;pointer&gt;</tt>'
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]2972operand must be a pointer to the <a href="#t_firstclass">first class</a> type
2973of the '<tt>&lt;value&gt;</tt>'
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2974operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
2975optimizer is not allowed to modify the number or order of execution of
2976this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
2977 href="#i_store">store</a></tt> instructions.</p>
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]2978<p>
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]2979The optional constant "align" argument specifies the alignment of the operation
Chris Lattner21003c82008-01-06 21:04:43 +0000[diff] [blame]2980(that is, the alignment of the memory address). A value of 0 or an
2981omitted "align" argument means that the operation has the preferential
2982alignment for the target. It is the responsibility of the code emitter
2983to ensure that the alignment information is correct. Overestimating
2984the alignment results in an undefined behavior. Underestimating the
2985alignment may produce less efficient code. An alignment of 1 is always
2986safe.
2987</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2988<h5>Semantics:</h5>
2989<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>'
2990at the location specified by the '<tt>&lt;pointer&gt;</tt>' operand.</p>
2991<h5>Example:</h5>
2992<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling63ffa142007-10-22 05:10:05 +0000[diff] [blame]2993 store i32 3, i32* %ptr <i>; yields {void}</i>
2994 %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]2995</pre>
2996</div>
2997
2998<!-- _______________________________________________________________________ -->
2999<div class="doc_subsubsection">
3000 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
3001</div>
3002
3003<div class="doc_text">
3004<h5>Syntax:</h5>
3005<pre>
3006 &lt;result&gt; = getelementptr &lt;ty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
3007</pre>
3008
3009<h5>Overview:</h5>
3010
3011<p>
3012The '<tt>getelementptr</tt>' instruction is used to get the address of a
3013subelement of an aggregate data structure.</p>
3014
3015<h5>Arguments:</h5>
3016
3017<p>This instruction takes a list of integer operands that indicate what
3018elements of the aggregate object to index to. The actual types of the arguments
3019provided depend on the type of the first pointer argument. The
3020'<tt>getelementptr</tt>' instruction is used to index down through the type
3021levels of a structure or to a specific index in an array. When indexing into a
3022structure, only <tt>i32</tt> integer constants are allowed. When indexing
3023into an array or pointer, only integers of 32 or 64 bits are allowed, and will
3024be sign extended to 64-bit values.</p>
3025
3026<p>For example, let's consider a C code fragment and how it gets
3027compiled to LLVM:</p>
3028
3029<div class="doc_code">
3030<pre>
3031struct RT {
3032 char A;
3033 int B[10][20];
3034 char C;
3035};
3036struct ST {
3037 int X;
3038 double Y;
3039 struct RT Z;
3040};
3041
3042int *foo(struct ST *s) {
3043 return &amp;s[1].Z.B[5][13];
3044}
3045</pre>
3046</div>
3047
3048<p>The LLVM code generated by the GCC frontend is:</p>
3049
3050<div class="doc_code">
3051<pre>
3052%RT = type { i8 , [10 x [20 x i32]], i8 }
3053%ST = type { i32, double, %RT }
3054
3055define i32* %foo(%ST* %s) {
3056entry:
3057 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
3058 ret i32* %reg
3059}
3060</pre>
3061</div>
3062
3063<h5>Semantics:</h5>
3064
3065<p>The index types specified for the '<tt>getelementptr</tt>' instruction depend
3066on the pointer type that is being indexed into. <a href="#t_pointer">Pointer</a>
3067and <a href="#t_array">array</a> types can use a 32-bit or 64-bit
3068<a href="#t_integer">integer</a> type but the value will always be sign extended
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3069to 64-bits. <a href="#t_struct">Structure</a> and <a href="#t_pstruct">packed
3070structure</a> types require <tt>i32</tt> <b>constants</b>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3071
3072<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
3073type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
3074}</tt>' type, a structure. The second index indexes into the third element of
3075the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
3076i8 }</tt>' type, another structure. The third index indexes into the second
3077element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
3078array. The two dimensions of the array are subscripted into, yielding an
3079'<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
3080to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
3081
3082<p>Note that it is perfectly legal to index partially through a
3083structure, returning a pointer to an inner element. Because of this,
3084the LLVM code for the given testcase is equivalent to:</p>
3085
3086<pre>
3087 define i32* %foo(%ST* %s) {
3088 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
3089 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
3090 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
3091 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
3092 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
3093 ret i32* %t5
3094 }
3095</pre>
3096
3097<p>Note that it is undefined to access an array out of bounds: array and
3098pointer indexes must always be within the defined bounds of the array type.
3099The one exception for this rules is zero length arrays. These arrays are
3100defined to be accessible as variable length arrays, which requires access
3101beyond the zero'th element.</p>
3102
3103<p>The getelementptr instruction is often confusing. For some more insight
3104into how it works, see <a href="GetElementPtr.html">the getelementptr
3105FAQ</a>.</p>
3106
3107<h5>Example:</h5>
3108
3109<pre>
3110 <i>; yields [12 x i8]*:aptr</i>
3111 %aptr = getelementptr {i32, [12 x i8]}* %sptr, i64 0, i32 1
3112</pre>
3113</div>
3114
3115<!-- ======================================================================= -->
3116<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
3117</div>
3118<div class="doc_text">
3119<p>The instructions in this category are the conversion instructions (casting)
3120which all take a single operand and a type. They perform various bit conversions
3121on the operand.</p>
3122</div>
3123
3124<!-- _______________________________________________________________________ -->
3125<div class="doc_subsubsection">
3126 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
3127</div>
3128<div class="doc_text">
3129
3130<h5>Syntax:</h5>
3131<pre>
3132 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3133</pre>
3134
3135<h5>Overview:</h5>
3136<p>
3137The '<tt>trunc</tt>' instruction truncates its operand to the type <tt>ty2</tt>.
3138</p>
3139
3140<h5>Arguments:</h5>
3141<p>
3142The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
3143be an <a href="#t_integer">integer</a> type, and a type that specifies the size
3144and type of the result, which must be an <a href="#t_integer">integer</a>
3145type. The bit size of <tt>value</tt> must be larger than the bit size of
3146<tt>ty2</tt>. Equal sized types are not allowed.</p>
3147
3148<h5>Semantics:</h5>
3149<p>
3150The '<tt>trunc</tt>' instruction truncates the high order bits in <tt>value</tt>
3151and converts the remaining bits to <tt>ty2</tt>. Since the source size must be
3152larger than the destination size, <tt>trunc</tt> cannot be a <i>no-op cast</i>.
3153It will always truncate bits.</p>
3154
3155<h5>Example:</h5>
3156<pre>
3157 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
3158 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
3159 %Y = trunc i32 122 to i1 <i>; yields i1:false</i>
3160</pre>
3161</div>
3162
3163<!-- _______________________________________________________________________ -->
3164<div class="doc_subsubsection">
3165 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
3166</div>
3167<div class="doc_text">
3168
3169<h5>Syntax:</h5>
3170<pre>
3171 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3172</pre>
3173
3174<h5>Overview:</h5>
3175<p>The '<tt>zext</tt>' instruction zero extends its operand to type
3176<tt>ty2</tt>.</p>
3177
3178
3179<h5>Arguments:</h5>
3180<p>The '<tt>zext</tt>' instruction takes a value to cast, which must be of
3181<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3182also be of <a href="#t_integer">integer</a> type. The bit size of the
3183<tt>value</tt> must be smaller than the bit size of the destination type,
3184<tt>ty2</tt>.</p>
3185
3186<h5>Semantics:</h5>
3187<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
3188bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
3189
3190<p>When zero extending from i1, the result will always be either 0 or 1.</p>
3191
3192<h5>Example:</h5>
3193<pre>
3194 %X = zext i32 257 to i64 <i>; yields i64:257</i>
3195 %Y = zext i1 true to i32 <i>; yields i32:1</i>
3196</pre>
3197</div>
3198
3199<!-- _______________________________________________________________________ -->
3200<div class="doc_subsubsection">
3201 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
3202</div>
3203<div class="doc_text">
3204
3205<h5>Syntax:</h5>
3206<pre>
3207 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3208</pre>
3209
3210<h5>Overview:</h5>
3211<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
3212
3213<h5>Arguments:</h5>
3214<p>
3215The '<tt>sext</tt>' instruction takes a value to cast, which must be of
3216<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3217also be of <a href="#t_integer">integer</a> type. The bit size of the
3218<tt>value</tt> must be smaller than the bit size of the destination type,
3219<tt>ty2</tt>.</p>
3220
3221<h5>Semantics:</h5>
3222<p>
3223The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
3224bit (highest order bit) of the <tt>value</tt> until it reaches the bit size of
3225the type <tt>ty2</tt>.</p>
3226
3227<p>When sign extending from i1, the extension always results in -1 or 0.</p>
3228
3229<h5>Example:</h5>
3230<pre>
3231 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
3232 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
3233</pre>
3234</div>
3235
3236<!-- _______________________________________________________________________ -->
3237<div class="doc_subsubsection">
3238 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
3239</div>
3240
3241<div class="doc_text">
3242
3243<h5>Syntax:</h5>
3244
3245<pre>
3246 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3247</pre>
3248
3249<h5>Overview:</h5>
3250<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
3251<tt>ty2</tt>.</p>
3252
3253
3254<h5>Arguments:</h5>
3255<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
3256 point</a> value to cast and a <a href="#t_floating">floating point</a> type to
3257cast it to. The size of <tt>value</tt> must be larger than the size of
3258<tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
3259<i>no-op cast</i>.</p>
3260
3261<h5>Semantics:</h5>
3262<p> The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
3263<a href="#t_floating">floating point</a> type to a smaller
3264<a href="#t_floating">floating point</a> type. If the value cannot fit within
3265the destination type, <tt>ty2</tt>, then the results are undefined.</p>
3266
3267<h5>Example:</h5>
3268<pre>
3269 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
3270 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
3271</pre>
3272</div>
3273
3274<!-- _______________________________________________________________________ -->
3275<div class="doc_subsubsection">
3276 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
3277</div>
3278<div class="doc_text">
3279
3280<h5>Syntax:</h5>
3281<pre>
3282 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3283</pre>
3284
3285<h5>Overview:</h5>
3286<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
3287floating point value.</p>
3288
3289<h5>Arguments:</h5>
3290<p>The '<tt>fpext</tt>' instruction takes a
3291<a href="#t_floating">floating point</a> <tt>value</tt> to cast,
3292and a <a href="#t_floating">floating point</a> type to cast it to. The source
3293type must be smaller than the destination type.</p>
3294
3295<h5>Semantics:</h5>
3296<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
3297<a href="#t_floating">floating point</a> type to a larger
3298<a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
3299used to make a <i>no-op cast</i> because it always changes bits. Use
3300<tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
3301
3302<h5>Example:</h5>
3303<pre>
3304 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
3305 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
3306</pre>
3307</div>
3308
3309<!-- _______________________________________________________________________ -->
3310<div class="doc_subsubsection">
3311 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
3312</div>
3313<div class="doc_text">
3314
3315<h5>Syntax:</h5>
3316<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3317 &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]3318</pre>
3319
3320<h5>Overview:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3321<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3322unsigned integer equivalent of type <tt>ty2</tt>.
3323</p>
3324
3325<h5>Arguments:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3326<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]3327scalar or vector <a href="#t_floating">floating point</a> value, and a type
3328to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
3329type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
3330vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3331
3332<h5>Semantics:</h5>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3333<p> The '<tt>fptoui</tt>' instruction converts its
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3334<a href="#t_floating">floating point</a> operand into the nearest (rounding
3335towards zero) unsigned integer value. If the value cannot fit in <tt>ty2</tt>,
3336the results are undefined.</p>
3337
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3338<h5>Example:</h5>
3339<pre>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3340 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]3341 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Reid Spencere6adee82007-07-31 14:40:14 +0000[diff] [blame]3342 %X = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3343</pre>
3344</div>
3345
3346<!-- _______________________________________________________________________ -->
3347<div class="doc_subsubsection">
3348 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
3349</div>
3350<div class="doc_text">
3351
3352<h5>Syntax:</h5>
3353<pre>
3354 &lt;result&gt; = fptosi &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3355</pre>
3356
3357<h5>Overview:</h5>
3358<p>The '<tt>fptosi</tt>' instruction converts
3359<a href="#t_floating">floating point</a> <tt>value</tt> to type <tt>ty2</tt>.
3360</p>
3361
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3362<h5>Arguments:</h5>
3363<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]3364scalar or vector <a href="#t_floating">floating point</a> value, and a type
3365to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
3366type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
3367vector integer type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3368
3369<h5>Semantics:</h5>
3370<p>The '<tt>fptosi</tt>' instruction converts its
3371<a href="#t_floating">floating point</a> operand into the nearest (rounding
3372towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
3373the results are undefined.</p>
3374
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3375<h5>Example:</h5>
3376<pre>
3377 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner681f1e82007-09-22 03:17:52 +0000[diff] [blame]3378 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3379 %X = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
3380</pre>
3381</div>
3382
3383<!-- _______________________________________________________________________ -->
3384<div class="doc_subsubsection">
3385 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
3386</div>
3387<div class="doc_text">
3388
3389<h5>Syntax:</h5>
3390<pre>
3391 &lt;result&gt; = uitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3392</pre>
3393
3394<h5>Overview:</h5>
3395<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
3396integer and converts that value to the <tt>ty2</tt> type.</p>
3397
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3398<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]3399<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be a
3400scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
3401to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
3402type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
3403floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3404
3405<h5>Semantics:</h5>
3406<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
3407integer quantity and converts it to the corresponding floating point value. If
3408the value cannot fit in the floating point value, the results are undefined.</p>
3409
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3410<h5>Example:</h5>
3411<pre>
3412 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
3413 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
3414</pre>
3415</div>
3416
3417<!-- _______________________________________________________________________ -->
3418<div class="doc_subsubsection">
3419 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
3420</div>
3421<div class="doc_text">
3422
3423<h5>Syntax:</h5>
3424<pre>
3425 &lt;result&gt; = sitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3426</pre>
3427
3428<h5>Overview:</h5>
3429<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed
3430integer and converts that value to the <tt>ty2</tt> type.</p>
3431
3432<h5>Arguments:</h5>
Nate Begeman78246ca2007-11-17 03:58:34 +0000[diff] [blame]3433<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be a
3434scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
3435to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
3436type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
3437floating point type with the same number of elements as <tt>ty</tt></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3438
3439<h5>Semantics:</h5>
3440<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed
3441integer quantity and converts it to the corresponding floating point value. If
3442the value cannot fit in the floating point value, the results are undefined.</p>
3443
3444<h5>Example:</h5>
3445<pre>
3446 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
3447 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
3448</pre>
3449</div>
3450
3451<!-- _______________________________________________________________________ -->
3452<div class="doc_subsubsection">
3453 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
3454</div>
3455<div class="doc_text">
3456
3457<h5>Syntax:</h5>
3458<pre>
3459 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3460</pre>
3461
3462<h5>Overview:</h5>
3463<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
3464the integer type <tt>ty2</tt>.</p>
3465
3466<h5>Arguments:</h5>
3467<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
3468must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
3469<tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> type.
3470
3471<h5>Semantics:</h5>
3472<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
3473<tt>ty2</tt> by interpreting the pointer value as an integer and either
3474truncating or zero extending that value to the size of the integer type. If
3475<tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
3476<tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
3477are the same size, then nothing is done (<i>no-op cast</i>) other than a type
3478change.</p>
3479
3480<h5>Example:</h5>
3481<pre>
3482 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
3483 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
3484</pre>
3485</div>
3486
3487<!-- _______________________________________________________________________ -->
3488<div class="doc_subsubsection">
3489 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
3490</div>
3491<div class="doc_text">
3492
3493<h5>Syntax:</h5>
3494<pre>
3495 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3496</pre>
3497
3498<h5>Overview:</h5>
3499<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to
3500a pointer type, <tt>ty2</tt>.</p>
3501
3502<h5>Arguments:</h5>
3503<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
3504value to cast, and a type to cast it to, which must be a
3505<a href="#t_pointer">pointer</a> type.
3506
3507<h5>Semantics:</h5>
3508<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
3509<tt>ty2</tt> by applying either a zero extension or a truncation depending on
3510the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
3511size of a pointer then a truncation is done. If <tt>value</tt> is smaller than
3512the size of a pointer then a zero extension is done. If they are the same size,
3513nothing is done (<i>no-op cast</i>).</p>
3514
3515<h5>Example:</h5>
3516<pre>
3517 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
3518 %X = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
3519 %Y = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
3520</pre>
3521</div>
3522
3523<!-- _______________________________________________________________________ -->
3524<div class="doc_subsubsection">
3525 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
3526</div>
3527<div class="doc_text">
3528
3529<h5>Syntax:</h5>
3530<pre>
3531 &lt;result&gt; = bitcast &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3532</pre>
3533
3534<h5>Overview:</h5>
3535<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
3536<tt>ty2</tt> without changing any bits.</p>
3537
3538<h5>Arguments:</h5>
3539<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be
3540a first class value, and a type to cast it to, which must also be a <a
3541 href="#t_firstclass">first class</a> type. The bit sizes of <tt>value</tt>
3542and the destination type, <tt>ty2</tt>, must be identical. If the source
3543type is a pointer, the destination type must also be a pointer.</p>
3544
3545<h5>Semantics:</h5>
3546<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
3547<tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
3548this conversion. The conversion is done as if the <tt>value</tt> had been
3549stored to memory and read back as type <tt>ty2</tt>. Pointer types may only be
3550converted to other pointer types with this instruction. To convert pointers to
3551other types, use the <a href="#i_inttoptr">inttoptr</a> or
3552<a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
3553
3554<h5>Example:</h5>
3555<pre>
3556 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
3557 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
3558 %Z = bitcast <2xint> %V to i64; <i>; yields i64: %V</i>
3559</pre>
3560</div>
3561
3562<!-- ======================================================================= -->
3563<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
3564<div class="doc_text">
3565<p>The instructions in this category are the "miscellaneous"
3566instructions, which defy better classification.</p>
3567</div>
3568
3569<!-- _______________________________________________________________________ -->
3570<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
3571</div>
3572<div class="doc_text">
3573<h5>Syntax:</h5>
3574<pre> &lt;result&gt; = icmp &lt;cond&gt; &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {i1}:result</i>
3575</pre>
3576<h5>Overview:</h5>
3577<p>The '<tt>icmp</tt>' instruction returns a boolean value based on comparison
Chris Lattner10368b62008-04-02 00:38:26 +0000[diff] [blame]3578of its two integer or pointer operands.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3579<h5>Arguments:</h5>
3580<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
3581the condition code indicating the kind of comparison to perform. It is not
3582a value, just a keyword. The possible condition code are:
3583<ol>
3584 <li><tt>eq</tt>: equal</li>
3585 <li><tt>ne</tt>: not equal </li>
3586 <li><tt>ugt</tt>: unsigned greater than</li>
3587 <li><tt>uge</tt>: unsigned greater or equal</li>
3588 <li><tt>ult</tt>: unsigned less than</li>
3589 <li><tt>ule</tt>: unsigned less or equal</li>
3590 <li><tt>sgt</tt>: signed greater than</li>
3591 <li><tt>sge</tt>: signed greater or equal</li>
3592 <li><tt>slt</tt>: signed less than</li>
3593 <li><tt>sle</tt>: signed less or equal</li>
3594</ol>
3595<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
3596<a href="#t_pointer">pointer</a> typed. They must also be identical types.</p>
3597<h5>Semantics:</h5>
3598<p>The '<tt>icmp</tt>' compares <tt>var1</tt> and <tt>var2</tt> according to
3599the condition code given as <tt>cond</tt>. The comparison performed always
3600yields a <a href="#t_primitive">i1</a> result, as follows:
3601<ol>
3602 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
3603 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
3604 </li>
3605 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
3606 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
3607 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
3608 <tt>true</tt> if <tt>var1</tt> is greater than <tt>var2</tt>.</li>
3609 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
3610 <tt>true</tt> if <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3611 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
3612 <tt>true</tt> if <tt>var1</tt> is less than <tt>var2</tt>.</li>
3613 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
3614 <tt>true</tt> if <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
3615 <li><tt>sgt</tt>: interprets the operands as signed values and yields
3616 <tt>true</tt> if <tt>var1</tt> is greater than <tt>var2</tt>.</li>
3617 <li><tt>sge</tt>: interprets the operands as signed values and yields
3618 <tt>true</tt> if <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3619 <li><tt>slt</tt>: interprets the operands as signed values and yields
3620 <tt>true</tt> if <tt>var1</tt> is less than <tt>var2</tt>.</li>
3621 <li><tt>sle</tt>: interprets the operands as signed values and yields
3622 <tt>true</tt> if <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
3623</ol>
3624<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
3625values are compared as if they were integers.</p>
3626
3627<h5>Example:</h5>
3628<pre> &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
3629 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
3630 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
3631 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
3632 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
3633 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
3634</pre>
3635</div>
3636
3637<!-- _______________________________________________________________________ -->
3638<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
3639</div>
3640<div class="doc_text">
3641<h5>Syntax:</h5>
3642<pre> &lt;result&gt; = fcmp &lt;cond&gt; &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {i1}:result</i>
3643</pre>
3644<h5>Overview:</h5>
3645<p>The '<tt>fcmp</tt>' instruction returns a boolean value based on comparison
3646of its floating point operands.</p>
3647<h5>Arguments:</h5>
3648<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
3649the condition code indicating the kind of comparison to perform. It is not
3650a value, just a keyword. The possible condition code are:
3651<ol>
3652 <li><tt>false</tt>: no comparison, always returns false</li>
3653 <li><tt>oeq</tt>: ordered and equal</li>
3654 <li><tt>ogt</tt>: ordered and greater than </li>
3655 <li><tt>oge</tt>: ordered and greater than or equal</li>
3656 <li><tt>olt</tt>: ordered and less than </li>
3657 <li><tt>ole</tt>: ordered and less than or equal</li>
3658 <li><tt>one</tt>: ordered and not equal</li>
3659 <li><tt>ord</tt>: ordered (no nans)</li>
3660 <li><tt>ueq</tt>: unordered or equal</li>
3661 <li><tt>ugt</tt>: unordered or greater than </li>
3662 <li><tt>uge</tt>: unordered or greater than or equal</li>
3663 <li><tt>ult</tt>: unordered or less than </li>
3664 <li><tt>ule</tt>: unordered or less than or equal</li>
3665 <li><tt>une</tt>: unordered or not equal</li>
3666 <li><tt>uno</tt>: unordered (either nans)</li>
3667 <li><tt>true</tt>: no comparison, always returns true</li>
3668</ol>
3669<p><i>Ordered</i> means that neither operand is a QNAN while
3670<i>unordered</i> means that either operand may be a QNAN.</p>
3671<p>The <tt>val1</tt> and <tt>val2</tt> arguments must be
3672<a href="#t_floating">floating point</a> typed. They must have identical
3673types.</p>
3674<h5>Semantics:</h5>
3675<p>The '<tt>fcmp</tt>' compares <tt>var1</tt> and <tt>var2</tt> according to
3676the condition code given as <tt>cond</tt>. The comparison performed always
3677yields a <a href="#t_primitive">i1</a> result, as follows:
3678<ol>
3679 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
3680 <li><tt>oeq</tt>: yields <tt>true</tt> if both operands are not a QNAN and
3681 <tt>var1</tt> is equal to <tt>var2</tt>.</li>
3682 <li><tt>ogt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
3683 <tt>var1</tt> is greather than <tt>var2</tt>.</li>
3684 <li><tt>oge</tt>: yields <tt>true</tt> if both operands are not a QNAN and
3685 <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3686 <li><tt>olt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
3687 <tt>var1</tt> is less than <tt>var2</tt>.</li>
3688 <li><tt>ole</tt>: yields <tt>true</tt> if both operands are not a QNAN and
3689 <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
3690 <li><tt>one</tt>: yields <tt>true</tt> if both operands are not a QNAN and
3691 <tt>var1</tt> is not equal to <tt>var2</tt>.</li>
3692 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
3693 <li><tt>ueq</tt>: yields <tt>true</tt> if either operand is a QNAN or
3694 <tt>var1</tt> is equal to <tt>var2</tt>.</li>
3695 <li><tt>ugt</tt>: yields <tt>true</tt> if either operand is a QNAN or
3696 <tt>var1</tt> is greater than <tt>var2</tt>.</li>
3697 <li><tt>uge</tt>: yields <tt>true</tt> if either operand is a QNAN or
3698 <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3699 <li><tt>ult</tt>: yields <tt>true</tt> if either operand is a QNAN or
3700 <tt>var1</tt> is less than <tt>var2</tt>.</li>
3701 <li><tt>ule</tt>: yields <tt>true</tt> if either operand is a QNAN or
3702 <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
3703 <li><tt>une</tt>: yields <tt>true</tt> if either operand is a QNAN or
3704 <tt>var1</tt> is not equal to <tt>var2</tt>.</li>
3705 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
3706 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
3707</ol>
3708
3709<h5>Example:</h5>
3710<pre> &lt;result&gt; = fcmp oeq float 4.0, 5.0 <i>; yields: result=false</i>
3711 &lt;result&gt; = icmp one float 4.0, 5.0 <i>; yields: result=true</i>
3712 &lt;result&gt; = icmp olt float 4.0, 5.0 <i>; yields: result=true</i>
3713 &lt;result&gt; = icmp ueq double 1.0, 2.0 <i>; yields: result=false</i>
3714</pre>
3715</div>
3716
3717<!-- _______________________________________________________________________ -->
3718<div class="doc_subsubsection"> <a name="i_phi">'<tt>phi</tt>'
3719Instruction</a> </div>
3720<div class="doc_text">
3721<h5>Syntax:</h5>
3722<pre> &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...<br></pre>
3723<h5>Overview:</h5>
3724<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in
3725the SSA graph representing the function.</p>
3726<h5>Arguments:</h5>
3727<p>The type of the incoming values is specified with the first type
3728field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
3729as arguments, with one pair for each predecessor basic block of the
3730current block. Only values of <a href="#t_firstclass">first class</a>
3731type may be used as the value arguments to the PHI node. Only labels
3732may be used as the label arguments.</p>
3733<p>There must be no non-phi instructions between the start of a basic
3734block and the PHI instructions: i.e. PHI instructions must be first in
3735a basic block.</p>
3736<h5>Semantics:</h5>
3737<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
3738specified by the pair corresponding to the predecessor basic block that executed
3739just prior to the current block.</p>
3740<h5>Example:</h5>
3741<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>
3742</div>
3743
3744<!-- _______________________________________________________________________ -->
3745<div class="doc_subsubsection">
3746 <a name="i_select">'<tt>select</tt>' Instruction</a>
3747</div>
3748
3749<div class="doc_text">
3750
3751<h5>Syntax:</h5>
3752
3753<pre>
3754 &lt;result&gt; = select i1 &lt;cond&gt;, &lt;ty&gt; &lt;val1&gt;, &lt;ty&gt; &lt;val2&gt; <i>; yields ty</i>
3755</pre>
3756
3757<h5>Overview:</h5>
3758
3759<p>
3760The '<tt>select</tt>' instruction is used to choose one value based on a
3761condition, without branching.
3762</p>
3763
3764
3765<h5>Arguments:</h5>
3766
3767<p>
3768The '<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.
3769</p>
3770
3771<h5>Semantics:</h5>
3772
3773<p>
3774If the boolean condition evaluates to true, the instruction returns the first
3775value argument; otherwise, it returns the second value argument.
3776</p>
3777
3778<h5>Example:</h5>
3779
3780<pre>
3781 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
3782</pre>
3783</div>
3784
3785
3786<!-- _______________________________________________________________________ -->
3787<div class="doc_subsubsection">
3788 <a name="i_call">'<tt>call</tt>' Instruction</a>
3789</div>
3790
3791<div class="doc_text">
3792
3793<h5>Syntax:</h5>
3794<pre>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]3795 &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]3796</pre>
3797
3798<h5>Overview:</h5>
3799
3800<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
3801
3802<h5>Arguments:</h5>
3803
3804<p>This instruction requires several arguments:</p>
3805
3806<ol>
3807 <li>
3808 <p>The optional "tail" marker indicates whether the callee function accesses
3809 any allocas or varargs in the caller. If the "tail" marker is present, the
3810 function call is eligible for tail call optimization. Note that calls may
3811 be marked "tail" even if they do not occur before a <a
3812 href="#i_ret"><tt>ret</tt></a> instruction.
3813 </li>
3814 <li>
3815 <p>The optional "cconv" marker indicates which <a href="#callingconv">calling
3816 convention</a> the call should use. If none is specified, the call defaults
3817 to using C calling conventions.
3818 </li>
3819 <li>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]3820 <p>'<tt>ty</tt>': the type of the call instruction itself which is also
3821 the type of the return value. Functions that return no value are marked
3822 <tt><a href="#t_void">void</a></tt>.</p>
3823 </li>
3824 <li>
3825 <p>'<tt>fnty</tt>': shall be the signature of the pointer to function
3826 value being invoked. The argument types must match the types implied by
3827 this signature. This type can be omitted if the function is not varargs
3828 and if the function type does not return a pointer to a function.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3829 </li>
3830 <li>
3831 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
3832 be invoked. In most cases, this is a direct function invocation, but
3833 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
3834 to function value.</p>
3835 </li>
3836 <li>
3837 <p>'<tt>function args</tt>': argument list whose types match the
3838 function signature argument types. All arguments must be of
3839 <a href="#t_firstclass">first class</a> type. If the function signature
3840 indicates the function accepts a variable number of arguments, the extra
3841 arguments can be specified.</p>
3842 </li>
3843</ol>
3844
3845<h5>Semantics:</h5>
3846
3847<p>The '<tt>call</tt>' instruction is used to cause control flow to
3848transfer to a specified function, with its incoming arguments bound to
3849the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
3850instruction in the called function, control flow continues with the
3851instruction after the function call, and the return value of the
Chris Lattner5e893ef2008-03-21 17:24:17 +0000[diff] [blame]3852function is bound to the result argument. If the callee returns multiple
3853values then the return values of the function are only accessible through
3854the '<tt><a href="#i_getresult">getresult</a></tt>' instruction.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3855
3856<h5>Example:</h5>
3857
3858<pre>
Nick Lewycky93082fc2007-09-08 13:57:50 +0000[diff] [blame]3859 %retval = call i32 @test(i32 %argc)
Chris Lattner5e893ef2008-03-21 17:24:17 +0000[diff] [blame]3860 call i32 (i8 *, ...)* @printf(i8 * %msg, i32 12, i8 42) <i>; yields i32</i>
3861 %X = tail call i32 @foo() <i>; yields i32</i>
3862 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
3863 call void %foo(i8 97 signext)
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]3864
3865 %struct.A = type { i32, i8 }
Chris Lattner5e893ef2008-03-21 17:24:17 +0000[diff] [blame]3866 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
3867 %gr = getresult %struct.A %r, 0 <i>; yields i32</i>
3868 %gr1 = getresult %struct.A %r, 1 <i>; yields i8</i>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3869</pre>
3870
3871</div>
3872
3873<!-- _______________________________________________________________________ -->
3874<div class="doc_subsubsection">
3875 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
3876</div>
3877
3878<div class="doc_text">
3879
3880<h5>Syntax:</h5>
3881
3882<pre>
3883 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
3884</pre>
3885
3886<h5>Overview:</h5>
3887
3888<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
3889the "variable argument" area of a function call. It is used to implement the
3890<tt>va_arg</tt> macro in C.</p>
3891
3892<h5>Arguments:</h5>
3893
3894<p>This instruction takes a <tt>va_list*</tt> value and the type of
3895the argument. It returns a value of the specified argument type and
3896increments the <tt>va_list</tt> to point to the next argument. The
3897actual type of <tt>va_list</tt> is target specific.</p>
3898
3899<h5>Semantics:</h5>
3900
3901<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
3902type from the specified <tt>va_list</tt> and causes the
3903<tt>va_list</tt> to point to the next argument. For more information,
3904see the variable argument handling <a href="#int_varargs">Intrinsic
3905Functions</a>.</p>
3906
3907<p>It is legal for this instruction to be called in a function which does not
3908take a variable number of arguments, for example, the <tt>vfprintf</tt>
3909function.</p>
3910
3911<p><tt>va_arg</tt> is an LLVM instruction instead of an <a
3912href="#intrinsics">intrinsic function</a> because it takes a type as an
3913argument.</p>
3914
3915<h5>Example:</h5>
3916
3917<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
3918
3919</div>
3920
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]3921<!-- _______________________________________________________________________ -->
3922<div class="doc_subsubsection">
3923 <a name="i_getresult">'<tt>getresult</tt>' Instruction</a>
3924</div>
3925
3926<div class="doc_text">
3927
3928<h5>Syntax:</h5>
3929<pre>
Chris Lattneree9da3f2008-03-21 17:20:51 +0000[diff] [blame]3930 &lt;resultval&gt; = getresult &lt;type&gt; &lt;retval&gt;, &lt;index&gt;
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]3931</pre>
Chris Lattneree9da3f2008-03-21 17:20:51 +0000[diff] [blame]3932
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]3933<h5>Overview:</h5>
3934
3935<p> The '<tt>getresult</tt>' instruction is used to extract individual values
Chris Lattneree9da3f2008-03-21 17:20:51 +0000[diff] [blame]3936from a '<tt><a href="#i_call">call</a></tt>'
3937or '<tt><a href="#i_invoke">invoke</a></tt>' instruction that returns multiple
3938results.</p>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]3939
3940<h5>Arguments:</h5>
3941
Chris Lattneree9da3f2008-03-21 17:20:51 +0000[diff] [blame]3942<p>The '<tt>getresult</tt>' instruction takes a call or invoke value as its
Chris Lattnerd8dd3522008-04-23 04:06:52 +0000[diff] [blame]3943first argument, or an undef value. The value must have <a
3944href="#t_struct">structure type</a>. The second argument is a constant
3945unsigned index value which must be in range for the number of values returned
3946by the call.</p>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]3947
3948<h5>Semantics:</h5>
3949
Chris Lattneree9da3f2008-03-21 17:20:51 +0000[diff] [blame]3950<p>The '<tt>getresult</tt>' instruction extracts the element identified by
3951'<tt>index</tt>' from the aggregate value.</p>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]3952
3953<h5>Example:</h5>
3954
3955<pre>
3956 %struct.A = type { i32, i8 }
3957
3958 %r = call %struct.A @foo()
Chris Lattneree9da3f2008-03-21 17:20:51 +0000[diff] [blame]3959 %gr = getresult %struct.A %r, 0 <i>; yields i32:%gr</i>
3960 %gr1 = getresult %struct.A %r, 1 <i>; yields i8:%gr1</i>
Devang Patela3cc5372008-03-10 20:49:15 +0000[diff] [blame]3961 add i32 %gr, 42
3962 add i8 %gr1, 41
3963</pre>
3964
3965</div>
3966
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3967<!-- *********************************************************************** -->
3968<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
3969<!-- *********************************************************************** -->
3970
3971<div class="doc_text">
3972
3973<p>LLVM supports the notion of an "intrinsic function". These functions have
3974well known names and semantics and are required to follow certain restrictions.
3975Overall, these intrinsics represent an extension mechanism for the LLVM
3976language that does not require changing all of the transformations in LLVM when
3977adding to the language (or the bitcode reader/writer, the parser, etc...).</p>
3978
3979<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
3980prefix is reserved in LLVM for intrinsic names; thus, function names may not
3981begin with this prefix. Intrinsic functions must always be external functions:
3982you cannot define the body of intrinsic functions. Intrinsic functions may
3983only be used in call or invoke instructions: it is illegal to take the address
3984of an intrinsic function. Additionally, because intrinsic functions are part
3985of the LLVM language, it is required if any are added that they be documented
3986here.</p>
3987
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]3988<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents
3989a family of functions that perform the same operation but on different data
3990types. Because LLVM can represent over 8 million different integer types,
3991overloading is used commonly to allow an intrinsic function to operate on any
3992integer type. One or more of the argument types or the result type can be
3993overloaded to accept any integer type. Argument types may also be defined as
3994exactly matching a previous argument's type or the result type. This allows an
3995intrinsic function which accepts multiple arguments, but needs all of them to
3996be of the same type, to only be overloaded with respect to a single argument or
3997the result.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]3998
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]3999<p>Overloaded intrinsics will have the names of its overloaded argument types
4000encoded into its function name, each preceded by a period. Only those types
4001which are overloaded result in a name suffix. Arguments whose type is matched
4002against another type do not. For example, the <tt>llvm.ctpop</tt> function can
4003take an integer of any width and returns an integer of exactly the same integer
4004width. This leads to a family of functions such as
4005<tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29 %val)</tt>.
4006Only one type, the return type, is overloaded, and only one type suffix is
4007required. Because the argument's type is matched against the return type, it
4008does not require its own name suffix.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4009
4010<p>To learn how to add an intrinsic function, please see the
4011<a href="ExtendingLLVM.html">Extending LLVM Guide</a>.
4012</p>
4013
4014</div>
4015
4016<!-- ======================================================================= -->
4017<div class="doc_subsection">
4018 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
4019</div>
4020
4021<div class="doc_text">
4022
4023<p>Variable argument support is defined in LLVM with the <a
4024 href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
4025intrinsic functions. These functions are related to the similarly
4026named macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
4027
4028<p>All of these functions operate on arguments that use a
4029target-specific value type "<tt>va_list</tt>". The LLVM assembly
4030language reference manual does not define what this type is, so all
4031transformations should be prepared to handle these functions regardless of
4032the type used.</p>
4033
4034<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
4035instruction and the variable argument handling intrinsic functions are
4036used.</p>
4037
4038<div class="doc_code">
4039<pre>
4040define i32 @test(i32 %X, ...) {
4041 ; Initialize variable argument processing
4042 %ap = alloca i8*
4043 %ap2 = bitcast i8** %ap to i8*
4044 call void @llvm.va_start(i8* %ap2)
4045
4046 ; Read a single integer argument
4047 %tmp = va_arg i8** %ap, i32
4048
4049 ; Demonstrate usage of llvm.va_copy and llvm.va_end
4050 %aq = alloca i8*
4051 %aq2 = bitcast i8** %aq to i8*
4052 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
4053 call void @llvm.va_end(i8* %aq2)
4054
4055 ; Stop processing of arguments.
4056 call void @llvm.va_end(i8* %ap2)
4057 ret i32 %tmp
4058}
4059
4060declare void @llvm.va_start(i8*)
4061declare void @llvm.va_copy(i8*, i8*)
4062declare void @llvm.va_end(i8*)
4063</pre>
4064</div>
4065
4066</div>
4067
4068<!-- _______________________________________________________________________ -->
4069<div class="doc_subsubsection">
4070 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
4071</div>
4072
4073
4074<div class="doc_text">
4075<h5>Syntax:</h5>
4076<pre> declare void %llvm.va_start(i8* &lt;arglist&gt;)<br></pre>
4077<h5>Overview:</h5>
4078<P>The '<tt>llvm.va_start</tt>' intrinsic initializes
4079<tt>*&lt;arglist&gt;</tt> for subsequent use by <tt><a
4080href="#i_va_arg">va_arg</a></tt>.</p>
4081
4082<h5>Arguments:</h5>
4083
4084<P>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
4085
4086<h5>Semantics:</h5>
4087
4088<P>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
4089macro available in C. In a target-dependent way, it initializes the
4090<tt>va_list</tt> element to which the argument points, so that the next call to
4091<tt>va_arg</tt> will produce the first variable argument passed to the function.
4092Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
4093last argument of the function as the compiler can figure that out.</p>
4094
4095</div>
4096
4097<!-- _______________________________________________________________________ -->
4098<div class="doc_subsubsection">
4099 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
4100</div>
4101
4102<div class="doc_text">
4103<h5>Syntax:</h5>
4104<pre> declare void @llvm.va_end(i8* &lt;arglist&gt;)<br></pre>
4105<h5>Overview:</h5>
4106
4107<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
4108which has been initialized previously with <tt><a href="#int_va_start">llvm.va_start</a></tt>
4109or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
4110
4111<h5>Arguments:</h5>
4112
4113<p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
4114
4115<h5>Semantics:</h5>
4116
4117<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
4118macro available in C. In a target-dependent way, it destroys the
4119<tt>va_list</tt> element to which the argument points. Calls to <a
4120href="#int_va_start"><tt>llvm.va_start</tt></a> and <a href="#int_va_copy">
4121<tt>llvm.va_copy</tt></a> must be matched exactly with calls to
4122<tt>llvm.va_end</tt>.</p>
4123
4124</div>
4125
4126<!-- _______________________________________________________________________ -->
4127<div class="doc_subsubsection">
4128 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
4129</div>
4130
4131<div class="doc_text">
4132
4133<h5>Syntax:</h5>
4134
4135<pre>
4136 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
4137</pre>
4138
4139<h5>Overview:</h5>
4140
4141<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
4142from the source argument list to the destination argument list.</p>
4143
4144<h5>Arguments:</h5>
4145
4146<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
4147The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
4148
4149
4150<h5>Semantics:</h5>
4151
4152<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
4153macro available in C. In a target-dependent way, it copies the source
4154<tt>va_list</tt> element into the destination <tt>va_list</tt> element. This
4155intrinsic is necessary because the <tt><a href="#int_va_start">
4156llvm.va_start</a></tt> intrinsic may be arbitrarily complex and require, for
4157example, memory allocation.</p>
4158
4159</div>
4160
4161<!-- ======================================================================= -->
4162<div class="doc_subsection">
4163 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
4164</div>
4165
4166<div class="doc_text">
4167
4168<p>
4169LLVM support for <a href="GarbageCollection.html">Accurate Garbage
4170Collection</a> requires the implementation and generation of these intrinsics.
4171These intrinsics allow identification of <a href="#int_gcroot">GC roots on the
4172stack</a>, as well as garbage collector implementations that require <a
4173href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a> barriers.
4174Front-ends for type-safe garbage collected languages should generate these
4175intrinsics to make use of the LLVM garbage collectors. For more details, see <a
4176href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
4177</p>
Christopher Lambcfe00962007-12-17 01:00:21 +0000[diff] [blame]4178
4179<p>The garbage collection intrinsics only operate on objects in the generic
4180 address space (address space zero).</p>
4181
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4182</div>
4183
4184<!-- _______________________________________________________________________ -->
4185<div class="doc_subsubsection">
4186 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
4187</div>
4188
4189<div class="doc_text">
4190
4191<h5>Syntax:</h5>
4192
4193<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4194 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4195</pre>
4196
4197<h5>Overview:</h5>
4198
4199<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
4200the code generator, and allows some metadata to be associated with it.</p>
4201
4202<h5>Arguments:</h5>
4203
4204<p>The first argument specifies the address of a stack object that contains the
4205root pointer. The second pointer (which must be either a constant or a global
4206value address) contains the meta-data to be associated with the root.</p>
4207
4208<h5>Semantics:</h5>
4209
4210<p>At runtime, a call to this intrinsics stores a null pointer into the "ptrloc"
4211location. At compile-time, the code generator generates information to allow
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4212the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
4213intrinsic may only be used in a function which <a href="#gc">specifies a GC
4214algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4215
4216</div>
4217
4218
4219<!-- _______________________________________________________________________ -->
4220<div class="doc_subsubsection">
4221 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
4222</div>
4223
4224<div class="doc_text">
4225
4226<h5>Syntax:</h5>
4227
4228<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4229 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4230</pre>
4231
4232<h5>Overview:</h5>
4233
4234<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
4235locations, allowing garbage collector implementations that require read
4236barriers.</p>
4237
4238<h5>Arguments:</h5>
4239
4240<p>The second argument is the address to read from, which should be an address
4241allocated from the garbage collector. The first object is a pointer to the
4242start of the referenced object, if needed by the language runtime (otherwise
4243null).</p>
4244
4245<h5>Semantics:</h5>
4246
4247<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
4248instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4249garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
4250may only be used in a function which <a href="#gc">specifies a GC
4251algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4252
4253</div>
4254
4255
4256<!-- _______________________________________________________________________ -->
4257<div class="doc_subsubsection">
4258 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
4259</div>
4260
4261<div class="doc_text">
4262
4263<h5>Syntax:</h5>
4264
4265<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4266 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4267</pre>
4268
4269<h5>Overview:</h5>
4270
4271<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
4272locations, allowing garbage collector implementations that require write
4273barriers (such as generational or reference counting collectors).</p>
4274
4275<h5>Arguments:</h5>
4276
4277<p>The first argument is the reference to store, the second is the start of the
4278object to store it to, and the third is the address of the field of Obj to
4279store to. If the runtime does not require a pointer to the object, Obj may be
4280null.</p>
4281
4282<h5>Semantics:</h5>
4283
4284<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
4285instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen40393542007-12-25 02:31:26 +0000[diff] [blame]4286garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
4287may only be used in a function which <a href="#gc">specifies a GC
4288algorithm</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4289
4290</div>
4291
4292
4293
4294<!-- ======================================================================= -->
4295<div class="doc_subsection">
4296 <a name="int_codegen">Code Generator Intrinsics</a>
4297</div>
4298
4299<div class="doc_text">
4300<p>
4301These intrinsics are provided by LLVM to expose special features that may only
4302be implemented with code generator support.
4303</p>
4304
4305</div>
4306
4307<!-- _______________________________________________________________________ -->
4308<div class="doc_subsubsection">
4309 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
4310</div>
4311
4312<div class="doc_text">
4313
4314<h5>Syntax:</h5>
4315<pre>
4316 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
4317</pre>
4318
4319<h5>Overview:</h5>
4320
4321<p>
4322The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
4323target-specific value indicating the return address of the current function
4324or one of its callers.
4325</p>
4326
4327<h5>Arguments:</h5>
4328
4329<p>
4330The argument to this intrinsic indicates which function to return the address
4331for. Zero indicates the calling function, one indicates its caller, etc. The
4332argument is <b>required</b> to be a constant integer value.
4333</p>
4334
4335<h5>Semantics:</h5>
4336
4337<p>
4338The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
4339the return address of the specified call frame, or zero if it cannot be
4340identified. The value returned by this intrinsic is likely to be incorrect or 0
4341for arguments other than zero, so it should only be used for debugging purposes.
4342</p>
4343
4344<p>
4345Note that calling this intrinsic does not prevent function inlining or other
4346aggressive transformations, so the value returned may not be that of the obvious
4347source-language caller.
4348</p>
4349</div>
4350
4351
4352<!-- _______________________________________________________________________ -->
4353<div class="doc_subsubsection">
4354 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
4355</div>
4356
4357<div class="doc_text">
4358
4359<h5>Syntax:</h5>
4360<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4361 declare i8 *@llvm.frameaddress(i32 &lt;level&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4362</pre>
4363
4364<h5>Overview:</h5>
4365
4366<p>
4367The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
4368target-specific frame pointer value for the specified stack frame.
4369</p>
4370
4371<h5>Arguments:</h5>
4372
4373<p>
4374The argument to this intrinsic indicates which function to return the frame
4375pointer for. Zero indicates the calling function, one indicates its caller,
4376etc. The argument is <b>required</b> to be a constant integer value.
4377</p>
4378
4379<h5>Semantics:</h5>
4380
4381<p>
4382The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
4383the frame address of the specified call frame, or zero if it cannot be
4384identified. The value returned by this intrinsic is likely to be incorrect or 0
4385for arguments other than zero, so it should only be used for debugging purposes.
4386</p>
4387
4388<p>
4389Note that calling this intrinsic does not prevent function inlining or other
4390aggressive transformations, so the value returned may not be that of the obvious
4391source-language caller.
4392</p>
4393</div>
4394
4395<!-- _______________________________________________________________________ -->
4396<div class="doc_subsubsection">
4397 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
4398</div>
4399
4400<div class="doc_text">
4401
4402<h5>Syntax:</h5>
4403<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4404 declare i8 *@llvm.stacksave()
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4405</pre>
4406
4407<h5>Overview:</h5>
4408
4409<p>
4410The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
4411the function stack, for use with <a href="#int_stackrestore">
4412<tt>llvm.stackrestore</tt></a>. This is useful for implementing language
4413features like scoped automatic variable sized arrays in C99.
4414</p>
4415
4416<h5>Semantics:</h5>
4417
4418<p>
4419This intrinsic returns a opaque pointer value that can be passed to <a
4420href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
4421<tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
4422<tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
4423state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
4424practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
4425that were allocated after the <tt>llvm.stacksave</tt> was executed.
4426</p>
4427
4428</div>
4429
4430<!-- _______________________________________________________________________ -->
4431<div class="doc_subsubsection">
4432 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
4433</div>
4434
4435<div class="doc_text">
4436
4437<h5>Syntax:</h5>
4438<pre>
4439 declare void @llvm.stackrestore(i8 * %ptr)
4440</pre>
4441
4442<h5>Overview:</h5>
4443
4444<p>
4445The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
4446the function stack to the state it was in when the corresponding <a
4447href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
4448useful for implementing language features like scoped automatic variable sized
4449arrays in C99.
4450</p>
4451
4452<h5>Semantics:</h5>
4453
4454<p>
4455See the description for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.
4456</p>
4457
4458</div>
4459
4460
4461<!-- _______________________________________________________________________ -->
4462<div class="doc_subsubsection">
4463 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
4464</div>
4465
4466<div class="doc_text">
4467
4468<h5>Syntax:</h5>
4469<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4470 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]4471</pre>
4472
4473<h5>Overview:</h5>
4474
4475
4476<p>
4477The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
4478a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
4479no
4480effect on the behavior of the program but can change its performance
4481characteristics.
4482</p>
4483
4484<h5>Arguments:</h5>
4485
4486<p>
4487<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
4488determining if the fetch should be for a read (0) or write (1), and
4489<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
4490locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
4491<tt>locality</tt> arguments must be constant integers.
4492</p>
4493
4494<h5>Semantics:</h5>
4495
4496<p>
4497This intrinsic does not modify the behavior of the program. In particular,
4498prefetches cannot trap and do not produce a value. On targets that support this
4499intrinsic, the prefetch can provide hints to the processor cache for better
4500performance.
4501</p>
4502
4503</div>
4504
4505<!-- _______________________________________________________________________ -->
4506<div class="doc_subsubsection">
4507 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
4508</div>
4509
4510<div class="doc_text">
4511
4512<h5>Syntax:</h5>
4513<pre>
Chris Lattner38bd5dd2007-09-21 17:30:40 +0000[diff] [blame]4514 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4515</pre>
4516
4517<h5>Overview:</h5>
4518
4519
4520<p>
4521The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
4522(PC) in a region of
4523code to simulators and other tools. The method is target specific, but it is
4524expected that the marker will use exported symbols to transmit the PC of the marker.
4525The marker makes no guarantees that it will remain with any specific instruction
4526after optimizations. It is possible that the presence of a marker will inhibit
4527optimizations. The intended use is to be inserted after optimizations to allow
4528correlations of simulation runs.
4529</p>
4530
4531<h5>Arguments:</h5>
4532
4533<p>
4534<tt>id</tt> is a numerical id identifying the marker.
4535</p>
4536
4537<h5>Semantics:</h5>
4538
4539<p>
4540This intrinsic does not modify the behavior of the program. Backends that do not
4541support this intrinisic may ignore it.
4542</p>
4543
4544</div>
4545
4546<!-- _______________________________________________________________________ -->
4547<div class="doc_subsubsection">
4548 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
4549</div>
4550
4551<div class="doc_text">
4552
4553<h5>Syntax:</h5>
4554<pre>
4555 declare i64 @llvm.readcyclecounter( )
4556</pre>
4557
4558<h5>Overview:</h5>
4559
4560
4561<p>
4562The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
4563counter register (or similar low latency, high accuracy clocks) on those targets
4564that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
4565As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
4566should only be used for small timings.
4567</p>
4568
4569<h5>Semantics:</h5>
4570
4571<p>
4572When directly supported, reading the cycle counter should not modify any memory.
4573Implementations are allowed to either return a application specific value or a
4574system wide value. On backends without support, this is lowered to a constant 0.
4575</p>
4576
4577</div>
4578
4579<!-- ======================================================================= -->
4580<div class="doc_subsection">
4581 <a name="int_libc">Standard C Library Intrinsics</a>
4582</div>
4583
4584<div class="doc_text">
4585<p>
4586LLVM provides intrinsics for a few important standard C library functions.
4587These intrinsics allow source-language front-ends to pass information about the
4588alignment of the pointer arguments to the code generator, providing opportunity
4589for more efficient code generation.
4590</p>
4591
4592</div>
4593
4594<!-- _______________________________________________________________________ -->
4595<div class="doc_subsubsection">
4596 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
4597</div>
4598
4599<div class="doc_text">
4600
4601<h5>Syntax:</h5>
4602<pre>
4603 declare void @llvm.memcpy.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
4604 i32 &lt;len&gt;, i32 &lt;align&gt;)
4605 declare void @llvm.memcpy.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
4606 i64 &lt;len&gt;, i32 &lt;align&gt;)
4607</pre>
4608
4609<h5>Overview:</h5>
4610
4611<p>
4612The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
4613location to the destination location.
4614</p>
4615
4616<p>
4617Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
4618intrinsics do not return a value, and takes an extra alignment argument.
4619</p>
4620
4621<h5>Arguments:</h5>
4622
4623<p>
4624The first argument is a pointer to the destination, the second is a pointer to
4625the source. The third argument is an integer argument
4626specifying the number of bytes to copy, and the fourth argument is the alignment
4627of the source and destination locations.
4628</p>
4629
4630<p>
4631If the call to this intrinisic has an alignment value that is not 0 or 1, then
4632the caller guarantees that both the source and destination pointers are aligned
4633to that boundary.
4634</p>
4635
4636<h5>Semantics:</h5>
4637
4638<p>
4639The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
4640location to the destination location, which are not allowed to overlap. It
4641copies "len" bytes of memory over. If the argument is known to be aligned to
4642some boundary, this can be specified as the fourth argument, otherwise it should
4643be set to 0 or 1.
4644</p>
4645</div>
4646
4647
4648<!-- _______________________________________________________________________ -->
4649<div class="doc_subsubsection">
4650 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
4651</div>
4652
4653<div class="doc_text">
4654
4655<h5>Syntax:</h5>
4656<pre>
4657 declare void @llvm.memmove.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
4658 i32 &lt;len&gt;, i32 &lt;align&gt;)
4659 declare void @llvm.memmove.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
4660 i64 &lt;len&gt;, i32 &lt;align&gt;)
4661</pre>
4662
4663<h5>Overview:</h5>
4664
4665<p>
4666The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the source
4667location to the destination location. It is similar to the
Chris Lattnerdba16ea2008-01-06 19:51:52 +0000[diff] [blame]4668'<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to overlap.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4669</p>
4670
4671<p>
4672Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
4673intrinsics do not return a value, and takes an extra alignment argument.
4674</p>
4675
4676<h5>Arguments:</h5>
4677
4678<p>
4679The first argument is a pointer to the destination, the second is a pointer to
4680the source. The third argument is an integer argument
4681specifying the number of bytes to copy, and the fourth argument is the alignment
4682of the source and destination locations.
4683</p>
4684
4685<p>
4686If the call to this intrinisic has an alignment value that is not 0 or 1, then
4687the caller guarantees that the source and destination pointers are aligned to
4688that boundary.
4689</p>
4690
4691<h5>Semantics:</h5>
4692
4693<p>
4694The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the source
4695location to the destination location, which may overlap. It
4696copies "len" bytes of memory over. If the argument is known to be aligned to
4697some boundary, this can be specified as the fourth argument, otherwise it should
4698be set to 0 or 1.
4699</p>
4700</div>
4701
4702
4703<!-- _______________________________________________________________________ -->
4704<div class="doc_subsubsection">
4705 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
4706</div>
4707
4708<div class="doc_text">
4709
4710<h5>Syntax:</h5>
4711<pre>
4712 declare void @llvm.memset.i32(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
4713 i32 &lt;len&gt;, i32 &lt;align&gt;)
4714 declare void @llvm.memset.i64(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
4715 i64 &lt;len&gt;, i32 &lt;align&gt;)
4716</pre>
4717
4718<h5>Overview:</h5>
4719
4720<p>
4721The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a particular
4722byte value.
4723</p>
4724
4725<p>
4726Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
4727does not return a value, and takes an extra alignment argument.
4728</p>
4729
4730<h5>Arguments:</h5>
4731
4732<p>
4733The first argument is a pointer to the destination to fill, the second is the
4734byte value to fill it with, the third argument is an integer
4735argument specifying the number of bytes to fill, and the fourth argument is the
4736known alignment of destination location.
4737</p>
4738
4739<p>
4740If the call to this intrinisic has an alignment value that is not 0 or 1, then
4741the caller guarantees that the destination pointer is aligned to that boundary.
4742</p>
4743
4744<h5>Semantics:</h5>
4745
4746<p>
4747The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting at
4748the
4749destination location. If the argument is known to be aligned to some boundary,
4750this can be specified as the fourth argument, otherwise it should be set to 0 or
47511.
4752</p>
4753</div>
4754
4755
4756<!-- _______________________________________________________________________ -->
4757<div class="doc_subsubsection">
4758 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
4759</div>
4760
4761<div class="doc_text">
4762
4763<h5>Syntax:</h5>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]4764<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]4765floating point or vector of floating point type. Not all targets support all
4766types however.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4767<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]4768 declare float @llvm.sqrt.f32(float %Val)
4769 declare double @llvm.sqrt.f64(double %Val)
4770 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
4771 declare fp128 @llvm.sqrt.f128(fp128 %Val)
4772 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4773</pre>
4774
4775<h5>Overview:</h5>
4776
4777<p>
4778The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]4779returning the same value as the libm '<tt>sqrt</tt>' functions would. Unlike
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4780<tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
Chris Lattnerf914a002008-01-29 07:00:44 +0000[diff] [blame]4781negative numbers other than -0.0 (which allows for better optimization, because
4782there is no need to worry about errno being set). <tt>llvm.sqrt(-0.0)</tt> is
4783defined to return -0.0 like IEEE sqrt.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4784</p>
4785
4786<h5>Arguments:</h5>
4787
4788<p>
4789The argument and return value are floating point numbers of the same type.
4790</p>
4791
4792<h5>Semantics:</h5>
4793
4794<p>
4795This function returns the sqrt of the specified operand if it is a nonnegative
4796floating point number.
4797</p>
4798</div>
4799
4800<!-- _______________________________________________________________________ -->
4801<div class="doc_subsubsection">
4802 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
4803</div>
4804
4805<div class="doc_text">
4806
4807<h5>Syntax:</h5>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]4808<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]4809floating point or vector of floating point type. Not all targets support all
4810types however.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4811<pre>
Dale Johannesenf9adbb62007-10-02 17:47:38 +0000[diff] [blame]4812 declare float @llvm.powi.f32(float %Val, i32 %power)
4813 declare double @llvm.powi.f64(double %Val, i32 %power)
4814 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
4815 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
4816 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4817</pre>
4818
4819<h5>Overview:</h5>
4820
4821<p>
4822The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
4823specified (positive or negative) power. The order of evaluation of
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]4824multiplications is not defined. When a vector of floating point type is
4825used, the second argument remains a scalar integer value.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4826</p>
4827
4828<h5>Arguments:</h5>
4829
4830<p>
4831The second argument is an integer power, and the first is a value to raise to
4832that power.
4833</p>
4834
4835<h5>Semantics:</h5>
4836
4837<p>
4838This function returns the first value raised to the second power with an
4839unspecified sequence of rounding operations.</p>
4840</div>
4841
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]4842<!-- _______________________________________________________________________ -->
4843<div class="doc_subsubsection">
4844 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
4845</div>
4846
4847<div class="doc_text">
4848
4849<h5>Syntax:</h5>
4850<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
4851floating point or vector of floating point type. Not all targets support all
4852types however.
4853<pre>
4854 declare float @llvm.sin.f32(float %Val)
4855 declare double @llvm.sin.f64(double %Val)
4856 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
4857 declare fp128 @llvm.sin.f128(fp128 %Val)
4858 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
4859</pre>
4860
4861<h5>Overview:</h5>
4862
4863<p>
4864The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.
4865</p>
4866
4867<h5>Arguments:</h5>
4868
4869<p>
4870The argument and return value are floating point numbers of the same type.
4871</p>
4872
4873<h5>Semantics:</h5>
4874
4875<p>
4876This function returns the sine of the specified operand, returning the
4877same values as the libm <tt>sin</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]4878conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]4879</div>
4880
4881<!-- _______________________________________________________________________ -->
4882<div class="doc_subsubsection">
4883 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
4884</div>
4885
4886<div class="doc_text">
4887
4888<h5>Syntax:</h5>
4889<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
4890floating point or vector of floating point type. Not all targets support all
4891types however.
4892<pre>
4893 declare float @llvm.cos.f32(float %Val)
4894 declare double @llvm.cos.f64(double %Val)
4895 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
4896 declare fp128 @llvm.cos.f128(fp128 %Val)
4897 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
4898</pre>
4899
4900<h5>Overview:</h5>
4901
4902<p>
4903The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.
4904</p>
4905
4906<h5>Arguments:</h5>
4907
4908<p>
4909The argument and return value are floating point numbers of the same type.
4910</p>
4911
4912<h5>Semantics:</h5>
4913
4914<p>
4915This function returns the cosine of the specified operand, returning the
4916same values as the libm <tt>cos</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]4917conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]4918</div>
4919
4920<!-- _______________________________________________________________________ -->
4921<div class="doc_subsubsection">
4922 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
4923</div>
4924
4925<div class="doc_text">
4926
4927<h5>Syntax:</h5>
4928<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
4929floating point or vector of floating point type. Not all targets support all
4930types however.
4931<pre>
4932 declare float @llvm.pow.f32(float %Val, float %Power)
4933 declare double @llvm.pow.f64(double %Val, double %Power)
4934 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
4935 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
4936 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
4937</pre>
4938
4939<h5>Overview:</h5>
4940
4941<p>
4942The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
4943specified (positive or negative) power.
4944</p>
4945
4946<h5>Arguments:</h5>
4947
4948<p>
4949The second argument is a floating point power, and the first is a value to
4950raise to that power.
4951</p>
4952
4953<h5>Semantics:</h5>
4954
4955<p>
4956This function returns the first value raised to the second power,
4957returning the
4958same values as the libm <tt>pow</tt> functions would, and handles error
Dan Gohmaneaba92e2007-10-17 18:05:13 +0000[diff] [blame]4959conditions in the same way.</p>
Dan Gohman361079c2007-10-15 20:30:11 +0000[diff] [blame]4960</div>
4961
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4962
4963<!-- ======================================================================= -->
4964<div class="doc_subsection">
4965 <a name="int_manip">Bit Manipulation Intrinsics</a>
4966</div>
4967
4968<div class="doc_text">
4969<p>
4970LLVM provides intrinsics for a few important bit manipulation operations.
4971These allow efficient code generation for some algorithms.
4972</p>
4973
4974</div>
4975
4976<!-- _______________________________________________________________________ -->
4977<div class="doc_subsubsection">
4978 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
4979</div>
4980
4981<div class="doc_text">
4982
4983<h5>Syntax:</h5>
4984<p>This is an overloaded intrinsic function. You can use bswap on any integer
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4985type that is an even number of bytes (i.e. BitWidth % 16 == 0).
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4986<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]4987 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
4988 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
4989 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]4990</pre>
4991
4992<h5>Overview:</h5>
4993
4994<p>
4995The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
4996values with an even number of bytes (positive multiple of 16 bits). These are
4997useful for performing operations on data that is not in the target's native
4998byte order.
4999</p>
5000
5001<h5>Semantics:</h5>
5002
5003<p>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5004The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5005and low byte of the input i16 swapped. Similarly, the <tt>llvm.bswap.i32</tt>
5006intrinsic returns an i32 value that has the four bytes of the input i32
5007swapped, 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]5008i32 will have its bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i48</tt>,
5009<tt>llvm.bswap.i64</tt> and other intrinsics extend this concept to
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5010additional even-byte lengths (6 bytes, 8 bytes and more, respectively).
5011</p>
5012
5013</div>
5014
5015<!-- _______________________________________________________________________ -->
5016<div class="doc_subsubsection">
5017 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
5018</div>
5019
5020<div class="doc_text">
5021
5022<h5>Syntax:</h5>
5023<p>This is an overloaded intrinsic. You can use llvm.ctpop on any integer bit
5024width. Not all targets support all bit widths however.
5025<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5026 declare i8 @llvm.ctpop.i8 (i8 &lt;src&gt;)
5027 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5028 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5029 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
5030 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5031</pre>
5032
5033<h5>Overview:</h5>
5034
5035<p>
5036The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
5037value.
5038</p>
5039
5040<h5>Arguments:</h5>
5041
5042<p>
5043The only argument is the value to be counted. The argument may be of any
5044integer type. The return type must match the argument type.
5045</p>
5046
5047<h5>Semantics:</h5>
5048
5049<p>
5050The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
5051</p>
5052</div>
5053
5054<!-- _______________________________________________________________________ -->
5055<div class="doc_subsubsection">
5056 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
5057</div>
5058
5059<div class="doc_text">
5060
5061<h5>Syntax:</h5>
5062<p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
5063integer bit width. Not all targets support all bit widths however.
5064<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5065 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
5066 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5067 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5068 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
5069 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5070</pre>
5071
5072<h5>Overview:</h5>
5073
5074<p>
5075The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
5076leading zeros in a variable.
5077</p>
5078
5079<h5>Arguments:</h5>
5080
5081<p>
5082The only argument is the value to be counted. The argument may be of any
5083integer type. The return type must match the argument type.
5084</p>
5085
5086<h5>Semantics:</h5>
5087
5088<p>
5089The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
5090in a variable. If the src == 0 then the result is the size in bits of the type
5091of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.
5092</p>
5093</div>
5094
5095
5096
5097<!-- _______________________________________________________________________ -->
5098<div class="doc_subsubsection">
5099 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
5100</div>
5101
5102<div class="doc_text">
5103
5104<h5>Syntax:</h5>
5105<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
5106integer bit width. Not all targets support all bit widths however.
5107<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5108 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
5109 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5110 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5111 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
5112 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5113</pre>
5114
5115<h5>Overview:</h5>
5116
5117<p>
5118The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
5119trailing zeros.
5120</p>
5121
5122<h5>Arguments:</h5>
5123
5124<p>
5125The only argument is the value to be counted. The argument may be of any
5126integer type. The return type must match the argument type.
5127</p>
5128
5129<h5>Semantics:</h5>
5130
5131<p>
5132The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
5133in a variable. If the src == 0 then the result is the size in bits of the type
5134of src. For example, <tt>llvm.cttz(2) = 1</tt>.
5135</p>
5136</div>
5137
5138<!-- _______________________________________________________________________ -->
5139<div class="doc_subsubsection">
5140 <a name="int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic</a>
5141</div>
5142
5143<div class="doc_text">
5144
5145<h5>Syntax:</h5>
5146<p>This is an overloaded intrinsic. You can use <tt>llvm.part.select</tt>
5147on any integer bit width.
5148<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5149 declare i17 @llvm.part.select.i17 (i17 %val, i32 %loBit, i32 %hiBit)
5150 declare i29 @llvm.part.select.i29 (i29 %val, i32 %loBit, i32 %hiBit)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5151</pre>
5152
5153<h5>Overview:</h5>
5154<p>The '<tt>llvm.part.select</tt>' family of intrinsic functions selects a
5155range of bits from an integer value and returns them in the same bit width as
5156the original value.</p>
5157
5158<h5>Arguments:</h5>
5159<p>The first argument, <tt>%val</tt> and the result may be integer types of
5160any bit width but they must have the same bit width. The second and third
5161arguments must be <tt>i32</tt> type since they specify only a bit index.</p>
5162
5163<h5>Semantics:</h5>
5164<p>The operation of the '<tt>llvm.part.select</tt>' intrinsic has two modes
5165of operation: forwards and reverse. If <tt>%loBit</tt> is greater than
5166<tt>%hiBits</tt> then the intrinsic operates in reverse mode. Otherwise it
5167operates in forward mode.</p>
5168<p>In forward mode, this intrinsic is the equivalent of shifting <tt>%val</tt>
5169right by <tt>%loBit</tt> bits and then ANDing it with a mask with
5170only the <tt>%hiBit - %loBit</tt> bits set, as follows:</p>
5171<ol>
5172 <li>The <tt>%val</tt> is shifted right (LSHR) by the number of bits specified
5173 by <tt>%loBits</tt>. This normalizes the value to the low order bits.</li>
5174 <li>The <tt>%loBits</tt> value is subtracted from the <tt>%hiBits</tt> value
5175 to determine the number of bits to retain.</li>
5176 <li>A mask of the retained bits is created by shifting a -1 value.</li>
5177 <li>The mask is ANDed with <tt>%val</tt> to produce the result.
5178</ol>
5179<p>In reverse mode, a similar computation is made except that the bits are
5180returned in the reverse order. So, for example, if <tt>X</tt> has the value
5181<tt>i16 0x0ACF (101011001111)</tt> and we apply
5182<tt>part.select(i16 X, 8, 3)</tt> to it, we get back the value
5183<tt>i16 0x0026 (000000100110)</tt>.</p>
5184</div>
5185
5186<div class="doc_subsubsection">
5187 <a name="int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic</a>
5188</div>
5189
5190<div class="doc_text">
5191
5192<h5>Syntax:</h5>
5193<p>This is an overloaded intrinsic. You can use <tt>llvm.part.set</tt>
5194on any integer bit width.
5195<pre>
Chandler Carrutha228e392007-08-04 01:51:18 +0000[diff] [blame]5196 declare i17 @llvm.part.set.i17.i9 (i17 %val, i9 %repl, i32 %lo, i32 %hi)
5197 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]5198</pre>
5199
5200<h5>Overview:</h5>
5201<p>The '<tt>llvm.part.set</tt>' family of intrinsic functions replaces a range
5202of bits in an integer value with another integer value. It returns the integer
5203with the replaced bits.</p>
5204
5205<h5>Arguments:</h5>
5206<p>The first argument, <tt>%val</tt> and the result may be integer types of
5207any bit width but they must have the same bit width. <tt>%val</tt> is the value
5208whose bits will be replaced. The second argument, <tt>%repl</tt> may be an
5209integer of any bit width. The third and fourth arguments must be <tt>i32</tt>
5210type since they specify only a bit index.</p>
5211
5212<h5>Semantics:</h5>
5213<p>The operation of the '<tt>llvm.part.set</tt>' intrinsic has two modes
5214of operation: forwards and reverse. If <tt>%lo</tt> is greater than
5215<tt>%hi</tt> then the intrinsic operates in reverse mode. Otherwise it
5216operates in forward mode.</p>
5217<p>For both modes, the <tt>%repl</tt> value is prepared for use by either
5218truncating it down to the size of the replacement area or zero extending it
5219up to that size.</p>
5220<p>In forward mode, the bits between <tt>%lo</tt> and <tt>%hi</tt> (inclusive)
5221are replaced with corresponding bits from <tt>%repl</tt>. That is the 0th bit
5222in <tt>%repl</tt> replaces the <tt>%lo</tt>th bit in <tt>%val</tt> and etc. up
5223to the <tt>%hi</tt>th bit.
5224<p>In reverse mode, a similar computation is made except that the bits are
5225reversed. That is, the <tt>0</tt>th bit in <tt>%repl</tt> replaces the
5226<tt>%hi</tt> bit in <tt>%val</tt> and etc. down to the <tt>%lo</tt>th bit.
5227<h5>Examples:</h5>
5228<pre>
5229 llvm.part.set(0xFFFF, 0, 4, 7) -&gt; 0xFF0F
5230 llvm.part.set(0xFFFF, 0, 7, 4) -&gt; 0xFF0F
5231 llvm.part.set(0xFFFF, 1, 7, 4) -&gt; 0xFF8F
5232 llvm.part.set(0xFFFF, F, 8, 3) -&gt; 0xFFE7
5233 llvm.part.set(0xFFFF, 0, 3, 8) -&gt; 0xFE07
5234</pre>
5235</div>
5236
5237<!-- ======================================================================= -->
5238<div class="doc_subsection">
5239 <a name="int_debugger">Debugger Intrinsics</a>
5240</div>
5241
5242<div class="doc_text">
5243<p>
5244The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
5245are described in the <a
5246href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
5247Debugging</a> document.
5248</p>
5249</div>
5250
5251
5252<!-- ======================================================================= -->
5253<div class="doc_subsection">
5254 <a name="int_eh">Exception Handling Intrinsics</a>
5255</div>
5256
5257<div class="doc_text">
5258<p> The LLVM exception handling intrinsics (which all start with
5259<tt>llvm.eh.</tt> prefix), are described in the <a
5260href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
5261Handling</a> document. </p>
5262</div>
5263
5264<!-- ======================================================================= -->
5265<div class="doc_subsection">
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5266 <a name="int_trampoline">Trampoline Intrinsic</a>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5267</div>
5268
5269<div class="doc_text">
5270<p>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5271 This intrinsic makes it possible to excise one parameter, marked with
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5272 the <tt>nest</tt> attribute, from a function. The result is a callable
5273 function pointer lacking the nest parameter - the caller does not need
5274 to provide a value for it. Instead, the value to use is stored in
5275 advance in a "trampoline", a block of memory usually allocated
5276 on the stack, which also contains code to splice the nest value into the
5277 argument list. This is used to implement the GCC nested function address
5278 extension.
5279</p>
5280<p>
5281 For example, if the function is
5282 <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]5283 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]5284<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5285 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
5286 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
5287 %p = call i8* @llvm.init.trampoline( i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8*), i8* %nval )
5288 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5289</pre>
Bill Wendlinge262b4c2007-09-22 09:23:55 +0000[diff] [blame]5290 <p>The call <tt>%val = call i32 %fp( i32 %x, i32 %y )</tt> is then equivalent
5291 to <tt>%val = call i32 %f( i8* %nval, i32 %x, i32 %y )</tt>.</p>
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5292</div>
5293
5294<!-- _______________________________________________________________________ -->
5295<div class="doc_subsubsection">
5296 <a name="int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a>
5297</div>
5298<div class="doc_text">
5299<h5>Syntax:</h5>
5300<pre>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5301declare 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]5302</pre>
5303<h5>Overview:</h5>
5304<p>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5305 This fills the memory pointed to by <tt>tramp</tt> with code
5306 and returns a function pointer suitable for executing it.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5307</p>
5308<h5>Arguments:</h5>
5309<p>
5310 The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
5311 pointers. The <tt>tramp</tt> argument must point to a sufficiently large
5312 and sufficiently aligned block of memory; this memory is written to by the
Duncan Sands35012212007-08-22 23:39:54 +0000[diff] [blame]5313 intrinsic. Note that the size and the alignment are target-specific - LLVM
5314 currently provides no portable way of determining them, so a front-end that
5315 generates this intrinsic needs to have some target-specific knowledge.
5316 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]5317</p>
5318<h5>Semantics:</h5>
5319<p>
5320 The block of memory pointed to by <tt>tramp</tt> is filled with target
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5321 dependent code, turning it into a function. A pointer to this function is
5322 returned, but needs to be bitcast to an
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5323 <a href="#int_trampoline">appropriate function pointer type</a>
Duncan Sands7407a9f2007-09-11 14:10:23 +0000[diff] [blame]5324 before being called. The new function's signature is the same as that of
5325 <tt>func</tt> with any arguments marked with the <tt>nest</tt> attribute
5326 removed. At most one such <tt>nest</tt> argument is allowed, and it must be
5327 of pointer type. Calling the new function is equivalent to calling
5328 <tt>func</tt> with the same argument list, but with <tt>nval</tt> used for the
5329 missing <tt>nest</tt> argument. If, after calling
5330 <tt>llvm.init.trampoline</tt>, the memory pointed to by <tt>tramp</tt> is
5331 modified, then the effect of any later call to the returned function pointer is
5332 undefined.
Duncan Sands38947cd2007-07-27 12:58:54 +0000[diff] [blame]5333</p>
5334</div>
5335
5336<!-- ======================================================================= -->
5337<div class="doc_subsection">
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]5338 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
5339</div>
5340
5341<div class="doc_text">
5342<p>
5343 These intrinsic functions expand the "universal IR" of LLVM to represent
5344 hardware constructs for atomic operations and memory synchronization. This
5345 provides an interface to the hardware, not an interface to the programmer. It
5346 is aimed at a low enough level to allow any programming models or APIs which
5347 need atomic behaviors to map cleanly onto it. It is also modeled primarily on
5348 hardware behavior. Just as hardware provides a "universal IR" for source
5349 languages, it also provides a starting point for developing a "universal"
5350 atomic operation and synchronization IR.
5351</p>
5352<p>
5353 These do <em>not</em> form an API such as high-level threading libraries,
5354 software transaction memory systems, atomic primitives, and intrinsic
5355 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
5356 application libraries. The hardware interface provided by LLVM should allow
5357 a clean implementation of all of these APIs and parallel programming models.
5358 No one model or paradigm should be selected above others unless the hardware
5359 itself ubiquitously does so.
5360
5361</p>
5362</div>
5363
5364<!-- _______________________________________________________________________ -->
5365<div class="doc_subsubsection">
5366 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
5367</div>
5368<div class="doc_text">
5369<h5>Syntax:</h5>
5370<pre>
5371declare void @llvm.memory.barrier( i1 &lt;ll&gt;, i1 &lt;ls&gt;, i1 &lt;sl&gt;, i1 &lt;ss&gt;,
5372i1 &lt;device&gt; )
5373
5374</pre>
5375<h5>Overview:</h5>
5376<p>
5377 The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
5378 specific pairs of memory access types.
5379</p>
5380<h5>Arguments:</h5>
5381<p>
5382 The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
5383 The first four arguments enables a specific barrier as listed below. The fith
5384 argument specifies that the barrier applies to io or device or uncached memory.
5385
5386</p>
5387 <ul>
5388 <li><tt>ll</tt>: load-load barrier</li>
5389 <li><tt>ls</tt>: load-store barrier</li>
5390 <li><tt>sl</tt>: store-load barrier</li>
5391 <li><tt>ss</tt>: store-store barrier</li>
5392 <li><tt>device</tt>: barrier applies to device and uncached memory also.
5393 </ul>
5394<h5>Semantics:</h5>
5395<p>
5396 This intrinsic causes the system to enforce some ordering constraints upon
5397 the loads and stores of the program. This barrier does not indicate
5398 <em>when</em> any events will occur, it only enforces an <em>order</em> in
5399 which they occur. For any of the specified pairs of load and store operations
5400 (f.ex. load-load, or store-load), all of the first operations preceding the
5401 barrier will complete before any of the second operations succeeding the
5402 barrier begin. Specifically the semantics for each pairing is as follows:
5403</p>
5404 <ul>
5405 <li><tt>ll</tt>: All loads before the barrier must complete before any load
5406 after the barrier begins.</li>
5407
5408 <li><tt>ls</tt>: All loads before the barrier must complete before any
5409 store after the barrier begins.</li>
5410 <li><tt>ss</tt>: All stores before the barrier must complete before any
5411 store after the barrier begins.</li>
5412 <li><tt>sl</tt>: All stores before the barrier must complete before any
5413 load after the barrier begins.</li>
5414 </ul>
5415<p>
5416 These semantics are applied with a logical "and" behavior when more than one
5417 is enabled in a single memory barrier intrinsic.
5418</p>
5419<p>
5420 Backends may implement stronger barriers than those requested when they do not
5421 support as fine grained a barrier as requested. Some architectures do not
5422 need all types of barriers and on such architectures, these become noops.
5423</p>
5424<h5>Example:</h5>
5425<pre>
5426%ptr = malloc i32
5427 store i32 4, %ptr
5428
5429%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
5430 call void @llvm.memory.barrier( i1 false, i1 true, i1 false, i1 false )
5431 <i>; guarantee the above finishes</i>
5432 store i32 8, %ptr <i>; before this begins</i>
5433</pre>
5434</div>
5435
Andrew Lenharthe44f3902008-02-21 06:45:13 +0000[diff] [blame]5436<!-- _______________________________________________________________________ -->
5437<div class="doc_subsubsection">
5438 <a name="int_atomic_lcs">'<tt>llvm.atomic.lcs.*</tt>' Intrinsic</a>
5439</div>
5440<div class="doc_text">
5441<h5>Syntax:</h5>
5442<p>
5443 This is an overloaded intrinsic. You can use <tt>llvm.atomic.lcs</tt> on any
5444 integer bit width. Not all targets support all bit widths however.</p>
5445
5446<pre>
5447declare i8 @llvm.atomic.lcs.i8( i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt; )
5448declare i16 @llvm.atomic.lcs.i16( i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt; )
5449declare i32 @llvm.atomic.lcs.i32( i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt; )
5450declare i64 @llvm.atomic.lcs.i64( i64* &lt;ptr&gt;, i64 &lt;cmp&gt;, i64 &lt;val&gt; )
5451
5452</pre>
5453<h5>Overview:</h5>
5454<p>
5455 This loads a value in memory and compares it to a given value. If they are
5456 equal, it stores a new value into the memory.
5457</p>
5458<h5>Arguments:</h5>
5459<p>
5460 The <tt>llvm.atomic.lcs</tt> intrinsic takes three arguments. The result as
5461 well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
5462 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
5463 this integer type. While any bit width integer may be used, targets may only
5464 lower representations they support in hardware.
5465
5466</p>
5467<h5>Semantics:</h5>
5468<p>
5469 This entire intrinsic must be executed atomically. It first loads the value
5470 in memory pointed to by <tt>ptr</tt> and compares it with the value
5471 <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the memory. The
5472 loaded value is yielded in all cases. This provides the equivalent of an
5473 atomic compare-and-swap operation within the SSA framework.
5474</p>
5475<h5>Examples:</h5>
5476
5477<pre>
5478%ptr = malloc i32
5479 store i32 4, %ptr
5480
5481%val1 = add i32 4, 4
5482%result1 = call i32 @llvm.atomic.lcs.i32( i32* %ptr, i32 4, %val1 )
5483 <i>; yields {i32}:result1 = 4</i>
5484%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
5485%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
5486
5487%val2 = add i32 1, 1
5488%result2 = call i32 @llvm.atomic.lcs.i32( i32* %ptr, i32 5, %val2 )
5489 <i>; yields {i32}:result2 = 8</i>
5490%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
5491
5492%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
5493</pre>
5494</div>
5495
5496<!-- _______________________________________________________________________ -->
5497<div class="doc_subsubsection">
5498 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
5499</div>
5500<div class="doc_text">
5501<h5>Syntax:</h5>
5502
5503<p>
5504 This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
5505 integer bit width. Not all targets support all bit widths however.</p>
5506<pre>
5507declare i8 @llvm.atomic.swap.i8( i8* &lt;ptr&gt;, i8 &lt;val&gt; )
5508declare i16 @llvm.atomic.swap.i16( i16* &lt;ptr&gt;, i16 &lt;val&gt; )
5509declare i32 @llvm.atomic.swap.i32( i32* &lt;ptr&gt;, i32 &lt;val&gt; )
5510declare i64 @llvm.atomic.swap.i64( i64* &lt;ptr&gt;, i64 &lt;val&gt; )
5511
5512</pre>
5513<h5>Overview:</h5>
5514<p>
5515 This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
5516 the value from memory. It then stores the value in <tt>val</tt> in the memory
5517 at <tt>ptr</tt>.
5518</p>
5519<h5>Arguments:</h5>
5520
5521<p>
5522 The <tt>llvm.atomic.ls</tt> intrinsic takes two arguments. Both the
5523 <tt>val</tt> argument and the result must be integers of the same bit width.
5524 The first argument, <tt>ptr</tt>, must be a pointer to a value of this
5525 integer type. The targets may only lower integer representations they
5526 support.
5527</p>
5528<h5>Semantics:</h5>
5529<p>
5530 This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
5531 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
5532 equivalent of an atomic swap operation within the SSA framework.
5533
5534</p>
5535<h5>Examples:</h5>
5536<pre>
5537%ptr = malloc i32
5538 store i32 4, %ptr
5539
5540%val1 = add i32 4, 4
5541%result1 = call i32 @llvm.atomic.swap.i32( i32* %ptr, i32 %val1 )
5542 <i>; yields {i32}:result1 = 4</i>
5543%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
5544%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
5545
5546%val2 = add i32 1, 1
5547%result2 = call i32 @llvm.atomic.swap.i32( i32* %ptr, i32 %val2 )
5548 <i>; yields {i32}:result2 = 8</i>
5549
5550%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
5551%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
5552</pre>
5553</div>
5554
5555<!-- _______________________________________________________________________ -->
5556<div class="doc_subsubsection">
5557 <a name="int_atomic_las">'<tt>llvm.atomic.las.*</tt>' Intrinsic</a>
5558
5559</div>
5560<div class="doc_text">
5561<h5>Syntax:</h5>
5562<p>
5563 This is an overloaded intrinsic. You can use <tt>llvm.atomic.las</tt> on any
5564 integer bit width. Not all targets support all bit widths however.</p>
5565<pre>
5566declare i8 @llvm.atomic.las.i8.( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
5567declare i16 @llvm.atomic.las.i16.( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
5568declare i32 @llvm.atomic.las.i32.( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
5569declare i64 @llvm.atomic.las.i64.( i64* &lt;ptr&gt;, i64 &lt;delta&gt; )
5570
5571</pre>
5572<h5>Overview:</h5>
5573<p>
5574 This intrinsic adds <tt>delta</tt> to the value stored in memory at
5575 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
5576</p>
5577<h5>Arguments:</h5>
5578<p>
5579
5580 The intrinsic takes two arguments, the first a pointer to an integer value
5581 and the second an integer value. The result is also an integer value. These
5582 integer types can have any bit width, but they must all have the same bit
5583 width. The targets may only lower integer representations they support.
5584</p>
5585<h5>Semantics:</h5>
5586<p>
5587 This intrinsic does a series of operations atomically. It first loads the
5588 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
5589 to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
5590</p>
5591
5592<h5>Examples:</h5>
5593<pre>
5594%ptr = malloc i32
5595 store i32 4, %ptr
5596%result1 = call i32 @llvm.atomic.las.i32( i32* %ptr, i32 4 )
5597 <i>; yields {i32}:result1 = 4</i>
5598%result2 = call i32 @llvm.atomic.las.i32( i32* %ptr, i32 2 )
5599 <i>; yields {i32}:result2 = 8</i>
5600%result3 = call i32 @llvm.atomic.las.i32( i32* %ptr, i32 5 )
5601 <i>; yields {i32}:result3 = 10</i>
5602%memval = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
5603</pre>
5604</div>
5605
Andrew Lenharth785610d2008-02-16 01:24:58 +0000[diff] [blame]5606
5607<!-- ======================================================================= -->
5608<div class="doc_subsection">
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5609 <a name="int_general">General Intrinsics</a>
5610</div>
5611
5612<div class="doc_text">
5613<p> This class of intrinsics is designed to be generic and has
5614no specific purpose. </p>
5615</div>
5616
5617<!-- _______________________________________________________________________ -->
5618<div class="doc_subsubsection">
5619 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
5620</div>
5621
5622<div class="doc_text">
5623
5624<h5>Syntax:</h5>
5625<pre>
5626 declare void @llvm.var.annotation(i8* &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
5627</pre>
5628
5629<h5>Overview:</h5>
5630
5631<p>
5632The '<tt>llvm.var.annotation</tt>' intrinsic
5633</p>
5634
5635<h5>Arguments:</h5>
5636
5637<p>
5638The first argument is a pointer to a value, the second is a pointer to a
5639global string, the third is a pointer to a global string which is the source
5640file name, and the last argument is the line number.
5641</p>
5642
5643<h5>Semantics:</h5>
5644
5645<p>
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]5646This intrinsic allows annotation of local variables with arbitrary strings.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5647This can be useful for special purpose optimizations that want to look for these
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]5648annotations. These have no other defined use, they are ignored by code
5649generation and optimization.
5650</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5651</div>
5652
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]5653<!-- _______________________________________________________________________ -->
5654<div class="doc_subsubsection">
Tanya Lattnerc9869b12007-09-21 23:57:59 +0000[diff] [blame]5655 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]5656</div>
5657
5658<div class="doc_text">
5659
5660<h5>Syntax:</h5>
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]5661<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
5662any integer bit width.
5663</p>
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]5664<pre>
Tanya Lattner09161fe2007-09-22 00:03:01 +0000[diff] [blame]5665 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
5666 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
5667 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
5668 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
5669 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]5670</pre>
5671
5672<h5>Overview:</h5>
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]5673
5674<p>
5675The '<tt>llvm.annotation</tt>' intrinsic.
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]5676</p>
5677
5678<h5>Arguments:</h5>
5679
5680<p>
5681The first argument is an integer value (result of some expression),
5682the second is a pointer to a global string, the third is a pointer to a global
5683string which is the source file name, and the last argument is the line number.
Tanya Lattnere545be72007-09-21 23:56:27 +0000[diff] [blame]5684It returns the value of the first argument.
Tanya Lattnerb306a9e2007-09-21 22:59:12 +0000[diff] [blame]5685</p>
5686
5687<h5>Semantics:</h5>
5688
5689<p>
5690This intrinsic allows annotations to be put on arbitrary expressions
5691with arbitrary strings. This can be useful for special purpose optimizations
5692that want to look for these annotations. These have no other defined use, they
5693are ignored by code generation and optimization.
5694</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5695
Anton Korobeynikove6e764f2008-01-15 22:31:34 +0000[diff] [blame]5696<!-- _______________________________________________________________________ -->
5697<div class="doc_subsubsection">
5698 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
5699</div>
5700
5701<div class="doc_text">
5702
5703<h5>Syntax:</h5>
5704<pre>
5705 declare void @llvm.trap()
5706</pre>
5707
5708<h5>Overview:</h5>
5709
5710<p>
5711The '<tt>llvm.trap</tt>' intrinsic
5712</p>
5713
5714<h5>Arguments:</h5>
5715
5716<p>
5717None
5718</p>
5719
5720<h5>Semantics:</h5>
5721
5722<p>
5723This intrinsics is lowered to the target dependent trap instruction. If the
5724target does not have a trap instruction, this intrinsic will be lowered to the
5725call of the abort() function.
5726</p>
5727</div>
5728
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5729<!-- *********************************************************************** -->
5730<hr>
5731<address>
5732 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
5733 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
5734 <a href="http://validator.w3.org/check/referer"><img
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]5735 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!"></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5736
5737 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
5738 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
5739 Last modified: $Date$
5740</address>
Chris Lattner08497ce2008-01-04 04:33:49 +0000[diff] [blame]5741
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]5742</body>
5743</html>