1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
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">
15 <div class="doc_title"> LLVM Language Reference Manual </div>
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>
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="#namedtypes">Named Types</a></li>
26 <li><a href="#globalvars">Global Variables</a></li>
27 <li><a href="#functionstructure">Functions</a></li>
28 <li><a href="#aliasstructure">Aliases</a></li>
29 <li><a href="#paramattrs">Parameter Attributes</a></li>
30 <li><a href="#fnattrs">Function Attributes</a></li>
31 <li><a href="#gc">Garbage Collector Names</a></li>
32 <li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
33 <li><a href="#datalayout">Data Layout</a></li>
36 <li><a href="#typesystem">Type System</a>
38 <li><a href="#t_classifications">Type Classifications</a></li>
39 <li><a href="#t_primitive">Primitive Types</a>
41 <li><a href="#t_floating">Floating Point Types</a></li>
42 <li><a href="#t_void">Void Type</a></li>
43 <li><a href="#t_label">Label Type</a></li>
46 <li><a href="#t_derived">Derived Types</a>
48 <li><a href="#t_integer">Integer Type</a></li>
49 <li><a href="#t_array">Array Type</a></li>
50 <li><a href="#t_function">Function Type</a></li>
51 <li><a href="#t_pointer">Pointer Type</a></li>
52 <li><a href="#t_struct">Structure Type</a></li>
53 <li><a href="#t_pstruct">Packed Structure Type</a></li>
54 <li><a href="#t_vector">Vector Type</a></li>
55 <li><a href="#t_opaque">Opaque Type</a></li>
58 <li><a href="#t_uprefs">Type Up-references</a></li>
61 <li><a href="#constants">Constants</a>
63 <li><a href="#simpleconstants">Simple Constants</a></li>
64 <li><a href="#complexconstants">Complex Constants</a></li>
65 <li><a href="#globalconstants">Global Variable and Function Addresses</a></li>
66 <li><a href="#undefvalues">Undefined Values</a></li>
67 <li><a href="#constantexprs">Constant Expressions</a></li>
68 <li><a href="#metadata">Embedded Metadata</a></li>
71 <li><a href="#othervalues">Other Values</a>
73 <li><a href="#inlineasm">Inline Assembler Expressions</a></li>
76 <li><a href="#instref">Instruction Reference</a>
78 <li><a href="#terminators">Terminator Instructions</a>
80 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
81 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
82 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
83 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
84 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
85 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
88 <li><a href="#binaryops">Binary Operations</a>
90 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
91 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
92 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
93 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
94 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
95 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
96 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
97 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
98 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
101 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
103 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
104 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
105 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
106 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
107 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
108 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
111 <li><a href="#vectorops">Vector Operations</a>
113 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
114 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
115 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
118 <li><a href="#aggregateops">Aggregate Operations</a>
120 <li><a href="#i_extractvalue">'<tt>extractvalue</tt>' Instruction</a></li>
121 <li><a href="#i_insertvalue">'<tt>insertvalue</tt>' Instruction</a></li>
124 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
126 <li><a href="#i_malloc">'<tt>malloc</tt>' Instruction</a></li>
127 <li><a href="#i_free">'<tt>free</tt>' Instruction</a></li>
128 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
129 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
130 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
131 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
134 <li><a href="#convertops">Conversion Operations</a>
136 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
137 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
138 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
139 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
140 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
141 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
142 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
143 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
144 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
145 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
146 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
147 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
150 <li><a href="#otherops">Other Operations</a>
152 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
153 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
154 <li><a href="#i_vicmp">'<tt>vicmp</tt>' Instruction</a></li>
155 <li><a href="#i_vfcmp">'<tt>vfcmp</tt>' Instruction</a></li>
156 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
157 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
158 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
159 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
164 <li><a href="#intrinsics">Intrinsic Functions</a>
166 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
168 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
169 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
170 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
173 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
175 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
176 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
177 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
180 <li><a href="#int_codegen">Code Generator Intrinsics</a>
182 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
183 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
184 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
185 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
186 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
187 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
188 <li><a href="#int_readcyclecounter"><tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
191 <li><a href="#int_libc">Standard C Library Intrinsics</a>
193 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
194 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
195 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
196 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
197 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
198 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
199 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
200 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
203 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
205 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
206 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
207 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
208 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
209 <li><a href="#int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic </a></li>
210 <li><a href="#int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic </a></li>
213 <li><a href="#int_overflow">Arithmetic with Overflow Intrinsics</a>
215 <li><a href="#int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt> Intrinsics</a></li>
216 <li><a href="#int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt> Intrinsics</a></li>
217 <li><a href="#int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt> Intrinsics</a></li>
218 <li><a href="#int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt> Intrinsics</a></li>
219 <li><a href="#int_smul_overflow">'<tt>llvm.smul.with.overflow.*</tt> Intrinsics</a></li>
220 <li><a href="#int_umul_overflow">'<tt>llvm.umul.with.overflow.*</tt> Intrinsics</a></li>
223 <li><a href="#int_debugger">Debugger intrinsics</a></li>
224 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
225 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
227 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
230 <li><a href="#int_atomics">Atomic intrinsics</a>
232 <li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
233 <li><a href="#int_atomic_cmp_swap"><tt>llvm.atomic.cmp.swap</tt></a></li>
234 <li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
235 <li><a href="#int_atomic_load_add"><tt>llvm.atomic.load.add</tt></a></li>
236 <li><a href="#int_atomic_load_sub"><tt>llvm.atomic.load.sub</tt></a></li>
237 <li><a href="#int_atomic_load_and"><tt>llvm.atomic.load.and</tt></a></li>
238 <li><a href="#int_atomic_load_nand"><tt>llvm.atomic.load.nand</tt></a></li>
239 <li><a href="#int_atomic_load_or"><tt>llvm.atomic.load.or</tt></a></li>
240 <li><a href="#int_atomic_load_xor"><tt>llvm.atomic.load.xor</tt></a></li>
241 <li><a href="#int_atomic_load_max"><tt>llvm.atomic.load.max</tt></a></li>
242 <li><a href="#int_atomic_load_min"><tt>llvm.atomic.load.min</tt></a></li>
243 <li><a href="#int_atomic_load_umax"><tt>llvm.atomic.load.umax</tt></a></li>
244 <li><a href="#int_atomic_load_umin"><tt>llvm.atomic.load.umin</tt></a></li>
247 <li><a href="#int_general">General intrinsics</a>
249 <li><a href="#int_var_annotation">
250 '<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
251 <li><a href="#int_annotation">
252 '<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
253 <li><a href="#int_trap">
254 '<tt>llvm.trap</tt>' Intrinsic</a></li>
255 <li><a href="#int_stackprotector">
256 '<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
263 <div class="doc_author">
264 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
265 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
268 <!-- *********************************************************************** -->
269 <div class="doc_section"> <a name="abstract">Abstract </a></div>
270 <!-- *********************************************************************** -->
272 <div class="doc_text">
273 <p>This document is a reference manual for the LLVM assembly language.
274 LLVM is a Static Single Assignment (SSA) based representation that provides
275 type safety, low-level operations, flexibility, and the capability of
276 representing 'all' high-level languages cleanly. It is the common code
277 representation used throughout all phases of the LLVM compilation
281 <!-- *********************************************************************** -->
282 <div class="doc_section"> <a name="introduction">Introduction</a> </div>
283 <!-- *********************************************************************** -->
285 <div class="doc_text">
287 <p>The LLVM code representation is designed to be used in three
288 different forms: as an in-memory compiler IR, as an on-disk bitcode
289 representation (suitable for fast loading by a Just-In-Time compiler),
290 and as a human readable assembly language representation. This allows
291 LLVM to provide a powerful intermediate representation for efficient
292 compiler transformations and analysis, while providing a natural means
293 to debug and visualize the transformations. The three different forms
294 of LLVM are all equivalent. This document describes the human readable
295 representation and notation.</p>
297 <p>The LLVM representation aims to be light-weight and low-level
298 while being expressive, typed, and extensible at the same time. It
299 aims to be a "universal IR" of sorts, by being at a low enough level
300 that high-level ideas may be cleanly mapped to it (similar to how
301 microprocessors are "universal IR's", allowing many source languages to
302 be mapped to them). By providing type information, LLVM can be used as
303 the target of optimizations: for example, through pointer analysis, it
304 can be proven that a C automatic variable is never accessed outside of
305 the current function... allowing it to be promoted to a simple SSA
306 value instead of a memory location.</p>
310 <!-- _______________________________________________________________________ -->
311 <div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
313 <div class="doc_text">
315 <p>It is important to note that this document describes 'well formed'
316 LLVM assembly language. There is a difference between what the parser
317 accepts and what is considered 'well formed'. For example, the
318 following instruction is syntactically okay, but not well formed:</p>
320 <div class="doc_code">
322 %x = <a href="#i_add">add</a> i32 1, %x
326 <p>...because the definition of <tt>%x</tt> does not dominate all of
327 its uses. The LLVM infrastructure provides a verification pass that may
328 be used to verify that an LLVM module is well formed. This pass is
329 automatically run by the parser after parsing input assembly and by
330 the optimizer before it outputs bitcode. The violations pointed out
331 by the verifier pass indicate bugs in transformation passes or input to
335 <!-- Describe the typesetting conventions here. -->
337 <!-- *********************************************************************** -->
338 <div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
339 <!-- *********************************************************************** -->
341 <div class="doc_text">
343 <p>LLVM identifiers come in two basic types: global and local. Global
344 identifiers (functions, global variables) begin with the @ character. Local
345 identifiers (register names, types) begin with the % character. Additionally,
346 there are three different formats for identifiers, for different purposes:</p>
349 <li>Named values are represented as a string of characters with their prefix.
350 For example, %foo, @DivisionByZero, %a.really.long.identifier. The actual
351 regular expression used is '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
352 Identifiers which require other characters in their names can be surrounded
353 with quotes. Special characters may be escaped using "\xx" where xx is the
354 ASCII code for the character in hexadecimal. In this way, any character can
355 be used in a name value, even quotes themselves.
357 <li>Unnamed values are represented as an unsigned numeric value with their
358 prefix. For example, %12, @2, %44.</li>
360 <li>Constants, which are described in a <a href="#constants">section about
361 constants</a>, below.</li>
364 <p>LLVM requires that values start with a prefix for two reasons: Compilers
365 don't need to worry about name clashes with reserved words, and the set of
366 reserved words may be expanded in the future without penalty. Additionally,
367 unnamed identifiers allow a compiler to quickly come up with a temporary
368 variable without having to avoid symbol table conflicts.</p>
370 <p>Reserved words in LLVM are very similar to reserved words in other
371 languages. There are keywords for different opcodes
372 ('<tt><a href="#i_add">add</a></tt>',
373 '<tt><a href="#i_bitcast">bitcast</a></tt>',
374 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
375 href="#t_void">void</a></tt>', '<tt><a href="#t_primitive">i32</a></tt>', etc...),
376 and others. These reserved words cannot conflict with variable names, because
377 none of them start with a prefix character ('%' or '@').</p>
379 <p>Here is an example of LLVM code to multiply the integer variable
380 '<tt>%X</tt>' by 8:</p>
384 <div class="doc_code">
386 %result = <a href="#i_mul">mul</a> i32 %X, 8
390 <p>After strength reduction:</p>
392 <div class="doc_code">
394 %result = <a href="#i_shl">shl</a> i32 %X, i8 3
398 <p>And the hard way:</p>
400 <div class="doc_code">
402 <a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
403 <a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
404 %result = <a href="#i_add">add</a> i32 %1, %1
408 <p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
409 important lexical features of LLVM:</p>
413 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
416 <li>Unnamed temporaries are created when the result of a computation is not
417 assigned to a named value.</li>
419 <li>Unnamed temporaries are numbered sequentially</li>
423 <p>...and it also shows a convention that we follow in this document. When
424 demonstrating instructions, we will follow an instruction with a comment that
425 defines the type and name of value produced. Comments are shown in italic
430 <!-- *********************************************************************** -->
431 <div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
432 <!-- *********************************************************************** -->
434 <!-- ======================================================================= -->
435 <div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
438 <div class="doc_text">
440 <p>LLVM programs are composed of "Module"s, each of which is a
441 translation unit of the input programs. Each module consists of
442 functions, global variables, and symbol table entries. Modules may be
443 combined together with the LLVM linker, which merges function (and
444 global variable) definitions, resolves forward declarations, and merges
445 symbol table entries. Here is an example of the "hello world" module:</p>
447 <div class="doc_code">
448 <pre><i>; Declare the string constant as a global constant...</i>
449 <a href="#identifiers">@.LC0</a> = <a href="#linkage_internal">internal</a> <a
450 href="#globalvars">constant</a> <a href="#t_array">[13 x i8]</a> c"hello world\0A\00" <i>; [13 x i8]*</i>
452 <i>; External declaration of the puts function</i>
453 <a href="#functionstructure">declare</a> i32 @puts(i8 *) <i>; i32(i8 *)* </i>
455 <i>; Definition of main function</i>
456 define i32 @main() { <i>; i32()* </i>
457 <i>; Convert [13 x i8]* to i8 *...</i>
459 href="#i_getelementptr">getelementptr</a> [13 x i8]* @.LC0, i64 0, i64 0 <i>; i8 *</i>
461 <i>; Call puts function to write out the string to stdout...</i>
463 href="#i_call">call</a> i32 @puts(i8 * %cast210) <i>; i32</i>
465 href="#i_ret">ret</a> i32 0<br>}<br>
469 <p>This example is made up of a <a href="#globalvars">global variable</a>
470 named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
471 function, and a <a href="#functionstructure">function definition</a>
472 for "<tt>main</tt>".</p>
474 <p>In general, a module is made up of a list of global values,
475 where both functions and global variables are global values. Global values are
476 represented by a pointer to a memory location (in this case, a pointer to an
477 array of char, and a pointer to a function), and have one of the following <a
478 href="#linkage">linkage types</a>.</p>
482 <!-- ======================================================================= -->
483 <div class="doc_subsection">
484 <a name="linkage">Linkage Types</a>
487 <div class="doc_text">
490 All Global Variables and Functions have one of the following types of linkage:
495 <dt><tt><b><a name="linkage_private">private</a></b></tt>: </dt>
497 <dd>Global values with private linkage are only directly accessible by
498 objects in the current module. In particular, linking code into a module with
499 an private global value may cause the private to be renamed as necessary to
500 avoid collisions. Because the symbol is private to the module, all
501 references can be updated. This doesn't show up in any symbol table in the
505 <dt><tt><b><a name="linkage_internal">internal</a></b></tt>: </dt>
507 <dd> Similar to private, but the value shows as a local symbol (STB_LOCAL in
508 the case of ELF) in the object file. This corresponds to the notion of the
509 '<tt>static</tt>' keyword in C.
512 <dt><tt><b><a name="available_externally">available_externally</a></b></tt>:
515 <dd>Globals with "<tt>available_externally</tt>" linkage are never emitted
516 into the object file corresponding to the LLVM module. They exist to
517 allow inlining and other optimizations to take place given knowledge of the
518 definition of the global, which is known to be somewhere outside the module.
519 Globals with <tt>available_externally</tt> linkage are allowed to be discarded
520 at will, and are otherwise the same as <tt>linkonce_odr</tt>. This linkage
521 type is only allowed on definitions, not declarations.</dd>
523 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
525 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
526 the same name when linkage occurs. This is typically used to implement
527 inline functions, templates, or other code which must be generated in each
528 translation unit that uses it. Unreferenced <tt>linkonce</tt> globals are
529 allowed to be discarded.
532 <dt><tt><b><a name="linkage_common">common</a></b></tt>: </dt>
534 <dd>"<tt>common</tt>" linkage is exactly the same as <tt>linkonce</tt>
535 linkage, except that unreferenced <tt>common</tt> globals may not be
536 discarded. This is used for globals that may be emitted in multiple
537 translation units, but that are not guaranteed to be emitted into every
538 translation unit that uses them. One example of this is tentative
539 definitions in C, such as "<tt>int X;</tt>" at global scope.
542 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
544 <dd>"<tt>weak</tt>" linkage is the same as <tt>common</tt> linkage, except
545 that some targets may choose to emit different assembly sequences for them
546 for target-dependent reasons. This is used for globals that are declared
547 "weak" in C source code.
550 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
552 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
553 pointer to array type. When two global variables with appending linkage are
554 linked together, the two global arrays are appended together. This is the
555 LLVM, typesafe, equivalent of having the system linker append together
556 "sections" with identical names when .o files are linked.
559 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt>: </dt>
561 <dd>The semantics of this linkage follow the ELF object file model: the
562 symbol is weak until linked, if not linked, the symbol becomes null instead
563 of being an undefined reference.
566 <dt><tt><b><a name="linkage_linkonce">linkonce_odr</a></b></tt>: </dt>
567 <dt><tt><b><a name="linkage_weak">weak_odr</a></b></tt>: </dt>
568 <dd>Some languages allow differing globals to be merged, such as two
569 functions with different semantics. Other languages, such as <tt>C++</tt>,
570 ensure that only equivalent globals are ever merged (the "one definition
571 rule" - "ODR"). Such languages can use the <tt>linkonce_odr</tt>
572 and <tt>weak_odr</tt> linkage types to indicate that the global will only
573 be merged with equivalent globals. These linkage types are otherwise the
574 same as their non-<tt>odr</tt> versions.
577 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
579 <dd>If none of the above identifiers are used, the global is externally
580 visible, meaning that it participates in linkage and can be used to resolve
581 external symbol references.
586 The next two types of linkage are targeted for Microsoft Windows platform
587 only. They are designed to support importing (exporting) symbols from (to)
588 DLLs (Dynamic Link Libraries).
592 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt>: </dt>
594 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
595 or variable via a global pointer to a pointer that is set up by the DLL
596 exporting the symbol. On Microsoft Windows targets, the pointer name is
597 formed by combining <code>__imp_</code> and the function or variable name.
600 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt>: </dt>
602 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
603 pointer to a pointer in a DLL, so that it can be referenced with the
604 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
605 name is formed by combining <code>__imp_</code> and the function or variable
611 <p>For example, since the "<tt>.LC0</tt>"
612 variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
613 variable and was linked with this one, one of the two would be renamed,
614 preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
615 external (i.e., lacking any linkage declarations), they are accessible
616 outside of the current module.</p>
617 <p>It is illegal for a function <i>declaration</i>
618 to have any linkage type other than "externally visible", <tt>dllimport</tt>
619 or <tt>extern_weak</tt>.</p>
620 <p>Aliases can have only <tt>external</tt>, <tt>internal</tt>, <tt>weak</tt>
621 or <tt>weak_odr</tt> linkages.</p>
624 <!-- ======================================================================= -->
625 <div class="doc_subsection">
626 <a name="callingconv">Calling Conventions</a>
629 <div class="doc_text">
631 <p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
632 and <a href="#i_invoke">invokes</a> can all have an optional calling convention
633 specified for the call. The calling convention of any pair of dynamic
634 caller/callee must match, or the behavior of the program is undefined. The
635 following calling conventions are supported by LLVM, and more may be added in
639 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
641 <dd>This calling convention (the default if no other calling convention is
642 specified) matches the target C calling conventions. This calling convention
643 supports varargs function calls and tolerates some mismatch in the declared
644 prototype and implemented declaration of the function (as does normal C).
647 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
649 <dd>This calling convention attempts to make calls as fast as possible
650 (e.g. by passing things in registers). This calling convention allows the
651 target to use whatever tricks it wants to produce fast code for the target,
652 without having to conform to an externally specified ABI (Application Binary
653 Interface). Implementations of this convention should allow arbitrary
654 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> to be
655 supported. This calling convention does not support varargs and requires the
656 prototype of all callees to exactly match the prototype of the function
660 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
662 <dd>This calling convention attempts to make code in the caller as efficient
663 as possible under the assumption that the call is not commonly executed. As
664 such, these calls often preserve all registers so that the call does not break
665 any live ranges in the caller side. This calling convention does not support
666 varargs and requires the prototype of all callees to exactly match the
667 prototype of the function definition.
670 <dt><b>"<tt>cc <<em>n</em>></tt>" - Numbered convention</b>:</dt>
672 <dd>Any calling convention may be specified by number, allowing
673 target-specific calling conventions to be used. Target specific calling
674 conventions start at 64.
678 <p>More calling conventions can be added/defined on an as-needed basis, to
679 support pascal conventions or any other well-known target-independent
684 <!-- ======================================================================= -->
685 <div class="doc_subsection">
686 <a name="visibility">Visibility Styles</a>
689 <div class="doc_text">
692 All Global Variables and Functions have one of the following visibility styles:
696 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
698 <dd>On targets that use the ELF object file format, default visibility means
699 that the declaration is visible to other
700 modules and, in shared libraries, means that the declared entity may be
701 overridden. On Darwin, default visibility means that the declaration is
702 visible to other modules. Default visibility corresponds to "external
703 linkage" in the language.
706 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
708 <dd>Two declarations of an object with hidden visibility refer to the same
709 object if they are in the same shared object. Usually, hidden visibility
710 indicates that the symbol will not be placed into the dynamic symbol table,
711 so no other module (executable or shared library) can reference it
715 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
717 <dd>On ELF, protected visibility indicates that the symbol will be placed in
718 the dynamic symbol table, but that references within the defining module will
719 bind to the local symbol. That is, the symbol cannot be overridden by another
726 <!-- ======================================================================= -->
727 <div class="doc_subsection">
728 <a name="namedtypes">Named Types</a>
731 <div class="doc_text">
733 <p>LLVM IR allows you to specify name aliases for certain types. This can make
734 it easier to read the IR and make the IR more condensed (particularly when
735 recursive types are involved). An example of a name specification is:
738 <div class="doc_code">
740 %mytype = type { %mytype*, i32 }
744 <p>You may give a name to any <a href="#typesystem">type</a> except "<a
745 href="t_void">void</a>". Type name aliases may be used anywhere a type is
746 expected with the syntax "%mytype".</p>
748 <p>Note that type names are aliases for the structural type that they indicate,
749 and that you can therefore specify multiple names for the same type. This often
750 leads to confusing behavior when dumping out a .ll file. Since LLVM IR uses
751 structural typing, the name is not part of the type. When printing out LLVM IR,
752 the printer will pick <em>one name</em> to render all types of a particular
753 shape. This means that if you have code where two different source types end up
754 having the same LLVM type, that the dumper will sometimes print the "wrong" or
755 unexpected type. This is an important design point and isn't going to
760 <!-- ======================================================================= -->
761 <div class="doc_subsection">
762 <a name="globalvars">Global Variables</a>
765 <div class="doc_text">
767 <p>Global variables define regions of memory allocated at compilation time
768 instead of run-time. Global variables may optionally be initialized, may have
769 an explicit section to be placed in, and may have an optional explicit alignment
770 specified. A variable may be defined as "thread_local", which means that it
771 will not be shared by threads (each thread will have a separated copy of the
772 variable). A variable may be defined as a global "constant," which indicates
773 that the contents of the variable will <b>never</b> be modified (enabling better
774 optimization, allowing the global data to be placed in the read-only section of
775 an executable, etc). Note that variables that need runtime initialization
776 cannot be marked "constant" as there is a store to the variable.</p>
779 LLVM explicitly allows <em>declarations</em> of global variables to be marked
780 constant, even if the final definition of the global is not. This capability
781 can be used to enable slightly better optimization of the program, but requires
782 the language definition to guarantee that optimizations based on the
783 'constantness' are valid for the translation units that do not include the
787 <p>As SSA values, global variables define pointer values that are in
788 scope (i.e. they dominate) all basic blocks in the program. Global
789 variables always define a pointer to their "content" type because they
790 describe a region of memory, and all memory objects in LLVM are
791 accessed through pointers.</p>
793 <p>A global variable may be declared to reside in a target-specifc numbered
794 address space. For targets that support them, address spaces may affect how
795 optimizations are performed and/or what target instructions are used to access
796 the variable. The default address space is zero. The address space qualifier
797 must precede any other attributes.</p>
799 <p>LLVM allows an explicit section to be specified for globals. If the target
800 supports it, it will emit globals to the section specified.</p>
802 <p>An explicit alignment may be specified for a global. If not present, or if
803 the alignment is set to zero, the alignment of the global is set by the target
804 to whatever it feels convenient. If an explicit alignment is specified, the
805 global is forced to have at least that much alignment. All alignments must be
808 <p>For example, the following defines a global in a numbered address space with
809 an initializer, section, and alignment:</p>
811 <div class="doc_code">
813 @G = addrspace(5) constant float 1.0, section "foo", align 4
820 <!-- ======================================================================= -->
821 <div class="doc_subsection">
822 <a name="functionstructure">Functions</a>
825 <div class="doc_text">
827 <p>LLVM function definitions consist of the "<tt>define</tt>" keyord,
828 an optional <a href="#linkage">linkage type</a>, an optional
829 <a href="#visibility">visibility style</a>, an optional
830 <a href="#callingconv">calling convention</a>, a return type, an optional
831 <a href="#paramattrs">parameter attribute</a> for the return type, a function
832 name, a (possibly empty) argument list (each with optional
833 <a href="#paramattrs">parameter attributes</a>), optional
834 <a href="#fnattrs">function attributes</a>, an optional section,
835 an optional alignment, an optional <a href="#gc">garbage collector name</a>,
836 an opening curly brace, a list of basic blocks, and a closing curly brace.
838 LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
839 optional <a href="#linkage">linkage type</a>, an optional
840 <a href="#visibility">visibility style</a>, an optional
841 <a href="#callingconv">calling convention</a>, a return type, an optional
842 <a href="#paramattrs">parameter attribute</a> for the return type, a function
843 name, a possibly empty list of arguments, an optional alignment, and an optional
844 <a href="#gc">garbage collector name</a>.</p>
846 <p>A function definition contains a list of basic blocks, forming the CFG
847 (Control Flow Graph) for
848 the function. Each basic block may optionally start with a label (giving the
849 basic block a symbol table entry), contains a list of instructions, and ends
850 with a <a href="#terminators">terminator</a> instruction (such as a branch or
851 function return).</p>
853 <p>The first basic block in a function is special in two ways: it is immediately
854 executed on entrance to the function, and it is not allowed to have predecessor
855 basic blocks (i.e. there can not be any branches to the entry block of a
856 function). Because the block can have no predecessors, it also cannot have any
857 <a href="#i_phi">PHI nodes</a>.</p>
859 <p>LLVM allows an explicit section to be specified for functions. If the target
860 supports it, it will emit functions to the section specified.</p>
862 <p>An explicit alignment may be specified for a function. If not present, or if
863 the alignment is set to zero, the alignment of the function is set by the target
864 to whatever it feels convenient. If an explicit alignment is specified, the
865 function is forced to have at least that much alignment. All alignments must be
870 <div class="doc_code">
872 define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
873 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
874 <ResultType> @<FunctionName> ([argument list])
875 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
876 [<a href="#gc">gc</a>] { ... }
883 <!-- ======================================================================= -->
884 <div class="doc_subsection">
885 <a name="aliasstructure">Aliases</a>
887 <div class="doc_text">
888 <p>Aliases act as "second name" for the aliasee value (which can be either
889 function, global variable, another alias or bitcast of global value). Aliases
890 may have an optional <a href="#linkage">linkage type</a>, and an
891 optional <a href="#visibility">visibility style</a>.</p>
895 <div class="doc_code">
897 @<Name> = alias [Linkage] [Visibility] <AliaseeTy> @<Aliasee>
905 <!-- ======================================================================= -->
906 <div class="doc_subsection"><a name="paramattrs">Parameter Attributes</a></div>
907 <div class="doc_text">
908 <p>The return type and each parameter of a function type may have a set of
909 <i>parameter attributes</i> associated with them. Parameter attributes are
910 used to communicate additional information about the result or parameters of
911 a function. Parameter attributes are considered to be part of the function,
912 not of the function type, so functions with different parameter attributes
913 can have the same function type.</p>
915 <p>Parameter attributes are simple keywords that follow the type specified. If
916 multiple parameter attributes are needed, they are space separated. For
919 <div class="doc_code">
921 declare i32 @printf(i8* noalias nocapture, ...)
922 declare i32 @atoi(i8 zeroext)
923 declare signext i8 @returns_signed_char()
927 <p>Note that any attributes for the function result (<tt>nounwind</tt>,
928 <tt>readonly</tt>) come immediately after the argument list.</p>
930 <p>Currently, only the following parameter attributes are defined:</p>
932 <dt><tt>zeroext</tt></dt>
933 <dd>This indicates to the code generator that the parameter or return value
934 should be zero-extended to a 32-bit value by the caller (for a parameter)
935 or the callee (for a return value).</dd>
937 <dt><tt>signext</tt></dt>
938 <dd>This indicates to the code generator that the parameter or return value
939 should be sign-extended to a 32-bit value by the caller (for a parameter)
940 or the callee (for a return value).</dd>
942 <dt><tt>inreg</tt></dt>
943 <dd>This indicates that this parameter or return value should be treated
944 in a special target-dependent fashion during while emitting code for a
945 function call or return (usually, by putting it in a register as opposed
946 to memory, though some targets use it to distinguish between two different
947 kinds of registers). Use of this attribute is target-specific.</dd>
949 <dt><tt><a name="byval">byval</a></tt></dt>
950 <dd>This indicates that the pointer parameter should really be passed by
951 value to the function. The attribute implies that a hidden copy of the
952 pointee is made between the caller and the callee, so the callee is unable
953 to modify the value in the callee. This attribute is only valid on LLVM
954 pointer arguments. It is generally used to pass structs and arrays by
955 value, but is also valid on pointers to scalars. The copy is considered to
956 belong to the caller not the callee (for example,
957 <tt><a href="#readonly">readonly</a></tt> functions should not write to
958 <tt>byval</tt> parameters). This is not a valid attribute for return
959 values. The byval attribute also supports specifying an alignment with the
960 align attribute. This has a target-specific effect on the code generator
961 that usually indicates a desired alignment for the synthesized stack
964 <dt><tt>sret</tt></dt>
965 <dd>This indicates that the pointer parameter specifies the address of a
966 structure that is the return value of the function in the source program.
967 This pointer must be guaranteed by the caller to be valid: loads and stores
968 to the structure may be assumed by the callee to not to trap. This may only
969 be applied to the first parameter. This is not a valid attribute for
972 <dt><tt>noalias</tt></dt>
973 <dd>This indicates that the pointer does not alias any global or any other
974 parameter. The caller is responsible for ensuring that this is the
975 case. On a function return value, <tt>noalias</tt> additionally indicates
976 that the pointer does not alias any other pointers visible to the
977 caller. For further details, please see the discussion of the NoAlias
979 <a href="http://llvm.org/docs/AliasAnalysis.html#MustMayNo">alias
982 <dt><tt>nocapture</tt></dt>
983 <dd>This indicates that the callee does not make any copies of the pointer
984 that outlive the callee itself. This is not a valid attribute for return
987 <dt><tt>nest</tt></dt>
988 <dd>This indicates that the pointer parameter can be excised using the
989 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
990 attribute for return values.</dd>
995 <!-- ======================================================================= -->
996 <div class="doc_subsection">
997 <a name="gc">Garbage Collector Names</a>
1000 <div class="doc_text">
1001 <p>Each function may specify a garbage collector name, which is simply a
1004 <div class="doc_code"><pre
1005 >define void @f() gc "name" { ...</pre></div>
1007 <p>The compiler declares the supported values of <i>name</i>. Specifying a
1008 collector which will cause the compiler to alter its output in order to support
1009 the named garbage collection algorithm.</p>
1012 <!-- ======================================================================= -->
1013 <div class="doc_subsection">
1014 <a name="fnattrs">Function Attributes</a>
1017 <div class="doc_text">
1019 <p>Function attributes are set to communicate additional information about
1020 a function. Function attributes are considered to be part of the function,
1021 not of the function type, so functions with different parameter attributes
1022 can have the same function type.</p>
1024 <p>Function attributes are simple keywords that follow the type specified. If
1025 multiple attributes are needed, they are space separated. For
1028 <div class="doc_code">
1030 define void @f() noinline { ... }
1031 define void @f() alwaysinline { ... }
1032 define void @f() alwaysinline optsize { ... }
1033 define void @f() optsize
1038 <dt><tt>alwaysinline</tt></dt>
1039 <dd>This attribute indicates that the inliner should attempt to inline this
1040 function into callers whenever possible, ignoring any active inlining size
1041 threshold for this caller.</dd>
1043 <dt><tt>noinline</tt></dt>
1044 <dd>This attribute indicates that the inliner should never inline this function
1045 in any situation. This attribute may not be used together with the
1046 <tt>alwaysinline</tt> attribute.</dd>
1048 <dt><tt>optsize</tt></dt>
1049 <dd>This attribute suggests that optimization passes and code generator passes
1050 make choices that keep the code size of this function low, and otherwise do
1051 optimizations specifically to reduce code size.</dd>
1053 <dt><tt>noreturn</tt></dt>
1054 <dd>This function attribute indicates that the function never returns normally.
1055 This produces undefined behavior at runtime if the function ever does
1056 dynamically return.</dd>
1058 <dt><tt>nounwind</tt></dt>
1059 <dd>This function attribute indicates that the function never returns with an
1060 unwind or exceptional control flow. If the function does unwind, its runtime
1061 behavior is undefined.</dd>
1063 <dt><tt>readnone</tt></dt>
1064 <dd>This attribute indicates that the function computes its result (or the
1065 exception it throws) based strictly on its arguments, without dereferencing any
1066 pointer arguments or otherwise accessing any mutable state (e.g. memory, control
1067 registers, etc) visible to caller functions. It does not write through any
1068 pointer arguments (including <tt><a href="#byval">byval</a></tt> arguments) and
1069 never changes any state visible to callers.</dd>
1071 <dt><tt><a name="readonly">readonly</a></tt></dt>
1072 <dd>This attribute indicates that the function does not write through any
1073 pointer arguments (including <tt><a href="#byval">byval</a></tt> arguments)
1074 or otherwise modify any state (e.g. memory, control registers, etc) visible to
1075 caller functions. It may dereference pointer arguments and read state that may
1076 be set in the caller. A readonly function always returns the same value (or
1077 throws the same exception) when called with the same set of arguments and global
1080 <dt><tt><a name="ssp">ssp</a></tt></dt>
1081 <dd>This attribute indicates that the function should emit a stack smashing
1082 protector. It is in the form of a "canary"—a random value placed on the
1083 stack before the local variables that's checked upon return from the function to
1084 see if it has been overwritten. A heuristic is used to determine if a function
1085 needs stack protectors or not.
1087 <p>If a function that has an <tt>ssp</tt> attribute is inlined into a function
1088 that doesn't have an <tt>ssp</tt> attribute, then the resulting function will
1089 have an <tt>ssp</tt> attribute.</p></dd>
1091 <dt><tt>sspreq</tt></dt>
1092 <dd>This attribute indicates that the function should <em>always</em> emit a
1093 stack smashing protector. This overrides the <tt><a href="#ssp">ssp</a></tt>
1096 <p>If a function that has an <tt>sspreq</tt> attribute is inlined into a
1097 function that doesn't have an <tt>sspreq</tt> attribute or which has
1098 an <tt>ssp</tt> attribute, then the resulting function will have
1099 an <tt>sspreq</tt> attribute.</p></dd>
1104 <!-- ======================================================================= -->
1105 <div class="doc_subsection">
1106 <a name="moduleasm">Module-Level Inline Assembly</a>
1109 <div class="doc_text">
1111 Modules may contain "module-level inline asm" blocks, which corresponds to the
1112 GCC "file scope inline asm" blocks. These blocks are internally concatenated by
1113 LLVM and treated as a single unit, but may be separated in the .ll file if
1114 desired. The syntax is very simple:
1117 <div class="doc_code">
1119 module asm "inline asm code goes here"
1120 module asm "more can go here"
1124 <p>The strings can contain any character by escaping non-printable characters.
1125 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
1130 The inline asm code is simply printed to the machine code .s file when
1131 assembly code is generated.
1135 <!-- ======================================================================= -->
1136 <div class="doc_subsection">
1137 <a name="datalayout">Data Layout</a>
1140 <div class="doc_text">
1141 <p>A module may specify a target specific data layout string that specifies how
1142 data is to be laid out in memory. The syntax for the data layout is simply:</p>
1143 <pre> target datalayout = "<i>layout specification</i>"</pre>
1144 <p>The <i>layout specification</i> consists of a list of specifications
1145 separated by the minus sign character ('-'). Each specification starts with a
1146 letter and may include other information after the letter to define some
1147 aspect of the data layout. The specifications accepted are as follows: </p>
1150 <dd>Specifies that the target lays out data in big-endian form. That is, the
1151 bits with the most significance have the lowest address location.</dd>
1153 <dd>Specifies that the target lays out data in little-endian form. That is,
1154 the bits with the least significance have the lowest address location.</dd>
1155 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1156 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
1157 <i>preferred</i> alignments. All sizes are in bits. Specifying the <i>pref</i>
1158 alignment is optional. If omitted, the preceding <tt>:</tt> should be omitted
1160 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1161 <dd>This specifies the alignment for an integer type of a given bit
1162 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1163 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1164 <dd>This specifies the alignment for a vector type of a given bit
1166 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1167 <dd>This specifies the alignment for a floating point type of a given bit
1168 <i>size</i>. The value of <i>size</i> must be either 32 (float) or 64
1170 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1171 <dd>This specifies the alignment for an aggregate type of a given bit
1174 <p>When constructing the data layout for a given target, LLVM starts with a
1175 default set of specifications which are then (possibly) overriden by the
1176 specifications in the <tt>datalayout</tt> keyword. The default specifications
1177 are given in this list:</p>
1179 <li><tt>E</tt> - big endian</li>
1180 <li><tt>p:32:64:64</tt> - 32-bit pointers with 64-bit alignment</li>
1181 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1182 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1183 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1184 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
1185 <li><tt>i64:32:64</tt> - i64 has ABI alignment of 32-bits but preferred
1186 alignment of 64-bits</li>
1187 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1188 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1189 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1190 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1191 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
1193 <p>When LLVM is determining the alignment for a given type, it uses the
1194 following rules:</p>
1196 <li>If the type sought is an exact match for one of the specifications, that
1197 specification is used.</li>
1198 <li>If no match is found, and the type sought is an integer type, then the
1199 smallest integer type that is larger than the bitwidth of the sought type is
1200 used. If none of the specifications are larger than the bitwidth then the the
1201 largest integer type is used. For example, given the default specifications
1202 above, the i7 type will use the alignment of i8 (next largest) while both
1203 i65 and i256 will use the alignment of i64 (largest specified).</li>
1204 <li>If no match is found, and the type sought is a vector type, then the
1205 largest vector type that is smaller than the sought vector type will be used
1206 as a fall back. This happens because <128 x double> can be implemented
1207 in terms of 64 <2 x double>, for example.</li>
1211 <!-- *********************************************************************** -->
1212 <div class="doc_section"> <a name="typesystem">Type System</a> </div>
1213 <!-- *********************************************************************** -->
1215 <div class="doc_text">
1217 <p>The LLVM type system is one of the most important features of the
1218 intermediate representation. Being typed enables a number of
1219 optimizations to be performed on the intermediate representation directly,
1220 without having to do
1221 extra analyses on the side before the transformation. A strong type
1222 system makes it easier to read the generated code and enables novel
1223 analyses and transformations that are not feasible to perform on normal
1224 three address code representations.</p>
1228 <!-- ======================================================================= -->
1229 <div class="doc_subsection"> <a name="t_classifications">Type
1230 Classifications</a> </div>
1231 <div class="doc_text">
1232 <p>The types fall into a few useful
1233 classifications:</p>
1235 <table border="1" cellspacing="0" cellpadding="4">
1237 <tr><th>Classification</th><th>Types</th></tr>
1239 <td><a href="#t_integer">integer</a></td>
1240 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
1243 <td><a href="#t_floating">floating point</a></td>
1244 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
1247 <td><a name="t_firstclass">first class</a></td>
1248 <td><a href="#t_integer">integer</a>,
1249 <a href="#t_floating">floating point</a>,
1250 <a href="#t_pointer">pointer</a>,
1251 <a href="#t_vector">vector</a>,
1252 <a href="#t_struct">structure</a>,
1253 <a href="#t_array">array</a>,
1254 <a href="#t_label">label</a>.
1258 <td><a href="#t_primitive">primitive</a></td>
1259 <td><a href="#t_label">label</a>,
1260 <a href="#t_void">void</a>,
1261 <a href="#t_floating">floating point</a>.</td>
1264 <td><a href="#t_derived">derived</a></td>
1265 <td><a href="#t_integer">integer</a>,
1266 <a href="#t_array">array</a>,
1267 <a href="#t_function">function</a>,
1268 <a href="#t_pointer">pointer</a>,
1269 <a href="#t_struct">structure</a>,
1270 <a href="#t_pstruct">packed structure</a>,
1271 <a href="#t_vector">vector</a>,
1272 <a href="#t_opaque">opaque</a>.
1278 <p>The <a href="#t_firstclass">first class</a> types are perhaps the
1279 most important. Values of these types are the only ones which can be
1280 produced by instructions, passed as arguments, or used as operands to
1284 <!-- ======================================================================= -->
1285 <div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
1287 <div class="doc_text">
1288 <p>The primitive types are the fundamental building blocks of the LLVM
1293 <!-- _______________________________________________________________________ -->
1294 <div class="doc_subsubsection"> <a name="t_floating">Floating Point Types</a> </div>
1296 <div class="doc_text">
1299 <tr><th>Type</th><th>Description</th></tr>
1300 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1301 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1302 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1303 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1304 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1309 <!-- _______________________________________________________________________ -->
1310 <div class="doc_subsubsection"> <a name="t_void">Void Type</a> </div>
1312 <div class="doc_text">
1314 <p>The void type does not represent any value and has no size.</p>
1323 <!-- _______________________________________________________________________ -->
1324 <div class="doc_subsubsection"> <a name="t_label">Label Type</a> </div>
1326 <div class="doc_text">
1328 <p>The label type represents code labels.</p>
1338 <!-- ======================================================================= -->
1339 <div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
1341 <div class="doc_text">
1343 <p>The real power in LLVM comes from the derived types in the system.
1344 This is what allows a programmer to represent arrays, functions,
1345 pointers, and other useful types. Note that these derived types may be
1346 recursive: For example, it is possible to have a two dimensional array.</p>
1350 <!-- _______________________________________________________________________ -->
1351 <div class="doc_subsubsection"> <a name="t_integer">Integer Type</a> </div>
1353 <div class="doc_text">
1356 <p>The integer type is a very simple derived type that simply specifies an
1357 arbitrary bit width for the integer type desired. Any bit width from 1 bit to
1358 2^23-1 (about 8 million) can be specified.</p>
1366 <p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1370 <table class="layout">
1373 <td><tt>i1</tt></td>
1374 <td>a single-bit integer.</td>
1376 <td><tt>i32</tt></td>
1377 <td>a 32-bit integer.</td>
1379 <td><tt>i1942652</tt></td>
1380 <td>a really big integer of over 1 million bits.</td>
1385 <p>Note that the code generator does not yet support large integer types
1386 to be used as function return types. The specific limit on how large a
1387 return type the code generator can currently handle is target-dependent;
1388 currently it's often 64 bits for 32-bit targets and 128 bits for 64-bit
1393 <!-- _______________________________________________________________________ -->
1394 <div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
1396 <div class="doc_text">
1400 <p>The array type is a very simple derived type that arranges elements
1401 sequentially in memory. The array type requires a size (number of
1402 elements) and an underlying data type.</p>
1407 [<# elements> x <elementtype>]
1410 <p>The number of elements is a constant integer value; elementtype may
1411 be any type with a size.</p>
1414 <table class="layout">
1416 <td class="left"><tt>[40 x i32]</tt></td>
1417 <td class="left">Array of 40 32-bit integer values.</td>
1420 <td class="left"><tt>[41 x i32]</tt></td>
1421 <td class="left">Array of 41 32-bit integer values.</td>
1424 <td class="left"><tt>[4 x i8]</tt></td>
1425 <td class="left">Array of 4 8-bit integer values.</td>
1428 <p>Here are some examples of multidimensional arrays:</p>
1429 <table class="layout">
1431 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1432 <td class="left">3x4 array of 32-bit integer values.</td>
1435 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1436 <td class="left">12x10 array of single precision floating point values.</td>
1439 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1440 <td class="left">2x3x4 array of 16-bit integer values.</td>
1444 <p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
1445 length array. Normally, accesses past the end of an array are undefined in
1446 LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
1447 As a special case, however, zero length arrays are recognized to be variable
1448 length. This allows implementation of 'pascal style arrays' with the LLVM
1449 type "{ i32, [0 x float]}", for example.</p>
1451 <p>Note that the code generator does not yet support large aggregate types
1452 to be used as function return types. The specific limit on how large an
1453 aggregate return type the code generator can currently handle is
1454 target-dependent, and also dependent on the aggregate element types.</p>
1458 <!-- _______________________________________________________________________ -->
1459 <div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
1460 <div class="doc_text">
1464 <p>The function type can be thought of as a function signature. It
1465 consists of a return type and a list of formal parameter types. The
1466 return type of a function type is a scalar type, a void type, or a struct type.
1467 If the return type is a struct type then all struct elements must be of first
1468 class types, and the struct must have at least one element.</p>
1473 <returntype list> (<parameter list>)
1476 <p>...where '<tt><parameter list></tt>' is a comma-separated list of type
1477 specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1478 which indicates that the function takes a variable number of arguments.
1479 Variable argument functions can access their arguments with the <a
1480 href="#int_varargs">variable argument handling intrinsic</a> functions.
1481 '<tt><returntype list></tt>' is a comma-separated list of
1482 <a href="#t_firstclass">first class</a> type specifiers.</p>
1485 <table class="layout">
1487 <td class="left"><tt>i32 (i32)</tt></td>
1488 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
1490 </tr><tr class="layout">
1491 <td class="left"><tt>float (i16 signext, i32 *) *
1493 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
1494 an <tt>i16</tt> that should be sign extended and a
1495 <a href="#t_pointer">pointer</a> to <tt>i32</tt>, returning
1498 </tr><tr class="layout">
1499 <td class="left"><tt>i32 (i8*, ...)</tt></td>
1500 <td class="left">A vararg function that takes at least one
1501 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1502 which returns an integer. This is the signature for <tt>printf</tt> in
1505 </tr><tr class="layout">
1506 <td class="left"><tt>{i32, i32} (i32)</tt></td>
1507 <td class="left">A function taking an <tt>i32</tt>, returning two
1508 <tt>i32</tt> values as an aggregate of type <tt>{ i32, i32 }</tt>
1514 <!-- _______________________________________________________________________ -->
1515 <div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
1516 <div class="doc_text">
1518 <p>The structure type is used to represent a collection of data members
1519 together in memory. The packing of the field types is defined to match
1520 the ABI of the underlying processor. The elements of a structure may
1521 be any type that has a size.</p>
1522 <p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1523 and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1524 field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1527 <pre> { <type list> }<br></pre>
1529 <table class="layout">
1531 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1532 <td class="left">A triple of three <tt>i32</tt> values</td>
1533 </tr><tr class="layout">
1534 <td class="left"><tt>{ float, i32 (i32) * }</tt></td>
1535 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1536 second element is a <a href="#t_pointer">pointer</a> to a
1537 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1538 an <tt>i32</tt>.</td>
1542 <p>Note that the code generator does not yet support large aggregate types
1543 to be used as function return types. The specific limit on how large an
1544 aggregate return type the code generator can currently handle is
1545 target-dependent, and also dependent on the aggregate element types.</p>
1549 <!-- _______________________________________________________________________ -->
1550 <div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
1552 <div class="doc_text">
1554 <p>The packed structure type is used to represent a collection of data members
1555 together in memory. There is no padding between fields. Further, the alignment
1556 of a packed structure is 1 byte. The elements of a packed structure may
1557 be any type that has a size.</p>
1558 <p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1559 and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1560 field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1563 <pre> < { <type list> } > <br></pre>
1565 <table class="layout">
1567 <td class="left"><tt>< { i32, i32, i32 } ></tt></td>
1568 <td class="left">A triple of three <tt>i32</tt> values</td>
1569 </tr><tr class="layout">
1571 <tt>< { float, i32 (i32)* } ></tt></td>
1572 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1573 second element is a <a href="#t_pointer">pointer</a> to a
1574 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1575 an <tt>i32</tt>.</td>
1580 <!-- _______________________________________________________________________ -->
1581 <div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
1582 <div class="doc_text">
1584 <p>As in many languages, the pointer type represents a pointer or
1585 reference to another object, which must live in memory. Pointer types may have
1586 an optional address space attribute defining the target-specific numbered
1587 address space where the pointed-to object resides. The default address space is
1590 <p>Note that LLVM does not permit pointers to void (<tt>void*</tt>) nor does
1591 it permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</tt> instead.</p>
1594 <pre> <type> *<br></pre>
1596 <table class="layout">
1598 <td class="left"><tt>[4 x i32]*</tt></td>
1599 <td class="left">A <a href="#t_pointer">pointer</a> to <a
1600 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
1603 <td class="left"><tt>i32 (i32 *) *</tt></td>
1604 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
1605 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
1609 <td class="left"><tt>i32 addrspace(5)*</tt></td>
1610 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
1611 that resides in address space #5.</td>
1616 <!-- _______________________________________________________________________ -->
1617 <div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
1618 <div class="doc_text">
1622 <p>A vector type is a simple derived type that represents a vector
1623 of elements. Vector types are used when multiple primitive data
1624 are operated in parallel using a single instruction (SIMD).
1625 A vector type requires a size (number of
1626 elements) and an underlying primitive data type. Vectors must have a power
1627 of two length (1, 2, 4, 8, 16 ...). Vector types are
1628 considered <a href="#t_firstclass">first class</a>.</p>
1633 < <# elements> x <elementtype> >
1636 <p>The number of elements is a constant integer value; elementtype may
1637 be any integer or floating point type.</p>
1641 <table class="layout">
1643 <td class="left"><tt><4 x i32></tt></td>
1644 <td class="left">Vector of 4 32-bit integer values.</td>
1647 <td class="left"><tt><8 x float></tt></td>
1648 <td class="left">Vector of 8 32-bit floating-point values.</td>
1651 <td class="left"><tt><2 x i64></tt></td>
1652 <td class="left">Vector of 2 64-bit integer values.</td>
1656 <p>Note that the code generator does not yet support large vector types
1657 to be used as function return types. The specific limit on how large a
1658 vector return type codegen can currently handle is target-dependent;
1659 currently it's often a few times longer than a hardware vector register.</p>
1663 <!-- _______________________________________________________________________ -->
1664 <div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1665 <div class="doc_text">
1669 <p>Opaque types are used to represent unknown types in the system. This
1670 corresponds (for example) to the C notion of a forward declared structure type.
1671 In LLVM, opaque types can eventually be resolved to any type (not just a
1672 structure type).</p>
1682 <table class="layout">
1684 <td class="left"><tt>opaque</tt></td>
1685 <td class="left">An opaque type.</td>
1690 <!-- ======================================================================= -->
1691 <div class="doc_subsection">
1692 <a name="t_uprefs">Type Up-references</a>
1695 <div class="doc_text">
1698 An "up reference" allows you to refer to a lexically enclosing type without
1699 requiring it to have a name. For instance, a structure declaration may contain a
1700 pointer to any of the types it is lexically a member of. Example of up
1701 references (with their equivalent as named type declarations) include:</p>
1704 { \2 * } %x = type { %x* }
1705 { \2 }* %y = type { %y }*
1710 An up reference is needed by the asmprinter for printing out cyclic types when
1711 there is no declared name for a type in the cycle. Because the asmprinter does
1712 not want to print out an infinite type string, it needs a syntax to handle
1713 recursive types that have no names (all names are optional in llvm IR).
1722 The level is the count of the lexical type that is being referred to.
1727 <table class="layout">
1729 <td class="left"><tt>\1*</tt></td>
1730 <td class="left">Self-referential pointer.</td>
1733 <td class="left"><tt>{ { \3*, i8 }, i32 }</tt></td>
1734 <td class="left">Recursive structure where the upref refers to the out-most
1741 <!-- *********************************************************************** -->
1742 <div class="doc_section"> <a name="constants">Constants</a> </div>
1743 <!-- *********************************************************************** -->
1745 <div class="doc_text">
1747 <p>LLVM has several different basic types of constants. This section describes
1748 them all and their syntax.</p>
1752 <!-- ======================================================================= -->
1753 <div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
1755 <div class="doc_text">
1758 <dt><b>Boolean constants</b></dt>
1760 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
1761 constants of the <tt><a href="#t_primitive">i1</a></tt> type.
1764 <dt><b>Integer constants</b></dt>
1766 <dd>Standard integers (such as '4') are constants of the <a
1767 href="#t_integer">integer</a> type. Negative numbers may be used with
1771 <dt><b>Floating point constants</b></dt>
1773 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
1774 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
1775 notation (see below). The assembler requires the exact decimal value of
1776 a floating-point constant. For example, the assembler accepts 1.25 but
1777 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
1778 constants must have a <a href="#t_floating">floating point</a> type. </dd>
1780 <dt><b>Null pointer constants</b></dt>
1782 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
1783 and must be of <a href="#t_pointer">pointer type</a>.</dd>
1787 <p>The one non-intuitive notation for constants is the hexadecimal form
1788 of floating point constants. For example, the form '<tt>double
1789 0x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
1790 4.5e+15</tt>'. The only time hexadecimal floating point constants are required
1791 (and the only time that they are generated by the disassembler) is when a
1792 floating point constant must be emitted but it cannot be represented as a
1793 decimal floating point number in a reasonable number of digits. For example,
1794 NaN's, infinities, and other
1795 special values are represented in their IEEE hexadecimal format so that
1796 assembly and disassembly do not cause any bits to change in the constants.</p>
1797 <p>When using the hexadecimal form, constants of types float and double are
1798 represented using the 16-digit form shown above (which matches the IEEE754
1799 representation for double); float values must, however, be exactly representable
1800 as IEE754 single precision.
1801 Hexadecimal format is always used for long
1802 double, and there are three forms of long double. The 80-bit
1803 format used by x86 is represented as <tt>0xK</tt>
1804 followed by 20 hexadecimal digits.
1805 The 128-bit format used by PowerPC (two adjacent doubles) is represented
1806 by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit
1807 format is represented
1808 by <tt>0xL</tt> followed by 32 hexadecimal digits; no currently supported
1809 target uses this format. Long doubles will only work if they match
1810 the long double format on your target. All hexadecimal formats are big-endian
1811 (sign bit at the left).</p>
1814 <!-- ======================================================================= -->
1815 <div class="doc_subsection">
1816 <a name="aggregateconstants"> <!-- old anchor -->
1817 <a name="complexconstants">Complex Constants</a></a>
1820 <div class="doc_text">
1821 <p>Complex constants are a (potentially recursive) combination of simple
1822 constants and smaller complex constants.</p>
1825 <dt><b>Structure constants</b></dt>
1827 <dd>Structure constants are represented with notation similar to structure
1828 type definitions (a comma separated list of elements, surrounded by braces
1829 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
1830 where "<tt>@G</tt>" is declared as "<tt>@G = external global i32</tt>". Structure constants
1831 must have <a href="#t_struct">structure type</a>, and the number and
1832 types of elements must match those specified by the type.
1835 <dt><b>Array constants</b></dt>
1837 <dd>Array constants are represented with notation similar to array type
1838 definitions (a comma separated list of elements, surrounded by square brackets
1839 (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74 ]</tt>". Array
1840 constants must have <a href="#t_array">array type</a>, and the number and
1841 types of elements must match those specified by the type.
1844 <dt><b>Vector constants</b></dt>
1846 <dd>Vector constants are represented with notation similar to vector type
1847 definitions (a comma separated list of elements, surrounded by
1848 less-than/greater-than's (<tt><></tt>)). For example: "<tt>< i32 42,
1849 i32 11, i32 74, i32 100 ></tt>". Vector constants must have <a
1850 href="#t_vector">vector type</a>, and the number and types of elements must
1851 match those specified by the type.
1854 <dt><b>Zero initialization</b></dt>
1856 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1857 value to zero of <em>any</em> type, including scalar and aggregate types.
1858 This is often used to avoid having to print large zero initializers (e.g. for
1859 large arrays) and is always exactly equivalent to using explicit zero
1863 <dt><b>Metadata node</b></dt>
1865 <dd>A metadata node is a structure-like constant with the type of an empty
1866 struct. For example: "<tt>{ } !{ i32 0, { } !"test" }</tt>". Unlike other
1867 constants that are meant to be interpreted as part of the instruction stream,
1868 metadata is a place to attach additional information such as debug info.
1874 <!-- ======================================================================= -->
1875 <div class="doc_subsection">
1876 <a name="globalconstants">Global Variable and Function Addresses</a>
1879 <div class="doc_text">
1881 <p>The addresses of <a href="#globalvars">global variables</a> and <a
1882 href="#functionstructure">functions</a> are always implicitly valid (link-time)
1883 constants. These constants are explicitly referenced when the <a
1884 href="#identifiers">identifier for the global</a> is used and always have <a
1885 href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1888 <div class="doc_code">
1892 @Z = global [2 x i32*] [ i32* @X, i32* @Y ]
1898 <!-- ======================================================================= -->
1899 <div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
1900 <div class="doc_text">
1901 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
1902 no specific value. Undefined values may be of any type and be used anywhere
1903 a constant is permitted.</p>
1905 <p>Undefined values indicate to the compiler that the program is well defined
1906 no matter what value is used, giving the compiler more freedom to optimize.
1910 <!-- ======================================================================= -->
1911 <div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1914 <div class="doc_text">
1916 <p>Constant expressions are used to allow expressions involving other constants
1917 to be used as constants. Constant expressions may be of any <a
1918 href="#t_firstclass">first class</a> type and may involve any LLVM operation
1919 that does not have side effects (e.g. load and call are not supported). The
1920 following is the syntax for constant expressions:</p>
1923 <dt><b><tt>trunc ( CST to TYPE )</tt></b></dt>
1924 <dd>Truncate a constant to another type. The bit size of CST must be larger
1925 than the bit size of TYPE. Both types must be integers.</dd>
1927 <dt><b><tt>zext ( CST to TYPE )</tt></b></dt>
1928 <dd>Zero extend a constant to another type. The bit size of CST must be
1929 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1931 <dt><b><tt>sext ( CST to TYPE )</tt></b></dt>
1932 <dd>Sign extend a constant to another type. The bit size of CST must be
1933 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1935 <dt><b><tt>fptrunc ( CST to TYPE )</tt></b></dt>
1936 <dd>Truncate a floating point constant to another floating point type. The
1937 size of CST must be larger than the size of TYPE. Both types must be
1938 floating point.</dd>
1940 <dt><b><tt>fpext ( CST to TYPE )</tt></b></dt>
1941 <dd>Floating point extend a constant to another type. The size of CST must be
1942 smaller or equal to the size of TYPE. Both types must be floating point.</dd>
1944 <dt><b><tt>fptoui ( CST to TYPE )</tt></b></dt>
1945 <dd>Convert a floating point constant to the corresponding unsigned integer
1946 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1947 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1948 of the same number of elements. If the value won't fit in the integer type,
1949 the results are undefined.</dd>
1951 <dt><b><tt>fptosi ( CST to TYPE )</tt></b></dt>
1952 <dd>Convert a floating point constant to the corresponding signed integer
1953 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1954 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1955 of the same number of elements. If the value won't fit in the integer type,
1956 the results are undefined.</dd>
1958 <dt><b><tt>uitofp ( CST to TYPE )</tt></b></dt>
1959 <dd>Convert an unsigned integer constant to the corresponding floating point
1960 constant. TYPE must be a scalar or vector floating point type. CST must be of
1961 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1962 of the same number of elements. If the value won't fit in the floating point
1963 type, the results are undefined.</dd>
1965 <dt><b><tt>sitofp ( CST to TYPE )</tt></b></dt>
1966 <dd>Convert a signed integer constant to the corresponding floating point
1967 constant. TYPE must be a scalar or vector floating point type. CST must be of
1968 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1969 of the same number of elements. If the value won't fit in the floating point
1970 type, the results are undefined.</dd>
1972 <dt><b><tt>ptrtoint ( CST to TYPE )</tt></b></dt>
1973 <dd>Convert a pointer typed constant to the corresponding integer constant
1974 TYPE must be an integer type. CST must be of pointer type. The CST value is
1975 zero extended, truncated, or unchanged to make it fit in TYPE.</dd>
1977 <dt><b><tt>inttoptr ( CST to TYPE )</tt></b></dt>
1978 <dd>Convert a integer constant to a pointer constant. TYPE must be a
1979 pointer type. CST must be of integer type. The CST value is zero extended,
1980 truncated, or unchanged to make it fit in a pointer size. This one is
1981 <i>really</i> dangerous!</dd>
1983 <dt><b><tt>bitcast ( CST to TYPE )</tt></b></dt>
1984 <dd>Convert a constant, CST, to another TYPE. The constraints of the operands
1985 are the same as those for the <a href="#i_bitcast">bitcast
1986 instruction</a>.</dd>
1988 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
1990 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
1991 constants. As with the <a href="#i_getelementptr">getelementptr</a>
1992 instruction, the index list may have zero or more indexes, which are required
1993 to make sense for the type of "CSTPTR".</dd>
1995 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
1997 <dd>Perform the <a href="#i_select">select operation</a> on
2000 <dt><b><tt>icmp COND ( VAL1, VAL2 )</tt></b></dt>
2001 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
2003 <dt><b><tt>fcmp COND ( VAL1, VAL2 )</tt></b></dt>
2004 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
2006 <dt><b><tt>vicmp COND ( VAL1, VAL2 )</tt></b></dt>
2007 <dd>Performs the <a href="#i_vicmp">vicmp operation</a> on constants.</dd>
2009 <dt><b><tt>vfcmp COND ( VAL1, VAL2 )</tt></b></dt>
2010 <dd>Performs the <a href="#i_vfcmp">vfcmp operation</a> on constants.</dd>
2012 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
2014 <dd>Perform the <a href="#i_extractelement">extractelement
2015 operation</a> on constants.</dd>
2017 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
2019 <dd>Perform the <a href="#i_insertelement">insertelement
2020 operation</a> on constants.</dd>
2023 <dt><b><tt>shufflevector ( VEC1, VEC2, IDXMASK )</tt></b></dt>
2025 <dd>Perform the <a href="#i_shufflevector">shufflevector
2026 operation</a> on constants.</dd>
2028 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
2030 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
2031 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
2032 binary</a> operations. The constraints on operands are the same as those for
2033 the corresponding instruction (e.g. no bitwise operations on floating point
2034 values are allowed).</dd>
2038 <!-- ======================================================================= -->
2039 <div class="doc_subsection"><a name="metadata">Embedded Metadata</a>
2042 <div class="doc_text">
2044 <p>Embedded metadata provides a way to attach arbitrary data to the
2045 instruction stream without affecting the behaviour of the program. There are
2046 two metadata primitives, strings and nodes. All metadata has the type of an
2047 empty struct and is identified in syntax by a preceding exclamation point
2051 <p>A metadata string is a string surrounded by double quotes. It can contain
2052 any character by escaping non-printable characters with "\xx" where "xx" is
2053 the two digit hex code. For example: "<tt>!"test\00"</tt>".
2056 <p>Metadata nodes are represented with notation similar to structure constants
2057 (a comma separated list of elements, surrounded by braces and preceeded by an
2058 exclamation point). For example: "<tt>!{ { } !"test\00", i32 10}</tt>".
2061 <p>Optimizations may rely on metadata to provide additional information about
2062 the program that isn't available in the instructions, or that isn't easily
2063 computable. Similarly, the code generator may expect a certain metadata format
2064 to be used to express debugging information.</p>
2067 <!-- *********************************************************************** -->
2068 <div class="doc_section"> <a name="othervalues">Other Values</a> </div>
2069 <!-- *********************************************************************** -->
2071 <!-- ======================================================================= -->
2072 <div class="doc_subsection">
2073 <a name="inlineasm">Inline Assembler Expressions</a>
2076 <div class="doc_text">
2079 LLVM supports inline assembler expressions (as opposed to <a href="#moduleasm">
2080 Module-Level Inline Assembly</a>) through the use of a special value. This
2081 value represents the inline assembler as a string (containing the instructions
2082 to emit), a list of operand constraints (stored as a string), and a flag that
2083 indicates whether or not the inline asm expression has side effects. An example
2084 inline assembler expression is:
2087 <div class="doc_code">
2089 i32 (i32) asm "bswap $0", "=r,r"
2094 Inline assembler expressions may <b>only</b> be used as the callee operand of
2095 a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we have:
2098 <div class="doc_code">
2100 %X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
2105 Inline asms with side effects not visible in the constraint list must be marked
2106 as having side effects. This is done through the use of the
2107 '<tt>sideeffect</tt>' keyword, like so:
2110 <div class="doc_code">
2112 call void asm sideeffect "eieio", ""()
2116 <p>TODO: The format of the asm and constraints string still need to be
2117 documented here. Constraints on what can be done (e.g. duplication, moving, etc
2118 need to be documented). This is probably best done by reference to another
2119 document that covers inline asm from a holistic perspective.
2124 <!-- *********************************************************************** -->
2125 <div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
2126 <!-- *********************************************************************** -->
2128 <div class="doc_text">
2130 <p>The LLVM instruction set consists of several different
2131 classifications of instructions: <a href="#terminators">terminator
2132 instructions</a>, <a href="#binaryops">binary instructions</a>,
2133 <a href="#bitwiseops">bitwise binary instructions</a>, <a
2134 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
2135 instructions</a>.</p>
2139 <!-- ======================================================================= -->
2140 <div class="doc_subsection"> <a name="terminators">Terminator
2141 Instructions</a> </div>
2143 <div class="doc_text">
2145 <p>As mentioned <a href="#functionstructure">previously</a>, every
2146 basic block in a program ends with a "Terminator" instruction, which
2147 indicates which block should be executed after the current block is
2148 finished. These terminator instructions typically yield a '<tt>void</tt>'
2149 value: they produce control flow, not values (the one exception being
2150 the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
2151 <p>There are six different terminator instructions: the '<a
2152 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
2153 instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
2154 the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
2155 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
2156 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
2160 <!-- _______________________________________________________________________ -->
2161 <div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
2162 Instruction</a> </div>
2163 <div class="doc_text">
2166 ret <type> <value> <i>; Return a value from a non-void function</i>
2167 ret void <i>; Return from void function</i>
2172 <p>The '<tt>ret</tt>' instruction is used to return control flow (and
2173 optionally a value) from a function back to the caller.</p>
2174 <p>There are two forms of the '<tt>ret</tt>' instruction: one that
2175 returns a value and then causes control flow, and one that just causes
2176 control flow to occur.</p>
2180 <p>The '<tt>ret</tt>' instruction optionally accepts a single argument,
2181 the return value. The type of the return value must be a
2182 '<a href="#t_firstclass">first class</a>' type.</p>
2184 <p>A function is not <a href="#wellformed">well formed</a> if
2185 it it has a non-void return type and contains a '<tt>ret</tt>'
2186 instruction with no return value or a return value with a type that
2187 does not match its type, or if it has a void return type and contains
2188 a '<tt>ret</tt>' instruction with a return value.</p>
2192 <p>When the '<tt>ret</tt>' instruction is executed, control flow
2193 returns back to the calling function's context. If the caller is a "<a
2194 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
2195 the instruction after the call. If the caller was an "<a
2196 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
2197 at the beginning of the "normal" destination block. If the instruction
2198 returns a value, that value shall set the call or invoke instruction's
2204 ret i32 5 <i>; Return an integer value of 5</i>
2205 ret void <i>; Return from a void function</i>
2206 ret { i32, i8 } { i32 4, i8 2 } <i>; Return a struct of values 4 and 2</i>
2209 <p>Note that the code generator does not yet fully support large
2210 return values. The specific sizes that are currently supported are
2211 dependent on the target. For integers, on 32-bit targets the limit
2212 is often 64 bits, and on 64-bit targets the limit is often 128 bits.
2213 For aggregate types, the current limits are dependent on the element
2214 types; for example targets are often limited to 2 total integer
2215 elements and 2 total floating-point elements.</p>
2218 <!-- _______________________________________________________________________ -->
2219 <div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
2220 <div class="doc_text">
2222 <pre> br i1 <cond>, label <iftrue>, label <iffalse><br> br label <dest> <i>; Unconditional branch</i>
2225 <p>The '<tt>br</tt>' instruction is used to cause control flow to
2226 transfer to a different basic block in the current function. There are
2227 two forms of this instruction, corresponding to a conditional branch
2228 and an unconditional branch.</p>
2230 <p>The conditional branch form of the '<tt>br</tt>' instruction takes a
2231 single '<tt>i1</tt>' value and two '<tt>label</tt>' values. The
2232 unconditional form of the '<tt>br</tt>' instruction takes a single
2233 '<tt>label</tt>' value as a target.</p>
2235 <p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
2236 argument is evaluated. If the value is <tt>true</tt>, control flows
2237 to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
2238 control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
2240 <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
2241 href="#i_ret">ret</a> i32 1<br>IfUnequal:<br> <a href="#i_ret">ret</a> i32 0<br></pre>
2243 <!-- _______________________________________________________________________ -->
2244 <div class="doc_subsubsection">
2245 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
2248 <div class="doc_text">
2252 switch <intty> <value>, label <defaultdest> [ <intty> <val>, label <dest> ... ]
2257 <p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
2258 several different places. It is a generalization of the '<tt>br</tt>'
2259 instruction, allowing a branch to occur to one of many possible
2265 <p>The '<tt>switch</tt>' instruction uses three parameters: an integer
2266 comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
2267 an array of pairs of comparison value constants and '<tt>label</tt>'s. The
2268 table is not allowed to contain duplicate constant entries.</p>
2272 <p>The <tt>switch</tt> instruction specifies a table of values and
2273 destinations. When the '<tt>switch</tt>' instruction is executed, this
2274 table is searched for the given value. If the value is found, control flow is
2275 transfered to the corresponding destination; otherwise, control flow is
2276 transfered to the default destination.</p>
2278 <h5>Implementation:</h5>
2280 <p>Depending on properties of the target machine and the particular
2281 <tt>switch</tt> instruction, this instruction may be code generated in different
2282 ways. For example, it could be generated as a series of chained conditional
2283 branches or with a lookup table.</p>
2288 <i>; Emulate a conditional br instruction</i>
2289 %Val = <a href="#i_zext">zext</a> i1 %value to i32
2290 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
2292 <i>; Emulate an unconditional br instruction</i>
2293 switch i32 0, label %dest [ ]
2295 <i>; Implement a jump table:</i>
2296 switch i32 %val, label %otherwise [ i32 0, label %onzero
2298 i32 2, label %ontwo ]
2302 <!-- _______________________________________________________________________ -->
2303 <div class="doc_subsubsection">
2304 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
2307 <div class="doc_text">
2312 <result> = invoke [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>] <ptr to function ty> <function ptr val>(<function args>) [<a href="#fnattrs">fn attrs</a>]
2313 to label <normal label> unwind label <exception label>
2318 <p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
2319 function, with the possibility of control flow transfer to either the
2320 '<tt>normal</tt>' label or the
2321 '<tt>exception</tt>' label. If the callee function returns with the
2322 "<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
2323 "normal" label. If the callee (or any indirect callees) returns with the "<a
2324 href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
2325 continued at the dynamically nearest "exception" label.</p>
2329 <p>This instruction requires several arguments:</p>
2333 The optional "cconv" marker indicates which <a href="#callingconv">calling
2334 convention</a> the call should use. If none is specified, the call defaults
2335 to using C calling conventions.
2338 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
2339 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>',
2340 and '<tt>inreg</tt>' attributes are valid here.</li>
2342 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
2343 function value being invoked. In most cases, this is a direct function
2344 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
2345 an arbitrary pointer to function value.
2348 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
2349 function to be invoked. </li>
2351 <li>'<tt>function args</tt>': argument list whose types match the function
2352 signature argument types. If the function signature indicates the function
2353 accepts a variable number of arguments, the extra arguments can be
2356 <li>'<tt>normal label</tt>': the label reached when the called function
2357 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
2359 <li>'<tt>exception label</tt>': the label reached when a callee returns with
2360 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
2362 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
2363 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
2364 '<tt>readnone</tt>' attributes are valid here.</li>
2369 <p>This instruction is designed to operate as a standard '<tt><a
2370 href="#i_call">call</a></tt>' instruction in most regards. The primary
2371 difference is that it establishes an association with a label, which is used by
2372 the runtime library to unwind the stack.</p>
2374 <p>This instruction is used in languages with destructors to ensure that proper
2375 cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
2376 exception. Additionally, this is important for implementation of
2377 '<tt>catch</tt>' clauses in high-level languages that support them.</p>
2381 %retval = invoke i32 @Test(i32 15) to label %Continue
2382 unwind label %TestCleanup <i>; {i32}:retval set</i>
2383 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
2384 unwind label %TestCleanup <i>; {i32}:retval set</i>
2389 <!-- _______________________________________________________________________ -->
2391 <div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
2392 Instruction</a> </div>
2394 <div class="doc_text">
2403 <p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
2404 at the first callee in the dynamic call stack which used an <a
2405 href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
2406 primarily used to implement exception handling.</p>
2410 <p>The '<tt>unwind</tt>' instruction causes execution of the current function to
2411 immediately halt. The dynamic call stack is then searched for the first <a
2412 href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
2413 execution continues at the "exceptional" destination block specified by the
2414 <tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
2415 dynamic call chain, undefined behavior results.</p>
2418 <!-- _______________________________________________________________________ -->
2420 <div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
2421 Instruction</a> </div>
2423 <div class="doc_text">
2432 <p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
2433 instruction is used to inform the optimizer that a particular portion of the
2434 code is not reachable. This can be used to indicate that the code after a
2435 no-return function cannot be reached, and other facts.</p>
2439 <p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
2444 <!-- ======================================================================= -->
2445 <div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
2446 <div class="doc_text">
2447 <p>Binary operators are used to do most of the computation in a
2448 program. They require two operands of the same type, execute an operation on them, and
2449 produce a single value. The operands might represent
2450 multiple data, as is the case with the <a href="#t_vector">vector</a> data type.
2451 The result value has the same type as its operands.</p>
2452 <p>There are several different binary operators:</p>
2454 <!-- _______________________________________________________________________ -->
2455 <div class="doc_subsubsection">
2456 <a name="i_add">'<tt>add</tt>' Instruction</a>
2459 <div class="doc_text">
2464 <result> = add <ty> <op1>, <op2> <i>; yields {ty}:result</i>
2469 <p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
2473 <p>The two arguments to the '<tt>add</tt>' instruction must be <a
2474 href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>, or
2475 <a href="#t_vector">vector</a> values. Both arguments must have identical
2480 <p>The value produced is the integer or floating point sum of the two
2483 <p>If an integer sum has unsigned overflow, the result returned is the
2484 mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2487 <p>Because LLVM integers use a two's complement representation, this
2488 instruction is appropriate for both signed and unsigned integers.</p>
2493 <result> = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
2496 <!-- _______________________________________________________________________ -->
2497 <div class="doc_subsubsection">
2498 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
2501 <div class="doc_text">
2506 <result> = sub <ty> <op1>, <op2> <i>; yields {ty}:result</i>
2511 <p>The '<tt>sub</tt>' instruction returns the difference of its two
2514 <p>Note that the '<tt>sub</tt>' instruction is used to represent the
2515 '<tt>neg</tt>' instruction present in most other intermediate
2516 representations.</p>
2520 <p>The two arguments to the '<tt>sub</tt>' instruction must be <a
2521 href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>,
2522 or <a href="#t_vector">vector</a> values. Both arguments must have identical
2527 <p>The value produced is the integer or floating point difference of
2528 the two operands.</p>
2530 <p>If an integer difference has unsigned overflow, the result returned is the
2531 mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2534 <p>Because LLVM integers use a two's complement representation, this
2535 instruction is appropriate for both signed and unsigned integers.</p>
2539 <result> = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
2540 <result> = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
2544 <!-- _______________________________________________________________________ -->
2545 <div class="doc_subsubsection">
2546 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
2549 <div class="doc_text">
2552 <pre> <result> = mul <ty> <op1>, <op2> <i>; yields {ty}:result</i>
2555 <p>The '<tt>mul</tt>' instruction returns the product of its two
2560 <p>The two arguments to the '<tt>mul</tt>' instruction must be <a
2561 href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>,
2562 or <a href="#t_vector">vector</a> values. Both arguments must have identical
2567 <p>The value produced is the integer or floating point product of the
2570 <p>If the result of an integer multiplication has unsigned overflow,
2571 the result returned is the mathematical result modulo
2572 2<sup>n</sup>, where n is the bit width of the result.</p>
2573 <p>Because LLVM integers use a two's complement representation, and the
2574 result is the same width as the operands, this instruction returns the
2575 correct result for both signed and unsigned integers. If a full product
2576 (e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands
2577 should be sign-extended or zero-extended as appropriate to the
2578 width of the full product.</p>
2580 <pre> <result> = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
2584 <!-- _______________________________________________________________________ -->
2585 <div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
2587 <div class="doc_text">
2589 <pre> <result> = udiv <ty> <op1>, <op2> <i>; yields {ty}:result</i>
2592 <p>The '<tt>udiv</tt>' instruction returns the quotient of its two
2597 <p>The two arguments to the '<tt>udiv</tt>' instruction must be
2598 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2599 values. Both arguments must have identical types.</p>
2603 <p>The value produced is the unsigned integer quotient of the two operands.</p>
2604 <p>Note that unsigned integer division and signed integer division are distinct
2605 operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
2606 <p>Division by zero leads to undefined behavior.</p>
2608 <pre> <result> = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2611 <!-- _______________________________________________________________________ -->
2612 <div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
2614 <div class="doc_text">
2617 <result> = sdiv <ty> <op1>, <op2> <i>; yields {ty}:result</i>
2622 <p>The '<tt>sdiv</tt>' instruction returns the quotient of its two
2627 <p>The two arguments to the '<tt>sdiv</tt>' instruction must be
2628 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2629 values. Both arguments must have identical types.</p>
2632 <p>The value produced is the signed integer quotient of the two operands rounded towards zero.</p>
2633 <p>Note that signed integer division and unsigned integer division are distinct
2634 operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
2635 <p>Division by zero leads to undefined behavior. Overflow also leads to
2636 undefined behavior; this is a rare case, but can occur, for example,
2637 by doing a 32-bit division of -2147483648 by -1.</p>
2639 <pre> <result> = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
2642 <!-- _______________________________________________________________________ -->
2643 <div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
2644 Instruction</a> </div>
2645 <div class="doc_text">
2648 <result> = fdiv <ty> <op1>, <op2> <i>; yields {ty}:result</i>
2652 <p>The '<tt>fdiv</tt>' instruction returns the quotient of its two
2657 <p>The two arguments to the '<tt>fdiv</tt>' instruction must be
2658 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2659 of floating point values. Both arguments must have identical types.</p>
2663 <p>The value produced is the floating point quotient of the two operands.</p>
2668 <result> = fdiv float 4.0, %var <i>; yields {float}:result = 4.0 / %var</i>
2672 <!-- _______________________________________________________________________ -->
2673 <div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
2675 <div class="doc_text">
2677 <pre> <result> = urem <ty> <op1>, <op2> <i>; yields {ty}:result</i>
2680 <p>The '<tt>urem</tt>' instruction returns the remainder from the
2681 unsigned division of its two arguments.</p>
2683 <p>The two arguments to the '<tt>urem</tt>' instruction must be
2684 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2685 values. Both arguments must have identical types.</p>
2687 <p>This instruction returns the unsigned integer <i>remainder</i> of a division.
2688 This instruction always performs an unsigned division to get the remainder.</p>
2689 <p>Note that unsigned integer remainder and signed integer remainder are
2690 distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
2691 <p>Taking the remainder of a division by zero leads to undefined behavior.</p>
2693 <pre> <result> = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2697 <!-- _______________________________________________________________________ -->
2698 <div class="doc_subsubsection">
2699 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
2702 <div class="doc_text">
2707 <result> = srem <ty> <op1>, <op2> <i>; yields {ty}:result</i>
2712 <p>The '<tt>srem</tt>' instruction returns the remainder from the
2713 signed division of its two operands. This instruction can also take
2714 <a href="#t_vector">vector</a> versions of the values in which case
2715 the elements must be integers.</p>
2719 <p>The two arguments to the '<tt>srem</tt>' instruction must be
2720 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2721 values. Both arguments must have identical types.</p>
2725 <p>This instruction returns the <i>remainder</i> of a division (where the result
2726 has the same sign as the dividend, <tt>op1</tt>), not the <i>modulo</i>
2727 operator (where the result has the same sign as the divisor, <tt>op2</tt>) of
2728 a value. For more information about the difference, see <a
2729 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
2730 Math Forum</a>. For a table of how this is implemented in various languages,
2731 please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
2732 Wikipedia: modulo operation</a>.</p>
2733 <p>Note that signed integer remainder and unsigned integer remainder are
2734 distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
2735 <p>Taking the remainder of a division by zero leads to undefined behavior.
2736 Overflow also leads to undefined behavior; this is a rare case, but can occur,
2737 for example, by taking the remainder of a 32-bit division of -2147483648 by -1.
2738 (The remainder doesn't actually overflow, but this rule lets srem be
2739 implemented using instructions that return both the result of the division
2740 and the remainder.)</p>
2742 <pre> <result> = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2746 <!-- _______________________________________________________________________ -->
2747 <div class="doc_subsubsection">
2748 <a name="i_frem">'<tt>frem</tt>' Instruction</a> </div>
2750 <div class="doc_text">
2753 <pre> <result> = frem <ty> <op1>, <op2> <i>; yields {ty}:result</i>
2756 <p>The '<tt>frem</tt>' instruction returns the remainder from the
2757 division of its two operands.</p>
2759 <p>The two arguments to the '<tt>frem</tt>' instruction must be
2760 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2761 of floating point values. Both arguments must have identical types.</p>
2765 <p>This instruction returns the <i>remainder</i> of a division.
2766 The remainder has the same sign as the dividend.</p>
2771 <result> = frem float 4.0, %var <i>; yields {float}:result = 4.0 % %var</i>
2775 <!-- ======================================================================= -->
2776 <div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
2777 Operations</a> </div>
2778 <div class="doc_text">
2779 <p>Bitwise binary operators are used to do various forms of
2780 bit-twiddling in a program. They are generally very efficient
2781 instructions and can commonly be strength reduced from other
2782 instructions. They require two operands of the same type, execute an operation on them,
2783 and produce a single value. The resulting value is the same type as its operands.</p>
2786 <!-- _______________________________________________________________________ -->
2787 <div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
2788 Instruction</a> </div>
2789 <div class="doc_text">
2791 <pre> <result> = shl <ty> <op1>, <op2> <i>; yields {ty}:result</i>
2796 <p>The '<tt>shl</tt>' instruction returns the first operand shifted to
2797 the left a specified number of bits.</p>
2801 <p>Both arguments to the '<tt>shl</tt>' instruction must be the same <a
2802 href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2803 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
2807 <p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod 2<sup>n</sup>,
2808 where n is the width of the result. If <tt>op2</tt> is (statically or dynamically) negative or
2809 equal to or larger than the number of bits in <tt>op1</tt>, the result is undefined.
2810 If the arguments are vectors, each vector element of <tt>op1</tt> is shifted by the
2811 corresponding shift amount in <tt>op2</tt>.</p>
2813 <h5>Example:</h5><pre>
2814 <result> = shl i32 4, %var <i>; yields {i32}: 4 << %var</i>
2815 <result> = shl i32 4, 2 <i>; yields {i32}: 16</i>
2816 <result> = shl i32 1, 10 <i>; yields {i32}: 1024</i>
2817 <result> = shl i32 1, 32 <i>; undefined</i>
2818 <result> = shl <2 x i32> < i32 1, i32 1>, < i32 1, i32 2> <i>; yields: result=<2 x i32> < i32 2, i32 4></i>
2821 <!-- _______________________________________________________________________ -->
2822 <div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
2823 Instruction</a> </div>
2824 <div class="doc_text">
2826 <pre> <result> = lshr <ty> <op1>, <op2> <i>; yields {ty}:result</i>
2830 <p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
2831 operand shifted to the right a specified number of bits with zero fill.</p>
2834 <p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
2835 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2836 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
2840 <p>This instruction always performs a logical shift right operation. The most
2841 significant bits of the result will be filled with zero bits after the
2842 shift. If <tt>op2</tt> is (statically or dynamically) equal to or larger than
2843 the number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
2844 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding shift
2845 amount in <tt>op2</tt>.</p>
2849 <result> = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
2850 <result> = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
2851 <result> = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
2852 <result> = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
2853 <result> = lshr i32 1, 32 <i>; undefined</i>
2854 <result> = lshr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 2> <i>; yields: result=<2 x i32> < i32 0x7FFFFFFF, i32 1></i>
2858 <!-- _______________________________________________________________________ -->
2859 <div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
2860 Instruction</a> </div>
2861 <div class="doc_text">
2864 <pre> <result> = ashr <ty> <op1>, <op2> <i>; yields {ty}:result</i>
2868 <p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
2869 operand shifted to the right a specified number of bits with sign extension.</p>
2872 <p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
2873 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2874 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
2877 <p>This instruction always performs an arithmetic shift right operation,
2878 The most significant bits of the result will be filled with the sign bit
2879 of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
2880 larger than the number of bits in <tt>op1</tt>, the result is undefined. If the
2881 arguments are vectors, each vector element of <tt>op1</tt> is shifted by the
2882 corresponding shift amount in <tt>op2</tt>.</p>
2886 <result> = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
2887 <result> = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
2888 <result> = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
2889 <result> = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
2890 <result> = ashr i32 1, 32 <i>; undefined</i>
2891 <result> = ashr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 3> <i>; yields: result=<2 x i32> < i32 -1, i32 0></i>
2895 <!-- _______________________________________________________________________ -->
2896 <div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
2897 Instruction</a> </div>
2899 <div class="doc_text">
2904 <result> = and <ty> <op1>, <op2> <i>; yields {ty}:result</i>
2909 <p>The '<tt>and</tt>' instruction returns the bitwise logical and of
2910 its two operands.</p>
2914 <p>The two arguments to the '<tt>and</tt>' instruction must be
2915 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2916 values. Both arguments must have identical types.</p>
2919 <p>The truth table used for the '<tt>and</tt>' instruction is:</p>
2922 <table border="1" cellspacing="0" cellpadding="4">
2954 <result> = and i32 4, %var <i>; yields {i32}:result = 4 & %var</i>
2955 <result> = and i32 15, 40 <i>; yields {i32}:result = 8</i>
2956 <result> = and i32 4, 8 <i>; yields {i32}:result = 0</i>
2959 <!-- _______________________________________________________________________ -->
2960 <div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
2961 <div class="doc_text">
2963 <pre> <result> = or <ty> <op1>, <op2> <i>; yields {ty}:result</i>
2966 <p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
2967 or of its two operands.</p>
2970 <p>The two arguments to the '<tt>or</tt>' instruction must be
2971 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2972 values. Both arguments must have identical types.</p>
2974 <p>The truth table used for the '<tt>or</tt>' instruction is:</p>
2977 <table border="1" cellspacing="0" cellpadding="4">
3008 <pre> <result> = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
3009 <result> = or i32 15, 40 <i>; yields {i32}:result = 47</i>
3010 <result> = or i32 4, 8 <i>; yields {i32}:result = 12</i>
3013 <!-- _______________________________________________________________________ -->
3014 <div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
3015 Instruction</a> </div>
3016 <div class="doc_text">
3018 <pre> <result> = xor <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3021 <p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
3022 or of its two operands. The <tt>xor</tt> is used to implement the
3023 "one's complement" operation, which is the "~" operator in C.</p>
3025 <p>The two arguments to the '<tt>xor</tt>' instruction must be
3026 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3027 values. Both arguments must have identical types.</p>
3031 <p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
3034 <table border="1" cellspacing="0" cellpadding="4">
3066 <pre> <result> = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
3067 <result> = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
3068 <result> = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
3069 <result> = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
3073 <!-- ======================================================================= -->
3074 <div class="doc_subsection">
3075 <a name="vectorops">Vector Operations</a>
3078 <div class="doc_text">
3080 <p>LLVM supports several instructions to represent vector operations in a
3081 target-independent manner. These instructions cover the element-access and
3082 vector-specific operations needed to process vectors effectively. While LLVM
3083 does directly support these vector operations, many sophisticated algorithms
3084 will want to use target-specific intrinsics to take full advantage of a specific
3089 <!-- _______________________________________________________________________ -->
3090 <div class="doc_subsubsection">
3091 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
3094 <div class="doc_text">
3099 <result> = extractelement <n x <ty>> <val>, i32 <idx> <i>; yields <ty></i>
3105 The '<tt>extractelement</tt>' instruction extracts a single scalar
3106 element from a vector at a specified index.
3113 The first operand of an '<tt>extractelement</tt>' instruction is a
3114 value of <a href="#t_vector">vector</a> type. The second operand is
3115 an index indicating the position from which to extract the element.
3116 The index may be a variable.</p>
3121 The result is a scalar of the same type as the element type of
3122 <tt>val</tt>. Its value is the value at position <tt>idx</tt> of
3123 <tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
3124 results are undefined.
3130 %result = extractelement <4 x i32> %vec, i32 0 <i>; yields i32</i>
3135 <!-- _______________________________________________________________________ -->
3136 <div class="doc_subsubsection">
3137 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
3140 <div class="doc_text">
3145 <result> = insertelement <n x <ty>> <val>, <ty> <elt>, i32 <idx> <i>; yields <n x <ty>></i>
3151 The '<tt>insertelement</tt>' instruction inserts a scalar
3152 element into a vector at a specified index.
3159 The first operand of an '<tt>insertelement</tt>' instruction is a
3160 value of <a href="#t_vector">vector</a> type. The second operand is a
3161 scalar value whose type must equal the element type of the first
3162 operand. The third operand is an index indicating the position at
3163 which to insert the value. The index may be a variable.</p>
3168 The result is a vector of the same type as <tt>val</tt>. Its
3169 element values are those of <tt>val</tt> except at position
3170 <tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
3171 exceeds the length of <tt>val</tt>, the results are undefined.
3177 %result = insertelement <4 x i32> %vec, i32 1, i32 0 <i>; yields <4 x i32></i>
3181 <!-- _______________________________________________________________________ -->
3182 <div class="doc_subsubsection">
3183 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
3186 <div class="doc_text">
3191 <result> = shufflevector <n x <ty>> <v1>, <n x <ty>> <v2>, <m x i32> <mask> <i>; yields <m x <ty>></i>
3197 The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
3198 from two input vectors, returning a vector with the same element type as
3199 the input and length that is the same as the shuffle mask.
3205 The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
3206 with types that match each other. The third argument is a shuffle mask whose
3207 element type is always 'i32'. The result of the instruction is a vector whose
3208 length is the same as the shuffle mask and whose element type is the same as
3209 the element type of the first two operands.
3213 The shuffle mask operand is required to be a constant vector with either
3214 constant integer or undef values.
3220 The elements of the two input vectors are numbered from left to right across
3221 both of the vectors. The shuffle mask operand specifies, for each element of
3222 the result vector, which element of the two input vectors the result element
3223 gets. The element selector may be undef (meaning "don't care") and the second
3224 operand may be undef if performing a shuffle from only one vector.
3230 %result = shufflevector <4 x i32> %v1, <4 x i32> %v2,
3231 <4 x i32> <i32 0, i32 4, i32 1, i32 5> <i>; yields <4 x i32></i>
3232 %result = shufflevector <4 x i32> %v1, <4 x i32> undef,
3233 <4 x i32> <i32 0, i32 1, i32 2, i32 3> <i>; yields <4 x i32></i> - Identity shuffle.
3234 %result = shufflevector <8 x i32> %v1, <8 x i32> undef,
3235 <4 x i32> <i32 0, i32 1, i32 2, i32 3> <i>; yields <4 x i32></i>
3236 %result = shufflevector <4 x i32> %v1, <4 x i32> %v2,
3237 <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7 > <i>; yields <8 x i32></i>
3242 <!-- ======================================================================= -->
3243 <div class="doc_subsection">
3244 <a name="aggregateops">Aggregate Operations</a>
3247 <div class="doc_text">
3249 <p>LLVM supports several instructions for working with aggregate values.
3254 <!-- _______________________________________________________________________ -->
3255 <div class="doc_subsubsection">
3256 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
3259 <div class="doc_text">
3264 <result> = extractvalue <aggregate type> <val>, <idx>{, <idx>}*
3270 The '<tt>extractvalue</tt>' instruction extracts the value of a struct field
3271 or array element from an aggregate value.
3278 The first operand of an '<tt>extractvalue</tt>' instruction is a
3279 value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a>
3280 type. The operands are constant indices to specify which value to extract
3281 in a similar manner as indices in a
3282 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
3288 The result is the value at the position in the aggregate specified by
3295 %result = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
3300 <!-- _______________________________________________________________________ -->
3301 <div class="doc_subsubsection">
3302 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
3305 <div class="doc_text">
3310 <result> = insertvalue <aggregate type> <val>, <ty> <val>, <idx> <i>; yields <n x <ty>></i>
3316 The '<tt>insertvalue</tt>' instruction inserts a value
3317 into a struct field or array element in an aggregate.
3324 The first operand of an '<tt>insertvalue</tt>' instruction is a
3325 value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a> type.
3326 The second operand is a first-class value to insert.
3327 The following operands are constant indices
3328 indicating the position at which to insert the value in a similar manner as
3330 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
3331 The value to insert must have the same type as the value identified
3338 The result is an aggregate of the same type as <tt>val</tt>. Its
3339 value is that of <tt>val</tt> except that the value at the position
3340 specified by the indices is that of <tt>elt</tt>.
3346 %result = insertvalue {i32, float} %agg, i32 1, 0 <i>; yields {i32, float}</i>
3351 <!-- ======================================================================= -->
3352 <div class="doc_subsection">
3353 <a name="memoryops">Memory Access and Addressing Operations</a>
3356 <div class="doc_text">
3358 <p>A key design point of an SSA-based representation is how it
3359 represents memory. In LLVM, no memory locations are in SSA form, which
3360 makes things very simple. This section describes how to read, write,
3361 allocate, and free memory in LLVM.</p>
3365 <!-- _______________________________________________________________________ -->
3366 <div class="doc_subsubsection">
3367 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
3370 <div class="doc_text">
3375 <result> = malloc <type>[, i32 <NumElements>][, align <alignment>] <i>; yields {type*}:result</i>
3380 <p>The '<tt>malloc</tt>' instruction allocates memory from the system
3381 heap and returns a pointer to it. The object is always allocated in the generic
3382 address space (address space zero).</p>
3386 <p>The '<tt>malloc</tt>' instruction allocates
3387 <tt>sizeof(<type>)*NumElements</tt>
3388 bytes of memory from the operating system and returns a pointer of the
3389 appropriate type to the program. If "NumElements" is specified, it is the
3390 number of elements allocated, otherwise "NumElements" is defaulted to be one.
3391 If a constant alignment is specified, the value result of the allocation is guaranteed to
3392 be aligned to at least that boundary. If not specified, or if zero, the target can
3393 choose to align the allocation on any convenient boundary.</p>
3395 <p>'<tt>type</tt>' must be a sized type.</p>
3399 <p>Memory is allocated using the system "<tt>malloc</tt>" function, and
3400 a pointer is returned. The result of a zero byte allocation is undefined. The
3401 result is null if there is insufficient memory available.</p>
3406 %array = malloc [4 x i8] <i>; yields {[%4 x i8]*}:array</i>
3408 %size = <a href="#i_add">add</a> i32 2, 2 <i>; yields {i32}:size = i32 4</i>
3409 %array1 = malloc i8, i32 4 <i>; yields {i8*}:array1</i>
3410 %array2 = malloc [12 x i8], i32 %size <i>; yields {[12 x i8]*}:array2</i>
3411 %array3 = malloc i32, i32 4, align 1024 <i>; yields {i32*}:array3</i>
3412 %array4 = malloc i32, align 1024 <i>; yields {i32*}:array4</i>
3415 <p>Note that the code generator does not yet respect the
3416 alignment value.</p>
3420 <!-- _______________________________________________________________________ -->
3421 <div class="doc_subsubsection">
3422 <a name="i_free">'<tt>free</tt>' Instruction</a>
3425 <div class="doc_text">
3430 free <type> <value> <i>; yields {void}</i>
3435 <p>The '<tt>free</tt>' instruction returns memory back to the unused
3436 memory heap to be reallocated in the future.</p>
3440 <p>'<tt>value</tt>' shall be a pointer value that points to a value
3441 that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
3446 <p>Access to the memory pointed to by the pointer is no longer defined
3447 after this instruction executes. If the pointer is null, the operation
3453 %array = <a href="#i_malloc">malloc</a> [4 x i8] <i>; yields {[4 x i8]*}:array</i>
3454 free [4 x i8]* %array
3458 <!-- _______________________________________________________________________ -->
3459 <div class="doc_subsubsection">
3460 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
3463 <div class="doc_text">
3468 <result> = alloca <type>[, i32 <NumElements>][, align <alignment>] <i>; yields {type*}:result</i>
3473 <p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
3474 currently executing function, to be automatically released when this function
3475 returns to its caller. The object is always allocated in the generic address
3476 space (address space zero).</p>
3480 <p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(<type>)*NumElements</tt>
3481 bytes of memory on the runtime stack, returning a pointer of the
3482 appropriate type to the program. If "NumElements" is specified, it is the
3483 number of elements allocated, otherwise "NumElements" is defaulted to be one.
3484 If a constant alignment is specified, the value result of the allocation is guaranteed
3485 to be aligned to at least that boundary. If not specified, or if zero, the target
3486 can choose to align the allocation on any convenient boundary.</p>
3488 <p>'<tt>type</tt>' may be any sized type.</p>
3492 <p>Memory is allocated; a pointer is returned. The operation is undefiend if
3493 there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
3494 memory is automatically released when the function returns. The '<tt>alloca</tt>'
3495 instruction is commonly used to represent automatic variables that must
3496 have an address available. When the function returns (either with the <tt><a
3497 href="#i_ret">ret</a></tt> or <tt><a href="#i_unwind">unwind</a></tt>
3498 instructions), the memory is reclaimed. Allocating zero bytes
3499 is legal, but the result is undefined.</p>
3504 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
3505 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
3506 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
3507 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
3511 <!-- _______________________________________________________________________ -->
3512 <div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
3513 Instruction</a> </div>
3514 <div class="doc_text">
3516 <pre> <result> = load <ty>* <pointer>[, align <alignment>]<br> <result> = volatile load <ty>* <pointer>[, align <alignment>]<br></pre>
3518 <p>The '<tt>load</tt>' instruction is used to read from memory.</p>
3520 <p>The argument to the '<tt>load</tt>' instruction specifies the memory
3521 address from which to load. The pointer must point to a <a
3522 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
3523 marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
3524 the number or order of execution of this <tt>load</tt> with other
3525 volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
3528 The optional constant "align" argument specifies the alignment of the operation
3529 (that is, the alignment of the memory address). A value of 0 or an
3530 omitted "align" argument means that the operation has the preferential
3531 alignment for the target. It is the responsibility of the code emitter
3532 to ensure that the alignment information is correct. Overestimating
3533 the alignment results in an undefined behavior. Underestimating the
3534 alignment may produce less efficient code. An alignment of 1 is always
3538 <p>The location of memory pointed to is loaded. If the value being loaded
3539 is of scalar type then the number of bytes read does not exceed the minimum
3540 number of bytes needed to hold all bits of the type. For example, loading an
3541 <tt>i24</tt> reads at most three bytes. When loading a value of a type like
3542 <tt>i20</tt> with a size that is not an integral number of bytes, the result
3543 is undefined if the value was not originally written using a store of the
3546 <pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
3548 href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
3549 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
3552 <!-- _______________________________________________________________________ -->
3553 <div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
3554 Instruction</a> </div>
3555 <div class="doc_text">
3557 <pre> store <ty> <value>, <ty>* <pointer>[, align <alignment>] <i>; yields {void}</i>
3558 volatile store <ty> <value>, <ty>* <pointer>[, align <alignment>] <i>; yields {void}</i>
3561 <p>The '<tt>store</tt>' instruction is used to write to memory.</p>
3563 <p>There are two arguments to the '<tt>store</tt>' instruction: a value
3564 to store and an address at which to store it. The type of the '<tt><pointer></tt>'
3565 operand must be a pointer to the <a href="#t_firstclass">first class</a> type
3566 of the '<tt><value></tt>'
3567 operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
3568 optimizer is not allowed to modify the number or order of execution of
3569 this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
3570 href="#i_store">store</a></tt> instructions.</p>
3572 The optional constant "align" argument specifies the alignment of the operation
3573 (that is, the alignment of the memory address). A value of 0 or an
3574 omitted "align" argument means that the operation has the preferential
3575 alignment for the target. It is the responsibility of the code emitter
3576 to ensure that the alignment information is correct. Overestimating
3577 the alignment results in an undefined behavior. Underestimating the
3578 alignment may produce less efficient code. An alignment of 1 is always
3582 <p>The contents of memory are updated to contain '<tt><value></tt>'
3583 at the location specified by the '<tt><pointer></tt>' operand.
3584 If '<tt><value></tt>' is of scalar type then the number of bytes
3585 written does not exceed the minimum number of bytes needed to hold all
3586 bits of the type. For example, storing an <tt>i24</tt> writes at most
3587 three bytes. When writing a value of a type like <tt>i20</tt> with a
3588 size that is not an integral number of bytes, it is unspecified what
3589 happens to the extra bits that do not belong to the type, but they will
3590 typically be overwritten.</p>
3592 <pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
3593 store i32 3, i32* %ptr <i>; yields {void}</i>
3594 %val = <a href="#i_load">load</a> i32* %ptr <i>; yields {i32}:val = i32 3</i>
3598 <!-- _______________________________________________________________________ -->
3599 <div class="doc_subsubsection">
3600 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
3603 <div class="doc_text">
3606 <result> = getelementptr <pty>* <ptrval>{, <ty> <idx>}*
3612 The '<tt>getelementptr</tt>' instruction is used to get the address of a
3613 subelement of an aggregate data structure. It performs address calculation only
3614 and does not access memory.</p>
3618 <p>The first argument is always a pointer, and forms the basis of the
3619 calculation. The remaining arguments are indices, that indicate which of the
3620 elements of the aggregate object are indexed. The interpretation of each index
3621 is dependent on the type being indexed into. The first index always indexes the
3622 pointer value given as the first argument, the second index indexes a value of
3623 the type pointed to (not necessarily the value directly pointed to, since the
3624 first index can be non-zero), etc. The first type indexed into must be a pointer
3625 value, subsequent types can be arrays, vectors and structs. Note that subsequent
3626 types being indexed into can never be pointers, since that would require loading
3627 the pointer before continuing calculation.</p>
3629 <p>The type of each index argument depends on the type it is indexing into.
3630 When indexing into a (packed) structure, only <tt>i32</tt> integer
3631 <b>constants</b> are allowed. When indexing into an array, pointer or vector,
3632 only integers of 16, 32 or 64 bits are allowed (also non-constants).16-bit
3633 values will be sign extended to 32-bits if required, and 32-bit values
3634 will be sign extended to 64-bits if required.</p>
3636 <p>For example, let's consider a C code fragment and how it gets
3637 compiled to LLVM:</p>
3639 <div class="doc_code">
3652 int *foo(struct ST *s) {
3653 return &s[1].Z.B[5][13];
3658 <p>The LLVM code generated by the GCC frontend is:</p>
3660 <div class="doc_code">
3662 %RT = <a href="#namedtypes">type</a> { i8 , [10 x [20 x i32]], i8 }
3663 %ST = <a href="#namedtypes">type</a> { i32, double, %RT }
3665 define i32* %foo(%ST* %s) {
3667 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
3675 <p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
3676 type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
3677 }</tt>' type, a structure. The second index indexes into the third element of
3678 the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
3679 i8 }</tt>' type, another structure. The third index indexes into the second
3680 element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
3681 array. The two dimensions of the array are subscripted into, yielding an
3682 '<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
3683 to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
3685 <p>Note that it is perfectly legal to index partially through a
3686 structure, returning a pointer to an inner element. Because of this,
3687 the LLVM code for the given testcase is equivalent to:</p>
3690 define i32* %foo(%ST* %s) {
3691 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
3692 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
3693 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
3694 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
3695 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
3700 <p>Note that it is undefined to access an array out of bounds: array
3701 and pointer indexes must always be within the defined bounds of the
3702 array type when accessed with an instruction that dereferences the
3703 pointer (e.g. a load or store instruction). The one exception for
3704 this rule is zero length arrays. These arrays are defined to be
3705 accessible as variable length arrays, which requires access beyond the
3706 zero'th element.</p>
3708 <p>The getelementptr instruction is often confusing. For some more insight
3709 into how it works, see <a href="GetElementPtr.html">the getelementptr
3715 <i>; yields [12 x i8]*:aptr</i>
3716 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
3717 <i>; yields i8*:vptr</i>
3718 %vptr = getelementptr {i32, <2 x i8>}* %svptr, i64 0, i32 1, i32 1
3719 <i>; yields i8*:eptr</i>
3720 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
3721 <i>; yields i16*:iptr</i>
3722 %iptr = getelementptr [10 x i16]* @arr, i16 0, i16 0
3726 <!-- ======================================================================= -->
3727 <div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
3729 <div class="doc_text">
3730 <p>The instructions in this category are the conversion instructions (casting)
3731 which all take a single operand and a type. They perform various bit conversions
3735 <!-- _______________________________________________________________________ -->
3736 <div class="doc_subsubsection">
3737 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
3739 <div class="doc_text">
3743 <result> = trunc <ty> <value> to <ty2> <i>; yields ty2</i>
3748 The '<tt>trunc</tt>' instruction truncates its operand to the type <tt>ty2</tt>.
3753 The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
3754 be an <a href="#t_integer">integer</a> type, and a type that specifies the size
3755 and type of the result, which must be an <a href="#t_integer">integer</a>
3756 type. The bit size of <tt>value</tt> must be larger than the bit size of
3757 <tt>ty2</tt>. Equal sized types are not allowed.</p>
3761 The '<tt>trunc</tt>' instruction truncates the high order bits in <tt>value</tt>
3762 and converts the remaining bits to <tt>ty2</tt>. Since the source size must be
3763 larger than the destination size, <tt>trunc</tt> cannot be a <i>no-op cast</i>.
3764 It will always truncate bits.</p>
3768 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
3769 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
3770 %Y = trunc i32 122 to i1 <i>; yields i1:false</i>
3774 <!-- _______________________________________________________________________ -->
3775 <div class="doc_subsubsection">
3776 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
3778 <div class="doc_text">
3782 <result> = zext <ty> <value> to <ty2> <i>; yields ty2</i>
3786 <p>The '<tt>zext</tt>' instruction zero extends its operand to type
3791 <p>The '<tt>zext</tt>' instruction takes a value to cast, which must be of
3792 <a href="#t_integer">integer</a> type, and a type to cast it to, which must
3793 also be of <a href="#t_integer">integer</a> type. The bit size of the
3794 <tt>value</tt> must be smaller than the bit size of the destination type,
3798 <p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
3799 bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
3801 <p>When zero extending from i1, the result will always be either 0 or 1.</p>
3805 %X = zext i32 257 to i64 <i>; yields i64:257</i>
3806 %Y = zext i1 true to i32 <i>; yields i32:1</i>
3810 <!-- _______________________________________________________________________ -->
3811 <div class="doc_subsubsection">
3812 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
3814 <div class="doc_text">
3818 <result> = sext <ty> <value> to <ty2> <i>; yields ty2</i>
3822 <p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
3826 The '<tt>sext</tt>' instruction takes a value to cast, which must be of
3827 <a href="#t_integer">integer</a> type, and a type to cast it to, which must
3828 also be of <a href="#t_integer">integer</a> type. The bit size of the
3829 <tt>value</tt> must be smaller than the bit size of the destination type,
3834 The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
3835 bit (highest order bit) of the <tt>value</tt> until it reaches the bit size of
3836 the type <tt>ty2</tt>.</p>
3838 <p>When sign extending from i1, the extension always results in -1 or 0.</p>
3842 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
3843 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
3847 <!-- _______________________________________________________________________ -->
3848 <div class="doc_subsubsection">
3849 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
3852 <div class="doc_text">
3857 <result> = fptrunc <ty> <value> to <ty2> <i>; yields ty2</i>
3861 <p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
3866 <p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
3867 point</a> value to cast and a <a href="#t_floating">floating point</a> type to
3868 cast it to. The size of <tt>value</tt> must be larger than the size of
3869 <tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
3870 <i>no-op cast</i>.</p>
3873 <p> The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
3874 <a href="#t_floating">floating point</a> type to a smaller
3875 <a href="#t_floating">floating point</a> type. If the value cannot fit within
3876 the destination type, <tt>ty2</tt>, then the results are undefined.</p>
3880 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
3881 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
3885 <!-- _______________________________________________________________________ -->
3886 <div class="doc_subsubsection">
3887 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
3889 <div class="doc_text">
3893 <result> = fpext <ty> <value> to <ty2> <i>; yields ty2</i>
3897 <p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
3898 floating point value.</p>
3901 <p>The '<tt>fpext</tt>' instruction takes a
3902 <a href="#t_floating">floating point</a> <tt>value</tt> to cast,
3903 and a <a href="#t_floating">floating point</a> type to cast it to. The source
3904 type must be smaller than the destination type.</p>
3907 <p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
3908 <a href="#t_floating">floating point</a> type to a larger
3909 <a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
3910 used to make a <i>no-op cast</i> because it always changes bits. Use
3911 <tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
3915 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
3916 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
3920 <!-- _______________________________________________________________________ -->
3921 <div class="doc_subsubsection">
3922 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
3924 <div class="doc_text">
3928 <result> = fptoui <ty> <value> to <ty2> <i>; yields ty2</i>
3932 <p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
3933 unsigned integer equivalent of type <tt>ty2</tt>.
3937 <p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
3938 scalar or vector <a href="#t_floating">floating point</a> value, and a type
3939 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
3940 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
3941 vector integer type with the same number of elements as <tt>ty</tt></p>
3944 <p> The '<tt>fptoui</tt>' instruction converts its
3945 <a href="#t_floating">floating point</a> operand into the nearest (rounding
3946 towards zero) unsigned integer value. If the value cannot fit in <tt>ty2</tt>,
3947 the results are undefined.</p>
3951 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
3952 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
3953 %X = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
3957 <!-- _______________________________________________________________________ -->
3958 <div class="doc_subsubsection">
3959 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
3961 <div class="doc_text">
3965 <result> = fptosi <ty> <value> to <ty2> <i>; yields ty2</i>
3969 <p>The '<tt>fptosi</tt>' instruction converts
3970 <a href="#t_floating">floating point</a> <tt>value</tt> to type <tt>ty2</tt>.
3974 <p> The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
3975 scalar or vector <a href="#t_floating">floating point</a> value, and a type
3976 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
3977 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
3978 vector integer type with the same number of elements as <tt>ty</tt></p>
3981 <p>The '<tt>fptosi</tt>' instruction converts its
3982 <a href="#t_floating">floating point</a> operand into the nearest (rounding
3983 towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
3984 the results are undefined.</p>
3988 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
3989 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
3990 %X = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
3994 <!-- _______________________________________________________________________ -->
3995 <div class="doc_subsubsection">
3996 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
3998 <div class="doc_text">
4002 <result> = uitofp <ty> <value> to <ty2> <i>; yields ty2</i>
4006 <p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
4007 integer and converts that value to the <tt>ty2</tt> type.</p>
4010 <p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be a
4011 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
4012 to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4013 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4014 floating point type with the same number of elements as <tt>ty</tt></p>
4017 <p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
4018 integer quantity and converts it to the corresponding floating point value. If
4019 the value cannot fit in the floating point value, the results are undefined.</p>
4023 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
4024 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
4028 <!-- _______________________________________________________________________ -->
4029 <div class="doc_subsubsection">
4030 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
4032 <div class="doc_text">
4036 <result> = sitofp <ty> <value> to <ty2> <i>; yields ty2</i>
4040 <p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed
4041 integer and converts that value to the <tt>ty2</tt> type.</p>
4044 <p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be a
4045 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
4046 to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4047 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4048 floating point type with the same number of elements as <tt>ty</tt></p>
4051 <p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed
4052 integer quantity and converts it to the corresponding floating point value. If
4053 the value cannot fit in the floating point value, the results are undefined.</p>
4057 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
4058 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
4062 <!-- _______________________________________________________________________ -->
4063 <div class="doc_subsubsection">
4064 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
4066 <div class="doc_text">
4070 <result> = ptrtoint <ty> <value> to <ty2> <i>; yields ty2</i>
4074 <p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
4075 the integer type <tt>ty2</tt>.</p>
4078 <p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
4079 must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
4080 <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> type.</p>
4083 <p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
4084 <tt>ty2</tt> by interpreting the pointer value as an integer and either
4085 truncating or zero extending that value to the size of the integer type. If
4086 <tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
4087 <tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
4088 are the same size, then nothing is done (<i>no-op cast</i>) other than a type
4093 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
4094 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
4098 <!-- _______________________________________________________________________ -->
4099 <div class="doc_subsubsection">
4100 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
4102 <div class="doc_text">
4106 <result> = inttoptr <ty> <value> to <ty2> <i>; yields ty2</i>
4110 <p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to
4111 a pointer type, <tt>ty2</tt>.</p>
4114 <p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
4115 value to cast, and a type to cast it to, which must be a
4116 <a href="#t_pointer">pointer</a> type.</p>
4119 <p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
4120 <tt>ty2</tt> by applying either a zero extension or a truncation depending on
4121 the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
4122 size of a pointer then a truncation is done. If <tt>value</tt> is smaller than
4123 the size of a pointer then a zero extension is done. If they are the same size,
4124 nothing is done (<i>no-op cast</i>).</p>
4128 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
4129 %X = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
4130 %Y = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
4134 <!-- _______________________________________________________________________ -->
4135 <div class="doc_subsubsection">
4136 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
4138 <div class="doc_text">
4142 <result> = bitcast <ty> <value> to <ty2> <i>; yields ty2</i>
4147 <p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
4148 <tt>ty2</tt> without changing any bits.</p>
4152 <p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be
4153 a non-aggregate first class value, and a type to cast it to, which must also be
4154 a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes of
4156 and the destination type, <tt>ty2</tt>, must be identical. If the source
4157 type is a pointer, the destination type must also be a pointer. This
4158 instruction supports bitwise conversion of vectors to integers and to vectors
4159 of other types (as long as they have the same size).</p>
4162 <p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
4163 <tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
4164 this conversion. The conversion is done as if the <tt>value</tt> had been
4165 stored to memory and read back as type <tt>ty2</tt>. Pointer types may only be
4166 converted to other pointer types with this instruction. To convert pointers to
4167 other types, use the <a href="#i_inttoptr">inttoptr</a> or
4168 <a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
4172 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
4173 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
4174 %Z = bitcast <2 x int> %V to i64; <i>; yields i64: %V</i>
4178 <!-- ======================================================================= -->
4179 <div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
4180 <div class="doc_text">
4181 <p>The instructions in this category are the "miscellaneous"
4182 instructions, which defy better classification.</p>
4185 <!-- _______________________________________________________________________ -->
4186 <div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
4188 <div class="doc_text">
4190 <pre> <result> = icmp <cond> <ty> <op1>, <op2> <i>; yields {i1} or {<N x i1>}:result</i>
4193 <p>The '<tt>icmp</tt>' instruction returns a boolean value or
4194 a vector of boolean values based on comparison
4195 of its two integer, integer vector, or pointer operands.</p>
4197 <p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
4198 the condition code indicating the kind of comparison to perform. It is not
4199 a value, just a keyword. The possible condition code are:
4202 <li><tt>eq</tt>: equal</li>
4203 <li><tt>ne</tt>: not equal </li>
4204 <li><tt>ugt</tt>: unsigned greater than</li>
4205 <li><tt>uge</tt>: unsigned greater or equal</li>
4206 <li><tt>ult</tt>: unsigned less than</li>
4207 <li><tt>ule</tt>: unsigned less or equal</li>
4208 <li><tt>sgt</tt>: signed greater than</li>
4209 <li><tt>sge</tt>: signed greater or equal</li>
4210 <li><tt>slt</tt>: signed less than</li>
4211 <li><tt>sle</tt>: signed less or equal</li>
4213 <p>The remaining two arguments must be <a href="#t_integer">integer</a> or
4214 <a href="#t_pointer">pointer</a>
4215 or integer <a href="#t_vector">vector</a> typed.
4216 They must also be identical types.</p>
4218 <p>The '<tt>icmp</tt>' compares <tt>op1</tt> and <tt>op2</tt> according to
4219 the condition code given as <tt>cond</tt>. The comparison performed always
4220 yields either an <a href="#t_primitive"><tt>i1</tt></a> or vector of <tt>i1</tt> result, as follows:
4223 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
4224 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
4226 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
4227 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.</li>
4228 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
4229 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
4230 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
4231 <tt>true</tt> if <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
4232 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
4233 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
4234 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
4235 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
4236 <li><tt>sgt</tt>: interprets the operands as signed values and yields
4237 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
4238 <li><tt>sge</tt>: interprets the operands as signed values and yields
4239 <tt>true</tt> if <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
4240 <li><tt>slt</tt>: interprets the operands as signed values and yields
4241 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
4242 <li><tt>sle</tt>: interprets the operands as signed values and yields
4243 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
4245 <p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
4246 values are compared as if they were integers.</p>
4247 <p>If the operands are integer vectors, then they are compared
4248 element by element. The result is an <tt>i1</tt> vector with
4249 the same number of elements as the values being compared.
4250 Otherwise, the result is an <tt>i1</tt>.
4254 <pre> <result> = icmp eq i32 4, 5 <i>; yields: result=false</i>
4255 <result> = icmp ne float* %X, %X <i>; yields: result=false</i>
4256 <result> = icmp ult i16 4, 5 <i>; yields: result=true</i>
4257 <result> = icmp sgt i16 4, 5 <i>; yields: result=false</i>
4258 <result> = icmp ule i16 -4, 5 <i>; yields: result=false</i>
4259 <result> = icmp sge i16 4, 5 <i>; yields: result=false</i>
4262 <p>Note that the code generator does not yet support vector types with
4263 the <tt>icmp</tt> instruction.</p>
4267 <!-- _______________________________________________________________________ -->
4268 <div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
4270 <div class="doc_text">
4272 <pre> <result> = fcmp <cond> <ty> <op1>, <op2> <i>; yields {i1} or {<N x i1>}:result</i>
4275 <p>The '<tt>fcmp</tt>' instruction returns a boolean value
4276 or vector of boolean values based on comparison
4277 of its operands.</p>
4279 If the operands are floating point scalars, then the result
4280 type is a boolean (<a href="#t_primitive"><tt>i1</tt></a>).
4282 <p>If the operands are floating point vectors, then the result type
4283 is a vector of boolean with the same number of elements as the
4284 operands being compared.</p>
4286 <p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
4287 the condition code indicating the kind of comparison to perform. It is not
4288 a value, just a keyword. The possible condition code are:</p>
4290 <li><tt>false</tt>: no comparison, always returns false</li>
4291 <li><tt>oeq</tt>: ordered and equal</li>
4292 <li><tt>ogt</tt>: ordered and greater than </li>
4293 <li><tt>oge</tt>: ordered and greater than or equal</li>
4294 <li><tt>olt</tt>: ordered and less than </li>
4295 <li><tt>ole</tt>: ordered and less than or equal</li>
4296 <li><tt>one</tt>: ordered and not equal</li>
4297 <li><tt>ord</tt>: ordered (no nans)</li>
4298 <li><tt>ueq</tt>: unordered or equal</li>
4299 <li><tt>ugt</tt>: unordered or greater than </li>
4300 <li><tt>uge</tt>: unordered or greater than or equal</li>
4301 <li><tt>ult</tt>: unordered or less than </li>
4302 <li><tt>ule</tt>: unordered or less than or equal</li>
4303 <li><tt>une</tt>: unordered or not equal</li>
4304 <li><tt>uno</tt>: unordered (either nans)</li>
4305 <li><tt>true</tt>: no comparison, always returns true</li>
4307 <p><i>Ordered</i> means that neither operand is a QNAN while
4308 <i>unordered</i> means that either operand may be a QNAN.</p>
4309 <p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be
4310 either a <a href="#t_floating">floating point</a> type
4311 or a <a href="#t_vector">vector</a> of floating point type.
4312 They must have identical types.</p>
4314 <p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
4315 according to the condition code given as <tt>cond</tt>.
4316 If the operands are vectors, then the vectors are compared
4318 Each comparison performed
4319 always yields an <a href="#t_primitive">i1</a> result, as follows:</p>
4321 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
4322 <li><tt>oeq</tt>: yields <tt>true</tt> if both operands are not a QNAN and
4323 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
4324 <li><tt>ogt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
4325 <tt>op1</tt> is greather than <tt>op2</tt>.</li>
4326 <li><tt>oge</tt>: yields <tt>true</tt> if both operands are not a QNAN and
4327 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
4328 <li><tt>olt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
4329 <tt>op1</tt> is less than <tt>op2</tt>.</li>
4330 <li><tt>ole</tt>: yields <tt>true</tt> if both operands are not a QNAN and
4331 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
4332 <li><tt>one</tt>: yields <tt>true</tt> if both operands are not a QNAN and
4333 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
4334 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
4335 <li><tt>ueq</tt>: yields <tt>true</tt> if either operand is a QNAN or
4336 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
4337 <li><tt>ugt</tt>: yields <tt>true</tt> if either operand is a QNAN or
4338 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
4339 <li><tt>uge</tt>: yields <tt>true</tt> if either operand is a QNAN or
4340 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
4341 <li><tt>ult</tt>: yields <tt>true</tt> if either operand is a QNAN or
4342 <tt>op1</tt> is less than <tt>op2</tt>.</li>
4343 <li><tt>ule</tt>: yields <tt>true</tt> if either operand is a QNAN or
4344 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
4345 <li><tt>une</tt>: yields <tt>true</tt> if either operand is a QNAN or
4346 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
4347 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
4348 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
4352 <pre> <result> = fcmp oeq float 4.0, 5.0 <i>; yields: result=false</i>
4353 <result> = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
4354 <result> = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
4355 <result> = fcmp ueq double 1.0, 2.0 <i>; yields: result=false</i>
4358 <p>Note that the code generator does not yet support vector types with
4359 the <tt>fcmp</tt> instruction.</p>
4363 <!-- _______________________________________________________________________ -->
4364 <div class="doc_subsubsection">
4365 <a name="i_vicmp">'<tt>vicmp</tt>' Instruction</a>
4367 <div class="doc_text">
4369 <pre> <result> = vicmp <cond> <ty> <op1>, <op2> <i>; yields {ty}:result</i>
4372 <p>The '<tt>vicmp</tt>' instruction returns an integer vector value based on
4373 element-wise comparison of its two integer vector operands.</p>
4375 <p>The '<tt>vicmp</tt>' instruction takes three operands. The first operand is
4376 the condition code indicating the kind of comparison to perform. It is not
4377 a value, just a keyword. The possible condition code are:</p>
4379 <li><tt>eq</tt>: equal</li>
4380 <li><tt>ne</tt>: not equal </li>
4381 <li><tt>ugt</tt>: unsigned greater than</li>
4382 <li><tt>uge</tt>: unsigned greater or equal</li>
4383 <li><tt>ult</tt>: unsigned less than</li>
4384 <li><tt>ule</tt>: unsigned less or equal</li>
4385 <li><tt>sgt</tt>: signed greater than</li>
4386 <li><tt>sge</tt>: signed greater or equal</li>
4387 <li><tt>slt</tt>: signed less than</li>
4388 <li><tt>sle</tt>: signed less or equal</li>
4390 <p>The remaining two arguments must be <a href="#t_vector">vector</a> or
4391 <a href="#t_integer">integer</a> typed. They must also be identical types.</p>
4393 <p>The '<tt>vicmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
4394 according to the condition code given as <tt>cond</tt>. The comparison yields a
4395 <a href="#t_vector">vector</a> of <a href="#t_integer">integer</a> result, of
4396 identical type as the values being compared. The most significant bit in each
4397 element is 1 if the element-wise comparison evaluates to true, and is 0
4398 otherwise. All other bits of the result are undefined. The condition codes
4399 are evaluated identically to the <a href="#i_icmp">'<tt>icmp</tt>'
4400 instruction</a>.</p>
4404 <result> = vicmp eq <2 x i32> < i32 4, i32 0>, < i32 5, i32 0> <i>; yields: result=<2 x i32> < i32 0, i32 -1 ></i>
4405 <result> = vicmp ult <2 x i8 > < i8 1, i8 2>, < i8 2, i8 2 > <i>; yields: result=<2 x i8> < i8 -1, i8 0 ></i>
4409 <!-- _______________________________________________________________________ -->
4410 <div class="doc_subsubsection">
4411 <a name="i_vfcmp">'<tt>vfcmp</tt>' Instruction</a>
4413 <div class="doc_text">
4415 <pre> <result> = vfcmp <cond> <ty> <op1>, <op2></pre>
4417 <p>The '<tt>vfcmp</tt>' instruction returns an integer vector value based on
4418 element-wise comparison of its two floating point vector operands. The output
4419 elements have the same width as the input elements.</p>
4421 <p>The '<tt>vfcmp</tt>' instruction takes three operands. The first operand is
4422 the condition code indicating the kind of comparison to perform. It is not
4423 a value, just a keyword. The possible condition code are:</p>
4425 <li><tt>false</tt>: no comparison, always returns false</li>
4426 <li><tt>oeq</tt>: ordered and equal</li>
4427 <li><tt>ogt</tt>: ordered and greater than </li>
4428 <li><tt>oge</tt>: ordered and greater than or equal</li>
4429 <li><tt>olt</tt>: ordered and less than </li>
4430 <li><tt>ole</tt>: ordered and less than or equal</li>
4431 <li><tt>one</tt>: ordered and not equal</li>
4432 <li><tt>ord</tt>: ordered (no nans)</li>
4433 <li><tt>ueq</tt>: unordered or equal</li>
4434 <li><tt>ugt</tt>: unordered or greater than </li>
4435 <li><tt>uge</tt>: unordered or greater than or equal</li>
4436 <li><tt>ult</tt>: unordered or less than </li>
4437 <li><tt>ule</tt>: unordered or less than or equal</li>
4438 <li><tt>une</tt>: unordered or not equal</li>
4439 <li><tt>uno</tt>: unordered (either nans)</li>
4440 <li><tt>true</tt>: no comparison, always returns true</li>
4442 <p>The remaining two arguments must be <a href="#t_vector">vector</a> of
4443 <a href="#t_floating">floating point</a> typed. They must also be identical
4446 <p>The '<tt>vfcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
4447 according to the condition code given as <tt>cond</tt>. The comparison yields a
4448 <a href="#t_vector">vector</a> of <a href="#t_integer">integer</a> result, with
4449 an identical number of elements as the values being compared, and each element
4450 having identical with to the width of the floating point elements. The most
4451 significant bit in each element is 1 if the element-wise comparison evaluates to
4452 true, and is 0 otherwise. All other bits of the result are undefined. The
4453 condition codes are evaluated identically to the
4454 <a href="#i_fcmp">'<tt>fcmp</tt>' instruction</a>.</p>
4458 <i>; yields: result=<2 x i32> < i32 0, i32 -1 ></i>
4459 <result> = vfcmp oeq <2 x float> < float 4, float 0 >, < float 5, float 0 >
4461 <i>; yields: result=<2 x i64> < i64 -1, i64 0 ></i>
4462 <result> = vfcmp ult <2 x double> < double 1, double 2 >, < double 2, double 2>
4466 <!-- _______________________________________________________________________ -->
4467 <div class="doc_subsubsection">
4468 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
4471 <div class="doc_text">
4475 <pre> <result> = phi <ty> [ <val0>, <label0>], ...<br></pre>
4477 <p>The '<tt>phi</tt>' instruction is used to implement the φ node in
4478 the SSA graph representing the function.</p>
4481 <p>The type of the incoming values is specified with the first type
4482 field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
4483 as arguments, with one pair for each predecessor basic block of the
4484 current block. Only values of <a href="#t_firstclass">first class</a>
4485 type may be used as the value arguments to the PHI node. Only labels
4486 may be used as the label arguments.</p>
4488 <p>There must be no non-phi instructions between the start of a basic
4489 block and the PHI instructions: i.e. PHI instructions must be first in
4494 <p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
4495 specified by the pair corresponding to the predecessor basic block that executed
4496 just prior to the current block.</p>
4500 Loop: ; Infinite loop that counts from 0 on up...
4501 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
4502 %nextindvar = add i32 %indvar, 1
4507 <!-- _______________________________________________________________________ -->
4508 <div class="doc_subsubsection">
4509 <a name="i_select">'<tt>select</tt>' Instruction</a>
4512 <div class="doc_text">
4517 <result> = select <i>selty</i> <cond>, <ty> <val1>, <ty> <val2> <i>; yields ty</i>
4519 <i>selty</i> is either i1 or {<N x i1>}
4525 The '<tt>select</tt>' instruction is used to choose one value based on a
4526 condition, without branching.
4533 The '<tt>select</tt>' instruction requires an 'i1' value or
4534 a vector of 'i1' values indicating the
4535 condition, and two values of the same <a href="#t_firstclass">first class</a>
4536 type. If the val1/val2 are vectors and
4537 the condition is a scalar, then entire vectors are selected, not
4538 individual elements.
4544 If the condition is an i1 and it evaluates to 1, the instruction returns the first
4545 value argument; otherwise, it returns the second value argument.
4548 If the condition is a vector of i1, then the value arguments must
4549 be vectors of the same size, and the selection is done element
4556 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
4559 <p>Note that the code generator does not yet support conditions
4560 with vector type.</p>
4565 <!-- _______________________________________________________________________ -->
4566 <div class="doc_subsubsection">
4567 <a name="i_call">'<tt>call</tt>' Instruction</a>
4570 <div class="doc_text">
4574 <result> = [tail] call [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>] <ty> [<fnty>*] <fnptrval>(<function args>) [<a href="#fnattrs">fn attrs</a>]
4579 <p>The '<tt>call</tt>' instruction represents a simple function call.</p>
4583 <p>This instruction requires several arguments:</p>
4587 <p>The optional "tail" marker indicates whether the callee function accesses
4588 any allocas or varargs in the caller. If the "tail" marker is present, the
4589 function call is eligible for tail call optimization. Note that calls may
4590 be marked "tail" even if they do not occur before a <a
4591 href="#i_ret"><tt>ret</tt></a> instruction.</p>
4594 <p>The optional "cconv" marker indicates which <a href="#callingconv">calling
4595 convention</a> the call should use. If none is specified, the call defaults
4596 to using C calling conventions.</p>
4600 <p>The optional <a href="#paramattrs">Parameter Attributes</a> list for
4601 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>',
4602 and '<tt>inreg</tt>' attributes are valid here.</p>
4606 <p>'<tt>ty</tt>': the type of the call instruction itself which is also
4607 the type of the return value. Functions that return no value are marked
4608 <tt><a href="#t_void">void</a></tt>.</p>
4611 <p>'<tt>fnty</tt>': shall be the signature of the pointer to function
4612 value being invoked. The argument types must match the types implied by
4613 this signature. This type can be omitted if the function is not varargs
4614 and if the function type does not return a pointer to a function.</p>
4617 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
4618 be invoked. In most cases, this is a direct function invocation, but
4619 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
4620 to function value.</p>
4623 <p>'<tt>function args</tt>': argument list whose types match the
4624 function signature argument types. All arguments must be of
4625 <a href="#t_firstclass">first class</a> type. If the function signature
4626 indicates the function accepts a variable number of arguments, the extra
4627 arguments can be specified.</p>
4630 <p>The optional <a href="#fnattrs">function attributes</a> list. Only
4631 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
4632 '<tt>readnone</tt>' attributes are valid here.</p>
4638 <p>The '<tt>call</tt>' instruction is used to cause control flow to
4639 transfer to a specified function, with its incoming arguments bound to
4640 the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
4641 instruction in the called function, control flow continues with the
4642 instruction after the function call, and the return value of the
4643 function is bound to the result argument.</p>
4648 %retval = call i32 @test(i32 %argc)
4649 call i32 (i8 *, ...)* @printf(i8 * %msg, i32 12, i8 42) <i>; yields i32</i>
4650 %X = tail call i32 @foo() <i>; yields i32</i>
4651 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
4652 call void %foo(i8 97 signext)
4654 %struct.A = type { i32, i8 }
4655 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
4656 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
4657 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
4658 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
4659 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
4664 <!-- _______________________________________________________________________ -->
4665 <div class="doc_subsubsection">
4666 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
4669 <div class="doc_text">
4674 <resultval> = va_arg <va_list*> <arglist>, <argty>
4679 <p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
4680 the "variable argument" area of a function call. It is used to implement the
4681 <tt>va_arg</tt> macro in C.</p>
4685 <p>This instruction takes a <tt>va_list*</tt> value and the type of
4686 the argument. It returns a value of the specified argument type and
4687 increments the <tt>va_list</tt> to point to the next argument. The
4688 actual type of <tt>va_list</tt> is target specific.</p>
4692 <p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
4693 type from the specified <tt>va_list</tt> and causes the
4694 <tt>va_list</tt> to point to the next argument. For more information,
4695 see the variable argument handling <a href="#int_varargs">Intrinsic
4698 <p>It is legal for this instruction to be called in a function which does not
4699 take a variable number of arguments, for example, the <tt>vfprintf</tt>
4702 <p><tt>va_arg</tt> is an LLVM instruction instead of an <a
4703 href="#intrinsics">intrinsic function</a> because it takes a type as an
4708 <p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
4710 <p>Note that the code generator does not yet fully support va_arg
4711 on many targets. Also, it does not currently support va_arg with
4712 aggregate types on any target.</p>
4716 <!-- *********************************************************************** -->
4717 <div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
4718 <!-- *********************************************************************** -->
4720 <div class="doc_text">
4722 <p>LLVM supports the notion of an "intrinsic function". These functions have
4723 well known names and semantics and are required to follow certain restrictions.
4724 Overall, these intrinsics represent an extension mechanism for the LLVM
4725 language that does not require changing all of the transformations in LLVM when
4726 adding to the language (or the bitcode reader/writer, the parser, etc...).</p>
4728 <p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
4729 prefix is reserved in LLVM for intrinsic names; thus, function names may not
4730 begin with this prefix. Intrinsic functions must always be external functions:
4731 you cannot define the body of intrinsic functions. Intrinsic functions may
4732 only be used in call or invoke instructions: it is illegal to take the address
4733 of an intrinsic function. Additionally, because intrinsic functions are part
4734 of the LLVM language, it is required if any are added that they be documented
4737 <p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents
4738 a family of functions that perform the same operation but on different data
4739 types. Because LLVM can represent over 8 million different integer types,
4740 overloading is used commonly to allow an intrinsic function to operate on any
4741 integer type. One or more of the argument types or the result type can be
4742 overloaded to accept any integer type. Argument types may also be defined as
4743 exactly matching a previous argument's type or the result type. This allows an
4744 intrinsic function which accepts multiple arguments, but needs all of them to
4745 be of the same type, to only be overloaded with respect to a single argument or
4748 <p>Overloaded intrinsics will have the names of its overloaded argument types
4749 encoded into its function name, each preceded by a period. Only those types
4750 which are overloaded result in a name suffix. Arguments whose type is matched
4751 against another type do not. For example, the <tt>llvm.ctpop</tt> function can
4752 take an integer of any width and returns an integer of exactly the same integer
4753 width. This leads to a family of functions such as
4754 <tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29 %val)</tt>.
4755 Only one type, the return type, is overloaded, and only one type suffix is
4756 required. Because the argument's type is matched against the return type, it
4757 does not require its own name suffix.</p>
4759 <p>To learn how to add an intrinsic function, please see the
4760 <a href="ExtendingLLVM.html">Extending LLVM Guide</a>.
4765 <!-- ======================================================================= -->
4766 <div class="doc_subsection">
4767 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
4770 <div class="doc_text">
4772 <p>Variable argument support is defined in LLVM with the <a
4773 href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
4774 intrinsic functions. These functions are related to the similarly
4775 named macros defined in the <tt><stdarg.h></tt> header file.</p>
4777 <p>All of these functions operate on arguments that use a
4778 target-specific value type "<tt>va_list</tt>". The LLVM assembly
4779 language reference manual does not define what this type is, so all
4780 transformations should be prepared to handle these functions regardless of
4783 <p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
4784 instruction and the variable argument handling intrinsic functions are
4787 <div class="doc_code">
4789 define i32 @test(i32 %X, ...) {
4790 ; Initialize variable argument processing
4792 %ap2 = bitcast i8** %ap to i8*
4793 call void @llvm.va_start(i8* %ap2)
4795 ; Read a single integer argument
4796 %tmp = va_arg i8** %ap, i32
4798 ; Demonstrate usage of llvm.va_copy and llvm.va_end
4800 %aq2 = bitcast i8** %aq to i8*
4801 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
4802 call void @llvm.va_end(i8* %aq2)
4804 ; Stop processing of arguments.
4805 call void @llvm.va_end(i8* %ap2)
4809 declare void @llvm.va_start(i8*)
4810 declare void @llvm.va_copy(i8*, i8*)
4811 declare void @llvm.va_end(i8*)
4817 <!-- _______________________________________________________________________ -->
4818 <div class="doc_subsubsection">
4819 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
4823 <div class="doc_text">
4825 <pre> declare void %llvm.va_start(i8* <arglist>)<br></pre>
4827 <p>The '<tt>llvm.va_start</tt>' intrinsic initializes
4828 <tt>*<arglist></tt> for subsequent use by <tt><a
4829 href="#i_va_arg">va_arg</a></tt>.</p>
4833 <p>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
4837 <p>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
4838 macro available in C. In a target-dependent way, it initializes the
4839 <tt>va_list</tt> element to which the argument points, so that the next call to
4840 <tt>va_arg</tt> will produce the first variable argument passed to the function.
4841 Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
4842 last argument of the function as the compiler can figure that out.</p>
4846 <!-- _______________________________________________________________________ -->
4847 <div class="doc_subsubsection">
4848 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
4851 <div class="doc_text">
4853 <pre> declare void @llvm.va_end(i8* <arglist>)<br></pre>
4856 <p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*<arglist></tt>,
4857 which has been initialized previously with <tt><a href="#int_va_start">llvm.va_start</a></tt>
4858 or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
4862 <p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
4866 <p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
4867 macro available in C. In a target-dependent way, it destroys the
4868 <tt>va_list</tt> element to which the argument points. Calls to <a
4869 href="#int_va_start"><tt>llvm.va_start</tt></a> and <a href="#int_va_copy">
4870 <tt>llvm.va_copy</tt></a> must be matched exactly with calls to
4871 <tt>llvm.va_end</tt>.</p>
4875 <!-- _______________________________________________________________________ -->
4876 <div class="doc_subsubsection">
4877 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
4880 <div class="doc_text">
4885 declare void @llvm.va_copy(i8* <destarglist>, i8* <srcarglist>)
4890 <p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
4891 from the source argument list to the destination argument list.</p>
4895 <p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
4896 The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
4901 <p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
4902 macro available in C. In a target-dependent way, it copies the source
4903 <tt>va_list</tt> element into the destination <tt>va_list</tt> element. This
4904 intrinsic is necessary because the <tt><a href="#int_va_start">
4905 llvm.va_start</a></tt> intrinsic may be arbitrarily complex and require, for
4906 example, memory allocation.</p>
4910 <!-- ======================================================================= -->
4911 <div class="doc_subsection">
4912 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
4915 <div class="doc_text">
4918 LLVM support for <a href="GarbageCollection.html">Accurate Garbage
4919 Collection</a> (GC) requires the implementation and generation of these
4921 These intrinsics allow identification of <a href="#int_gcroot">GC roots on the
4922 stack</a>, as well as garbage collector implementations that require <a
4923 href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a> barriers.
4924 Front-ends for type-safe garbage collected languages should generate these
4925 intrinsics to make use of the LLVM garbage collectors. For more details, see <a
4926 href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
4929 <p>The garbage collection intrinsics only operate on objects in the generic
4930 address space (address space zero).</p>
4934 <!-- _______________________________________________________________________ -->
4935 <div class="doc_subsubsection">
4936 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
4939 <div class="doc_text">
4944 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
4949 <p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
4950 the code generator, and allows some metadata to be associated with it.</p>
4954 <p>The first argument specifies the address of a stack object that contains the
4955 root pointer. The second pointer (which must be either a constant or a global
4956 value address) contains the meta-data to be associated with the root.</p>
4960 <p>At runtime, a call to this intrinsic stores a null pointer into the "ptrloc"
4961 location. At compile-time, the code generator generates information to allow
4962 the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
4963 intrinsic may only be used in a function which <a href="#gc">specifies a GC
4969 <!-- _______________________________________________________________________ -->
4970 <div class="doc_subsubsection">
4971 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
4974 <div class="doc_text">
4979 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
4984 <p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
4985 locations, allowing garbage collector implementations that require read
4990 <p>The second argument is the address to read from, which should be an address
4991 allocated from the garbage collector. The first object is a pointer to the
4992 start of the referenced object, if needed by the language runtime (otherwise
4997 <p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
4998 instruction, but may be replaced with substantially more complex code by the
4999 garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
5000 may only be used in a function which <a href="#gc">specifies a GC
5006 <!-- _______________________________________________________________________ -->
5007 <div class="doc_subsubsection">
5008 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
5011 <div class="doc_text">
5016 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
5021 <p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
5022 locations, allowing garbage collector implementations that require write
5023 barriers (such as generational or reference counting collectors).</p>
5027 <p>The first argument is the reference to store, the second is the start of the
5028 object to store it to, and the third is the address of the field of Obj to
5029 store to. If the runtime does not require a pointer to the object, Obj may be
5034 <p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
5035 instruction, but may be replaced with substantially more complex code by the
5036 garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
5037 may only be used in a function which <a href="#gc">specifies a GC
5044 <!-- ======================================================================= -->
5045 <div class="doc_subsection">
5046 <a name="int_codegen">Code Generator Intrinsics</a>
5049 <div class="doc_text">
5051 These intrinsics are provided by LLVM to expose special features that may only
5052 be implemented with code generator support.
5057 <!-- _______________________________________________________________________ -->
5058 <div class="doc_subsubsection">
5059 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
5062 <div class="doc_text">
5066 declare i8 *@llvm.returnaddress(i32 <level>)
5072 The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
5073 target-specific value indicating the return address of the current function
5074 or one of its callers.
5080 The argument to this intrinsic indicates which function to return the address
5081 for. Zero indicates the calling function, one indicates its caller, etc. The
5082 argument is <b>required</b> to be a constant integer value.
5088 The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
5089 the return address of the specified call frame, or zero if it cannot be
5090 identified. The value returned by this intrinsic is likely to be incorrect or 0
5091 for arguments other than zero, so it should only be used for debugging purposes.
5095 Note that calling this intrinsic does not prevent function inlining or other
5096 aggressive transformations, so the value returned may not be that of the obvious
5097 source-language caller.
5102 <!-- _______________________________________________________________________ -->
5103 <div class="doc_subsubsection">
5104 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
5107 <div class="doc_text">
5111 declare i8 *@llvm.frameaddress(i32 <level>)
5117 The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
5118 target-specific frame pointer value for the specified stack frame.
5124 The argument to this intrinsic indicates which function to return the frame
5125 pointer for. Zero indicates the calling function, one indicates its caller,
5126 etc. The argument is <b>required</b> to be a constant integer value.
5132 The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
5133 the frame address of the specified call frame, or zero if it cannot be
5134 identified. The value returned by this intrinsic is likely to be incorrect or 0
5135 for arguments other than zero, so it should only be used for debugging purposes.
5139 Note that calling this intrinsic does not prevent function inlining or other
5140 aggressive transformations, so the value returned may not be that of the obvious
5141 source-language caller.
5145 <!-- _______________________________________________________________________ -->
5146 <div class="doc_subsubsection">
5147 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
5150 <div class="doc_text">
5154 declare i8 *@llvm.stacksave()
5160 The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
5161 the function stack, for use with <a href="#int_stackrestore">
5162 <tt>llvm.stackrestore</tt></a>. This is useful for implementing language
5163 features like scoped automatic variable sized arrays in C99.
5169 This intrinsic returns a opaque pointer value that can be passed to <a
5170 href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
5171 <tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
5172 <tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
5173 state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
5174 practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
5175 that were allocated after the <tt>llvm.stacksave</tt> was executed.
5180 <!-- _______________________________________________________________________ -->
5181 <div class="doc_subsubsection">
5182 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
5185 <div class="doc_text">
5189 declare void @llvm.stackrestore(i8 * %ptr)
5195 The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
5196 the function stack to the state it was in when the corresponding <a
5197 href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
5198 useful for implementing language features like scoped automatic variable sized
5205 See the description for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.
5211 <!-- _______________________________________________________________________ -->
5212 <div class="doc_subsubsection">
5213 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
5216 <div class="doc_text">
5220 declare void @llvm.prefetch(i8* <address>, i32 <rw>, i32 <locality>)
5227 The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
5228 a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
5230 effect on the behavior of the program but can change its performance
5237 <tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
5238 determining if the fetch should be for a read (0) or write (1), and
5239 <tt>locality</tt> is a temporal locality specifier ranging from (0) - no
5240 locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
5241 <tt>locality</tt> arguments must be constant integers.
5247 This intrinsic does not modify the behavior of the program. In particular,
5248 prefetches cannot trap and do not produce a value. On targets that support this
5249 intrinsic, the prefetch can provide hints to the processor cache for better
5255 <!-- _______________________________________________________________________ -->
5256 <div class="doc_subsubsection">
5257 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
5260 <div class="doc_text">
5264 declare void @llvm.pcmarker(i32 <id>)
5271 The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
5273 code to simulators and other tools. The method is target specific, but it is
5274 expected that the marker will use exported symbols to transmit the PC of the
5276 The marker makes no guarantees that it will remain with any specific instruction
5277 after optimizations. It is possible that the presence of a marker will inhibit
5278 optimizations. The intended use is to be inserted after optimizations to allow
5279 correlations of simulation runs.
5285 <tt>id</tt> is a numerical id identifying the marker.
5291 This intrinsic does not modify the behavior of the program. Backends that do not
5292 support this intrinisic may ignore it.
5297 <!-- _______________________________________________________________________ -->
5298 <div class="doc_subsubsection">
5299 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
5302 <div class="doc_text">
5306 declare i64 @llvm.readcyclecounter( )
5313 The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
5314 counter register (or similar low latency, high accuracy clocks) on those targets
5315 that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
5316 As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
5317 should only be used for small timings.
5323 When directly supported, reading the cycle counter should not modify any memory.
5324 Implementations are allowed to either return a application specific value or a
5325 system wide value. On backends without support, this is lowered to a constant 0.
5330 <!-- ======================================================================= -->
5331 <div class="doc_subsection">
5332 <a name="int_libc">Standard C Library Intrinsics</a>
5335 <div class="doc_text">
5337 LLVM provides intrinsics for a few important standard C library functions.
5338 These intrinsics allow source-language front-ends to pass information about the
5339 alignment of the pointer arguments to the code generator, providing opportunity
5340 for more efficient code generation.
5345 <!-- _______________________________________________________________________ -->
5346 <div class="doc_subsubsection">
5347 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
5350 <div class="doc_text">
5353 <p>This is an overloaded intrinsic. You can use llvm.memcpy on any integer bit
5354 width. Not all targets support all bit widths however.</p>
5356 declare void @llvm.memcpy.i8(i8 * <dest>, i8 * <src>,
5357 i8 <len>, i32 <align>)
5358 declare void @llvm.memcpy.i16(i8 * <dest>, i8 * <src>,
5359 i16 <len>, i32 <align>)
5360 declare void @llvm.memcpy.i32(i8 * <dest>, i8 * <src>,
5361 i32 <len>, i32 <align>)
5362 declare void @llvm.memcpy.i64(i8 * <dest>, i8 * <src>,
5363 i64 <len>, i32 <align>)
5369 The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
5370 location to the destination location.
5374 Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
5375 intrinsics do not return a value, and takes an extra alignment argument.
5381 The first argument is a pointer to the destination, the second is a pointer to
5382 the source. The third argument is an integer argument
5383 specifying the number of bytes to copy, and the fourth argument is the alignment
5384 of the source and destination locations.
5388 If the call to this intrinisic has an alignment value that is not 0 or 1, then
5389 the caller guarantees that both the source and destination pointers are aligned
5396 The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
5397 location to the destination location, which are not allowed to overlap. It
5398 copies "len" bytes of memory over. If the argument is known to be aligned to
5399 some boundary, this can be specified as the fourth argument, otherwise it should
5405 <!-- _______________________________________________________________________ -->
5406 <div class="doc_subsubsection">
5407 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
5410 <div class="doc_text">
5413 <p>This is an overloaded intrinsic. You can use llvm.memmove on any integer bit
5414 width. Not all targets support all bit widths however.</p>
5416 declare void @llvm.memmove.i8(i8 * <dest>, i8 * <src>,
5417 i8 <len>, i32 <align>)
5418 declare void @llvm.memmove.i16(i8 * <dest>, i8 * <src>,
5419 i16 <len>, i32 <align>)
5420 declare void @llvm.memmove.i32(i8 * <dest>, i8 * <src>,
5421 i32 <len>, i32 <align>)
5422 declare void @llvm.memmove.i64(i8 * <dest>, i8 * <src>,
5423 i64 <len>, i32 <align>)
5429 The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the source
5430 location to the destination location. It is similar to the
5431 '<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to overlap.
5435 Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
5436 intrinsics do not return a value, and takes an extra alignment argument.
5442 The first argument is a pointer to the destination, the second is a pointer to
5443 the source. The third argument is an integer argument
5444 specifying the number of bytes to copy, and the fourth argument is the alignment
5445 of the source and destination locations.
5449 If the call to this intrinisic has an alignment value that is not 0 or 1, then
5450 the caller guarantees that the source and destination pointers are aligned to
5457 The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the source
5458 location to the destination location, which may overlap. It
5459 copies "len" bytes of memory over. If the argument is known to be aligned to
5460 some boundary, this can be specified as the fourth argument, otherwise it should
5466 <!-- _______________________________________________________________________ -->
5467 <div class="doc_subsubsection">
5468 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
5471 <div class="doc_text">
5474 <p>This is an overloaded intrinsic. You can use llvm.memset on any integer bit
5475 width. Not all targets support all bit widths however.</p>
5477 declare void @llvm.memset.i8(i8 * <dest>, i8 <val>,
5478 i8 <len>, i32 <align>)
5479 declare void @llvm.memset.i16(i8 * <dest>, i8 <val>,
5480 i16 <len>, i32 <align>)
5481 declare void @llvm.memset.i32(i8 * <dest>, i8 <val>,
5482 i32 <len>, i32 <align>)
5483 declare void @llvm.memset.i64(i8 * <dest>, i8 <val>,
5484 i64 <len>, i32 <align>)
5490 The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a particular
5495 Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
5496 does not return a value, and takes an extra alignment argument.
5502 The first argument is a pointer to the destination to fill, the second is the
5503 byte value to fill it with, the third argument is an integer
5504 argument specifying the number of bytes to fill, and the fourth argument is the
5505 known alignment of destination location.
5509 If the call to this intrinisic has an alignment value that is not 0 or 1, then
5510 the caller guarantees that the destination pointer is aligned to that boundary.
5516 The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting at
5518 destination location. If the argument is known to be aligned to some boundary,
5519 this can be specified as the fourth argument, otherwise it should be set to 0 or
5525 <!-- _______________________________________________________________________ -->
5526 <div class="doc_subsubsection">
5527 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
5530 <div class="doc_text">
5533 <p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
5534 floating point or vector of floating point type. Not all targets support all
5537 declare float @llvm.sqrt.f32(float %Val)
5538 declare double @llvm.sqrt.f64(double %Val)
5539 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
5540 declare fp128 @llvm.sqrt.f128(fp128 %Val)
5541 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
5547 The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
5548 returning the same value as the libm '<tt>sqrt</tt>' functions would. Unlike
5549 <tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
5550 negative numbers other than -0.0 (which allows for better optimization, because
5551 there is no need to worry about errno being set). <tt>llvm.sqrt(-0.0)</tt> is
5552 defined to return -0.0 like IEEE sqrt.
5558 The argument and return value are floating point numbers of the same type.
5564 This function returns the sqrt of the specified operand if it is a nonnegative
5565 floating point number.
5569 <!-- _______________________________________________________________________ -->
5570 <div class="doc_subsubsection">
5571 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
5574 <div class="doc_text">
5577 <p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
5578 floating point or vector of floating point type. Not all targets support all
5581 declare float @llvm.powi.f32(float %Val, i32 %power)
5582 declare double @llvm.powi.f64(double %Val, i32 %power)
5583 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
5584 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
5585 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
5591 The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
5592 specified (positive or negative) power. The order of evaluation of
5593 multiplications is not defined. When a vector of floating point type is
5594 used, the second argument remains a scalar integer value.
5600 The second argument is an integer power, and the first is a value to raise to
5607 This function returns the first value raised to the second power with an
5608 unspecified sequence of rounding operations.</p>
5611 <!-- _______________________________________________________________________ -->
5612 <div class="doc_subsubsection">
5613 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
5616 <div class="doc_text">
5619 <p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
5620 floating point or vector of floating point type. Not all targets support all
5623 declare float @llvm.sin.f32(float %Val)
5624 declare double @llvm.sin.f64(double %Val)
5625 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
5626 declare fp128 @llvm.sin.f128(fp128 %Val)
5627 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
5633 The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.
5639 The argument and return value are floating point numbers of the same type.
5645 This function returns the sine of the specified operand, returning the
5646 same values as the libm <tt>sin</tt> functions would, and handles error
5647 conditions in the same way.</p>
5650 <!-- _______________________________________________________________________ -->
5651 <div class="doc_subsubsection">
5652 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
5655 <div class="doc_text">
5658 <p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
5659 floating point or vector of floating point type. Not all targets support all
5662 declare float @llvm.cos.f32(float %Val)
5663 declare double @llvm.cos.f64(double %Val)
5664 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
5665 declare fp128 @llvm.cos.f128(fp128 %Val)
5666 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
5672 The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.
5678 The argument and return value are floating point numbers of the same type.
5684 This function returns the cosine of the specified operand, returning the
5685 same values as the libm <tt>cos</tt> functions would, and handles error
5686 conditions in the same way.</p>
5689 <!-- _______________________________________________________________________ -->
5690 <div class="doc_subsubsection">
5691 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
5694 <div class="doc_text">
5697 <p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
5698 floating point or vector of floating point type. Not all targets support all
5701 declare float @llvm.pow.f32(float %Val, float %Power)
5702 declare double @llvm.pow.f64(double %Val, double %Power)
5703 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
5704 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
5705 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
5711 The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
5712 specified (positive or negative) power.
5718 The second argument is a floating point power, and the first is a value to
5719 raise to that power.
5725 This function returns the first value raised to the second power,
5727 same values as the libm <tt>pow</tt> functions would, and handles error
5728 conditions in the same way.</p>
5732 <!-- ======================================================================= -->
5733 <div class="doc_subsection">
5734 <a name="int_manip">Bit Manipulation Intrinsics</a>
5737 <div class="doc_text">
5739 LLVM provides intrinsics for a few important bit manipulation operations.
5740 These allow efficient code generation for some algorithms.
5745 <!-- _______________________________________________________________________ -->
5746 <div class="doc_subsubsection">
5747 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
5750 <div class="doc_text">
5753 <p>This is an overloaded intrinsic function. You can use bswap on any integer
5754 type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
5756 declare i16 @llvm.bswap.i16(i16 <id>)
5757 declare i32 @llvm.bswap.i32(i32 <id>)
5758 declare i64 @llvm.bswap.i64(i64 <id>)
5764 The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
5765 values with an even number of bytes (positive multiple of 16 bits). These are
5766 useful for performing operations on data that is not in the target's native
5773 The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
5774 and low byte of the input i16 swapped. Similarly, the <tt>llvm.bswap.i32</tt>
5775 intrinsic returns an i32 value that has the four bytes of the input i32
5776 swapped, so that if the input bytes are numbered 0, 1, 2, 3 then the returned
5777 i32 will have its bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i48</tt>,
5778 <tt>llvm.bswap.i64</tt> and other intrinsics extend this concept to
5779 additional even-byte lengths (6 bytes, 8 bytes and more, respectively).
5784 <!-- _______________________________________________________________________ -->
5785 <div class="doc_subsubsection">
5786 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
5789 <div class="doc_text">
5792 <p>This is an overloaded intrinsic. You can use llvm.ctpop on any integer bit
5793 width. Not all targets support all bit widths however.</p>
5795 declare i8 @llvm.ctpop.i8(i8 <src>)
5796 declare i16 @llvm.ctpop.i16(i16 <src>)
5797 declare i32 @llvm.ctpop.i32(i32 <src>)
5798 declare i64 @llvm.ctpop.i64(i64 <src>)
5799 declare i256 @llvm.ctpop.i256(i256 <src>)
5805 The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
5812 The only argument is the value to be counted. The argument may be of any
5813 integer type. The return type must match the argument type.
5819 The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
5823 <!-- _______________________________________________________________________ -->
5824 <div class="doc_subsubsection">
5825 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
5828 <div class="doc_text">
5831 <p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
5832 integer bit width. Not all targets support all bit widths however.</p>
5834 declare i8 @llvm.ctlz.i8 (i8 <src>)
5835 declare i16 @llvm.ctlz.i16(i16 <src>)
5836 declare i32 @llvm.ctlz.i32(i32 <src>)
5837 declare i64 @llvm.ctlz.i64(i64 <src>)
5838 declare i256 @llvm.ctlz.i256(i256 <src>)
5844 The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
5845 leading zeros in a variable.
5851 The only argument is the value to be counted. The argument may be of any
5852 integer type. The return type must match the argument type.
5858 The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
5859 in a variable. If the src == 0 then the result is the size in bits of the type
5860 of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.
5866 <!-- _______________________________________________________________________ -->
5867 <div class="doc_subsubsection">
5868 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
5871 <div class="doc_text">
5874 <p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
5875 integer bit width. Not all targets support all bit widths however.</p>
5877 declare i8 @llvm.cttz.i8 (i8 <src>)
5878 declare i16 @llvm.cttz.i16(i16 <src>)
5879 declare i32 @llvm.cttz.i32(i32 <src>)
5880 declare i64 @llvm.cttz.i64(i64 <src>)
5881 declare i256 @llvm.cttz.i256(i256 <src>)
5887 The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
5894 The only argument is the value to be counted. The argument may be of any
5895 integer type. The return type must match the argument type.
5901 The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
5902 in a variable. If the src == 0 then the result is the size in bits of the type
5903 of src. For example, <tt>llvm.cttz(2) = 1</tt>.
5907 <!-- _______________________________________________________________________ -->
5908 <div class="doc_subsubsection">
5909 <a name="int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic</a>
5912 <div class="doc_text">
5915 <p>This is an overloaded intrinsic. You can use <tt>llvm.part.select</tt>
5916 on any integer bit width.</p>
5918 declare i17 @llvm.part.select.i17 (i17 %val, i32 %loBit, i32 %hiBit)
5919 declare i29 @llvm.part.select.i29 (i29 %val, i32 %loBit, i32 %hiBit)
5923 <p>The '<tt>llvm.part.select</tt>' family of intrinsic functions selects a
5924 range of bits from an integer value and returns them in the same bit width as
5925 the original value.</p>
5928 <p>The first argument, <tt>%val</tt> and the result may be integer types of
5929 any bit width but they must have the same bit width. The second and third
5930 arguments must be <tt>i32</tt> type since they specify only a bit index.</p>
5933 <p>The operation of the '<tt>llvm.part.select</tt>' intrinsic has two modes
5934 of operation: forwards and reverse. If <tt>%loBit</tt> is greater than
5935 <tt>%hiBits</tt> then the intrinsic operates in reverse mode. Otherwise it
5936 operates in forward mode.</p>
5937 <p>In forward mode, this intrinsic is the equivalent of shifting <tt>%val</tt>
5938 right by <tt>%loBit</tt> bits and then ANDing it with a mask with
5939 only the <tt>%hiBit - %loBit</tt> bits set, as follows:</p>
5941 <li>The <tt>%val</tt> is shifted right (LSHR) by the number of bits specified
5942 by <tt>%loBits</tt>. This normalizes the value to the low order bits.</li>
5943 <li>The <tt>%loBits</tt> value is subtracted from the <tt>%hiBits</tt> value
5944 to determine the number of bits to retain.</li>
5945 <li>A mask of the retained bits is created by shifting a -1 value.</li>
5946 <li>The mask is ANDed with <tt>%val</tt> to produce the result.</li>
5948 <p>In reverse mode, a similar computation is made except that the bits are
5949 returned in the reverse order. So, for example, if <tt>X</tt> has the value
5950 <tt>i16 0x0ACF (101011001111)</tt> and we apply
5951 <tt>part.select(i16 X, 8, 3)</tt> to it, we get back the value
5952 <tt>i16 0x0026 (000000100110)</tt>.</p>
5955 <div class="doc_subsubsection">
5956 <a name="int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic</a>
5959 <div class="doc_text">
5962 <p>This is an overloaded intrinsic. You can use <tt>llvm.part.set</tt>
5963 on any integer bit width.</p>
5965 declare i17 @llvm.part.set.i17.i9 (i17 %val, i9 %repl, i32 %lo, i32 %hi)
5966 declare i29 @llvm.part.set.i29.i9 (i29 %val, i9 %repl, i32 %lo, i32 %hi)
5970 <p>The '<tt>llvm.part.set</tt>' family of intrinsic functions replaces a range
5971 of bits in an integer value with another integer value. It returns the integer
5972 with the replaced bits.</p>
5975 <p>The first argument, <tt>%val</tt>, and the result may be integer types of
5976 any bit width, but they must have the same bit width. <tt>%val</tt> is the value
5977 whose bits will be replaced. The second argument, <tt>%repl</tt> may be an
5978 integer of any bit width. The third and fourth arguments must be <tt>i32</tt>
5979 type since they specify only a bit index.</p>
5982 <p>The operation of the '<tt>llvm.part.set</tt>' intrinsic has two modes
5983 of operation: forwards and reverse. If <tt>%lo</tt> is greater than
5984 <tt>%hi</tt> then the intrinsic operates in reverse mode. Otherwise it
5985 operates in forward mode.</p>
5987 <p>For both modes, the <tt>%repl</tt> value is prepared for use by either
5988 truncating it down to the size of the replacement area or zero extending it
5989 up to that size.</p>
5991 <p>In forward mode, the bits between <tt>%lo</tt> and <tt>%hi</tt> (inclusive)
5992 are replaced with corresponding bits from <tt>%repl</tt>. That is the 0th bit
5993 in <tt>%repl</tt> replaces the <tt>%lo</tt>th bit in <tt>%val</tt> and etc. up
5994 to the <tt>%hi</tt>th bit.</p>
5996 <p>In reverse mode, a similar computation is made except that the bits are
5997 reversed. That is, the <tt>0</tt>th bit in <tt>%repl</tt> replaces the
5998 <tt>%hi</tt> bit in <tt>%val</tt> and etc. down to the <tt>%lo</tt>th bit.</p>
6003 llvm.part.set(0xFFFF, 0, 4, 7) -> 0xFF0F
6004 llvm.part.set(0xFFFF, 0, 7, 4) -> 0xFF0F
6005 llvm.part.set(0xFFFF, 1, 7, 4) -> 0xFF8F
6006 llvm.part.set(0xFFFF, F, 8, 3) -> 0xFFE7
6007 llvm.part.set(0xFFFF, 0, 3, 8) -> 0xFE07
6012 <!-- ======================================================================= -->
6013 <div class="doc_subsection">
6014 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
6017 <div class="doc_text">
6019 LLVM provides intrinsics for some arithmetic with overflow operations.
6024 <!-- _______________________________________________________________________ -->
6025 <div class="doc_subsubsection">
6026 <a name="int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics</a>
6029 <div class="doc_text">
6033 <p>This is an overloaded intrinsic. You can use <tt>llvm.sadd.with.overflow</tt>
6034 on any integer bit width.</p>
6037 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
6038 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6039 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
6044 <p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
6045 a signed addition of the two arguments, and indicate whether an overflow
6046 occurred during the signed summation.</p>
6050 <p>The arguments (%a and %b) and the first element of the result structure may
6051 be of integer types of any bit width, but they must have the same bit width. The
6052 second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6053 and <tt>%b</tt> are the two values that will undergo signed addition.</p>
6057 <p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
6058 a signed addition of the two variables. They return a structure — the
6059 first element of which is the signed summation, and the second element of which
6060 is a bit specifying if the signed summation resulted in an overflow.</p>
6064 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6065 %sum = extractvalue {i32, i1} %res, 0
6066 %obit = extractvalue {i32, i1} %res, 1
6067 br i1 %obit, label %overflow, label %normal
6072 <!-- _______________________________________________________________________ -->
6073 <div class="doc_subsubsection">
6074 <a name="int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics</a>
6077 <div class="doc_text">
6081 <p>This is an overloaded intrinsic. You can use <tt>llvm.uadd.with.overflow</tt>
6082 on any integer bit width.</p>
6085 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
6086 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6087 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
6092 <p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
6093 an unsigned addition of the two arguments, and indicate whether a carry occurred
6094 during the unsigned summation.</p>
6098 <p>The arguments (%a and %b) and the first element of the result structure may
6099 be of integer types of any bit width, but they must have the same bit width. The
6100 second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6101 and <tt>%b</tt> are the two values that will undergo unsigned addition.</p>
6105 <p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
6106 an unsigned addition of the two arguments. They return a structure — the
6107 first element of which is the sum, and the second element of which is a bit
6108 specifying if the unsigned summation resulted in a carry.</p>
6112 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6113 %sum = extractvalue {i32, i1} %res, 0
6114 %obit = extractvalue {i32, i1} %res, 1
6115 br i1 %obit, label %carry, label %normal
6120 <!-- _______________________________________________________________________ -->
6121 <div class="doc_subsubsection">
6122 <a name="int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics</a>
6125 <div class="doc_text">
6129 <p>This is an overloaded intrinsic. You can use <tt>llvm.ssub.with.overflow</tt>
6130 on any integer bit width.</p>
6133 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
6134 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6135 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
6140 <p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
6141 a signed subtraction of the two arguments, and indicate whether an overflow
6142 occurred during the signed subtraction.</p>
6146 <p>The arguments (%a and %b) and the first element of the result structure may
6147 be of integer types of any bit width, but they must have the same bit width. The
6148 second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6149 and <tt>%b</tt> are the two values that will undergo signed subtraction.</p>
6153 <p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
6154 a signed subtraction of the two arguments. They return a structure — the
6155 first element of which is the subtraction, and the second element of which is a bit
6156 specifying if the signed subtraction resulted in an overflow.</p>
6160 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6161 %sum = extractvalue {i32, i1} %res, 0
6162 %obit = extractvalue {i32, i1} %res, 1
6163 br i1 %obit, label %overflow, label %normal
6168 <!-- _______________________________________________________________________ -->
6169 <div class="doc_subsubsection">
6170 <a name="int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt>' Intrinsics</a>
6173 <div class="doc_text">
6177 <p>This is an overloaded intrinsic. You can use <tt>llvm.usub.with.overflow</tt>
6178 on any integer bit width.</p>
6181 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
6182 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6183 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
6188 <p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
6189 an unsigned subtraction of the two arguments, and indicate whether an overflow
6190 occurred during the unsigned subtraction.</p>
6194 <p>The arguments (%a and %b) and the first element of the result structure may
6195 be of integer types of any bit width, but they must have the same bit width. The
6196 second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6197 and <tt>%b</tt> are the two values that will undergo unsigned subtraction.</p>
6201 <p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
6202 an unsigned subtraction of the two arguments. They return a structure — the
6203 first element of which is the subtraction, and the second element of which is a bit
6204 specifying if the unsigned subtraction resulted in an overflow.</p>
6208 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6209 %sum = extractvalue {i32, i1} %res, 0
6210 %obit = extractvalue {i32, i1} %res, 1
6211 br i1 %obit, label %overflow, label %normal
6216 <!-- _______________________________________________________________________ -->
6217 <div class="doc_subsubsection">
6218 <a name="int_smul_overflow">'<tt>llvm.smul.with.overflow.*</tt>' Intrinsics</a>
6221 <div class="doc_text">
6225 <p>This is an overloaded intrinsic. You can use <tt>llvm.smul.with.overflow</tt>
6226 on any integer bit width.</p>
6229 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
6230 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6231 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
6236 <p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
6237 a signed multiplication of the two arguments, and indicate whether an overflow
6238 occurred during the signed multiplication.</p>
6242 <p>The arguments (%a and %b) and the first element of the result structure may
6243 be of integer types of any bit width, but they must have the same bit width. The
6244 second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6245 and <tt>%b</tt> are the two values that will undergo signed multiplication.</p>
6249 <p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
6250 a signed multiplication of the two arguments. They return a structure —
6251 the first element of which is the multiplication, and the second element of
6252 which is a bit specifying if the signed multiplication resulted in an
6257 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6258 %sum = extractvalue {i32, i1} %res, 0
6259 %obit = extractvalue {i32, i1} %res, 1
6260 br i1 %obit, label %overflow, label %normal
6265 <!-- _______________________________________________________________________ -->
6266 <div class="doc_subsubsection">
6267 <a name="int_umul_overflow">'<tt>llvm.umul.with.overflow.*</tt>' Intrinsics</a>
6270 <div class="doc_text">
6274 <p>This is an overloaded intrinsic. You can use <tt>llvm.umul.with.overflow</tt>
6275 on any integer bit width.</p>
6278 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
6279 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6280 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
6285 <p><i><b>Warning:</b> '<tt>llvm.umul.with.overflow</tt>' is badly broken. It is
6286 actively being fixed, but it should not currently be used!</i></p>
6288 <p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
6289 a unsigned multiplication of the two arguments, and indicate whether an overflow
6290 occurred during the unsigned multiplication.</p>
6294 <p>The arguments (%a and %b) and the first element of the result structure may
6295 be of integer types of any bit width, but they must have the same bit width. The
6296 second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
6297 and <tt>%b</tt> are the two values that will undergo unsigned
6302 <p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
6303 an unsigned multiplication of the two arguments. They return a structure —
6304 the first element of which is the multiplication, and the second element of
6305 which is a bit specifying if the unsigned multiplication resulted in an
6310 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6311 %sum = extractvalue {i32, i1} %res, 0
6312 %obit = extractvalue {i32, i1} %res, 1
6313 br i1 %obit, label %overflow, label %normal
6318 <!-- ======================================================================= -->
6319 <div class="doc_subsection">
6320 <a name="int_debugger">Debugger Intrinsics</a>
6323 <div class="doc_text">
6325 The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
6326 are described in the <a
6327 href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
6328 Debugging</a> document.
6333 <!-- ======================================================================= -->
6334 <div class="doc_subsection">
6335 <a name="int_eh">Exception Handling Intrinsics</a>
6338 <div class="doc_text">
6339 <p> The LLVM exception handling intrinsics (which all start with
6340 <tt>llvm.eh.</tt> prefix), are described in the <a
6341 href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
6342 Handling</a> document. </p>
6345 <!-- ======================================================================= -->
6346 <div class="doc_subsection">
6347 <a name="int_trampoline">Trampoline Intrinsic</a>
6350 <div class="doc_text">
6352 This intrinsic makes it possible to excise one parameter, marked with
6353 the <tt>nest</tt> attribute, from a function. The result is a callable
6354 function pointer lacking the nest parameter - the caller does not need
6355 to provide a value for it. Instead, the value to use is stored in
6356 advance in a "trampoline", a block of memory usually allocated
6357 on the stack, which also contains code to splice the nest value into the
6358 argument list. This is used to implement the GCC nested function address
6362 For example, if the function is
6363 <tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
6364 pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as follows:</p>
6366 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
6367 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
6368 %p = call i8* @llvm.init.trampoline( i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8*), i8* %nval )
6369 %fp = bitcast i8* %p to i32 (i32, i32)*
6371 <p>The call <tt>%val = call i32 %fp( i32 %x, i32 %y )</tt> is then equivalent
6372 to <tt>%val = call i32 %f( i8* %nval, i32 %x, i32 %y )</tt>.</p>
6375 <!-- _______________________________________________________________________ -->
6376 <div class="doc_subsubsection">
6377 <a name="int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a>
6379 <div class="doc_text">
6382 declare i8* @llvm.init.trampoline(i8* <tramp>, i8* <func>, i8* <nval>)
6386 This fills the memory pointed to by <tt>tramp</tt> with code
6387 and returns a function pointer suitable for executing it.
6391 The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
6392 pointers. The <tt>tramp</tt> argument must point to a sufficiently large
6393 and sufficiently aligned block of memory; this memory is written to by the
6394 intrinsic. Note that the size and the alignment are target-specific - LLVM
6395 currently provides no portable way of determining them, so a front-end that
6396 generates this intrinsic needs to have some target-specific knowledge.
6397 The <tt>func</tt> argument must hold a function bitcast to an <tt>i8*</tt>.
6401 The block of memory pointed to by <tt>tramp</tt> is filled with target
6402 dependent code, turning it into a function. A pointer to this function is
6403 returned, but needs to be bitcast to an
6404 <a href="#int_trampoline">appropriate function pointer type</a>
6405 before being called. The new function's signature is the same as that of
6406 <tt>func</tt> with any arguments marked with the <tt>nest</tt> attribute
6407 removed. At most one such <tt>nest</tt> argument is allowed, and it must be
6408 of pointer type. Calling the new function is equivalent to calling
6409 <tt>func</tt> with the same argument list, but with <tt>nval</tt> used for the
6410 missing <tt>nest</tt> argument. If, after calling
6411 <tt>llvm.init.trampoline</tt>, the memory pointed to by <tt>tramp</tt> is
6412 modified, then the effect of any later call to the returned function pointer is
6417 <!-- ======================================================================= -->
6418 <div class="doc_subsection">
6419 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
6422 <div class="doc_text">
6424 These intrinsic functions expand the "universal IR" of LLVM to represent
6425 hardware constructs for atomic operations and memory synchronization. This
6426 provides an interface to the hardware, not an interface to the programmer. It
6427 is aimed at a low enough level to allow any programming models or APIs
6428 (Application Programming Interfaces) which
6429 need atomic behaviors to map cleanly onto it. It is also modeled primarily on
6430 hardware behavior. Just as hardware provides a "universal IR" for source
6431 languages, it also provides a starting point for developing a "universal"
6432 atomic operation and synchronization IR.
6435 These do <em>not</em> form an API such as high-level threading libraries,
6436 software transaction memory systems, atomic primitives, and intrinsic
6437 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
6438 application libraries. The hardware interface provided by LLVM should allow
6439 a clean implementation of all of these APIs and parallel programming models.
6440 No one model or paradigm should be selected above others unless the hardware
6441 itself ubiquitously does so.
6446 <!-- _______________________________________________________________________ -->
6447 <div class="doc_subsubsection">
6448 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
6450 <div class="doc_text">
6453 declare void @llvm.memory.barrier( i1 <ll>, i1 <ls>, i1 <sl>, i1 <ss>,
6459 The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
6460 specific pairs of memory access types.
6464 The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
6465 The first four arguments enables a specific barrier as listed below. The fith
6466 argument specifies that the barrier applies to io or device or uncached memory.
6470 <li><tt>ll</tt>: load-load barrier</li>
6471 <li><tt>ls</tt>: load-store barrier</li>
6472 <li><tt>sl</tt>: store-load barrier</li>
6473 <li><tt>ss</tt>: store-store barrier</li>
6474 <li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
6478 This intrinsic causes the system to enforce some ordering constraints upon
6479 the loads and stores of the program. This barrier does not indicate
6480 <em>when</em> any events will occur, it only enforces an <em>order</em> in
6481 which they occur. For any of the specified pairs of load and store operations
6482 (f.ex. load-load, or store-load), all of the first operations preceding the
6483 barrier will complete before any of the second operations succeeding the
6484 barrier begin. Specifically the semantics for each pairing is as follows:
6487 <li><tt>ll</tt>: All loads before the barrier must complete before any load
6488 after the barrier begins.</li>
6490 <li><tt>ls</tt>: All loads before the barrier must complete before any
6491 store after the barrier begins.</li>
6492 <li><tt>ss</tt>: All stores before the barrier must complete before any
6493 store after the barrier begins.</li>
6494 <li><tt>sl</tt>: All stores before the barrier must complete before any
6495 load after the barrier begins.</li>
6498 These semantics are applied with a logical "and" behavior when more than one
6499 is enabled in a single memory barrier intrinsic.
6502 Backends may implement stronger barriers than those requested when they do not
6503 support as fine grained a barrier as requested. Some architectures do not
6504 need all types of barriers and on such architectures, these become noops.
6511 %result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
6512 call void @llvm.memory.barrier( i1 false, i1 true, i1 false, i1 false )
6513 <i>; guarantee the above finishes</i>
6514 store i32 8, %ptr <i>; before this begins</i>
6518 <!-- _______________________________________________________________________ -->
6519 <div class="doc_subsubsection">
6520 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
6522 <div class="doc_text">
6525 This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
6526 any integer bit width and for different address spaces. Not all targets
6527 support all bit widths however.</p>
6530 declare i8 @llvm.atomic.cmp.swap.i8.p0i8( i8* <ptr>, i8 <cmp>, i8 <val> )
6531 declare i16 @llvm.atomic.cmp.swap.i16.p0i16( i16* <ptr>, i16 <cmp>, i16 <val> )
6532 declare i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* <ptr>, i32 <cmp>, i32 <val> )
6533 declare i64 @llvm.atomic.cmp.swap.i64.p0i64( i64* <ptr>, i64 <cmp>, i64 <val> )
6538 This loads a value in memory and compares it to a given value. If they are
6539 equal, it stores a new value into the memory.
6543 The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result as
6544 well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
6545 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
6546 this integer type. While any bit width integer may be used, targets may only
6547 lower representations they support in hardware.
6552 This entire intrinsic must be executed atomically. It first loads the value
6553 in memory pointed to by <tt>ptr</tt> and compares it with the value
6554 <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the memory. The
6555 loaded value is yielded in all cases. This provides the equivalent of an
6556 atomic compare-and-swap operation within the SSA framework.
6564 %val1 = add i32 4, 4
6565 %result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 4, %val1 )
6566 <i>; yields {i32}:result1 = 4</i>
6567 %stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
6568 %memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
6570 %val2 = add i32 1, 1
6571 %result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 5, %val2 )
6572 <i>; yields {i32}:result2 = 8</i>
6573 %stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
6575 %memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
6579 <!-- _______________________________________________________________________ -->
6580 <div class="doc_subsubsection">
6581 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
6583 <div class="doc_text">
6587 This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
6588 integer bit width. Not all targets support all bit widths however.</p>
6590 declare i8 @llvm.atomic.swap.i8.p0i8( i8* <ptr>, i8 <val> )
6591 declare i16 @llvm.atomic.swap.i16.p0i16( i16* <ptr>, i16 <val> )
6592 declare i32 @llvm.atomic.swap.i32.p0i32( i32* <ptr>, i32 <val> )
6593 declare i64 @llvm.atomic.swap.i64.p0i64( i64* <ptr>, i64 <val> )
6598 This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
6599 the value from memory. It then stores the value in <tt>val</tt> in the memory
6605 The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both the
6606 <tt>val</tt> argument and the result must be integers of the same bit width.
6607 The first argument, <tt>ptr</tt>, must be a pointer to a value of this
6608 integer type. The targets may only lower integer representations they
6613 This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
6614 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
6615 equivalent of an atomic swap operation within the SSA framework.
6623 %val1 = add i32 4, 4
6624 %result1 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val1 )
6625 <i>; yields {i32}:result1 = 4</i>
6626 %stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
6627 %memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
6629 %val2 = add i32 1, 1
6630 %result2 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val2 )
6631 <i>; yields {i32}:result2 = 8</i>
6633 %stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
6634 %memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
6638 <!-- _______________________________________________________________________ -->
6639 <div class="doc_subsubsection">
6640 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
6643 <div class="doc_text">
6646 This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.add</tt> on any
6647 integer bit width. Not all targets support all bit widths however.</p>
6649 declare i8 @llvm.atomic.load.add.i8..p0i8( i8* <ptr>, i8 <delta> )
6650 declare i16 @llvm.atomic.load.add.i16..p0i16( i16* <ptr>, i16 <delta> )
6651 declare i32 @llvm.atomic.load.add.i32..p0i32( i32* <ptr>, i32 <delta> )
6652 declare i64 @llvm.atomic.load.add.i64..p0i64( i64* <ptr>, i64 <delta> )
6657 This intrinsic adds <tt>delta</tt> to the value stored in memory at
6658 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
6663 The intrinsic takes two arguments, the first a pointer to an integer value
6664 and the second an integer value. The result is also an integer value. These
6665 integer types can have any bit width, but they must all have the same bit
6666 width. The targets may only lower integer representations they support.
6670 This intrinsic does a series of operations atomically. It first loads the
6671 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
6672 to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
6679 %result1 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 4 )
6680 <i>; yields {i32}:result1 = 4</i>
6681 %result2 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 2 )
6682 <i>; yields {i32}:result2 = 8</i>
6683 %result3 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 5 )
6684 <i>; yields {i32}:result3 = 10</i>
6685 %memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
6689 <!-- _______________________________________________________________________ -->
6690 <div class="doc_subsubsection">
6691 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
6694 <div class="doc_text">
6697 This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.sub</tt> on
6698 any integer bit width and for different address spaces. Not all targets
6699 support all bit widths however.</p>
6701 declare i8 @llvm.atomic.load.sub.i8.p0i32( i8* <ptr>, i8 <delta> )
6702 declare i16 @llvm.atomic.load.sub.i16.p0i32( i16* <ptr>, i16 <delta> )
6703 declare i32 @llvm.atomic.load.sub.i32.p0i32( i32* <ptr>, i32 <delta> )
6704 declare i64 @llvm.atomic.load.sub.i64.p0i32( i64* <ptr>, i64 <delta> )
6709 This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
6710 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
6715 The intrinsic takes two arguments, the first a pointer to an integer value
6716 and the second an integer value. The result is also an integer value. These
6717 integer types can have any bit width, but they must all have the same bit
6718 width. The targets may only lower integer representations they support.
6722 This intrinsic does a series of operations atomically. It first loads the
6723 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
6724 result to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
6731 %result1 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 4 )
6732 <i>; yields {i32}:result1 = 8</i>
6733 %result2 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 2 )
6734 <i>; yields {i32}:result2 = 4</i>
6735 %result3 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 5 )
6736 <i>; yields {i32}:result3 = 2</i>
6737 %memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
6741 <!-- _______________________________________________________________________ -->
6742 <div class="doc_subsubsection">
6743 <a name="int_atomic_load_and">'<tt>llvm.atomic.load.and.*</tt>' Intrinsic</a><br>
6744 <a name="int_atomic_load_nand">'<tt>llvm.atomic.load.nand.*</tt>' Intrinsic</a><br>
6745 <a name="int_atomic_load_or">'<tt>llvm.atomic.load.or.*</tt>' Intrinsic</a><br>
6746 <a name="int_atomic_load_xor">'<tt>llvm.atomic.load.xor.*</tt>' Intrinsic</a><br>
6749 <div class="doc_text">
6752 These are overloaded intrinsics. You can use <tt>llvm.atomic.load_and</tt>,
6753 <tt>llvm.atomic.load_nand</tt>, <tt>llvm.atomic.load_or</tt>, and
6754 <tt>llvm.atomic.load_xor</tt> on any integer bit width and for different
6755 address spaces. Not all targets support all bit widths however.</p>
6757 declare i8 @llvm.atomic.load.and.i8.p0i8( i8* <ptr>, i8 <delta> )
6758 declare i16 @llvm.atomic.load.and.i16.p0i16( i16* <ptr>, i16 <delta> )
6759 declare i32 @llvm.atomic.load.and.i32.p0i32( i32* <ptr>, i32 <delta> )
6760 declare i64 @llvm.atomic.load.and.i64.p0i64( i64* <ptr>, i64 <delta> )
6765 declare i8 @llvm.atomic.load.or.i8.p0i8( i8* <ptr>, i8 <delta> )
6766 declare i16 @llvm.atomic.load.or.i16.p0i16( i16* <ptr>, i16 <delta> )
6767 declare i32 @llvm.atomic.load.or.i32.p0i32( i32* <ptr>, i32 <delta> )
6768 declare i64 @llvm.atomic.load.or.i64.p0i64( i64* <ptr>, i64 <delta> )
6773 declare i8 @llvm.atomic.load.nand.i8.p0i32( i8* <ptr>, i8 <delta> )
6774 declare i16 @llvm.atomic.load.nand.i16.p0i32( i16* <ptr>, i16 <delta> )
6775 declare i32 @llvm.atomic.load.nand.i32.p0i32( i32* <ptr>, i32 <delta> )
6776 declare i64 @llvm.atomic.load.nand.i64.p0i32( i64* <ptr>, i64 <delta> )
6781 declare i8 @llvm.atomic.load.xor.i8.p0i32( i8* <ptr>, i8 <delta> )
6782 declare i16 @llvm.atomic.load.xor.i16.p0i32( i16* <ptr>, i16 <delta> )
6783 declare i32 @llvm.atomic.load.xor.i32.p0i32( i32* <ptr>, i32 <delta> )
6784 declare i64 @llvm.atomic.load.xor.i64.p0i32( i64* <ptr>, i64 <delta> )
6789 These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
6790 the value stored in memory at <tt>ptr</tt>. It yields the original value
6796 These intrinsics take two arguments, the first a pointer to an integer value
6797 and the second an integer value. The result is also an integer value. These
6798 integer types can have any bit width, but they must all have the same bit
6799 width. The targets may only lower integer representations they support.
6803 These intrinsics does a series of operations atomically. They first load the
6804 value stored at <tt>ptr</tt>. They then do the bitwise operation
6805 <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the original
6806 value stored at <tt>ptr</tt>.
6812 store i32 0x0F0F, %ptr
6813 %result0 = call i32 @llvm.atomic.load.nand.i32.p0i32( i32* %ptr, i32 0xFF )
6814 <i>; yields {i32}:result0 = 0x0F0F</i>
6815 %result1 = call i32 @llvm.atomic.load.and.i32.p0i32( i32* %ptr, i32 0xFF )
6816 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
6817 %result2 = call i32 @llvm.atomic.load.or.i32.p0i32( i32* %ptr, i32 0F )
6818 <i>; yields {i32}:result2 = 0xF0</i>
6819 %result3 = call i32 @llvm.atomic.load.xor.i32.p0i32( i32* %ptr, i32 0F )
6820 <i>; yields {i32}:result3 = FF</i>
6821 %memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
6826 <!-- _______________________________________________________________________ -->
6827 <div class="doc_subsubsection">
6828 <a name="int_atomic_load_max">'<tt>llvm.atomic.load.max.*</tt>' Intrinsic</a><br>
6829 <a name="int_atomic_load_min">'<tt>llvm.atomic.load.min.*</tt>' Intrinsic</a><br>
6830 <a name="int_atomic_load_umax">'<tt>llvm.atomic.load.umax.*</tt>' Intrinsic</a><br>
6831 <a name="int_atomic_load_umin">'<tt>llvm.atomic.load.umin.*</tt>' Intrinsic</a><br>
6834 <div class="doc_text">
6837 These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
6838 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
6839 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
6840 address spaces. Not all targets
6841 support all bit widths however.</p>
6843 declare i8 @llvm.atomic.load.max.i8.p0i8( i8* <ptr>, i8 <delta> )
6844 declare i16 @llvm.atomic.load.max.i16.p0i16( i16* <ptr>, i16 <delta> )
6845 declare i32 @llvm.atomic.load.max.i32.p0i32( i32* <ptr>, i32 <delta> )
6846 declare i64 @llvm.atomic.load.max.i64.p0i64( i64* <ptr>, i64 <delta> )
6851 declare i8 @llvm.atomic.load.min.i8.p0i8( i8* <ptr>, i8 <delta> )
6852 declare i16 @llvm.atomic.load.min.i16.p0i16( i16* <ptr>, i16 <delta> )
6853 declare i32 @llvm.atomic.load.min.i32..p0i32( i32* <ptr>, i32 <delta> )
6854 declare i64 @llvm.atomic.load.min.i64..p0i64( i64* <ptr>, i64 <delta> )
6859 declare i8 @llvm.atomic.load.umax.i8.p0i8( i8* <ptr>, i8 <delta> )
6860 declare i16 @llvm.atomic.load.umax.i16.p0i16( i16* <ptr>, i16 <delta> )
6861 declare i32 @llvm.atomic.load.umax.i32.p0i32( i32* <ptr>, i32 <delta> )
6862 declare i64 @llvm.atomic.load.umax.i64.p0i64( i64* <ptr>, i64 <delta> )
6867 declare i8 @llvm.atomic.load.umin.i8..p0i8( i8* <ptr>, i8 <delta> )
6868 declare i16 @llvm.atomic.load.umin.i16.p0i16( i16* <ptr>, i16 <delta> )
6869 declare i32 @llvm.atomic.load.umin.i32..p0i32( i32* <ptr>, i32 <delta> )
6870 declare i64 @llvm.atomic.load.umin.i64..p0i64( i64* <ptr>, i64 <delta> )
6875 These intrinsics takes the signed or unsigned minimum or maximum of
6876 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
6877 original value at <tt>ptr</tt>.
6882 These intrinsics take two arguments, the first a pointer to an integer value
6883 and the second an integer value. The result is also an integer value. These
6884 integer types can have any bit width, but they must all have the same bit
6885 width. The targets may only lower integer representations they support.
6889 These intrinsics does a series of operations atomically. They first load the
6890 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or max
6891 <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They yield
6892 the original value stored at <tt>ptr</tt>.
6899 %result0 = call i32 @llvm.atomic.load.min.i32.p0i32( i32* %ptr, i32 -2 )
6900 <i>; yields {i32}:result0 = 7</i>
6901 %result1 = call i32 @llvm.atomic.load.max.i32.p0i32( i32* %ptr, i32 8 )
6902 <i>; yields {i32}:result1 = -2</i>
6903 %result2 = call i32 @llvm.atomic.load.umin.i32.p0i32( i32* %ptr, i32 10 )
6904 <i>; yields {i32}:result2 = 8</i>
6905 %result3 = call i32 @llvm.atomic.load.umax.i32.p0i32( i32* %ptr, i32 30 )
6906 <i>; yields {i32}:result3 = 8</i>
6907 %memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
6911 <!-- ======================================================================= -->
6912 <div class="doc_subsection">
6913 <a name="int_general">General Intrinsics</a>
6916 <div class="doc_text">
6917 <p> This class of intrinsics is designed to be generic and has
6918 no specific purpose. </p>
6921 <!-- _______________________________________________________________________ -->
6922 <div class="doc_subsubsection">
6923 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
6926 <div class="doc_text">
6930 declare void @llvm.var.annotation(i8* <val>, i8* <str>, i8* <str>, i32 <int> )
6936 The '<tt>llvm.var.annotation</tt>' intrinsic
6942 The first argument is a pointer to a value, the second is a pointer to a
6943 global string, the third is a pointer to a global string which is the source
6944 file name, and the last argument is the line number.
6950 This intrinsic allows annotation of local variables with arbitrary strings.
6951 This can be useful for special purpose optimizations that want to look for these
6952 annotations. These have no other defined use, they are ignored by code
6953 generation and optimization.
6957 <!-- _______________________________________________________________________ -->
6958 <div class="doc_subsubsection">
6959 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
6962 <div class="doc_text">
6965 <p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
6966 any integer bit width.
6969 declare i8 @llvm.annotation.i8(i8 <val>, i8* <str>, i8* <str>, i32 <int> )
6970 declare i16 @llvm.annotation.i16(i16 <val>, i8* <str>, i8* <str>, i32 <int> )
6971 declare i32 @llvm.annotation.i32(i32 <val>, i8* <str>, i8* <str>, i32 <int> )
6972 declare i64 @llvm.annotation.i64(i64 <val>, i8* <str>, i8* <str>, i32 <int> )
6973 declare i256 @llvm.annotation.i256(i256 <val>, i8* <str>, i8* <str>, i32 <int> )
6979 The '<tt>llvm.annotation</tt>' intrinsic.
6985 The first argument is an integer value (result of some expression),
6986 the second is a pointer to a global string, the third is a pointer to a global
6987 string which is the source file name, and the last argument is the line number.
6988 It returns the value of the first argument.
6994 This intrinsic allows annotations to be put on arbitrary expressions
6995 with arbitrary strings. This can be useful for special purpose optimizations
6996 that want to look for these annotations. These have no other defined use, they
6997 are ignored by code generation and optimization.
7001 <!-- _______________________________________________________________________ -->
7002 <div class="doc_subsubsection">
7003 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
7006 <div class="doc_text">
7010 declare void @llvm.trap()
7016 The '<tt>llvm.trap</tt>' intrinsic
7028 This intrinsics is lowered to the target dependent trap instruction. If the
7029 target does not have a trap instruction, this intrinsic will be lowered to the
7030 call of the abort() function.
7034 <!-- _______________________________________________________________________ -->
7035 <div class="doc_subsubsection">
7036 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
7038 <div class="doc_text">
7041 declare void @llvm.stackprotector( i8* <guard>, i8** <slot> )
7046 The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and stores
7047 it onto the stack at <tt>slot</tt>. The stack slot is adjusted to ensure that
7048 it is placed on the stack before local variables.
7052 The <tt>llvm.stackprotector</tt> intrinsic requires two pointer arguments. The
7053 first argument is the value loaded from the stack guard
7054 <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt> that
7055 has enough space to hold the value of the guard.
7059 This intrinsic causes the prologue/epilogue inserter to force the position of
7060 the <tt>AllocaInst</tt> stack slot to be before local variables on the
7061 stack. This is to ensure that if a local variable on the stack is overwritten,
7062 it will destroy the value of the guard. When the function exits, the guard on
7063 the stack is checked against the original guard. If they're different, then
7064 the program aborts by calling the <tt>__stack_chk_fail()</tt> function.
7068 <!-- *********************************************************************** -->
7071 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
7072 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
7073 <a href="http://validator.w3.org/check/referer"><img
7074 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
7076 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
7077 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
7078 Last modified: $Date$